No. ... Division Range Shelf..... f>' ~Dtoib \p$ Received ^E9^^:4 l8jb& /" PRESENTP:D TO THE I Libf| of HieUniversitj of California, I %^z^ lOFT&^ll XS jJ^BWll il^^^n? >r 7*1 MAB^rO. \! VARRONIANUS. LICET OMNIA ITALICA PRO ROMANIS HABEAM. QuiNTU. DrfntcB at tfic 23ni&ersttfi Press. VABRONIANUS:' A CRITICAL AND HISTORICAL INTRODUCTION TO THE ETHNOGRAPHY OF ANCIENT ITALY AKD TO THE PHILOLOGICAL STUDY OF THE LATIN LANGUAGE. BY JOHN WILLIAM DONALDSON, D.D., HEAD MASTER OF BURY SCHOOL; AND FOBMERLY FELLOW AND CLASSICAL LECTUBER OF TBIN1TY COLLEGE, CAMBRIDGE. SECOND EDITION, REVISED AND CONSIDERABLY ENLARGED. LONDON: JOHN W. PARKER AND SON. CAMBRIDGE: JOHN DEIGHTON. 1852. TO ^_ THE RIGHT REVEREND CONNOP THIRLWALL, D.D. LORD BISHOP OF ST. DAVID'S, PRESIDENT OF THE PHILOLOGICAL SOCIETY, ETC. ETC. MY LORD, IN repeating my dedication of this work to your Lordship, I may perhaps feel more confidence, than when I first inscribed it with your distinguished name, not only because it has, to a certain extent, obtained the approval of the public to which it appealed, but still more, because I am enabled to revise it with such additional knowledge as I have acquired in the interval since its first appearance. But the renewal of my labours in this field has increased my conviction of the difficulties, which attend a scientific examination of the Latin Language ; and I have introduced so much new matter, that I must feel anxious to know, whether the conclusions, at which I have arrived, are likely to be sanctioned by your Lordship and other competent judges. However this may be, the republication of this book has at least given me an opportunity of renewing the ex- pression of my respect and esteem for your Lordship, and of declaring my undiminished appreciation of the services, which you have rendered to the students of classical philology in this country. I have the honour to be, MY LORD, Your Lordship's faithful servant, J. W. DONALDSON. PREFACE TO THE SECOND EDITION. work, as it originally appeared, was a first attempt to discuss the comparative philology of the Latin Language on the broad basis of general Ethnogra- phy, and to show historically how the classical idiom of ancient Rome resulted from the absorption or centrali- sation of the other dialects spoken in the Peninsula. My motto was : licet omnia Italicapro Romanis habeam ; and I did not content myself with a survey of the Ita- lian races, but endeavoured to prove that the elements of this cisalpine population might be recognised in the Scythia of Herodotus, either in juxta-position or in some degree of fusion ; and thus, that they might be traced back to the primary settlements of the Indo-Germanic family. In maintaining the composite structure of the Latin language, I assert also that the different elements, of which it is made up, are to be found in the fragmentary languages which have come down to us. When Lepsius proposed (de Tabulis EuguUnis, pp. 102, 105) to defend the thesis : Latinam linguam non esse mixtam, he must have had in view, either an opposition to the doctrine that Latin may be divided into a Greek and non-Greek part, which Lassen calls one-sided and erroneous, for we might as well speak of the German and non-German, or the Indian and non-Indian parts of Latin (Rhein. Mus. 1833, p. 361); or else a confutation of one of those untenable theories, which represent this language as an viii PREFACE TO THE SECOND EDITION. imperfectly combined assemblage of heterogeneous in- gredients. Admitting that in Italy, as in other penin- sulas and islands of Europe, there must have been a Celtic substratum, this book undertakes to prove that the old Italian tribes were either Sclavonians, Low- Germans, or that well-fused compound of these two, the Lithuanians. Thus all the elements were homo- geneous, and a perfect combination or absorption of idioms was a natural result of the political centrali- sation occasioned by the conquests of the Imperial City on the Tiber. In order to arrive at this conclusion, it was necessary to examine all the details of Italian ethnography ; and I am quite sure that, if Niebuhr thought a long series of essays on the old tribes of the Peninsula a proper intro- duction to his researches in Roman history, a similar investigation, supported by an analysis of the linguistic fragments, must be a still more indispensable preliminary to a treatise on Latin philology. To complete the ethnographical portion of this work, I have drawn up a map of ancient Italy, which may also serve as a specimen of the best method, as it appears to me, of representing in a geographical form the results of philological and historical researches respecting the origin and changes of population in a particular district. Maps like those of Berghaus do indeed exhibit the area and boundaries of a nation or language at a given time; but the only ethnographical map, which can really assist the student's memory, is one which shows to the eye the origin and affinities of the different ele- ments in the population of a country. To effect this, I have not only given, if I may say so, a section of the PREFACE TO THE SECOND EDITION. ix various strata, but I have so chosen the colours, as to indicate their structure and relationship. As I believe that the Greeks and Celts like the Teutones and Cim- bri of history were scions ultimately of the same stock, I have represented them by cognate colours red and pink ; and then, taking yellow to mark the Sclavonians and Hue to indicate the Gothic tribes, the fusion of these races in the Lithuanian or Latin is shown to the eye by a stratum of green, which is a mixture of blue and yellow. The former edition of this book, though complete with reference to its immediate object, was merely a review of existing knowledge, extended by suggestions and materials for further researches. The present repu- blication endeavours to fill up the outline, which was thus presented. It will be found, therefore, that there is much more of enlargement than of alteration in the book as it now appears. Scarcely any chapter is without considerable and important additions, and I have thought it right to insert four new chapters, containing a full discussion of some subjects, which received only an inci- dental notice in the former edition. In fact, I have not intentionally omitted an examination of any important or difficult question connected with the ethnography of ancient Italy, or with the higher departments of Latin etymology and grammar 1 . With regard to the great 1 In regard to all discussions in the present Volume, which bear im- mediately on the practical study of the Latin language, I should wish this work to be considered as a sequel to the Latin Grammar and Exercises which were published a few months since. Teachers will, I hope, find that I have fully explained and justified my departure from the tra- ditionary, and, as it appears to me, erroneous method so long pursued in our classical schools. x PREFACE TO THE SECOND EDITION. philological problem, the origin of the Etruscans and the nature of their language, I think that I have so far extended and confirmed the theory, which I laid before the British Association in 1851, that it may now claim formal recognition as a discovery resting firmly on in- ductive evidence. In reprinting this volume, I have felt much distrust of my ability to do all that I wished with the book ; but I have no want of confidence in the soundness of the principles, which support it, or in the certainty of the results, to which it leads ; and I believe that, whatever may be its defects, this work will contribute, in some degree, to facilitate and promote an important branch of those studies, to which I have devoted the best years of my life. J. W. D. BURY ST. EDMUND'S, November 6, 1852. PREFACE TO THE FIRST EDITION. NO person who is conversant with the subject will ven- ture to assert that Latin scholarship is at present flourishing in England 1 . On the contrary, it must be ad- mitted that, while we have lost that practical familiarity with the Latin language, which was possessed some forty years ago by every Englishman with any pretensions to scholarship, we have not supplied the deficiency by making ourselves acquainted with the results of modern philology, so far as they have been brought to bear upon the lan- guage and literature of ancient Rome. The same impulse, which has increased and extended our knowledge of Greek, has checked and impoverished our Latinity. The dis- covery that the Greek is, after all, an easier language than the Latin, and that it may be learned without the aid of its sister idiom, while it has certainly enabled many to penetrate into the arcana of Greek criticism who must otherwise have stopt at the threshold, has at the same time prevented many from facing the difficulties which surround the less attractive literature of Rome, and, by removing one reason for learning Latin, has induced the student to overlook the other and higher considerations which must always confer upon this language its value, its importance, and its dignity. A return to the Latin scholarship of our ancestors can only be effected by a revival of certain old-fashioned methods and usages, which have been abandoned, perhaps more hastily than wisely, in favour of new habits and new 1 See the Postscript at the end of this Preface. xii PREFACE TO THE FIRST EDITION. theories. No arguments can make it fashionable for scholars to clothe their thoughts in a classic garb : example will do more than precept ; and when some English phi- lologer of sufficient authority shall acquire and exert the faculty of writing Latin with terse and simple elegance, he will not want imitators and followers. With regard, however, to our ignorance of modern Latin philology, it must be owned that our younger students have at least one excuse namely, that they have no manual of instruc- tion ; no means of learning what has been done and is still doing in the higher departments of Italian philology ; and if we may judge from the want of information on these subjects which is so frequently conspicuous in the works of our learned authors, our literary travellers, and our classical commentators, this deficiency is deeply rooted, and has been long and sensibly felt. Even those among us who have access to the stores of German literature, would seek in vain for a single book which might serve as the groundwork of their studies in this department. The most comprehensive Roman histories, and the most elabo- rate Latin grammars, do not satisfy the curiosity of the inquisitive student; and though there is already before the world a great mass of materials, these are scattered through the voluminous works of German and Italian scholars, and are, therefore, of little use to him who is not prepared to select for himself what is really valuable, and to throw aside the crude speculations and vague conjec- tures by which such researches are too often encumbered and deformed. These considerations, and the advice of some friends, who have supposed that I might not be unprepared for such an office, have induced me to undertake the work which is now presented to the English student. How far I have accomplished my design must be left to the judg- PREFACE TO THE FIRST EDITION. xiii ment of others. It has been my wish to produce, within as short a compass as possible, a complete and systematic treatise on the origin of the Romans, and the structure and affinities of their language, a work which, while it might be practically useful to the intelligent and educated traveller in Italy, no less than to the reader of Niebuhr and Arnold, might at the same time furnish a few specimens and samples of those deeper researches, the full prosecu- tion of which is reserved for a chosen few. The most cursory inspection of the table of contents will show what is the plan of the book, and what informa- tion it professes to give. Most earnestly do I hope that it may contribute in some degree to awaken among my countrymen a more thoughtful and manly spirit of Latin philology. In proportion as it effects this object, I shall feel myself excused in having thus ventured to commit to a distant press a work necessarily composed amid the dis- tractions and interruptions of a laborious and engrossing profession. J. W. D. THE SCHOOL HALL, BURY ST. EDMUND'S, 25th March, 1844. POSTSCRIPT, 1852. On the Causes and Remedies of the present neglect of Latin Scholarship in England. IN the first sentence of the preceding Preface, I have stated my belief that Latin Scholarship is not flourishing in England, and this statement was repeated in the pre- face to the Latin Grammar, which was published in January last. On each appearance of this assertion, I was obliged to defend it from direct attacks on the part of those who felt themselves aggrieved by it. My first assailant was the principal of an educational establishment connected with University College, London, who regarded himself as a champion of " crude-form" philology. My second op- ponent was the Master of an endowed Grammar School, who came forward as a vindicator of old-fashioned La- tinity. But they both agreed in the personality of their opposition to a censure of English Scholarship, which they conceived to be in some measure directed against them- selves. The former controversialist gave no indication of superior knowledge or ability, and as a clamosus mercium undique compilatarum venditator, his egotism and presump- tion would have been simply ridiculous, had not his dis- regard of those principles, which regulate the conduct of honourable men, suggested some considerations affecting himself of a graver and more painful nature. The second defender of English Latinity needs no testimony from me to his respectability and moral worth, and he is an excellent Greek scholar, if brilliant success at the University may be taken as a criterion ; but his pamphlet was chiefly re- markable as showing how unconsciously our best men can POSTSCRIPT. xv put forth and maintain obsolete and erroneous doctrines in Latin grammar and philology. Whatever other effect these discussions may have produced, they have at least failed to change my opinions respecting the Latin Scho- larship of this country. But when I adhere to and repeat those opinions, I do not wish to inquire whether any other persons are disposed to contradict or censure me ; I do not ask, with Macaulay's Horatius, "What noble Lucumo comes next To taste our Roman cheer?" Personal considerations do not enter into a general criticism which includes a whole department of classical learning. Even if I could, without presumption, enumerate those whom I consider as exceptions to the laxity of our Latin Scholarship, I should be deterred by the fear of omitting many whose attainments are unknown to me ; and I feel assured that, while there are always some who will defend the faults which they exemplify, all those, who are really good scholars, will readily admit the comparative neglect into which the study of the Latin language has fallen among us ; and with regard to those who are less conscious of it, I shall hope to point out some of the causes and remedies of our deficiency in this respect, without provoking a contest, which, like those already re- ferred to, might enable me to gain an easy triumph at the expense of some individual. Latin Scholarship is in a low state among us, because we have abandoned the old inducements to this study, without taking up the new applications which give it an increased interest and value. For the fact, it is sufficient to mention that, although our public schools impart a fa- cility in the composition of Latin verse, which is rarely attained on the continent, and though this is highly valu- able as a practical habit of skill and accuracy, examiners xvi POSTSCRIPT. at the Universities and bishops at their ordinations have publicly complained that they very rarely meet with a young man who can write tolerably good Latin prose. And among our maturer scholars, while some cannot write a page without inaccuracy, there are certainly not many whose Latin style will bear a comparison with that of Ernesti, Ruhnken, Garatoni, F. A. Wolf, and Wyttenbach. Then again, although the present generation of our scholars can point to publications of the Greek authors and lexicographers, at least equal to the best specimens of the kind which have appeared on the continent, we have produced no edition of a Latin work, which can be men- tioned in the same breath with Orelli's Horace, Lachmann's Lucretius, RitschFs Plautus, and the Varro and Festus of C. O. Miiller; still less can we claim to have done any thing for the classical study of the Roman law, which deserves to be placed beside the labours of Haubold, Dirksen, Hugo, and Savigny. There can be no doubt that the proper remedy for this comparative neglect of Latin Scholarship, is to in- crease or revive the demand for a knowledge of Latin, and to point out to amateur or dilettanti students the real interest and practical value of this branch of classical learning. This will amount to a resumption on the one hand, of "certain old-fashioned methods and usages" (above, p. ix.), and will involve, on the other hand, a proper cultivation of modern Latin philology in all its ap- plications. An increased or revived demand for Latin Scholarship will be promoted, if the Universities allow it to be seen that the rewards and honours, which they have to bestow, are at least as attainable by this means, as by an accurate and critical acquaintance with Attic Greek. At present it is well known, that, although the examinations at Oxford POSTSCRIPT. xvn and Cambridge presume an equal attention to Latin and Greek on the part of the candidates for classical honours, practically it is not expected or required that the former language should have been studied with the same minute and scrupulous regard to its texture and idioms. This is shown, in part, by the direct or presumed references to the works of those critics who have written on the Greek language, and by the absence of any similar appeal to the writings of the great Latin scholars. It is required, for example, that the competitor should be familiar with what Porson, Elmsley, and Hermann have written on the text of Euripides, but it is not implied that he must have studied the notes of Drakenborch on Livy, or the miscel- laneous observations of Gronovius. During my long resi- dence at one of the Universities, I knew more than one case in which a high place in the Tripos was perilled by an error in Greek syntax or metre, and I was informed of one instance in which the most distinguished classical honours were awarded to a youth, whose knowledge of Latin was so confused and uncertain that he had con- strued ventos as the passive participle of venio. When University students know that their examiners value and exact as scholarlike and critical an acquaintance with the best Latin, as with the best Greek authors, they will not fail to bring their industry and talents to bear on the neg- lected literature of Eome. It might be desirable that our Universities should require the use of the Latin language in all books of a strictly learned character, which are pub- lished at their expense. At any rate, great advantages would be gained if all theological works of a higher class were clothed in this classic garb. Religious newspapers and other periodicals conducted by unlearned and anony- mous writers, who are only anxious to fan the flame of one-sided prejudice, would lose much of their fuel, if b xviii POSTSCRIPT. original and well-informed divines, who are anxious to elicit the truth, which lies mid-way between the opinions of extreme parties, were content to write ad clerum in the first instance. And I should rejoice, if among the con- templated reforms of our Universities, we could revive the discipline of our divinity schools, strenuously refusing the honours of the highest faculty to all who cannot maintain a disputation in precise and accurate Latinity 1 . To increase a more general interest in the philological study of the Latin language, we must begin by engaging professed scholars in a proper regard for Roman literature. This will be best effected, if they can be induced to be- lieve that there is still the same room for the display of their abilities and learning in the revision and illustration of the Latin authors, as in their favourite field of Greek criticism. Not to speak of Cicero, many of whose works expect a competent editor acquainted with the highest philology of the day, there is ample opportunity for criti- cism of the best kind in the proper interpretation of Plautus, Lucretius, Propertius, Virgil, Livy, and Tacitus. Then again we may hope that the general ethnographer and philologer will be more and more persuaded that ancient Italy furnishes the most difficult as well as the most important subject for his speculations. If the new combinations in this work are as valid and conclusive as I believe them to be, a true explanation of even the com- 1 As undergraduates were expected to hold Latin disputations in the schools, the Universities must have assumed that they would come up perfectly able to carry on a conversation in Latin. The Grammar schools were instituted expressly for this purpose (see New Crat. 83), and the old statutes of Bury School direct that "the scholars shall speak con- tinually Latin as well without the school as within." The presumption that Latin will be sufficiently learned before the commencement of a college career is farther indicated by the fact, that neither of our great Universities has a Professor of Humanity or Latin. POSTSCRIPT. xix monest and most striking peculiarities of Latin word-forms was hitherto undiscovered. In those great seats of learn- ing, where the luxury of study may be enjoyed for its own sake, it is to be regretted that we have no lectures on the Romance languages, which are so deserving of the attention of all those whose ancestors, in part or wholly, adopted them, and which lend a new interest to the study of the Latin language, their immediate parent. Above all, the cultivation of Eoman literature will never be restored to its proper place in the estimation of learned English- men, until we have revived the classical spirit, which for- merly prevailed in this country, and which, on the continent, still directs and influences the study of the civil law. On this subject, I shall take the liberty of quoting the words of a writer, with whom I do not often agree, and whose Latin scholarship is by no means an exception to the general rule of laxity and incompleteness, but who has enjoyed, as I have, the advantage of a regular and pro- longed course of legal study ; and I am the more induced to quote his words, because, as he has been a public teacher both of Latin and of law, his admissions may be received as partly affecting himself : " That in this country, where we profess to cultivate ancient learning, we should so long have neglected the study of the Roman law, the best and only original part of their literature, and should have gone on in the dark, admiring and thinking that we understood the writings of Cicero, our model of Latinity, is a proof, the strongest possible, of the degradation into which classical studies have sunk in our higher places of education. In one University, lectures on the civil law have ceased to be given, though there is still a Professor ; and in the other (Cambridge), though lectures are given, and degrees are taken in civil law, it is well known in how little estimation both the subject itself and the de- 62 xx POSTSCRIPT. grees are held by those who follow what may be called the regular studies of the University. Instead of the lectures on civil law being considered as auxiliary to and part of the Latin studies of the University, which they ought to be and might be, an attendance on the course of civil law, and a residence in the Hall where the lectures are delivered, are generally viewed rather as a convenient means of obtaining a degree. Such being the case, it would not be an easy matter for the Professor to restore the study of the civil law to its proper dignity, and to make it an integral part of the University course 1 ." It cannot be denied that there is some general truth in these remarks; but the writer overestimates the difficulty of remedying the defects of which he complains. Whenever the subject of civil law shall be taken up by some genuine Latin scholar fully impressed with its dignity and impor- tance, he will form a school for himself; and to say nothing of my own University, I may be permitted to re- mark, that the fabric of juristic learning, which an eminent civilian at Oxford has built upon a solid foundation of classical scholarship, not unconnected with a careful study of Niebuhr, may lead us to believe that there are already some persons in England who can bring to the study of the Roman law the thoughtful erudition of Gibbon and the philological acuteness of Savigny. On the whole, though I feel myself obliged on this occasion to repeat the preface to Varronianus, as it origi- nally stood, I venture to indulge in the hope that, if I live long enough to write again on this subject, I shall be able to speak in more flattering terms of the Latin Scholarship of England. Central Society of Education. Third Publication, p. 220. CONTENTS. CHAPTER I. THE OLD ITALIAN TRIBES CONSIDERED AS RELATED TO EACH OTHER. SECT. PAGB 1 Elements of the population of Rome .... 1 2 The LATINS a composite tribe 3 3 The Oscans, &c. 3 4 Alba and Lavinium ...*.. 5 5 Trojan colony in Latium ...... 6 6 The SABINES how related to the Umbrians and Oscans . 7 7 The Umbrians their ancient greatness .... 8 8 Reduced to insignificance by successive contacts with the Tyrrheno-Pelasgians and Etruscans ... 9 9 The PELASGIANS the differences of their position in Italy and Greece respectively . . . . . .10 10 They preserve their national integrity in Etruria . . 11 11 Meaning and extent of the name "Tyrrhenian" . . 11 12 The ETRUSCANS the author's theory respecting their origin 14 13 The names ETRUSCUS and RASENA cannot be brought to an agreement with TYRSENUS . . . . . 16 14 It is explicitly stated by ancient writers that the Etruscans came from Rsetia ....... 17 15 This view of the case is after all the most reasonable . 18 16 It is confirmed by all available evidence, and especially by the contrast between the town and country languages of Etruria ........ 19 17 Farther inferences derivable from (a) the traditionary his- tory of the Luceres . . . . . .21 18 (6) Fragmentary records of the early constitution of Rome . 23 19 (c) Etymology of some mythical proper names . . 24 20 General conclusion as to the mutual relations of the old Italian tribes 26 CHAPTER II. THE FOREIGN AFFINITIES OF THE ANCIENT ITALIANS, 1 Etymology of the word IleXatryos .... 28 2 How the Pelasgians came into Europe . 30 63 xxii CONTENTS. SECT. PAGE 3 Inferences derivable from the contrast of Pelasgian and Hel- lenic architecture . . . . . . . 31 4 Supported by deductions from the contrasted mythology of the two races ....... 36 6 Thracians, Getse, and Scythians ..... 39 6 Scythians and Medes ...... 40 7 Iranian origin of the Sarmatians, Scythians, and Getse, may be shown (1) generally, and (2) by an examination of the remains of the Scythian language .... 40 8 Mode of discriminating the ethnical elements in this chain of nations . . . . . . . . 42 9 Peculiarities of the Scythian language suggested by Aristo- phanes ........ 44 10 Names of the Scythian rivers derived and explained . ' : 45 11 Names of the Scythian divinities . . . 48 12 Other Scythian words explained ..... 52 13 Successive peopling of Asia and Europe : fate of the Mon- golian race ........ 55 14 The Pelasgians were of Sclavonian origin . . .'.> 58 15 Foreign affinities of the Umbrians, &c. .... 59 16 Reasons for believing that they were the same race as the Lithuanians ...... 59 17 Farther confirmation from etymology . . . -'-.. 61 18 Celtic tribes intermixed with the Sclavonians and Lithuanians in Italy and elsewhere .... - ^ 260 6 The Greek letters used by the Romans . . . . 267 7 The numeral signs ..... ...y ) . 272 CONTENTS. xxv CHAPTER VIII. THE LATIN CASE-SYSTEM. SECT. PAGE 1 Fulness and deficiencies of the Latin case-system . . 274 2 General scheme of the case-endings .... 275 3 Differences of crude-form ...... 276 4 Hypothetical forms of the nominative and accusative plural 278 5 Existing forms the genitive ..... 280 6 The dative and locative 282 7 The accusative singular ...... 283 8 The ablative 284 9 The neuter forms 284 10 The vocative 286 11 Adverbs considered as cases of nouns .... 287 12 Adverbial expression for the day of the month . . 292 CHAPTER IX. DECLENSIONS OF THE LATIN NOUN. 1 The usual arrangement is erroneous .... 293 2 General rules for the classification of Latin nouns . ; 294 3 First or -a Declension ...... 295 4 Second or -o Declension ...... 296 5 Third Declension or consonantal nouns .... 296 6 A. First class or purely consonantal nouns . . . 297 7 B. Second class or semi-consonantal nouns . . 301 CHAPTER X. PRONOUNS AND PRONOMINAL WORDS. 1 General definitions 307 2 Personal Pronouns ....... 307 3 Indicative Pronouns . . . . . . 310 4 Distinctive Pronouns . . . . . . .315 5 Relative, interrogative, and indefinite Pronouns . . 318 6 Numerals and Degrees of Comparison .... 327 7 Prepositions ........ 329 8 Negative particles .. . . . . . . 337 xxvi CONTENTS. CHAPTER XI. THE THEORY OF THE LATIN VERB. SECT. PAGE 1 The Latin verb generally defective .... 341 2 The personal inflexions their consistent anomalies . .341 3 Doctrine of the Latin tenses ..... 343 4 The substantive verbs . . . . . y 345 6 Paucity of organic formations in the regular Latin verb . 350 6 General scheme of tenses in the Latin verb . . , , . 351 7 Verbs which may be regarded as parathetic compounds . 352 8 Tenses of the vowel-verbs which are combinations of the same kind ........ 353 9 Organic derivation of the tenses in the consonant verb 355 10 Auxiliary tenses of the passive voice ... * 355 11 The modal distinctions their syntax , . . 356 12 Forms of the infinitive and participle how connected in derivation and meaning ...... 359 13 The gerundium and gerundivum shown to be active and present 361 14 The participle in -tdrus ...... 365 15 The Perfect Subjunctive ... ... 365 16 The past tense of the infinitive active . . . . 369 CHAPTER XII. THE LATIN CONJUGATIONS. 1 The conjugations are regulated by the same principles as the declensions . . . . . . . 372 2 The first or -a conjugation .... 1,.,^^ . 373 3 The second or -e conjugation ..... 377 4 The third or -i conjugation ...... 382 5 The fourth or consonant conjugation. A. Mute verbs . 384 6 B. Liquid verbs ....... 388 7 C. Semi-consonantal verbs ..... 390 8 Irregular verbs. A. Additions to the present tense . . 391 9 B. Abbreviated forms . . . . . 397 10 Defective verbs 399 CHAPTER XIII. DERIVATION AND COMPOSITION. 1 A. Derivation* General principles . . '-".fl . 400 2 Derivation is merely extended or ulterior inflexion . . 401 CONTENTS. xxvii SECT. PAGH 3 I. Derived nouns ....... 402 4 (a) Forms with the first Pronominal Element only . . 402 5 (b) Forms with the second Pronominal Element only . 403 6 (c) Forms with the third Pronominal Element only . . 405 7 (a) Terminations compounded of the first and other Prono- minal Elements ....... 405 8 (/3) Terminations compounded of the second and other Pronominal Elements ...... 406 9 (y) The third Pronominal Element, compounded with others and reduplicated . . . . . 417 10 II. Derived verbs .419 11 B. Composition. Discrimination of compound words . 424 12 Classification of Latin compounds .... 426 CHAPTER XIV. CONSTITUTION AND PATHOLOGY OF THE LATIN LANGUAGE. 1 Genius of the Latin Language ..... 432 2 Abbreviations observable in the written forms . . . 433 3 Ancient testimonies to the difference between the spoken and the written language ...... 437 4 The poetry of the Augustan age does not represent the ge- nuine Latin pronunciation ..... 439 5 Which is rather to be derived from an examination of the comic metres ....... 440 6 The French language is the best modern representative of the spoken Latin 444 7 The modern Italian not equally so ; and why . . 447 8 Different dialects of the French language . . . 448 9 But all these dialects were closely related to the Latin 451 10 Leading distinctions between the Roman and Romance idioms 453 11 Importance and value of the Latin Language . . 458 ERRATA. Page 25, line 25, for suiters read suitors. 75, 34, for granst read graust. 364, 7, Add " That these attributive usages really correspond to active infi- nitives even in those cases, in which the gerundive might be referred to a passive verb, as in: vir minime contemnendus, &c., appears from Greek phrases like : ov TTO.VV juotpas evdaifJLOviffai TT/JW'TIJS (Soph. (Ed. Col. 142)." 382, penult, for Metium read Mettum. Those who look to such minutiffl will observe an inconsistency in the spelling of verbs in -ise or -ize ; I write them uniformly with s ; the printer seems to prefer z, and I have not always insisted on my own orthography. VARRONIANUS, CHAPTER I. THE OLD ITALIAN TRIBES CONSIDERED AS RELATED TO EACH OTHER. 1. Elements of the population of Home. 2. The LATINS a composite tribe. 3. The Oscans, &c. 4. Alba and Lavinium. 5. Trojan colony in Latium. 6. The SABINES how related to the Umbrians and Oscans. 7. The Um- brians their ancient greatness. 8. Reduced to insignificance by successive contacts with the Tyrrheno-Pelasgians and Etruscans. 9. The PELASGIANS the differences of their position in Italy and Greece respectively. 10. They preserve their national integrity in Etruria. 11. Meaning and ethnical extent of the name "Tyrrhenian." 12. The ETRUSCANS the author's theory respecting their origin. 13. The names Etruscus and Rasena cannot be brought to an agreement with Tyrsenus. 14. It is explicitly stated by ancient writers that the Etruscans came from Reetia. 1 5. This view of the case is after all the most reasonable. 16. It is confirmed by all available evidence, and espe- cially by the contrast between the town and country language of ancient Etruria. 17. Further inferences derivable from (a) the traditionary history of the Luceres. 18. (b) Fragmentary records of the early constitution of Rome. 19. (c) Etymology of some mythical proper names. 20. General conclusion as to the mutual relations of the old Italian tribes. $ 1. Elements of the population of Rome. HHHE sum of all that is known of the earliest history of Rome JL is comprised in the following enumeration of particulars. A tribe of Latin origin, more or less connected with Alba, settled on the Palatine hill, and in the process of time united itself, by the right of intermarriage and other ties, with a band of Sabine warriors, who had taken up their abode on the Quirinal and Capitoline hills. These two towns admitted into fellowship with themselves a third community, established on the Cselian and Esquiline hills, which seems to have consisted of Pelasgians, either from the Solonian plain lying between Rome and Lavi- nium, or from the opposite side of the river near Caere ; and the whole body became one city, governed by a king, or magister populi, and a senate ; the latter being the representatives of the 1 2 THE OLD ITALIAN TRIBES [On. I. three original elements of the state, the Latin or Oscan Ramnes, the Sabine Titienses or Quirites, and the Pelasgian Luceres. It appears, moreover, that the Etruscans, on the other side of the Tiber, eventually influenced the destinies of Rome in no slight degree, and the last three kings mentioned in the legendary tra- ditions were of Etruscan origin. In other words, Rome was, during the period referred to by their reigns, subjected to a powerful Etruscan dynasty, from the tyranny of which it had, on two occasions, the good fortune to escape. What Servius planned was for the most part carried into effect by the consular constitu- tion, which followed the expulsion of the last Tarquinius. As these facts are established by satisfactory evidence, and as we have nothing else on which we can depend with certainty, it follows that in order to investigate the ethnical affinities of the Roman people, and the origin and growth of their language, we must in the first instance inquire who were the Latins, the Sa- bines, the Pelasgians, and the Etruscans, and what were their relations one with another. After this we shall be able with greater accuracy to examine their respective connexions with the several elements in the original population of Europe. The general result will be this : that the Septimontium, or seven Hills of Rome, contained a miniature representation of the ethnography of the whole Peninsula. Leaving out of the ques- tion the Celtic substratum, which cannot be ascertained, but which was probably most pure in the mountaineers of the Apennines, the original population of Italy from the Po to the straits of Rhegium was, like that of ancient Greece, Pelasgo-Sclavonian. This population remained unadulterated up to the dawn of ancient history in the central plains to the West namely, in Etruria and Latium, but in the rest of Italy it was superseded or ab- sorbed or qualified in different degrees of fusion by a population of Gothic or Low- German origin, which, although undoubtedly of later introduction in the Peninsula, was so mixed up with the Celtic or primary tribes that it claimed to be aboriginal. When this Low-German race remained tolerably pure, or at least only infected with Celtic ingredients, it bore the names of Umbrians or Ombricans in the North, and of Opicans or Oscans in the South. When it was intermixed with Sclavonic elements to about the same extent as the Lithuanians or Old Prussians in the North of Europe, this Low- German population became 1.] AS RELATED TO EACH OTHER. 3 known as Latins and Sabines. And the Etruscans or Eascna were a later and uninfected importation of Low Germans fresh from the North, who conquered and were partly absorbed into the pure Tyrrhenians, or Pelasgo-Sclavonians to the right of the Tiber. J 2. The LATINS a composite tribe. The investigations of Niebuhr and others have made it sufficiently certain that the Pelasgians formed a very important element in the population of ancient Latium. This appears not merely from the primitive traditions, but also, and more strongly, from the mythology, language, and architecture of the country. It has likewise been proved that this Pelasgian population was at an early period partially conquered by a tribe of mountaineers, who are called Oscans, and who descended on Latium from the basins of the Nar and the Velinus. The influence of these foreign invaders was most sensibly and durably felt in the language of the country ; which in its earliest form presents phenomena not unlike those which have marked the idiom spoken in this island since the Norman conquest. The words relating to husbandry and peaceful life are Pelasgian, and the terms of war and the chase are Oscan 1 . As it is this foreign element which forms the distinction between the Latins and the Pelasgians, let us in the first place inquire into the origin and affinities of these Oscan conquerors, in order that we may more easily disentangle the complexities of the subject. 3. The Oscans, fyc. The Oscans were known at different times and in different places under the various names of Opicans, Opscans, Ausonians, 1 Niebuhr, H. R. I. p. 82. Miiller, Etrusker, I. p. 17. This observa- tion must not be pressed too far; for it does not in fact amount to more than prima facie evidence. The Opican or Oscan language belongs to the Tndo-German-ic family no less than the Pelasgian ; the latter, however, was one ingredient in the language of ancient Greece, and it does not appear that any Hellenic tribes were connected with the Oscans; con- sequently it is fair to say that, as one element in the Latin language resembles the Greek, while the other does not, the Grsecising element is Pelasgian. 12 4 THE OLD ITALIAN TRIBES [On. I. and Auruncans. The primary denomination was Op-icus or Oqu-icus, derived from Ops or Opis = Oqu-is, the Italian name of the goddess Earth ; and these people were therefore, in accordance with their name, the Autochthones, or aboriginal inhabitants of the district where they are first found. The other denominations are derived from the same word, Op-s= Oqu-is, by the addition of the endings -si-cus, -sunus, and -sun-icus. The guttural is assimilated in Oscus, the labial is absorbed in AUGWV, and the s has become r, according to the regular pro- cess, in Auruncus 1 . 1 See Niebuhr, I. 69, note. Buttmann, Lexilogus, I. p. 68, note 1. (p. 154, Fishlake). The investigation of these names leads to a variety of important and interesting results. It has been shown elsewhere that in the oldest languages of the Indo- Germanic family the names of the cow or ox and the earth are commutable (N. Crat. 470). Not to refer to the obvious but not so certain analogy between \TTLS, the ox-god, and the anLrj yaTa, it can be shown to demonstration that the steer or ox, which was to the last the symbol of the old Italians, as appears by their coins, entered into the meaning of their two national designations, Italus and Opicus. With regard to the former it is well known, that italos, or itulus, or with the digamma vitulus, meant an ox or steer (Niebuhr, I. 18 sqq.) 3 and Vitellium appears on coins as a synonym for Italia. This takes us at once to the Gothic vithrus, O. N. vedr, O. S. withar, Anglo-S. vether, O. H. G. vidar, N. H. G. widder (properly the castrated animal), English wether ; and as these are referred to sheep rather than oxen, we must conclude that the name is an epithet which is applicable to either animal. With regard to the other root, qv in ^Equus carries us back to the principle of combined but divergent articulations, to which I first called attention (N. Crat. 110), and on which the late Mr. Garnett wrote some valuable papers (Philol. Soc. II. p. 233, 257 al.), and we may infer that the roots ap- or op- present a labial only instead of an original com- bination of labial and guttural, while we find the opposite divergence in the guttural forms vac-ca, veh-o, Sanscr. vaha, Gr. o^o?, e\;a>, Goth, auh-sa, O. N. ox, Anglo-S. oxa, O. H. G. ohso, N. H. G. ochs, Engl. ox. The labial form is sometimes strengthened by an inserted anusvdra, or homo- geneous liquid ; thus by the side of on-upa and op-s we have 6-fj.-z>, which refers to the dispossession of the Celtic inhabitants of Umbria and Etruria, as belonging to the same traditions which led Antiochus to write that the Sicilians were driven over into Sicily by the Opicans (H. R. I. p. 82) : for Antiochus is speaking exclusively of what took place in the southern extremity of Italy, and the Pelasgians and Ombrici mentioned by Philistus were the Tyrrhenians and Umbrians of the north. 6 THE OLD ITALIAN TRIBES [On. I. querors living in the midst of the Pelasgians whom they had subdued, the other a Pelasgian nation not yet conquered by the invaders. These two nations formed at first two distinct confe- deracies : of the former Alba was the head, while the place of congress for the latter was Lavinium. At the latter place, the Penates, or old Pelasgian Cabeiri, were worshipped ; and even after the Pelasgian league was broken up by the power of Alba, and when Alba became the capital of the united nation of the Latins and sent a colony to Lavinium, the religious sanctity of the place was still maintained, the Penates were still wor- shipped there, and deputies still met in the temple of Yenus. The influence of Alba was, however, so great, that even after its fall, when the Pelasgian Latins partially recovered their inde- pendence, there remained a large admixture of foreign elements in the whole population of Latium, and that which was purely Pelasgian in their character and institutions became gradually less and less perceptible, till nothing remained on the south of the Tiber which could claim exemption from the predominating influence of the Oscans. That the name Lavinium is only a dialectical variety of Latinium has long been admitted. The original form of the name Latinus, which afterwards furnished a denomination for the language of the civilised world, was Latvinus; and while the Pelasgian Latins preserved the labial only, the mixed people retained only the dental 1 . 5. Trojan Colony in Latium. The tradition speaks of the Pelasgian Latins as a colony of Trojans who settled on the coast under ^Eneas, the son of Anchises. Without entering at length into an examination of this poetical legend, it may be mentioned here that the names jtEneas and Anchises refer, wherever they are found, to the Pelasgian or Cabeiric worship of water in general, and of the flowing stream in particular, and therefore indicate the presence of a Pelasgian population. We have other reasons for inferring the existence of Pelasgians on the coast of Asia Minor, in Thes- 1 The same has been the case in the Pelasgian forms, liber, libra, bis, ruber, &c., compared with their Hellenic equivalents, e-Xeu0epos, Xtrpa, dis, e-pv6pos, &C. 5.] AS RELATED TO EACH OTHER. 7 saly, Boeotia, Arcadia, and the west of Italy. It is therefore quite natural that we should find in these localities the name of jiEneas as that of a river or river-god. The word itself denotes "the ever-flowing" (atveias or auveasi aevvaos, del or aiel vecov, cf. dowlas, dfj.vvcoi>, N. Crat. 262), and in accordance with this we have the rivers Anias, JEnios, j^Enus, and Anio. In the same way, because the stream is the child of its fountain, Anchises the father of ^Eneas, whose mother is Aphrodite, the goddess of the sea-foam, denotes the outpouring of water (cryx/crtys, dyxycri's, ay%6) and sanctus. According to this, the name Sabini is nearly equivalent to Sacran-L The tables also mention the picus Martins of the Sabines, from which the Piceni derived their name (piquier Martier, V. 6. 9, 14) ; comp. Strabo, V. p. 240. 5 Niebuhr, I. note 430. 7.] AS RELATED TO EACH OTHER. 9 floods which had destroyed many an earlier race of men 1 . This is about as valuable as other Greek etymologies. The ethno- graphical import of the name will be examined in the following chapter, and we certainly do not need a forced etymology to prove that the Umbrians must have been among the earliest inhabitants of Italy. Cato said that their city Ameria was founded 381 years before Rome 2 . All that we read about them implies that they were a great and an ancient nation 3 . There are distinct traditions to prove that the country, after- wards called Etruria, was originally in the occupation of the Umbrians. The name of the primitive occupants of that country was preserved by the Tuscan river Umbro, and the tract of land through which it flowed into the sea was to the last called Umbria*. It is expressly stated that Cortona was once Um- brian 5 ; and Gamers, the ancient name of Clusium 6 , points at once to the Camertes, a great Urnbriam tribe *. It is certain also that the Umbrians occupied Picenum, till they were expelled from that region by their brethren the Sabines 8 . 8. Reduced to insignificance by successive contacts with the Tyrrheno-Pelasgians and Etruscans. Since history, then, exhibits this once great nation expelled from the best part of its original possessions, driven beyond the Apennines, deprived of all natural barriers to the north, and reduced to insignificance, we are led at once to inquire into the cause of this phenomenon. Livy speaks of the Umbrians as dependent allies of the Tuscans 9 ; and Strabo tells us that the Etruscans and Umbrians maintained a stubborn contest for the possession of the district between the Apennines and the mouth of the Po 10 . The people which thus ruled them or strove with them in the latter period of their history, when they were 1 See Plin. //. N. III. 19: "Umbrorum gens antiquissima Italise existimatur, ut quos Ombrios a Grsecis putent dictos, quod inundatione terrarum imbribus superfuissent." 2 Pliny, III. 14, 19. 3 Florus, I. 17, Dionys. I. 19. 4 Pliny, III. 5. (8). 5 Dionys. I. 20. 6 Liv. X. 25. 7 Liv. IX. 36. s Pliny, III. 13, 14. 9 In Books IX. and X. 10 P. 216. 10 THE OLD ITALIAN TRIBES [On. I. living within the circumscribed limits of their ultimate posses- sions, was that which deprived them of a national existence within the fairest portion of their originally wide domains. It will be shown that the national integrity of the Umbrians was impaired by their successive contacts with the Tyrrheno- Pelasgians, and the Etruscans properly so called ; and it will be convenient to consider, as separate questions, these qualifying elements in the population of ancient Umbria. 9. The PELASGIANS the differences of their position in Italy and Greece respectively. Without stopping to inquire at present who the Pelasgians were out of Italy, let us take them up where they first make their appearance at the mouth of the Po. We find that their area commences with this district, and that having crossed the Apennines, they wrested from the Umbrians the great city Gamers, from whence they carried on war all around. Continu- ally pressing towards the south, and as they advanced, conquering the indigenous tribes, or driving them up into the highlands, they eventually made themselves masters of all the level plains and of the coasts. Though afterwards, as we have seen, invaded in their turn, and in part conquered by the Oscan aborigines, they were for a long time in possession of Latium ; and, under the widely diffused name of GEnotrians, they held all the south of Italy, till they were conquered or dispossessed by the spread of the great Sabellian race. To these Pelasgians were due the most important elements in the ancient civilisation of Italy. It was not their destiny to be exposed throughout their settlements, like their brethren in Greece, to the overruling influence of ruder and more warlike tribes. This was to a certain extent the case in the south ; where they were not only overborne by the power of their Sabellian conquerors, but also Hellenised by the Greek colonies which were at an early period established among them. But in Etruria and Latium the Pelasgian nationality was never extinguished : even among the Latins it survived the severest shocks of Oscan invasion. In Etruria it remained to the end the one prevailing characteristic of the people ; and Rome herself, though she owed her military greatness to the Sabellian ingredient in her compo- sition, was, to the days of her decline, Pelasgian in all the essen- tials of her language, her religion, and her law. 10.] AS RELATED TO EACH OTHER. 11 10. Preserve their national integrity in Etruria. It is easy to see why the Pelasgians retained their national integrity on the north-western coast so much more perfectly than in the south and east. It was because they entered Etruria in a body, and established there the bulk of their nation. All their other settlements were of the nature of colonies ; and the density of the population, and its proportion to the number of the con- quered mingled with it, varied, of course inversely, with the dis- tance from the main body of the people. In Etruria the Pelas- gians were most thickly settled, and next to Etruria in Latium. Consequently, while the Etruscans retained their conquest, and compelled the Sabines, the most vigorous of the dispossessed Umbrians, to direct their energies southwards, and while the Latins were only partially reconquered by the aboriginal tribes, the Pelasgians of the south resigned their national existence, and were merged in the concourse of Sabeliian conquerors and Greek colonists. 11. Meaning and extent of the name " TYRRHENIAN." From the time of Herodotus 1 there has been no doubt that the Pelasgians in Greece and Italy were the same race, and that 1 I. 57. The following is the substance of what Herodotus has told us respecting the Tyrrhenians and Pelasgians; and his information, though much compressed, is still very valuable. He seems tacitly to draw a distinction between the Pelasgians and the Tyrrhenians, whom he really identifies with one another. With regard to the latter he relates the Lydian story (I. 94 : i> cdveav /3ap/3a/ja>p a-vxy&v immediately follows. I cannot doubt that we ought to read av^rai es s, rav TleXaa-ymv /udXtoTa Trpoo-KfxcoprjKorav avra KOI a\\a>v edvtwv /3ap- xvav. The epithet TroXXwv has crept into the text from a mar- ginal explanation of o-u^wai/, and T&V eOveav TrdXX&v has consequently taken the place of the abbreviation r. 1 Proceed, of the Phil. Soc. I. p. 161, sqq. 2 Apud Dion. Hal. I. 25: lva%e yevvarop TTOL Kprjvatv Trarpos 'flKeai/oC, /ne'ya Trpfvpevav *Apyovs re yvais, "Upas re ndyois KOI Tvpvrjvo'io-i HeXao-yois. See' also Schol ApolL Rh. I. 580. 3 Ueber die Tyrrhenischen Pelasger in Etrurien. Leipsig, 1842. Dr. Lepsius maintains the identity of the Tyrrheno-Pelasgians with the Etruscans ; and in the former edition I accepted his view, which was true as far as it went : but subsequent research has convinced me that we must recognise aRsetian element superinduced on the previously exist- ing combination of Tyrrheno-Pelasgian and Umbrian ingredients. We are indebted to this scholar for some of the most important contributions which Italian philology has ever received. In histreatise on the Eugubine Tables, which he published in the year 1833, as an exercise for his degree, he evinced an extent of knowledge, an accuracy of scholarship, and a maturity of judgment, such as we rarely meet with in so young a man. His collection of Umbrian and Oscan inscriptions (Lipsise, 1841) has sup- plied the greatest want felt by those who are interested in the old languages of Italy ; and some fruitful results have proceeded from those $11.] AS RELATED TO EACH OTHER. 13 Tu perjvos signifies " tower-builder," and that this term has been properly explained even by Dionysius 1 , as referring to the TV pa eiV T7)$ OIKOVVTCOV KO.T(TKVa(TaVTO. TV per cis yap KOI Trapa Tvpprjvols ai evrei^ioi KOI (TTcyaval oiKrjfreis 6vop.d- govrat, ao-irep Trap "E\\r)o-iv. Tzetzes, ad Lycophr. 717: Tvpo~ts TO TeT^oy, on Tvpcnjvol Trp&Tov ecpevpov TTJV ret^oTrottai/. Comp. Etyvn. M, S. V. Tvpavvos. 2 Ol. II. 70: Ti\av Atos o&ov Trapa Kpovov Ti>p 2 ; 1 Salmasius de Hellenistica, p. 342- 2 De Dialect, Macedon. p. 9. 1.] THE ANCIENT ITALIANS. 29 Hermann finds the root in TreXayos, from TreXa^V; Wachs- muth 2 and Miiller 3 , considering TreXapyos to be the original form of the word, give as its etymology 7reXo>, "to till," and Hypos, " the field," looking upon the nation as originally de- voted to husbandry. The most common derivation is that which writes TlGXapyoi, and interprets it " the storks," either from the wandering habits of this race 4 , or from their linen dress 5 , or from their barbarous speech 6 . Every one of these etymologies admits of an easy confutation. The best answer to them all is to point out a better analysis of the word. Buttmann? suggested long ago that the last two syllables were an ethnical designation, connected with the name Asca-nius, common in Phrygia, Lydia, and Bithynia, and with the name of Asia itself. He also cor- rectly pointed to the relationship between Ashkenaz, the son of Gomer, and Javan, the biblical progenitor of the lonians ('Ia'Foi/es) (Gen. x. 3). Now the first syllable of the word Pel- asgus is clearly the same as that of Pel-ops. There are two Niobes in Greek mythology, daughters, the one of Phoroneus, the other of Tantalus the latter is the sister of Pelops, the former the mother of Pelasgus. The syllable ITeX- stands in the same relation to /ueX- that 7rea does to /mera. The original form of the root signifying "blackness" was /c^eX- 8 ; but the labial generally predominated over the guttural element. Of the labial forms, that with the tenuis more usually came to signify " livid " than " black ;" as we see in the words TreXios, TreXi&w, &c. Apollodorus expressly says 9 that IleXic^ was so called be- cause his face was rendered livid (TreXtos) by a kick from a horse; and it is obvious that TleX-ov//, which signifies "dark- 1 Opusc. II. p. 174 : " TreXayos enim, a verbo 7reXaV/ dictum, ut ab Latinis Venilia, mare notat : a qua origine etiam TreXacryot, advence." 2 Hellenische Altertliumsk. I. p. 29, Trans, p. 39. He also, half in jest, refers to 7rXae/, " to lead astray," p. 36. 3 "Von 7reXo> (TroXts-, 7roXe'o>, der Sparte lleXcop, und HeXw/ata, das Fest der BewoJinung) und apyos." Orchom. p. 125. 4 Strabo, V. p. 221 ; VIII. p. 397. 6 Bekker, Anecd. p. 229 : bta ras J This has been done by Metzger, in Thiersch's tract, iiber das Erechtheum. 3 Olymp. XIII. 21, sqq. : airav 6* evpovros epyov' Tal Aiuvixrov TTodev fe(pavfv TIS yap Imreiois fv evTccrcriv perpa T) Bftav vadiviv olavav /SacrtXea diftvpov f7T0r}Kf ; That the aeVos-, or derw/za, meant the tympanum, or gable, and not any figures within or upon it, has been fully shown by Brondsted, Voyages et Recherches en Grece,II. p. 154 ; and by Welcker, Alte Denkmaler, I. p. 3, sqq. The pediment was originally open ; the deep relief, or rather complete figures, which appear in it, indicate the original practice, when it might be said in the language of Euripides (Fr. Hypsip.} : Idov Trpos aidep* e'a/it\Ac3z/rai Kopai ypanrovs [eV aiejToto-t 7rpoo-[B\c7rfiv rinrovs. And the ground was subsequently painted blue to recal the darkness of the space under the roof. 4 The commercial dealings were a fact ; the mythology of Bellerophon was a poetical record of it. 3 34 THE FOREIGN AFFINITIES OF [On. II. Dorian fagade existed in Asia Minor long before the Dorian and Ionian colonies were established there, it is a fair conclusion that the Dorian and Ionian architecture, like the distinctions of dialect, was due to the reaction of the Dorian and Ionian colonies on the mother-land. And thus we see that all the architecture of Greece, the more refined porch as well as the ruder masses of Cyclopian masonry, was imported from the sunny land to which we trace the name of the Pelasgians. We may go a step farther, and say that the more recent architecture of Asia Minor, which was afterwards naturalized in Greece, was due to the Semitic tribes which extended inland from Lydia to Assyria and Egypt, whereas the Cyclopian architecture was strictly Indo- Germanic. The primary distinction between the Pelasgo- Achaean and the Doro-Ionian architecture consisted in the materials which they respectively adopted, the former being the adaptation of huge masses of uncemented stone, the latter the result of the best arrangement of beams and joists. The materials of the Cyclopian walls require no comment, but a few remarks may be necessary to show that the Doro-Ionian architecture originated in wood- carpentry. The simplest form of this architecture is the apteral temple in antis. This has no column or portico, the porch being supported by Trapaara^ or antce , i. e. projections of the side walls 1 . We then come to the prostyle, with a vestibule sup- ported by columns beyond the antce ; then to the amphiprostyle, with such a termination at each end ; and finally to the peripteral temple, surrounded by columns, like the Parthenon. The com- plete form is the best exemplification of the tectonics or carpentry in which the architecture originated. If we compare the Doric building, as restored from the inverted column on the gate of the lions, with the remains of Lycian architecture 2 , we shall see that the foundation consisted of trunks of trees, laid level and crossed at right angles by the trunks of other trees. On these last, as we see in the gate of the lions, the plinth of the column rested, and on this the torus. The shaft of the column was the trunk of a tree, and its capital originally nothing more than a plinth. On the top of the column was placed the architrave 1 On the sense of Trapaa-ras, or Trao-rds, I may refer to my note on the Antigone, 1173, p. 225, where I have collected all the authorities. 2 See Thiersch, iiber das Erechtheum, p, 149, sqq. $3.] THE ANCIENT ITALIANS. 35 or main beam of the entablature, and on this rested the frieze with holes immediately above the columns for the reception of the upper joists of the building. When these joists were in- serted, their ends, ornamented by channels cut in the wood, were termed triglyphs, and the spaces between the triglyphs, which were flat wood, and upon which it was customary to nail up spoils taken in the chase, garlands, and sculptures, were called metopes, or intervals between the holes 1 . The frieze was sur- mounted by the cornice, which originated in transverse beams supporting the a/uuAA^riJ/oe? of the sloping roof, and the fagade was finished off by the pediment, tympanum, or aeVwjua, which was originally an open gable formed by the sloping rafters. Now every detail in this form of edifice points to wood-work or carpentry, which always constituted the material of pure Semitic architecture. The complete details which have been preserved of the temple of Solomon, which was a masterpiece of Phce- 1 It has been the opinion of many learned architects that the metopes, or spaces between the beam-ends, were originally hollow. This is an opinion contrary to the evidences furnished by the Gr,eek language and by the Greek authors, and is plainly overthrown by the Mycenaean monu- ment, which shows us that the frieze was originally a solid piece with holes for the beam-ends. The word OTTJ) means " an opening or hole," i. e. the bed of a beam ; hence the Roman architects called the triglyphs cava columbaria, or "pigeon-holes." The word fj-eronrj must signify "a space between OTTO/," as TO p.raixiJ.tov means " a space between two armies ;" consequently the metope could not have been itself a cavity. Besides, spoils taken in the chase, garlands, and sculptures, were nailed up to the frieze, which must therefore have been solid. The triglyphs were the ornamented ends of the beams, cut short on a line with the frieze : but these beams could not have projected in the same plane in the sides and at the ends of the building. Supposing then that those which ran the whole length of the building terminated in the frieze of the portico, the cross-beams must have rested upon them and served as supports to the end of the roof. Consequently the frieze on the sides of the building must either have had hollow spaces instead of beams, which was of course the original form, or they were filled by imaginary beam-ends, i. e. mere triglyphs. When the fagade of a temple was imitated on the Greek stage, it seems that the OTTCU or beds of the beams were left open, i. e. there were large holes through which a man might crawl. This enables us to understand such passages as the following: Euripid. Iph. T. 113: Spa 8e y 10-0) rpi-y\v(p(i)V OTTOI KCVOV Se'/xas ^Qelvai. Aristoph. Vesp. 126: 6 ' ee8idpaz> OTTCOJ/. 32 36 THE FOREIGN AFFINITIES OF [On. II. nician workmanship, show how the most costly and elaborate building could be erected without the assistance of the stonemason 1 , and the ivory palaces of Solomon 2 were also specimens of the same application of art with that which appeared in the chrysele- phantine statues of Phidias. The very fact that the Doro-Ionian architecture, in its original and oldest type, not only admitted but required polychrome decorations, indicates that the materials employed must have been wood and metal, not stone, in the first instance. And the result of the whole discussion is to confirm our previous inference, that the Pelasgians were an Indo-Germanic tribe, who passed by the north of the Euxine into Europe, and re- crossed into Asia Minor by the Hellespont, where they came into direct contact with Semitic art and civilization. All tradition con- firms this, and the ready adoption by the Hellenes of the Asiatic, as opposed to the Cyclopian architecture, cannot be regarded as altogether unconnected with the ethnographical fact that the Dorians or Hellenes were a tribe which passed through Asia Minor in a strong but narrow stream on their way from the mountains of Garamania to the highlands of western Germany and northern Greece 3 . $ 4. Supported by deductions from the contrasted mythology of the two races. These views of the Cyclopian architecture, as distinctively characterizing the Pelasgians, are confirmed by all that we know of their religious system. The worship of the Pelasgians was not only elementary ; it not only consisted in an adoration of the great objects of nature for this was common to it with other primitive tribes ; .but it was especially a sun-worship, like that of the Medes, from whom, as we shall see, they trace their legi- timate descent. Thus, while the so-called aborigines of Italy worshipped Saturnus-Ops, the divinity of the earth 4 , the Pelasgo- Tyrrhenians who dwelt beside them worshipped Tina or Janus, the God of light. The two tribes, who constituted the original populus, being especially warriors, worshipped the God of war ; * For the details of Solomon's Temple, see Thenius, uber die Slicker der Konige, Anhang. p. 25, sqq. 2 Psalm xlv. 8; cf. 1 Kings xxii. 39; Amos in. 15. 3 New Crat. 92. 4 See Zumpt's Essay on this subject. H'J THE ANCIENT ITALIANS. 37 as Romulus was mythically the son of Mars, we may conclude that Mars or Mamers was the God of the Ramnes ; and then Quirinus 1 would be the spear-god of the Titles. Just in the same way, the Hellenes, who, as I have shown in another place, were a warlike tribe of high German character 2 , brought into Greece their war-god Apollo*, a sort of refined Woden ; but eventually allowed some of his attributes to be absorbed by the God of light, who was worshipped by the Pelasgians 4 . The Hyacinthia, which were retained by the Dorians in Laconia and applied to the worship of their own Apollo, were a festival of Achaean or Pelasgian origin, and symbolically expressed the triumph of the sun's disk over the rainy months of winter 5 . All the Pelasgian religion, wherever it can be discerned under the incrustations of later Hellenism, points to the same worship of the sun. Jupiter and Danae, of whose union the Argive Perseus was the fruit, represent the golden showers of the fructifying sky descending on the dry earth (Savdrj yfj) 6 . The Argive goddess Juno is called /3ou>iri^ as being a representative of the moon-goddess, who bore her disk between two horns, and who is thus identified with lo, " the earth," the daughter of Inachus 7 . In the same way Europa, the " broad-faced" moon, is borne across the sea from east to west by Jupiter in the form of a bull, that is, the sun in Taurus in conjunction with the moon rises from the eastern waves. Here she assumes the functions of ''Apre|ut9 Tavpoiro\o we might with still greater safety bring the Thracians and the Aga-thyrsi into the same etymology. The Bithynians were Thracians ; and there were Medo-Bithynians (MaiSol Wvos Qpaicrjs, Steph. Byz. p. 527) as well as Parthians (of SKU&K TOVS Js or Boreas. There is little difficulty, however, in showing that the name is identical with that by which the river is known at the present time, the Dnie-per or Dana-paris, with the last part of which we may compare the name Porata or Pruth. It is well known that the northern Greeks were in the habit of substituting the medial, not only for the tenuis, but even for the aspirate ; thus we have flvpyov for Trvpyos, ^epei'iKrj for QepevtKtj, Saveiv for Qavelv, and Boo--7ropo$ for 3>ojcr-00|0os. Accordingly, their pronunciation of the word Dana-paris (Paris-danas} would be Dana-baris, or, by an interchange of the two synonymous elements, Baris-danas 2 . But the Greek ear was so familiar with the sequence cr9-, that the sd- would inevitably fall into this collocation ; and, with a change of vowels, for the same purpose of giving the barbarous name a Greek sound, the compound would become the Hellenic form BopvaOevrj?, a word which has hitherto eluded etymological analysis. The Tana-is was the most easterly of Scythian, and indeed of European rivers. The explanation of the name is implied in what has been already stated. No difficulty can arise from the appearance of a tenuis instead of the medial, which generally 1 The identification of the Ingul-etz with the Pan-ticapes depends upon the position of the Hylcea, or " woodland " district, which must have been on the right bank of the Borysthenes, for the other side of the river is both woodless and waterless (see Lindner Skythien, Stuttgart, 1841, p. 40, sqq.). The name Inyul is borne by another river, which may be identified with the Hypa-caris. 2 A similar change has taken place in the name Berezina. 48 THE FOREIGN AFFINITIES OF [CH. II. appears in the first part of this name ; for the Danube, which is most consistently spelt with the medial, is called the Tun-owe in the Niebelung en-lied (v. 6116). The Tanais seems to have been the same river which the Cossacks still call the Donaetz or Tanaetz. We find the wprd Dana-s in composition not only with the synonyms Is-, Ap-, Paris, and Ter, but also with Rha-, which occurs in the names of the Asiatic A-ra-xes, and in that of the Rha- s or Wolga. Thus, we have the E-ri-danus in Italy, the Rha-danau in Prussia, the Rho-danus in France, and the name 'Poy-^ot/, quoted by Ptolemy. In England the name Dana occurs by itself as " the Don." 11. Names of the Scythian divinities. Let us now pass to the names of the Scythian gods, which may be referred without any difficulty to the roots of the Indo- Germanic family of languages. Herodotus informs us (iv. 59), that the names by which the Scythians designated the Greek divinities, 'lo-Tirj, ZeJ?, F^, 'ATroXXeoj/, Ovpavirj 'A , 01 OIKOVCTIV dp.(p\ vdfjLa H\OVTVOS nopov. The position of the article before /uLowwira shows that the words 'Apipacnrov 'nnrofid/uLova are to be taken in close connexion, and apart from the epithet fjiovvwTra ; and I see in this fragment of symbolical mythology a trace of that Hyperborean sun-worship, which the Pelasgians brought from Media into Greece and Italy. For Arim-aspas is most naturally explained as Ahurim-a$pa, or Orim-a$pa, the " horse " or " horseman of light," thus explain- ing the term 'nrTrofidiuwv, and the epithet IJLOVVW\^ will refer to the circular disk which surmounted the head of the Sun-god, and so gave rise to a belief in Cyclopian or monophthalmic deities. With this view, the meaning of the fable is clear. The one-eyed, equestrian people dwelling in the Hyperborean regions, which are regarded as the inaccessible and ever-guarded sanctuary of the Sun, can only represent the Sun-god himself mounted on his heavenly courser (the aurvat a^pa, "cheval rapide," of the a: Burnouf, pp. cxxxiv. 371); and the Gryfin, which 12.] THE ANCIENT ITALIANS. 53 Pursues the Arimaspian, who by stealth Had from his wakeful custody purloined The guarded gold is the Kcp/3-epos or 1VD, which vainly seeks to prevent the golden light of day from being borne to the southern regions by the horseman of light 1 . In a communication read before the Royal Asiatic Society in January 1851, I have pointed out a similar error of Herodotus respecting the horse of Darius and his groom Oibares; and I have shown that, while this last name refers to the verb vyabara, or the noun asbara, which must have occurred in the original inscription, Darius, as in his other inscriptions, must have referred his power not to the ingenuity of a servant, but to the gracious help of Ahura-mazda, " the lord of light," and his celestial steed the Sun. Another compound, which may with equal facility be referred to the Indo- Germanic family of languages, is the name by which the Scythians designated the Amazons. Oiopirara, according to Herodotus, is equivalent to dvdpoKTovo? oiop yap KaXeovai TOV av^pa, TO $e TraTct, Kreiveiv. Now oiop is clearly the Sanscrit vira, the Zend vairya, vira (Burnouf, Ya$na t p. 236), the Latin vir, Gothic vair-s, Welsh gwyr, and the Lithuanian vyras. The root pat in Sanscrit does not signify primarily " to kill," but " to fall ;" though the causative form pdtyati constantly means " he kills ;" i. e. " causes to fall." It seems more pro- bable, however, that the Scythian articulation has substituted a tenuis for the v-sound, as in the case of sparga for svarga, men- tioned above, and that the verb is to be sought in the common Sanscrit root vadha, " to strike," " to kill," " to destroy." Pliny (Hist. Nat. VI. 17) tells us that the Scythian name for Mount Caucasus was Grau-casis, i. e. nive candidus. The first part of this word is clearly connected with gelu, glades, Kpvos, Kpv-o-Ta\\os, kalt, cold, grau, and grey; and casis, " white," may be compared with cas-tus, cas-nar (senex Osco- rum lingua, Fest. ; comp. Yarro, L. L. VII. 29), canus, &c. 1 Ariosto mixes up the horse of the Arimaspian with the Gryfin which pursued him, and in his joking way speaks of the composite animal as still extant in the northern regions : Orlando Fur. IV. 18 : chiamasi Ippogrifo, Che ne i monti Rifei vengon, ma rari. 54 THE FOREIGN AFFINITIES OF [On. II. In the tract about rivers, printed among Plutarch's Frag- ments, we have the following Scythian words, with interpreta- tions annexed. He does not translate aXtVSa, which he describes as a sort of cabbage growing near the Tanais (c. XIY. 2) : we may compare the word with Temarunda. He tells us, however, that fipi%d/3a means Kpiov /uertuTroj/ (c. XIV. 4), that pu (phru) in a, " a virgin," seems unmistakeably connected with that of ap-rjs, dpe-Tij, ap- crqv, denoting distinctive manliness. It may be doubtful whether the Scythian word eVajoees, " the unmanly," (Herod. I. 105) is compounded of a and nri, or of an- and ar. But it is clear that the root ar in the Indo-Germanic language was originally var, and the Scythian oiop, as we have just seen, is the Sanscrit vira. It is not at all improbable that the anlaut may have been dropt in the other word apa, just as in ''Aprjs, "A/o-re/uus. At any rate there is no doubt as to the connexion between vir and virgo or virago : compare the synonyms Varro and Nero, wehren and nehrung ; &c. The mythology of Minerva and the etymology of castus may suffice to tell us how the ideas of protection, re- sistance, and virginity, are combined: and it is clear that the two former constitute the fundamental meaning of vir and a (N. Crat. 285). 12.] THE ANCIENT ITALIANS. 55 Herodotus (IV. 52) mentions a fountain the name of which was 'ZKvOicrTi peit 'E%a/UL7rcuos, Kara $e Tijv'EXXrjvwv yXwacrai^, 'Ipal o$oi. Hitter ( Vorhalle, p. 345) conjectures that the ori- ginal form of ' E^a/m-Trcti-os must have been Hexen-Pfad, i. e. Asen-Pfad, which he compares with Siri-pad, and which de- notes, he thinks, the sacred ominous road by which the Cim- merian Buddhists travelled towards the west. Bockh (Corpus Inscript. II. p. Ill) supposes the right interpretation to be ewea o$oi; so that e%dv is "nine." The numeral "nine" is pre- served in a very mutilated state in all languages, both Semitic and Indo-Germanic, and it would not be difficult to point out a possible explanation of the word e%av 9 if the reading eiWa ocol were really certain. But there is more reason to suppose that the other interpretation is correct, and that e%av corresponds to the Zend asja, aschavan, ashaun, ashaon, " holy," so that the termination will be the Persian pai, Zend pate, " a path," and the compound will correspond to the Persian Mah-pai, Satter- pai, and will denote " Holy-road" or Hali-dom : cf. the Persian names Bcrya-TraTos and Btrya-TraT^s (Zeuss, p. 295). This examination includes all the Scythian words which have come down to us with an interpretation ; and in all of them it has been shown that they are connected, in the signification assigned to them, with the roots or elements which we find in the Indo-Germanic languages generally, and especially in the Medo-Persian idioms. If we add this result of philology to the traditionary facts which have been recorded of the international relations of the Getse, Scythse, Sauromata3, and Medes, we must conclude that the inhabitants of the northern side of the Euxine, who were known to the Greeks under the general name of Scy- thians, were members of the Indo-Germanic family, and not Mongolians, as Niebuhr has supposed 1 . 13. Successive peopling of Asia and Europe : fate of the Mongolian race. The true theory with regard to the successive peopling of Asia and Europe seems to be the following 2 . Believing that 1 Kleine Schriften, I. p. 361. 2 The author's views are given in the New Cratylus (2nd Ed.) 64, sqq. and in the Transactions of the British Association for 1851, p. 138, sqq. 56 THE FOREIGN AFFINITIES OF [On. II. the human race originated in the table-land of Armenia 1 , I give the name of Central to the two sister-races, the Semitic and See also Winning's Manual, p. 124, sqq. Rask, uber das Alter und die Echtheit der Zen d-Spr ache, p. 69, sqq., Hageu's Tr. And, for the affinity of the inhabitants of Northern Asia in particular, see Prichard on the Ethnography of High Asia (Journal of R. G. S. IX. 2, p. 192, sqq.). 1 The general reasons for this opinion are given in the New Cratylus, 64. But I am inclined to attach much more importance than some other ethnographers to the geography of Eden, as given in the book of Genesis ; and I believe that the first seats of the human race are strictly denned by the four rivers there mentioned. Delitsch, in his recent Commentary on Genesis (p. 101, sqq.), has given a summary of all the leading views on the subject of these four rivers. In my opinion, the sacred writer wishes to indicate the immediate neighbourhood of the Caspian sea, a part of whose area may have corresponded originally to the once happy home of the family of man. At any rate, it is clear that physical changes have taken place in this region, and the book of Genesis implies that Eden no longer exists. Be this as it may, there can be no doubt that the sacred writer directs our view to a district from which there is a divergence of four great rivers. It does not follow that they all rose in this country, but this is true of the two which we have no difficulty in identifying, namely, the J1H2 or Euphrates, and the ^Wl or Tigris. The sources of these rivers point to the south of Armenia, and as no other rivers of great consequence, or answering to the definitions of the book of Genesis, take their rise in this district, we are naturally led to seek the other two O^JO, or main branches, in the two great rivers, the Oxus, and the Rha or Wolga, which terminate in the Caspian sea, and by this enormous confluence form the boundary of Armenia on the side opposite to the sources of the other rivers. It is worthy of remark that Pliny (VI. 18) makes the Oxus rise in the lake or sea in which it now terminates ; and the same mode of speaking may be conceded to the sacred writer. Now it can be shown that the Oxus and the Wolga, which are the two greatest rivers in the district, the only two, in fact, which can be compared with the Tigris and Euphrates, answer exactly to the descrip- tion given of the p'n\a and the jl'^S. With regard to the former, not only does the river Oxus bear the name of Jihon as well as Amoo, but the description tf^ \njjfte fltf 111011 can only apply to this river which ran from the mountains of India (Strabo, p. 510) through the lake of Aral into the Caspian, and so furnished a northern boundary to the whole of the country which the Hebrews called Gush. The name of the Jl'tt^S* which signifies " water poured forth," or " over-flowing," corre- sponds to the meaning of Rha (peo>, &c.), and to the character of the Volga as described by its Tartar name Ethel, " the bountiful." The reasons 13.] THE ANCIENT ITALIANS. 57 the Indo-Germanic, which formed themselves in Mesopotamia and Iran, and became the twin-mothers of human population, and the joint source and home of intellectual culture. To this central group, I oppose the Sporadic, as including all those nations and languages which were scattered over the globe by the first and farthest wanderers from the birth-place of our race. The process of successive peopling may be thus described. While the Indo-Germanic or Japhetic race was developing itself within the limits of Iran, and while the Semitic family was spreading from Mesopotamia to Arabia and Egypt, a great popu- lation of Tchudes, or Mongolians, Celts and Turanians, had ex- tended its migrations from the Arctic to the Indian Ocean, and from Greenland over the whole north of America, Asia, and Europe, even as far as Britain, France, and Spain. In propor- tion, however, as these Celto-Turanians were widely spread, so in proportion were they thinly scattered ; their habits were nomadic, and they never formed themselves into large or power- ful communities. Consequently, when the Iranians broke forth from their narrow limits, in compacter bodies, and with superior physical and intellectual organisation, they easily mastered or drove before them these rude barbarians of the old world ; and in the great breadth of territory which they occupied, the Tu- ranians have formed only four great and independent states the Mantchus in China, the Turks in Europe, and the Aztecs and the Peruvians in America. The student of ethnography must bear in mind some essential differences between the spread of those Sporadic tribes, which derived their origin from Iran, and to which the aboriginal popu- lation of Europe, Asia, and America is due, and those which emigrated from Mesopotamia and Arabia, and furnished a sub- stratum of dispersed inhabitants for Africa. For while the which led Reland, Rosemmiiller, and Rauiner, to identify this river with the Phasis, apply with still greater force, if we go farther north, and seek their justification in the great stream which skirts the Ural mountains. The mineral wealth of this district is well known, and the fact, that the land of Chawildh is found also in Arabia, does not prevent us from identifying this name with that of the Chwalissi who dwelt on the west of the Ural by the Volga, and to whom the Caspian owes its modern Russian name of Chwalinskoye More. 58 THE FOREIGN AFFINITIES OF [On. II. Sporadic Syro-Arabians in Africa exhibit, as we go farther from the center of their dispersion, a successive degeneration in the passage of the Aramaic languages from the Abyssinian to the Galla and Berber, from this again to the Caffre, from the Caffre to the Hottentot, and from the Hottentot to the clucking of the savage Bushman, and while there is no later infusion of civilized Semitic elements until the conquest of North Africa by the Arabs ; on the other hand, the Celto-Turanian tribes were overrun or absorbed at a very early period by successive or parallel streams of Sclavonians, Lithuanians, and Saxo-Goths, flowing freely and freshly from the north of Iran ; and the latest of these emigrants, the High-Germans, found many traces of similarity in the Celtic tribes with which they ultimately came in contact. Whatever might have been the degradation of the Ugro-Turanian races in those regions where they were most thinly scattered, it is obvious that the Scythia of Herodotus, which was the highway of the earliest march of Indo-Germanic migration into Europe, could not have been, as Niebuhr supposed, mainly peopled by a Tchudic or Mongolian stock. And though the name of S-colotce or Asa-Galatce, by which some of the Scytha3 called themselves, may be regarded as pointing to a Celtic or Turanian intermixture, the great mass of the hordes which dwelt to the north of the Euxine must have consisted of Indo-Germanic tribes who con- quered or ejected the Turanians; and I have no hesitation in referring these invaders, together with the Pelasgians of Greece and Italy, to different branches of the Sclavonian, Lithuanian, Saxo-Gothic, or generally Low Iranian stock. 14. The Pelasgians were of Sclavonian origin. It has been proved that the Sarmatians belonged to the parent stock of the Sclavonians ; and we find in the Sclavonian dialects ample illustrations of those general principles by which the Scy- thian languages seem to have been characterised. Making, then, a fresh start from this point, we shall find an amazing number of coincidences between the Sclavonian languages and the Pelas- gian element of Greek and Latin : most of these have been pointed out elsewhere * ; at present it is only necessary to call New Crat. 88. 14.] THE ANCIENT ITALIANS. 59 attention to the fact. So that, whichever way we look at it, we shall find new reasons for considering the Pelasgians as a branch of the great Sarmatian or Sclavonian race. The Thracians, Getae, Scythas, and Sauromatse, were so many links in a long chain connecting the Pelasgians with Media ; the Sauromatse were at least in part Sclavonians ; and the Pelasgian language, as it appears in the oldest forms of Latin, and in certain Greek archaisms, was unquestionably most nearly allied to the Sclavo- nian : we cannot, therefore, doubt that this was the origin of the Pelasgian people, especially as there is no evidence or argument to the contrary. J 15. Foreign affinities of the Umbrians, fyc. But, to return to Italy, who were the old inhabitants of that peninsula ? Whom did the Pelasgians in the first instance con- quer or drive to the mountains ? What was the origin of that hardy race, which, descending once more to the plain, subjugated Latium, founded Rome, and fixed the destiny of the world ? The Umbrians, Oscans, Latins, or Sabines for, in their historical appearances, we must consider them as only different members of the same family are never mentioned as foreigners. We know, however, that they must have had their Transpadane affinities as well as their Pelasgian rivals. It is only because their Celtic substratum was in Italy before the Pelasgians arrived there, that they are called aborigines. The difference between them and the Pelasgians is in effect this : in examining the ethnical affinities of the latter we have tradition as well as comparative grammar to aid us; whereas the establishment of the Umbrian pedigree depends upon philology alone. 16. Reasons for believing that they were the same race as the Lithuanians. Among the oldest languages of the Indo-Germanic family not the least remarkable is the Lithuanian, which stands first among the Sclavonian dialects 1 , and bears a nearer resemblance to Sanscrit than any European idiom. It is spoken, in different 1 See Pott, Et. Forscli. I. p. xxxiii. and his Commentatio de Borusso- Lithuanicce tarn in Slavicis quam Letticis llnguis principatu. Halis Saxoimin, 18371841. 60 THE FOREIGN AFFINITIES OF [On. II, dialects, by people who live around the south-east corner of the Baltic. One branch of this language is the old Prussian, which used to be indigenous in the Sam-land or " Fen-country" be- tween the Memel and the Pregel, along the shore of the Curische If of, and the Lithuanians are often called Samo-Getce or " Fen- Goths." Other writers have pointed out the numerous and strik- ing coincidences between the people who spoke this language and the Italian aborigines 1 . Thus the connexion between the Sabine Cures, Quirinus, Quirites, &c. and the old Prussian names Cures, Cour-land, Curische Haf, &c. has been remarked ; it has been shown that the wolf (hirpus), which was an object of mystic reverence among the Sabines, and was connected with many of their ceremonies and some of their legends, is also regarded as ominous of good luck among the Lettons and Courlanders ; the Sabine legend of the rape of the virgins, in the early history of Rome, was invented to explain their marriage ceremonies, which are still preserved among the Courlanders and Lithuanians, where the bride is carried off from her father's house with an appear- ance of force ; even the immortal name of Rome is found in the Prussian Romowo ; and the connexion of the words Roma, Romulus, ruma lupce, and ruminalis ficus, is explained by the Lithuanian raumu, gen. raumens, signifying "a dug" or " udder 2 ." 1 Perhaps the oldest observation of this affinity is that which is quoted by Pott (Commentatio, I. p. 6), from a work published at Leyden in 1642 by Michalo Lituanus (in rep. Pol. &c. p. 246) : " nos Lithuani ex Italico sanguine oriundi sumus, quod ita esse liquet ex nostro sermone semi-latino et ex ritibus Romanorum vetustis, qui non ita pridem apud nos desiere, &c. Etenim et ignis (Lith. ugnis f.) et undo, (wandu m.), aer (uras), sol (sdule) . . . unus (widnas) . . . et pleraque alia, idem significant Lithuano sermone quod et Latino." 2 See Festus, pp. 266-8, Miiller ; and Pott, Etymol. Forsch. II. p. 283. According to this etymology, the name Romanus ultimately identifies it- self with the ethnical denomination Hirpinus. The derivation of the word Roma is, after all, very uncertain ; and there are many who might prefer to connect it with Groma, the name given to the forum, or point of inter- section of the main streets in the original Roma quadrata, which was also, by a very significant augury, called mundus (see Festus, p. 266 ; Dionys. I. 88 ; Bunsen, JBeschreib. d. Stadt Rom, III. p. 81 ; and below, Ch. VII. 6). The word groma or gruma, however, is not without its Lithuanian affini- ties. I cannot agree with Muller (Etrusk. II. p. 152), Pott (Etym. Forsch. II. 101), and Benfey (Wurzel-Lexikon, II. p. 143), who follow the old $16.] THE ANCIENT ITALIANS. 61 Besides these, a great number of words and forms of words in the Sabine language are explicable most readily from a comparison with the Lithuanian; and the general impression which these arguments leave upon our mind is, that the Latins and Sabines were of the same race as the Lithuanians or old Prussians. $17. Further confirmation from etymology* Let us add to this comparison one feature which has not yet been observed. The Lithuanians were not only called by this name 1 , which involves both the aspirated dental th and the vo- calised labial u, but also by the names Livonian and Lettonian, which omit respectively one or other of these articulations. Now it has been mentioned before, that the name of the Latins ex- hibits the same phenomenon ; for as they were called both Latins and Lavines, it follows that their original name must have been Latuinians, which is only another way of spelling and pro- nouncing Lithuanians. If, therefore, the warrior-tribe, which descended upon Latium from Reate and conquered the Pelasgians, gave their name to the country, we see that these aborigines were actually called Lithuanians ; and it has been shown that they and the Sabines were virtually the same stock. Consequently, the old Prussians brought even their name into Italy. And what does this name signify ? Simply, " freemen 2 ;" for the root grammarians, and connect this word with the Greek yvapa, it is much more reasonable to suppose, with Klenze (Abhandl. p. 135, note), that it is a genuine Latin term ; and I would suggest that it may be connected with grumus, Lithuan. kruwa, Lettish kraut : comp. Kp/ia, globus, gleba, &c. The name may have been given to the point of intersection of the main via and limes, because a heap of stones was there erected as a mark (cf. Charis. I. p. 19). Even in our day it is common to mark the junction of several roads by a cross, an obelisk, or some other erection ; to which the grumus, or " barrow," was the first rude approximation. If so, it may still be connected with ruma ; just as paoTor signifies both "a hillock" and "a breast;" and the omission of the initial g before a liquid is very common in Latin, comp. narro with ywplfa, nosco with ytyi/eoo-Kco, and norma with yvaptp-os. 1 The known forms of the name are Litwa, Lietuwa, Litauen, Lietu- wininkas, Atr/3ot, Lethowini, Lituini, Letwini, Lethuini, Lettowii, Litwani, Letthones, and Letthi. 2 By a singular change, the name of the kindred Sclavonians, which in the oldest remains of the language signifies either " celebrated," " illus- 62 THE FOREIGN AFFINITIES OF [On. II. signifying " free," in all the European languages consisted of and a combination of dental and labial, with, of course, a vowel interposed. In most languages the labial is vocalised into u, and prefixed to the dental ; as in Greek c-XevOe-pos, Lithuan. llau- dis, Germ, leute, &C. 1 In the Latin liber the labial alone re- mains. $18. Celtic tribes intermixed with the Sclayonians and Lithuanians in Italy and elsewhere. The name of the Umbrians, the most northerly of the indi- genous Italians, leads to some other considerations of great im- portance. It can scarcely be doubted that in their northern as well as their southern settlements the Lithuanians were a good deal intermixed with Celto-Finnish tribes in the first instance, and subjected to Sclavonian influences afterwards. That this was the case with the Lithuanians, we learn from their authentic and comparatively modern history. The proper names cited by Zeuss (p. 229) show that there was a Celtic ingredient in the popula- tion of Rsetia and Noricum. It appears, too, that in Italy there was a substratum of Celts before the Lithuanians arrived there ; this is expressly recorded of the Umbrians by M. Antonius and Bocchus (apud Solin. c. 2.) and by Servius (ad Virg. jtEneid. XII. 753), and the fact is clearly indicated by the name of the country, Umbria, and its principal river Umbro. If the oldest inhabit- ants of this country were Celtic, they must have been an offshoot of the Celtic race which occupied the contiguous district of Ligu- trious" (from plava, "glory," root pZw, Sanscr. prw, Gr. *Xv-: see 'Safafik, and Palacky's litest. Denkm. der Bohm. Spr. pp. 63, 140), or " intelligibly speaking," as opposed to barbarian (from slovo, " a word "), has furnished the modern designation of "a slave," esclave, schiavo. The Bulgarians, whom Gibbon classes with the Sclavonians (VII. p. 279, ed. Milman), have been still more unfortunate in the secondary application of their name (Gibbon, X. p. 177). 1 Dr Latham says (Germania of Tacitus, Epilegom. p. cxi.): "the root L-t = people is German (Leute), yet no one argues that the Lat-ins, Lith- uanians, and a host of other populations, must, for that reason, be German." If the people called themselves by this name, it may be fairly inferred that it was to them a significant term, and may therefore be taken as a mark of affinity : no Indo-Gerraanic philologer will deny that the Lithuanians and Germans were cognate races. 18.] THE ANCIENT ITALIANS 63 ria. Now not only are the Ambrones said to have been a Celtic race (Ambrones, says Festus, fuerunt gens qucedam Gallicd), but this was also the generic name of the Ligurians (<70a? yap auTovs OVTGDS ovoimdfyvcri Kara yevos Aiyves, Plut. Vit. Marii, c. XIX.). Whatever weight we may attach to the statement in Festus, that they were driven from their original settlements by an inundation of the sea, we cannot fail to see the resemblance between the name of the Ambrones and that of the river Umbro ; and no Englishman is ignorant that the North-umbrians are so called with reference to an Ymbra-land through which the river Humber flowed. Dr Latham ( Tac. German. Epilegom. p. ex.) has suggested a connexion between a number of different tribes which bore names more or less resembling this, and he thinks that there is some reference in this name to the settlement of the race bearing it near the lower part of some river. Thus the Am- brones seem to have been on the Lower Rhine, the Umbri on the Lower Po, the Cumbrians of Cumberland on the Solway, and the Gambrivii and Si-gambri on the Lower Rhine. Dr Latham also conjectures that Humber may be the Gallic and East British form of the Welsh Aber and the Gaelic Inver=" mouth of a river." It appears to me that the Sigambri and Gambrivii belonged to a German, not to a Celtic stock, and I am disposed to refer the name of Cumber-land to the form Cymmry. Nor do I think it reasonable to suppose that Umber or Ambro is a dialectical variety of Aber or Inver. But whether we are or are not to connect the word with amhainn or amhna, " a river," found in Gar-umna, it cannot be doubted that the name of Urn- bria points to a continuous population of Ligurians or Ambrones extending from the Cottian Alps to the Tiber ; and there is every reason to believe that this was only part of a Celtic population which occupied originally the three peninsulas of Greece, Italy, and Spain, together with the great islands of Sicily, Sardinia, and Corsica. The first inhabitants of Spain and Sicily are called Iberians by every ancient writer, and they are identified with the Sicanians ; and Philistus must have referred to these when he said that the Sicilians were Ligurians who had been driven southwards by the Umbrians and Pelasgians (Dionys. Hal. I. 22), meaning of course the Low-German and Sclavonian tribes, who subsequently occupied north Italy. With regard to Greece, there is no reason why the Leleges, whom we have other grounds for 64 THE FOREIGN AFFINITIES OF [Car. II. considering as Celtic, should not be regarded as exhibiting the name of the Ligyes with that reduplication of the initial I- which is so universal in Welsh 1 . 19. The SarmatcB probably a branch of the Lithuanian family. If it is necessary to go one step farther, and identify this Lithuanian race with some one of the tribes which form so many 1 Professor F. TV. Newman, in his little work entitled Regal Rome, maintains that the old languages of Italy, especially the Umbrian and Sabine, contained a striking predominance of Celtic ingredients, and he wishes to show that this is still evident even in the Latin of Cicero. His proof rests on vocabularies (pp. 19 26), especially in regard to the military, political, and religious words, which he supposes that the R/omans derived from the Sabines (p. 61). With regard to these lists I have to observe, that while all that is valid in the comparison merely gives the Indo-Germanic affinities of the Celtic languages a fact beyond dispute Mr. Newman has taken no pains to discriminate between the marks of an original identity of root, and those words which the Celts of Britain derived from their Roman conquerors. In general, Mr. Newman's philology is neither solid nor scientific. It is not at all creditable to a professed student of languages to compare the participial word cliens (clie-nt-s) with the Gaelic clann, cloinne, " children." If anything is certain about the former, it is clear that it contains the verb-root cli- or clu- with a merely formative termination in nt, which does not belong to the root. Again, when every one knows the Latin meaning of tripudium, referring to the triple ictus, what is the use of deriving it from the Gaelic tir " earth," and put " to push ?" If quir-i[t]-s with a regular Indo-Germanic ending, is naturally derived from quiris "a spear," what miserable ety- mology it is to compare the former with curaidh " a champion," from cur "power," and the latter with coir "just, honourable, noble." And all regard for simple reasoning is neglected by a writer, who analyses augur = avi-ger into the Gaulish auca " a bird," and the Welsh cur " care." I am influenced only by a regard for the interests of sound learning when I express the strong feelings of dissatisfaction with which I have read most of Mr. F. W. Newman's books. With great natural abilities and the power of giving a specious and plausible representation of the views which he adopts, his self-reliance has led him to attempt a wide and very important range of subjects, with very inadequate preparation for their proper discussion ; and thus in history, philology, biblical criticism, and political economy, he has contrived to exhibit himself as a rash and mischievous writer, and has done considerable damage to the good cause of independent thought and original investigation. 19.] THE ANCIENT ITALIANS. 65 links of the chain between Media and Thrace, it would be only reasonable to select the Saurornatce, whose name receives its in- terpretation from the Lithuanian language (Szaure-Mateni, i. e. "Northern Medes"). The Sauromatse and the Scytha3 were undoubtedly kindred tribes ; but still there were some marked differences between them, insomuch that Herodotus reckons the Sarmatae as a separate nation. Between the Pelasgians and the Umbrians, &c., there existed the same affinities, with similar dif- ferences ; and the fairest conclusion seems to be this, that as the Latins or Lithuanians were a combination of Gothic and Sclavo- nian ingredients, so were the Sauromataa ; that as the indigenous tribes of Italy were pure Gothic, mixed with Celtic, so were the Scythsa or Asa-Goths. At the same time it must be remarked, that the term Sarmatian has a wider as well as a narrower signi- fication. In its more extended meaning it is synonymous with Sclavonian, and therefore includes the Pelasgians. In its nar- rower use, it is expressive of that admixture of Sclavonian and Low-German elements which characterizes the Lithuanian or Samo-Getic languages, and in which the Sclavonian is so predo- minant that the Gothic element is almost overpowered. Revert- ing to the Asiatic settlements of these races, we may say, as we pass from West to East across the northern frontiers of the plateau of Iran, that the true Sclavonians extended from the borders of Assyria to those of Hyrcania and Parthia ; that they there abutted on the debateable land or oscillating boundary-line between the Sclavonian and Gothic races, and so became Massa- GetaB or Lithuanians ; and that the Sacae, Saxons, or genuine Gothic and Low-German tribes, the Daci, Danes, and Northmen of Europe, occupied Sogdiana to the banks of the laxartes. If we suppose, what we have a right to suppose, that this line was preserved as the march of emigration wheeled round the north of the Caspian the Sclavonians to the left, the Lithuanians in the centre, and the pure Goths to the right, we shall have a simple explanation of all the facts in the ethnography of Eastern Europe. For these are still the relative positions of the different races. The right wing becomes in the course of this geographical evolu- tion the most northerly or the most westerly, while the left wing or pivot of the movement becomes most southerly or most easterly, and the centre remains between the two. Thus the pure Low- Germans and the Lithuanians never come into Greece, which 5 66 THE FOREIGN AFFINITIES OF [On. II. is peopled by the Sclavonians. Lithuanian and Sclavonian are mingled in Italy. But although, as we shall see, a branch of the pure Gothic invaded that peninsula, it felt, to the end of its early history, that it had approached a distinct line of demarcation wherever it touched, without Lithuanian intervention, on the borders of pure Sclavonism. 20. Gothic or Low- German affinities of the ancient Etruscans shown by their ethnographic opposition to the VENETI. This brings us to the crowning problem in Italian ethnogra- phy, the establishment of the foreign affinities of the ancient Etruscans. Wherever the advancing tide of Sclavonian emigra- tion came to a check before the established settlements of a purely Gothic or Low-German tribe, wherever, consequently, the Sclavonians felt a need for a distinctive appellation, we find that they called themselves Serbs, Sorbs, or Servians, a name apparently denoting their agricultural habits, or else Slow-jane, Slow-jene, or Sclavonian, a name implying, according to the most recent interpretation, that they opposed their own language as intelligible to the foreign jargon of their neighbours. By these names they were known in the distant lands to which the wars of the ninth and tenth centuries transported them as cap- tives ; and as a foreign and barbarous slave was a Scythian in the older days of Athens, a Davus or Dacian and a Geta or Goth in the later comedies, so all prisoners were called indifferently Slave or Syrf, a circumstance which proves the identity and prevalence of these national designations. But while these were the names which the Sclavonians assumed on their own western boundary-lines, and by which they were known in foreign coun- tries, they received the name of Wends, Winiden, 0. H. G. Winidd, A. S. Veonodas, from the Gothic tribes on whom they immediately abutted. By this name, or that of Finns, which is merely a different pronunciation, the Goths of the north desig- nated their eastern neighbours, whether of Sclavonian or Turanian race. By this name the Saxons distinguished the Sclavonians in Lusatia. The traveller's song in the Codex Exoniensis expressly opposes the Goths to the Wineds wherever found; "I was," says the author (vv. 113, sqq.) "with Huns and with Hreth- Goths, with Swedes and with South-Danes, with Wends I was 20.] THE ANCIENT ITALIANS. 67 and with Waerns, and with Wikings, with Gefths I was and with Wineds." Although the strong but narrow stream of High-Ger- man conquest disturbed the continuous frontier of the Sclavonian and Low-German tribes, we find, as late as Charlemagne's time, that Sclavonians were recognized in central Germany under the designations of Moinu-winidi and Ratanz-winidi, from the names of the rivers which formed their geographical limits. The same denomination was applied in much earlier times to the Sclavo- nians settled in Bavaria, who were called the Vinde-lici, or Wineds settled on the Licus or Lech. Farther east on the Danube the March-field furnished another boundary to the Scla- vonians, whose city there was called Vind-o-bonum. We must of course admit the same term in the name of the Veneti at the head of the Adriatic. And thus we trace this distinctive appel- lation from Scandinavia to the north of Italy, in a line nearly corresponding to the parallel of longitude. The ethnographic importance of the name Wined can scarcely be overrated : for it not only tells us that the tribes to the east of the line upon which it is found were generally pure Sclavonian, but it tells us as plainly that the tribes to the west, who imposed the name, were equally pure branches of the Gothic, Saxon, or Low-German race. Indeed, the latter fact is more certain than the former. For if, as I believe, the term Wined merely indicates, in the mouth of a Low-German, the end or wend-point of his distinctive territory, our inference must be that whatever the Wineds were, they indicated the boundary-line of some branch of the Gothic race. Now we have such a boundary line in Bavaria ; therefore the Rcetians who faced the Vindelici or Lech -Wineds were Low-Germans. We have a similar line in the north of Italy ; therefore there must have been Low-Germans in opposition and contiguity at the western frontier of the Veneti or Wineds on the Po. But we have seen that the Etruscans, properly so called, were Rcetians, who at one time occupied a continuous area stretching from western Germany across the Tyrol into the plains of Lombardy. It follows therefore, as an ethnographical fact, that the Etruscans must have been a Low- German, Gothic, or Saxon tribe. 52 68 THE FOREIGN AFFINITIES OF [On. II. j 21. Reasons for comparing the old Etruscan with the Old Norse. These combinations would be sufficient, if we had nothing else, to establish prima facie the Gothic affinities of the old Etruscans. But they are only the first step in a cumulative argument, which, when complete, raises our conclusion to the rank of a philological demonstration. Some of the details must be reserved for the chapter on the Etruscan language ; but the general effect of the reasoning shall be given here. If the ancient Etruscans were Low-Germans, they must present the most striking marks of resemblance when they are compared with the oldest and least alloyed branches of that family. In the center of Europe the Low-German element was absorbed by the High-German, and the latter became a qualifying ingredient in all the Teutonic tribes of the mainland, who were not similarly affected by Sclavonism. As I have elsewhere sug- gested (New Crat. 78), the Lithuanians were Low-Germans thoroughly Sclavonized ; the Saxons or Ingcevones were Low- Germans untainted by Sclavonism, and but slightly influenced by High- Germanism ; the Franks or Isccevones were Low- Germans over whom the High-Germans had exercised considerable control ; and the Thuringians or Herminones were pure High-Germans, in the full vigour, of their active opposition to the tribes among which they had settled. For Low-German unaffected by any qualifying element we must go to the Scandinavian or Norse branch of the race, which contains the Danish, Swedish, Nor- wegian, Faroic, and Icelandic tribes. The oldest or standard form of the languages spoken by these tribes is the Old Norse or Icelandic, which not only exists as a spoken tongue, but is also found in a very flourishing and ancient literature. The present inhabitants of Iceland trace their descent from emigrants who settled there in the ninth century ; and from circumstances con- nected with their isolated position the language has remained the unaltered representative of the oldest known form of Scandinavian or pure Gothic. It is therefore with this Old Norse or Icelandic, the language of the Sagas and Runes, that we must compare the old Etruscan, if we wish to approximate to the common mother of both, on the hypothesis that they are both traceable to the same stock. But the reader must from the first be guarded 21.] THE ANCIENT ITALIANS. 69 against the ridiculous idea that I identify the Etruscan with the Icelandic. The proposition which I maintain is this : that the Icelandic in the uncultivated north represents in the ninth century of our aera the language of a race of men, who might have claimed a common pedigree with those Raeto-Etruscans of the south, who became partakers in the Pelasgian civilization about 1600 years before that epoch. Moreover the Icelandic or Old Norse remains pure to the last, whereas the Etruscan is from the first alloyed by an interpenetration of Umbrian and Pelasgian ingredients. Consequently, it will justify all our reasonable expectations, if we find clear traces of the Old Norse in the dis- tinctive designations of the Etruscans, that is, in those names which they imported into Italy, and if we can make the Scandi- navian languages directly available for the explanation Q such of their words and phrases as are clearly alien from the other old idioms of Italy. This, and more than this, I shall be able to do. $ 22. Old Norse explanations of Etruscan proper names. It has been shown in the preceding chapter that the con- querors of the Umbrians and Tyrrheno-Pelasgians in Northern Italy called themselves Ras-ena. Niebuhr has suggested that this word contains the root ras- with the termination -ena found in Pors-ena, &c., and I have hinted that the same root is found in the distinctive designation of this race, Et-rus-ci or Het-rus-ci, which presumes an original Het-rusi, whence Het- rur-ia for Het-rusia. The old Norse will tell us the meaning both of the root and of the prefix : for in Icelandic hetia is " a warrior, hero, or soldier," and in the same language ras implies rapidity of motion, as at rasa, " to run." So that Ras-ena and Het-rusi imply a warrior-tribe, distinguished by their sudden onset and rapid career. Thus a warrior is TTO^OS WK Je, predaceous animals are Owes, and the old Scandinavian pirates have left the eagle or the war-galley on the armorial bearings of those families which claim a descent from them, as an indication of the same characteristic. This would be admitted as a reasonable con- jecture even if it had nothing else to recommend it. However, it does so happen that we have a distinct record of a migratory conquest by the Scandinavians in the heart of Europe rather before the colonization of Iceland, in which they called themselves by the same name as these Rasena or Het-rus-i. It has been 70 THE FOREIGN AFFINITIES OF. [Cn. II. shown by Zeuss (die DeutscJien, pp. 547, sqq.) that the language of these conquerors, who descended the Dnieper, the Volga, and the Don, was old Norse, and that their leader Chacan bears the Norse name Hakon ; and Symeon Magister, who wrote A. D. 1140, has given the same Scandinavian explanation of their name Ros, which I have suggested for Ras-ena ; for he says (Scriptor. post Theophan. ed Paris, p. 490): ol Po>? oi KO! Apofjurat \ey6fjL6voi, " the Ros who are called the racers or runners ;" and (p. 465): 'Po>s $e ot Apofurai (pepwvv^oi opofUTai oe airo TOV o^eo)? Tpe^eiv ctvTois TTpocreyevcTo, " the Ros are called the runners, and they are so called from the rapidity of their motion 1 ." Here the conjecture, which I proposed to the British Association, is confirmed by an authority subsequently observed : and no^ne will deny the obvious value of this corroboration. It may therefore be laid down as a matter of fact that the distinctive ethnical designation of the old Etruscans is Scandina- vian ; and we shall see that their mythological or heroic names are explicable in the same way. Niebuhr remarked, without attaching any importance to the observation, that there was a singular resemblance between the Scandinavian mythology and that of the Etruscans : " according to their religion, as in that of the Scandinavians, a limit and end was fixed to the life even of the highest gods" (H. R. I. note 421). Now in the Scan- dinavian mythology there is no name more prominent than that of Thor or Tor, and this prefix is a certain indication of the presence of the Northmen in any country in which it is found. Hickes says : " Prsep. Thor vel Tor in compositis denotat diffi- cultatem, arduitatem, et quid efficiendi molestiam, pessumdans significationem vocis cui prseponitur, ut in Tor-cere ' annonse difficultas et caritas,' Tor-fcera, ' iter difficile et impeditum,' Tor- feiginn, ' acquisitu difficilis,' Tor-gcetu, ' rarus nactu,' &c. Ex quibus constat, ut nomen deastri Tyr veterum septentrionalium 1 Zeuss suggests that the original old Norse form was Rcesar from the sing. Rcesir = Spo/uYj;? = cursor. He asks : " gehort hieher auch Rcesir in den Liedern hatifiges Synonymum fur Kontingr, etwa der Schnelle, Edle ?" and quotes Skaldskaparm. p. 191, for Rcesir as a man's name. The name Ros or Rus, as applied to the Scandinavians, is presumed in the designation P-rusi = po-Rus-i "adjoining the Ros:" cf. Po-morani, "the dwellers on the sea" (po-more). 22.] THE ANCIENT ITALIANS. 71 Mercurii in compositione gloriam, laudem, et excellentiam denotet : sic nomen idoli Tfior euphonice Tor eorum Jovis et Herculis, qui cum malleo suo omnia domuit et superavit, in com- positione significat et insinuat difficultatem quasi Herculeam vel rem adeo arduam et difficilem, ut Thori opem posceret, qua superari quiret." The lexicographer has here confused between the name of the god Thor (Grimm, D. M. p. 146, et passim) and a prefix equivalent to the Sanscrit dur- Greek Svv- (N. Crat. 180). But whatever may be the true explanation of this initial syllable, there can be no doubt that it belongs to the oldest and most genuine forms of the Low- German languages ; and when we find the name Tar-chon or Tar-quin among the mythical and local terms of the ancient Etruscans, we cannot but be struck by the old Norse character impressed upon them. We at once recognise the Scandinavian origin of the town of Thor-igny in the north-west of Normandy, where the termina- tion is the same as that of many towns in the same district, as Formigny, Juvigny, &c., and corresponds to the Danish ter- mination -inge, as Bellinge, Helsinge, &c. (Etienne Borring, sur la limite meridionale de la Monarchic Danoise. Paris, 1849, p. 9). It is worthy of remark that the word ing-, which is appropriated by the Ing-cevones, Ang-li, JEngl-lish, and other Low-German tribes, seems to signify "a man" or "a warrior" (Grimm, D. M. I. p. 320), and as quinna is the Icelandic for mulier, Tor-ing and Tar-quin might be antithetical terms ; and the latter would find a Low-German representative in Tor-quil. The other mythical name of the old Etruscans, which comes in close connexion with Tar-quin, is Tana-quil ; and Tar-quin or Tor-quil and Tana-quil might represent a pair of deities worshipped at Tarquinii, the plural name of which indicates, like Athence and Thebce, the union of two communities and two worships, the Pelasgian Tina or Tana, i. e. Janus, being placed on an equal footing with the Scandinavian Thor. This is in- verted in the tradition which weds the Greek Demaratus to the indigenous Tana-quil. At any rate, we cannot but be struck with the Scandinavian sound of Tana-quil, which reminds us of Tana-quisl, the old Norse name of the Tanais, which, although the name of a river, is feminine (Grimm, D. Gr. III. p. 385). These coincidences become the more striking, when we re- member that we are comparing the old Norse, of which we know 72 THE FOREIGN AFFINITIES OF [Cn. II. nothing before the eighth century of our sera, with the old Etruscan, which flourished nearly as many centuries before the birth of Christ. And when we add to all these evidences of direct history, ethnography, and mythology, the fact, which will be exhibited in a subsequent Chapter, that the Scandinavian languages supply an immediate and consistent interpretation of those parts of the Etruscan inscriptions which are otherwise inexplicable, no reasonable man will refuse to admit that the linguistic and ethnological problem suggested by the old inha- bitants of Etruria has at length received the only solution, which is in accordance with all the data, and in harmony with the nature and extent of the materials and with the other conditions of the case. $ 23. Contacts and contrasts of the Semitic and the Sclavonian. It appears that the original settlements of the Sclavonian race were in that part of Northern Media which immediately abuts on Assyria, and therefore on the cradle of the Semitic family 1 . From this we should expect that the Sclavonian dia- 1 It can scarcely be necessary to point out the difference between the ethnological argument by which I have traced the Pelasgo-Sclavonians to an original settlement in the immediate vicinity of upper Mesopotamia, and Mrs. Hamilton Gray's conjectural derivation of the Rasena from. Resen on the Tigris (History of Etruria, I. pp. 21, sqq.). To say nothing of the fact that I do not regard the Rasena as Pelasgian, I must observe that it is one thing to indicate a chain of ethnical affinities which extended itself link by link through many centuries, and another thing to assume a direct emigration from Resen to Egypt, and from Egypt to Etruria. The hypo- thesis of an Egyptian origin of the Etruscans is as old as the time of Bo- narota, but we know enough of the Semitic languages to be perfectly aware that the Rasena did not come immediately from Assyria or Egypt. Be- sides, if this had been the case, they would have retained the name of their native Resen until they reached Italy. In tracking the High- Germans and Hellenes from Caramania to Greece and central Europe, we find in the dry-bed of history continuous indications of their starting-point and route (New Cratylus, 92). And the Sauro-matce preserve in all their settlements a name referring to their "Median home." But Mrs. Gray's Rasena forget their native Resen in the alluvial plains of Egypt, and mi- raculously recover this ethnographical recollection in Umbria and among the Apennines. This is not in accordance with observed facts. Wan- dering tribes call themselves by the name of their tutelary hero, or by 23.] THE ANCIENT ITALIANS. 73 lects would furnish us with the point of transition from the Indo- Germanic to the Semitic languages ; and an accurate examination of the question tends to show that this expectation is well founded. 'But etymological affinities may exist by the side of the greatest contrast in regard to the state or condition of two languages ; and thus we find that, while the Semitic and Sclavonian come very close in etymology, they are unlike in syntactical develop- ment in those points which most distinguish the Sclavonian from other Indo-Germanic idioms. As I have elsewhere discussed this subject at sufficient length 1 , I shall here only recapitulate the general results of the inquiry, (l) The salient points of resemblance between the etymological structure of the Semitic and Sclavonian languages are (a) a number of common words which are more or less peculiar to both: as 2l'ED dhob, jjJ debr, "good," compared with the Russian dob-ro ; "-fTr derek, jj derej, "a road," compared with the Russian doroga, TIT"]! gd-dol, " great," compared with the Russian dolgie, &c. ; (6) a tendency to the agglutination of concrete structures in both. If roots were originally monosyllabic, the triliteral roots of the Semitic languages cannot be otherwise accounted for than by supposing that they are pollarded forms of words consisting of monosyllabic roots combined with a prefix, affix, or both. As then the Sclavonian languages exhibit words in this state of accretion, and as the Semitic petrefactions would most naturally emanate from this state, we must reckon this among the proofs of their etymological affinity ; (c) the correspondences furnished by the comparative anatomy of the Semitic and Sclavonian verb. some significant epithet applicable either to themselves or to their original country, and they keep this throughout their progress. There is no parallel to Mrs. Gray's assumed fact, that a body of men set forth from a great city, lost their name on the route, and resumed it in their ulterior settlements. On the whole, I must designate the conjecture about Resen as a lady-like surmise ; very imaginative and poetical ; but representing rather the conversational ingenuity of the drawing-room than the well- considered criticism of the library. On the contacts between the Semitic and Sclavonian tribes in their original settlements, the reader may consult the authorities quoted by Prichard, Natural History of Man, p. 142, and Mill, Myth. Interpr. of Luke, p. 66, note. * Report of the British Association for 1851, pp. 146, sqq. 74 THE FOREIGN AFFINITIES OF [On. II. "We find in both a parsimony of tense forms by the side of a lavish abundance of derived or conjugational forms; (d) the complete coincidence of the Semitic and Sclavonian languages in regard to their unimpaired development of the original sibilants ~ for it is only in these languages that we find the three sounds of sain and zemlja, of tsade and tsi, of $amech and slovo : and while the formation of palatals has proceeded to its full extent in Sclavonian and Arabic, the permanence of the pure sibilant in Hebrew is shown by the fact, that, with a full array of breathings, there is no diminution in the use of the sibilants in anlaut or as initials. (2) The most striking difference between the Semitic and Sclavonian languages and it is one which marks the earliest of the former no less than the most modern repre- sentatives of the latter consists in the fact, that while the Semitic languages are all in a syntactical condition, having lost most of their inflexions, and exhibiting all the machinery of definite articles, prepositional determinatives of the oblique cases, and other uses of particles to compensate defects of etymological structure, the Sclavonic languages have never arrived at this syntactical or logical distinctness, and have never abandoned their formative appendages and the other symptoms of etymological life and activity. These differences are due to the fact that while the Sclavonic tribes have remained pure up to the present time, and have been remarkable for their slow adoption of the art of writing and their inferior literary cultivation, the Semitic nations were from the earliest times exposed to the frequent intermixture of cognate races, and were the first possessors of an alphabet and of written records. We have therefore, in the antithesis or contrast of the Sclavonic and Semitic, a proof of the effects which external circumstances may produce on the state or condition of a language ; and the resemblances, to which I have called attention, must be taken as an indication of the perma- nence of that affinity which results from the geographical contact and intermixture of two races at a very early period. 24. Predominant Sclavonism of the old Italian languages. As the result of the ethnological speculations of this Chapter has been to show that the Pelasgian or Sclavonian was one of the earliest and certainly the most permanently influential element THE ANCIENT ITALIANS. 75 in the old languages of Italy, we should expect to find in these languages those characteristics of Sclavonism which evince the primitive contact and actual contrast of the Semitic and Sclavo- nian idioms. And this expectation is amply justified by the facts of the case. For while, on the one hand, we observe in the old Latin, Umbrian, and Oscan, verbal resemblances to the Semitic, which cannot be accidental, because they belong to some of the oldest forms in the respective languages ; and while both the Semitic and the old Italian are remarkable, like the Sclavonian, for their superabundance of sibilants, we observe that in spite of the cultivation of Greek literature by the Romans, and in spite of the adoption of the Greek ritual by the Sclavonians, these lan- guages have never attained to the use of a definite article, which is the key-stone of Greek syntax, and without which the Semitic languages could not construct a single sentence. The prepon- derance of the sibilants in the old Italian languages will be dis- cussed in the next Chapter, and we shall see in the proper place that in anlaut, or as an initial, the s always appears in Latin where it is omitted altogether, or represented only by an aspi- rate in Greek. Of the coincidences between the pure Latin and genuine Semitic words, it will be sufficient to give a few examples out of many which might be adduced, (a) The verb aveo or haveo is at least as closely connected with IHiJ or Hltf as with any Indo-Germanic synonym. (6) The words se-curis and sa-gitla have occasioned great difficulty to philologers. The former, according to Bopp, (Vergl. Gr. p. 1097) is a participial noun fromseco, and sec-uris=se-cusis must be compared with the Sanscrit forms in -usht=Gr. -u?a. This however is hardly more than a conjecture, for we have no other Latin noun to support the analogy. It is more probable that the initial syllable in both words is one of those prepositional affixes which we find in cr-KGTrapvov compared with /COTTTOJ, s-ponte compared with pondus, &c., and then we shall be able to see the resemblance between se- curis and the Hebrew ]pi|, Lett, granst " to hack or gnaw," and between sa-gitta and the Hebrew \>J1 from \>OT, which again is not unconnected with b~$, and the Latin ccedo. (c) It has been proposed to derive mare, Sclav, more, from the Sanscr. maru, " the waste" (Zeitschr. f. Vergl. Sprf. I. p. 33) ; but it appears much more reasonable to compare these words with the Hebrew , in which case the affix re will be connected with a word 76 THE FOREIGN AFFINITIES OF [On. II. denoting "flowing:" cf. teme with tema-runda (above 11). (d) The Hebrew TT^ gives us the root reg-, " to reach out/' with the prepositional affix ba, from abhi, as fully as the Latin p-recor, posco=p-roc-sco, Sanscrit p-rach-chdmi, &c. (e) It is only in the Pelasgian $o\ixs> ^e Sclavonic dolgye, and the Latin in-dulgeo, that we find a complete reproduction of the Semitic bvr|. (/) As the impersonal use of debeo nearly accords with that of oportet, and as the latter is manifestly connected with opus (Doderlein, Lat. Syn. u. Et. V. 324), it may be after all more reasonable to connect deb-eo with the important root dob, "a suitable time" (Polish), dob-ro, "good" (Polish and Russian), which furnishes us with one of the most remarkable instances of a connexion between the Sclavonian and Semitic lan- guages (cf. the Hebrew liD dhob, and the Arabic oj, debr), than to fall back upon either of the favourite derivations from SevevOai or dehibeo. The adjective debilis differs so entirely in meaning and application from the verb debeo, to which it is re- ferred, that I cannot concede the identity of origin. As there is reason to believe that the termination -bills is connected with the substantive verb fio (written bo in the agglutinate forms), a refer- ence to the usage of de-sum and de-fio would best explain the origin and meaning of de-bi-lis. How the sense of " owing " or " obliga- tion" borne by deb-eo is connected with that of " fitness," " good- ness," and " propriety," may be seen at once by an examination of such idioms, as Sacaios el/ui TOUTO 7roieii>, " I am bound to do this," ei /my aSiKw, " I ought," &c. (g) A comparison of heri and -)(0es enables us to see that the Latin humus and the Greek ^ajuat must meet in the root of ^a^aa-Xo?. This combined form is therefore the Pelasgo-Sclavonic original, and as such we recognise it in the kethuma of the Cervetri inscription. Now this again is a near approximation to the Hebrew HDl^. (ti) The Roman use of regio, dirigo, &c., in reference to road-making, is the best explanation of the obvious connexion between the Rus- sian doroga and the Hebrew TfTT, in which the initial dental must be explained in the same way as that in $pw = /BXeVo), a-Opeco, &c., compared with o-paa) and the Hebrew ilN*"1 (Maskil le- Sopher, p. 38): for we have in Greek r-pe^ and S-pafw (^oaTr-er^s) by the side of o-peyw, and e-p^o-fjiai. These ex- amples might be extended to any limit : but they are sufficient to 24.] THE ANCIENT ITALIANS. 77 show how permanently the stamp of a Sclavonian origin and consequent Semitic affinity was impressed even on the composite Latin language. And this will enhance the interest with which the philosophical ethnographer must always regard the desperate struggle for empire between the Romans, as the ultimate repre- sentatives of Pelasgian Italy, and that great Punic colony, which maintained a Semitic language and Semitic civilization on the south coast of the Mediterranean, CHAPTER III. THE UMBRIAN LANGUAGE AS EXHIBITED IN THE EUGUBINE TABLES. 1. The Eugubine Tables. 2. Peculiarities by which the old Italian alphabets were distinguished. 3. The sibilants. 4. Some remarks on the other letters. 5. Umbrian grammatical forms. 6. Selections from the Eugubine Tables, with explanations: Tab. I. a, 1. 7. Tab. I. a, 2-6. 8. Tab. I. b. 13, sqq. 9. Extracts from the Litany in Tab. VI. a. 10. Umbrian words which ap- proximate to their Latin synonyms. 11. The Todi inscription contains four words of the same class. 1. The Eugubine Tables. FROM the preceding investigations it appears that the original inhabitants of ancient Italy may be divided into three classes. It is not necessary to speak here of the Celts, who formed the substratum in all the insular and peninsular districts of Europe, or of the Greeks, who colonized part of the country ; but con- fining our attention to the more important ingredients of the population, we find only three Sclavonians, Lithuanians or Scla- vonized Goths, and pure Goths or Low-Germans. To the first belonged the various ramifications of the Pelasgian race ; to the second, the Umbrians, Oscans, and, the connecting link between them, the Sabines ; to the third, the Etruscans or Rasena, as dis- tinguished from the Tyrrhenians. The next step will be to examine in detail some of the frag- mentary remains of the languages spoken by these ancient tribes. The Umbrian claims the precedence, not only on account of the copiousness and importance of the relics of the language, but also because the Umbrians must be considered as the most important and original of all those ancient Italian tribes with whom the Pelasgians became intermixed either as conquerors or as vassals. The Eugubine Tables, which contain a living specimen of the Umbrian language, were discovered in the year 1444 in a sub- terraneous chamber at La Schieggia, in the neighbourhood of the ancient city of Iguvium (now Gubbio or Ugubio), which lay at the foot of the Apennines, near the via Flaminia (Plin. H. N. XXIII. 49). On the mountain, which commanded the city, stood the temple of Jupiter Apenninus ; and from its connexion with the 1 J UMBRIAN LANGUAGE IN THE EUGUBINE TABLES. 79 worship of this deity the city derived its name: Iguvium, Umbr. liovium, i. e. lovium, A?OJ', Ato? 77-0X19- The Tables, which are seven in number, and are in perfect preservation, relate chiefly to matters of religion. From the change of s in those of the Tables which are written in the Etruscan or Umbrian character, into r in those which are engraved in Roman letters, Lepsius infers (de Tabb. Eugub. p. 86, sqq.) that the former were written not later than A.U.C. 400 ; for it appears that even in proper names the original s began to be changed into r about A.U.C. 400 (see Cic. ad Famil. IX. 21. comp. Liv. III. cap. 4, 8. Pompon. in Digg. I. 2, 2, 36. Schneider, Lat. Gr. I. 1, p. 341, note); and it is reasonable to suppose that the same change took place at a still earlier period in common words. By a similar argu- ment, derived chiefly from the insertion of h between two vowels in the Tabulae Latine scriptce, Lepsius infers (p. 93) that these were written about the middle of the sixth century A.U.C., i. e. at least two centuries after the Tabulce Umbrice scriptce. But here I think he is mistaken : for the etymology of the words shows that the longer forms must have been more ancient than their abbreviations. And, in general, it is not very consistent with scientific philology to speak of an arbitrary distractio voca- lium, when we are surprised by the appearance of an elongated syllable. 2. Peculiarities by which the old Italian Alphabets were distinguished. Before, however, we turn our attention to these Tables and the forms of words which are found in them, it will be advisable to make a few remarks on the alphabet which was used in ancient Italy. The general facts with regard to the adaptation of the Semitic alphabet to express the sounds of the Pelasgian language have been discussed elsewhere 1 . It has there been shown that the original sixteen characters of the Semitic syllabarium were the following twelve : 1 N. Crat. 100. 80 THE UMBEIAN LANGUAGE [On. III. Breathings. Labials. Palatals. Dentals. Medials. Aspirates. Tenues. X'k 16 Iff Id ru fel n) for orum, liki-tud for lice-to, kvaisstur for qucestor, &c. The Umbrians and Oscans distinguished between u and v. The latter was a consonant, and was probably pronounced like our w. It was written as a consonant after K ; but the vowel u was preferred, as in Latin, after Q. The letters L and B were of rare occurrence in the Umbrian language. The former never stands at the beginning of a word, the latter never at the end of one. In the Oscan language we meet with L more frequently. As the Etruscan alphabet had no medials, those of the Eugu- bine Tables which are written in Etruscan characters substitute K for G, e. g. Krapuvi for Gr above. But the Oscan and Um- brian inscriptions when written in Latin characters distinguish between the tenuis and medial gutturals, according to the marks introduced by Sp. Carvilius, viz. c, G. In the Oscan alphabet D is represented as an inverted R ; and the affinity between these letters in the Latin language is well known. The labial P, which never terminates a word in Latin, stands at the end of many mutilated forms both in Umbrian and Oscan, as in the Umbrian vitlup for vitulibus (vitulis), and the Oscan nep for neque. In general, it is to be remarked that the letters p, F, R, s, D, and T, all occur as terminations of Umbrian or Oscan words. 5. Umbrian Grammatical Forms. The grammatical forms of the Umbrian language are very instructive. In Umbrian we see the secondary letter r, that im- portant element in the formation of Latin words, not only regu- larly used in the formation of the cases and numbers of nouns which in Latin retain their original s, but also appearing in plural verb-forms by the side of the primitive s, which is retained in the singular, though the Latin has substituted the r in both 62 THE UMBRIAN LANGUAGE [Cn. III. numbers. The following are the three declensions o Umbrian nouns, according to the scheme given by Aufrecht and Kirchhoff (Umbr. Sprachdenkm. pp. 115, sqq. ; see also Miiller, Getting. Gel. Anz. 1838, p. 58) : I. DECL. Sing. Norn. Gen. Dat. Accus. Abl. 1. Locat. 2. Locat. 3. Locat. Plur. Norn. Gen. Dat.j Abl.J Accus. 1. Locat. 2. Locat. Tuta, a city. tuta, tutu. tuta-s, tutar. tute. tutam. tuta. tutamem. tutemem. tute. tutas, tutar. tutarum. tutes. tutaf. tutafem. tutere? II. DECL. Puplus, a people. puplus. puple-s, pupler. puple. puplu-m. puplu. puplumem. puplus. puplum. puples. pupluf. puplufem. puplere? III. DECL. Sing. Norn. Gen. Dat. Accus. Abl. Locat. Plur. Norn. Gen. Dat.) Abl.J Accus. Locat. Ucri-s, a mountain. ucar. ucres. ucre. ucrem. ucri. ucremem. ucres. ucrium ? ucres. ucref. ucrefem ? Nume, a name. numen. numnes. numne.' numen. numne. numenem? numena ? numenum? numnes? numena? numenem? The Umbrian pronouns are the demonstratives eso, or ero 9 and esto, corresponding to the Latin is and iste, and the relative or interrogative poe, corresponding to the labial element in qui and quis. The demonstratives are generally construed as adjec- tives ; but, with the affix -hunt or -&, ero may become substantive. 5.] IN THE EUGUBINE TABLES. 85 Thus we have er-ont, or ere-Jc, as an indicative pronoun. The affix -Jc is that which plays so important a part in Latin. The affix -hunt or -hont (Goth. Mndana, Etrusc. hinthiu or hintha) appears in the comparative and superlative adverbs Jmnt-ra or hond-ra, (Goth, hindar, 0. N. hindra), and hond-omu, Goth. hindumist, signifying "farther," "lower," or "farthest," "lowest;" so that hond may correspond to our yon or yonder : and as k expresses proximity, ere-k and er-ont will gain the meaning of " here " and " there," from their terminations respectively ; so that esu-k, es-tu, and er-ont, may have corresponded in distinctive meaning to the Latin hie, iste, Hie, the first part being the same in each, and identical with the initial syllable of is-te. The verbs generally occur in the imperative mood, as might be expected, since the Tables contain chiefly prayers and injunc- tions about praying. In these imperatives we mostly recognise a singular in -tu, and a plural in -tutu ; &sfu-tu (VI. a, 30, &c.), smdfu-tutu (VI. b, 61), corresponding to es-to, es-tote. Verbs of the -a conjugation seem occasionally to make their imperative in -a, like the Latin. See I. b, 33 : pune purtinsus, karetu ; pufe apruf fakurent, puze erus tera ; ape erus terust, pustru kupifiatu : where, though the meaning of particular words may be doubtful, the construction is plain enough : postquam por- rexeris, calato ; ubi apros fecerint, uti preces det ; quando preces dederit, poster o (= retro} conspicito. We often have the perf. subj. both singular and plural, as may be seen in the ex- ample just quoted. The pres. subj. too occasionally appears, the person-ending in the singular being generally omitted, as in arsie for arsies = ad-sies, and habia for habeas. The Oscan infinitive in wn, as a-ferum = circum-ferre, is also used in Umbrian ; and we often find the auxiliary perfect both in the singular and in the plural. See VI. b, 30 : perse touer peskier vasetom est, pese- tom est, peretum est, frosetom est, daetom est, touer peskier virseto avirseto vas est : i. e. quod tui sacrificii vacatum est, peccatum est, neglectum est, rejectum est, projectum est, tui sacrificii visa invisa vacatio est *. And we have not only sJcrehto 1 It seems that vas must be the root of vas-etom, and probably both refer to the evacuation or nullification of the sacrifice; cf. vas-tus, &c. with the Greek eWei/ow: virseto avirseto is compared vrith Cato's "ut tu morbos visos invisosque prohibessis" (R. R. 141). 86 THE UMBKIAN LANGUAGE [On. III. est, but also skreihtor sent (VI. a, 15). The active participle seems to end both in -ens, like the Latin, and also in -is, like that of the Greek verbs in -^i. The following are the forms of sum, fui, and habeo which are found in the Tables . SUM (root ES). Fu-. PRES. INDIC. (A. I.) 3. sing. est. 3. plur. sent. PRES. SUBJ. (A. III.) 2. sing, sir, si, sei, sie. 3. sing. si. 3. plur. sins. fuia. PERF. SUBJ. (C. III.) 3. sing, fuiest, fust. 3. plur. furent. IMPER. (B. I.) 2, 3. sing. futu. 2. plur. fututo. INFIN. (D.) eru or erom, (V. 26, 29, VII. b, 2.) HABEO. PRES. INDIC. (A. I.) 3. sing. habe[t] (I. b, 18 ; VI. b, 54). PRES. SUBJ. (C. I.) 2. sing. habia[s] (V. a, 17). PERF. SUBJ. (C. III.) 2. sing, habiest (VI. b, 50) ; habus (habueris) (VI. b, 40). 3. plur. haburent (VII. a, 52). IMPERAT. (B.) 2. sing, habitu (VI. a, 19) ; or habetu (II. a, 23). 2. plur. habituto (VI. b, 51); or habetutu (I. b, 15). 6. Selections from the Eugubine Tables, with explanations. In interpreting the remains of the Umbrian language, it seems advisable, in the present state of our knowledge, that we should confine our attention to those passages which fall within 6.] IN THE EUGUBINE TABLES. 87 the reach of a scientific philological examination. Grotefend 1 , indeed, has frankly and boldly presented us with a Latin version of all the Eugubine Tables ; but although he has here and there fallen upon some happy conjectures, his performance is for the most part mere guesswork of the vaguest kind, and therefore, for all purposes of scholarship, uninstructive and unsatisfactory. Lassen, by attempting less, has really effected more 2 . There is, however, no one who has done more to prepare the way for a scientific examination of these Umbrian documents than Lepsius, who examined all the preliminary questions connected with the subject in an inaugural dissertation published in 1833, 3 and who has subsequently edited a most accurate collection of facsimiles, which appeared in 1841. 4 The materials furnished by Lepsius have been elaborately discussed in a special work by Aufrecht and Kirchhoff, published in 1849 ; 5 and though this treatise is defective in arrangement and inconvenient for purposes of re- ference, it deserves the praise of never attempting too much, and it is generally distinguished by a careful regard for the principles of sound philology. The following extracts are selected from the admirable transcripts of Lepsius 6 , and the arrangement of the Tables is that which he has adopted. The first four Tables, and part of 1 Rudimenta Linguce Umbricce, Particulse VIII. Hannov. 1835-1839. 2 Beitr'dge zur Deutung der EugubiniscJien Tafeln, in the RJiein. Mus. for 1833, 4. Of earlier interpretations it is scarcely necessary to speak. It may, however, amuse the reader to know that the recent attempt of a worthy herald, in the sister-island, to Drove that Irish of a certain kind was spoken by the ancient Umbrians and Tuscans, has its parallel in a book published at Ypres in 1614, by Adriaen Schrieck, who finds the ancient language of his own country in the seventh Eugubine Table! (Van 't Begliin der eerster Volcken van Europen, t'Ypre, 1614). The Irish book, however, is the more elaborately ridiculous of the two. It has been exposed, with considerable ability and humour, in the Quarterly Revieiv, Vol. LXXVI. pp. 45, sqq. 3 De Tdbulis EuguUnis. Berolini, 1833. * Inscriptiones Umbricce et Oscce. Lips. 1841. 5 Die Umbrischen Sprachdenkmaler ; ein VersucJi zur Deutung derselben. Berlin, 1849. 6 In citing the edition of Lepsius as now constituting the standard text, we must not forget the excellence of Bonarota's transcriptions, to which Lepsius himself has borne testimony. De Tabb. Eug. p. 14. 88 THE UMBRIAN LANGUAGE [On. III. the fifth, are written in the Etruscan or Umbrian character. The others are in Latin letters. Tab. I. a, 1. This Table and its reverse contain the rules for twelve sacrifices to be performed by the Fratres Atiersii in honour of the twelve gods. The same rules are given in Tables VI. and VII. and in nearly the same words, the differences being merely dialectical ; but the latter Tables add the liturgy to be used on the occasion, and also dwell at greater length on the auguries to be employed, &c. The first Table begins as follows: Este persMum aves anzeriates enetu, 2. pernaies pusnaes. And in VI. a, 11, we have : Este persklo aveis aseriater enetu. There can be little doubt as to the meaning of these words. Este, which is of constant recurrence in the Tables, is the Umbrian adverb corresponding to ita, which is only a weaker form of it. If we may infer that persklum or persklo pre^- culum, we may render this word "a prayer." Grotefend de- rives the noun from pur go, and translates it by " lustrum" But pur-go is a compound of purus and ago (comp. castigo, &c.), whereas the root pers-, signifying " pray," is of constant occur- rence in Umbrian ; and every one, however slightly conversant with etymology, understands the metathesis in a case of this kind. It is the same root as prec- or proc- in Lat., pere$- in Zend, practi- in Sanscr., frag- en or forsch-en in Germ., &c. It is clear that aves anzeriates or aveis aseriater are ab- latives absolute. As we have avif seritu or aseriatu (VI. b, 48, 49. I. b, 11, &c.) by the side of salvam seritu (VI. a, 51, &c.), and as this last is manifestly salvam servato, it is pretty clear that aves anzeriates must be equivalent to avibus observatis (= in-servatis). Enetu is clearly the imperative of ineo, for in-ito ; the pre- position had the form en = in in old Latin ; thus we find in the Columna Rostrata : enque eodem macistratod : and the same was the case in Oscan, which gives us em-bratur for im-perator. The adjectives per-naies, pus-naes, are derived from per-ne, post-ne, which are locative forms of the prepositions prce and post, and signify "at the southern and northern side of the temple." The birds are so defined with reference to the practice 6.] IN THE EUGUBINE TABLES. 89 of the augurs in such cases. See Yarro, L. L. VII. 7, p. 119, Muller : " quocirca coelum, qua attuimur, dictum templum Ejus templi partes iv. dicuntur, sinistra ab oriente, dextra ab occasu, antica ad meridiem, postica ad septentrionem." The meaning of the whole passage will therefore be: Ita litationem avibus observatis inito anticis, posticis ; i.e. "Thus enter upon the supplication, the birds having been observed, those in the south, as well as those in the north." 7. Tab. L a, 2 6. Tab. I. a, 2. Pre-veres treplanes, 3. luve Krapuvi tre[f] buf fetu, arma ustentu, 4. vatuva ferine feitu, lieris vinu, heri[_s] puni, 5. ukriper Fisiu, tutaper Ikuvina, feitu sevum, 6. kutef pesnimu ; afepes arves. Coinp. VI. a, 22. Pre-vereir treblaneir luue Grabovei buf treiffetu. VI. b, 1. Arviofetu, uatuo ferine fetu, poni fetu, 3. okriper Fisiu, totaper liovina. The words pre-veres (vereir) treplanes (treblaneir) are easily explained in connexion with (7) pus-veres treplanes, (11) pre- veres tesenakes, (14) pus-veres tesenakes, (20) pre-veres veliiies, (24) pus-veres vehiies. It is obvious that these passages begin with the prepositions pre, " before," and pus -post, " behind," and that they fix a locality. The prepositions per, signifying " for," and co or ku, signifying " with" or " at," are placed after the word which they govern : thus we have tuta-per Ikuvina = " pro urbe Iguvina" vocu-com loviu "cum" or "infoco Jovio^ But the prepositions pre and pus precede, and it seems that they both govern the ablative, contrary to the Latin usage, which places an accus. after ante and post. The word veres (vereir) is the abl. plur. of a noun verus (cf. I. b, 9), corresponding in root and signification to the Latin fores. Compare also porta with the German Pforte. The v answers to the/, as vocus, vas, &c. for focus, fas, &c. Lassen (Rhein. Mus. 1833, pp. 380, sqq.) refers treplanes, tesenakes, vehiies, to the numerals tres, decem, and viginti. Grotefend, more pro- bably, understands the adjectives as describing the carriages 90 THE UMBRIAN LANGUAGE [Cn.III. used at the particular feasts. Cato (R. R. c. 135) mentions the trebla as a rustic carriage. Tensa is the well-known name of the sumptuous processional chariot in which the images of the gods were carried to the pulvinar at the ludi Circenses (Festus, p. 364, Miiller) 1 ; and veia was the Oscan synonym for plan- strum (Festus, p. 368, Mliller). It is, therefore, not unreason- able to suppose, that the fores treblance furnished an entrance to the Ocris or citadel for treblce ; that through the fores tesenakes the statues of the gods were conveyed to their pulvinar in tensce ; and that the fores vehice allowed the larger chariots to enter in triumphal or festive procession. In the Latin Table the adj. derived from tesnaor tensa ends in -ox, -ocis, like velox; in the Umbrian it ends in -ax, -ads, like capax. Aufrecht and Kirchhoff, to whom the true explanation of verus is due, sup- pose a quadrangular citadel with one side closed, and the other three opening with gates called by the names of the cities to which they led. But this mode of designation is not borne out by the names of the three gates, if there were only three, in the Roma Quadrata on the Palatine. These gates were called the Porta Romanula, Janualis, and Mucionis, and lay to the W., N.W., and N. (Muller, Etrusk. II. p. 147). Whatever the names meant, it is clear that they are not designations of towns to which the gates led. As there were no cities called Trebla and Tesena, and as Veil was too far off to give a name to one of the gates of Iguvium, it is much more reasonable to suppose that the entrances refer to the names of carriages with which they are so easily identified. To say nothing of the analogy of the French porte cochere, which actually denotes une porte assez grande pour donner entree aux coches ou voitures, it is well known that the ancients measured road-ways by the kind of carriages which traversed them, or by the number of such carriages which could pass abreast. Thus we have o$o? diu.a%iTos for a wide road (Find. -ZV. VI. 56) ; d^a^iro^ alone is used in the same sense (id. P. IV. 247) ; and Thucydides defines the breadth of a wall by saying that : ovo ctju.a%ai evavTiat aAA^Aais TOI)S \iOovi (I. 93). 1 For the metathesis tesna or tesena for tensa we may compare mesene jftusare in an inscription found near Amiternum (Leps. Tab. XXVII. 4 6 with menseflusare in the Latin inscription quoted by Muratori (p. 587). $ 7.] IN THE EUGUBINE TABLES. 91 The epithet Krapuvius, or in the Latin Table Gra-bovius, according to Lassen signifies " nourisher or feeder of cattle." The first syllable, he supposes, contains the root gra-, implying growth and nourishment, and found in the Sanscr. grd-ma (signifying either "a herd of feeding cattle" grex or vicus inter pascud), in the Lat. gra-men, in the Goth, gras, and in the Old Norse groa = virescere. Lassen, too, suggests that Gradivus contains the same root. This comparison ought perhaps to have led him to the true explanation of both words. For it is manifest that Gra-divus = grams or grandis Divus ; and it is equally certain that no genuine Latin compound begins with a verbal root. If, therefore, Gra-bovius contains the root of bos, bovis, the first syllable must be the element of the adjective gravis or grandis ; so that Grabovius will be a compound of the same kind as /caX- \i7rap9evos (see Lobeck, Paralip. p. 372). Pott, however, (Et. Forsch. II. p. 201) considers Grab-ovius as another form of Gravi-Jovius. Tre or treif buf is either boves tres or bobus tribus. If we have here the accus. plural, we must conclude that this case in the Umbrian language ends in -af, -of, -uf, -ef, -if, -eif, according to the stem ; and the labial termination has been compared with the Sanscrit and Zend change of s into u at the end of a word (Wilkins, 51. Bopp, 76). This is the opinion of Lassen (Rhein. Mus. 1833, p. 377). According to Lepsius and Grote- fend, on the other hand, all these words are ablatives, because the termination is more easily explained on this hypothesis, and because verbs signifying " to sacrifice " are construed with the ablative in good Latin (Virg. Eclog. III. 77. Hor. Carm. I. 4, 11). The latter reason is confuted by the tables themselves ; for it is quite clear that abrons is an accusative, like the Gothic vulfans, and yet we have both abrons fakurent (VII. a, 43) and abrof fetu (VII. a, 3). See also Pott, Et. Forsch. II. p. 202. With regard to the form, it is not explained by the Sanscrit ana- logies cited by Lassen, for these spring from the visargdh after a, as in Ramah, Ramau, Ramo. There is a much simpler way of bringing abrof and abrons into harmony. For the plural is formed from the singular by adding s to the latter. If then the accusative singular assumed the form n from m, this would be retained before s, as in abron-s ; but if abrom-s passed by visar- gdh into abrom-h, this, according to the Celtic articulation, would 92 THE UMBRIAN LANGUAGE [Cte. III. regularly become abrof; for in Celtic mh and bh are regularly- changed into v-f. And we have seen above (p. 63) very good reasons for recognising Celtic influences in Umbria. Feitu (fetu) is simply facito, the guttural being softened down, as in ditu for dicito (VI. b, 10, &C.) 1 . Arvia seems to be the same as the Latin arvina, i. e. " the hard fat which lies between the skin and the flesh " (Servius ad Virg. ^En. VII. 627); and ustentu is probably obstineto, which was the old Latin for ostendito (Festus, p. 197, Mull.). Vatuva ferine feitu must mean " offer up unsalted meal" (fatuam farinam or fatua farina), according to Nonius Mar- cellus, IV. 291 (quoting Varro, de Vit. Pop. Rom. Lib. I.): quod calend. Jun. et publice et privatim fatuam pultem diis mactat. Grotefend supposes that ferine must mean raw flesh, and not /anna, because "bread" (puni) is mentioned in the pas- sage. But in minute directions like these, a difference would be marked between the meal (aXevpa) and the bread (a^oros); just as the hard fat (arvina) is distinguished from the soft fat (adi- pes), if the interpretation suggested below is to be admitted. Heris vinu, Tieris puni, "either with bread or wine." Heris, as a particle of choice, is derived from the Sanscr. root hri, " to take ;" Lat. Mr, " a hand," &c. ; and may be compared with vel, which is connected with the root of volo, as this is with the root of aipeco. In fact, heris appears to be the parti- ciple of the verb, of which the imperative is lieritu (VI. a, 27, &c.). This verb occurs in the Oscan also ( Tab. Bantin. 12, &c.). That ocriper (ucriper) Fisiu means " for the Fisian mount" may be demonstrated from Festus, p. 181, Miiller : " Ocrem antiqui, ut Ateius philologus in libro Glossematorum refert, montem confragosum vocabant, ut aput Livium : Sed qui sunt hi, qui ascendunt altum ocrim ? et : celsosque ocris, arvaque putria et mare magnum, et : namque Tcenari celsos ocris. et : haut ut quern Chiro in Pelio docuit ocri. Unde fortasse etiam ocreae sint dicts3 insequaliter tuberatse." From this word are derived the names of some Umbrian towns, e. g. Ocriculum and Interocrea (cf. Interamna}. The epithet Fisius indicates that the mountain was dedicated to the god Fisius or Fisovius Sansius (Fidius Sancus), a name under which the old Italians According to Pott and Lepsius this imperative stands for /to =fiat. . 7.] IN THE EUGUBINE TABLES. 93 worshipped Jupiter in their mountain-temples. Lassen (p. 388) refers to this temple the following lines of Claudian (de VI. Cons. Honor. 503, 4) : Exsuperans delubra lovis, saxoque minantes Apenninigenis cultas pastoribus aras. He also quotes from the Peutinger inscription : " Jo vis Penninus, idem Agubio," where Iguvium is obviously referred to. Lepsius thinks that ocris Fisius was the citadel of Iguvium. Tota-per (tuta-per) Ikuvina, " for the city of Iguvium." It was always understood by previous interpreters that tuta or tota was nothing more than the fern, of the Lat. totus. But Lepsius has clearly proved that it is both an Oscan and an Umbrian substantive, signifying " city," from which the adj. tuti-cus is derived, as in the name of the magistrate meddix tuticus, i. e. consul urbanus : consequently tuta-per Ikuvina is simply <( pro urbe Iguvina." This substantive, tota or tuta, is, no doubt, derived from the adject, totus ; for the idea of a city is that of " fulness," " collection," " entirety." Similarly, the Greek 71-0X19 must contain the root TTO\- (TTO\-VS) or TrXe- (7rXeo$)> signifying the aggregation of the inhabitants in one spot. The derivation of the adjective to-tus is by no means easy ; but if we compare it with in-vi-tus (from vel-le), we may be disposed to connect it with the root of the words tel-lus, tol-lo, ter-ra, ter-minus (re'X-os, Tep-/ma), &c. l Op-pidum, an- other name for "city," is only "a plain" (ob-ped-um 67ri- ireSov)', and oppido, " entirely" = in toto,is synonymous with plane. The student will take care not to confuse between this to-tus and the reduplicated form to-tus (comp. to-t- t quo-tus, &c.), which is sufficiently distinguished from it in the line of Lucretius (VI. 652) : Nee tota pars homo terra'i quota totiw unus. Sevum and kutef are two adverbs. The former signifies " with reverence," and contains the root sev- (sev-erus) or o-e/3- (a-efico) 2 . The latter is derived from cav-eo, cautus, with the affix -f= W /my OVTWS e^eiv : and the same principle may be applied to explain ofy ijKicrrci, ov yap a//eti>o*>, &c. In a case like this the Romans seem to have used nee as qualifying and converting the whole word, in preference to non. Muller supposes that negritu t quoted by Festus (p. 165) as signifying cegritudo in augurial language, stands for nec-ritu. I think it must be a corruption for ne-gritu\do'] : see below, Ch. VII. 5. Heritu is the imper. of hri, " to take away," Sanscrit hri = caper e, toller e^ demere, auferre, rapere, abripere, Welsh hwra. The whole passage then may be rendered : J. Gr. precor precatione, quoniam in ocri Fisio ignis ortus est, in urbe Iguvina sacerdotes dissecantes submissi sunt, ita ne tu adimas. 1 Prof. Newman (Regal Rome, p. 26) says that neg-ligo is to be com- pared with nach-lassen, and exhibits the German nach " after " a particle unknown to Latin. I believe he is not responsible for this puerile deri- vation, which evinces a complete ignorance of the part which nee or neg plays in Latin words, and of the connexion of this particle with nach. We shall see when we come to the Etruscan language that nak occurs in an inscription with the sense " in " or " down in ;" and in this or a similar sense na or nach is used in all the Sclavonian and German dialects to say nothing of po-ne, si-ne, &c. in Latin. The guttural at the end of ou-F, ou-^i, does not differ from that in ne-c, ne-que ; and as the Sanscrit ava-k, which is obviously connected with the Greek OV-K = va-fa-K (New Crat. 189) signifies deorsum, we can easily reconcile the different signi- fications of these particles. 10.] IN THE EUGUBINE TABLES. 99 10. Umbrian words which approximate to their Latin synonyms. This may suffice as far as the direct interpretation of the Tables is concerned. In conclusion, it may be well to give a list of those words in the Umbrian language which approach most closely to their Latin equivalents. And first, with respect to the numerals, which are the least mutable elements in every language, it is clear that tuves (duves), tuva (duva), and tris, treia, correspond to duo and tres, tria. Similarly tupler (dupler) and tripler represent duplus and triplus, and tuplak (III. 14) is duplice. It is obvious, too, that petur is " four," as in Oscan ; see VI. b, 10 : du-pursus, petur-pursus, i. e. bifariam, quadrifariam. As to the ordinals, prumum is pri- mum, etre (etrama) is alter, and tertie (tertiama) is tertius. The other words may be given in alphabetical order : Abrof(apruf) (VII. a,3)=apros. Ager (Tab. XXVII. 21). Ahes-no (III. 8, \$) = ahenus. Alfu (I. b, Z9)=albus (aA0oY). Amb-, prefix. Ander (anter) (VI. b, 47- I. b, 8) = inter (sim. in Oscan). Angla or ankla (VI. a, \)=aquila (comp. anguis with e^t?, unda with uSwjo, &c., see New Crat., p. 303). Anglome (VI. a, 9) = angulus. An-tentu (passim) = in-tendito. Ar-fertur (VI. a, 3) = affertor. Arputrati (V. a, 12) = arbitratu. Ar-veitu (I. b, 6) = advehito (cf. arms and arms). Asa (VI. a, 9, et passim) = ara. Asiane (I. a, 25) = Asiano. Atru (I. b, 29) = ater. Aveis (VI. a, 1) = avibus, &c. Benes (I. b, 50) = venies. Fakust (IV. 31) =fecerit. Bue (VI. a, 26, et passim) = bove. Famerias Pumper-las (VIII. a, 2) Cesna (V. b, 9) = ccena. = families Pompilice. 72 Der-sikurent (VI. b, 62) caverint. Der or ter, later ders or dirs, from deda, a reduplicated form of da = dare. It is sometimes found under the forms duve or tuve, especially in composition with pur, as in pur-tuvi-tu =pro-dito orpor-ricito (II. a, 24). Dekuria or tekuria (II. b, 1) = decuria, i. e. decu-viria. Destru or ^s^rw (I. a, 29) = dexter. Dife or ft'p* (II. a, 17) = decere. Ditu (VI. b, 10) = dicito. Dupla (VI. b, 18), so also numer tupler (V. a, 19) comp. numer prever (V. a, 18) and numer tripler (V. a, 21). (VII. b, 2) = 100 THE UMBRIAN LANGUAGE [On. IIL Far (V. b, 10) =far. Fato (VI. b, ll)=/a Feraklu (Miiller, EtrusJc. I. p. 57, note) ferculum. Ferehtru (III. 16) =feretrum. Ferine (I. a, 4) = farina. Fertu (VI. b, 50) = Crater (V. b, 11). .Fos (VI. a, 23) = Funtler e (I. b, 24) =fontulo. Habetu (II. a, 23) = Aafofo. Here = velle, connected with " the hand/' pre-HExv-o, &c. (A^w? Oa. 162) ; hence /&0H = vel (I. a, 22) ; also in the sense of taking away, &c. like the Sanscr. hri, "Welsh hwra (above p. 98). Homonus (V. b, 10) = homines. Ife (II. b, 12) ibi. Jvenka (I. b, 40) =juvenca. Kanetu (IV. 29) = canito. Kapire (I. a, 29) = ca/wfe, " with a sacrificial jug." Kaprwm (II. a, 1). Karetu (I. b, 33) * cfofo. Karne (II. a, 1). Kastruo (VI. a, 30, et passim) = castra. Katlo (II. a, 38)=catulus. Komohota (VI. a, 54) = commota. Kuratu (V. a, 24) : s#0 r comix. Kvestur (V. a, 23) = qucestor. Maletu (II. a, 18) = molito. Manu (II. a, 32) = manus. Mehe (VI. a, 5) = mihi. Mestru (V. a, 24) = magister v. major. Mugatu (VI. a, 6) = mugito. Muneklu (V. a, \J)=munusculum. Muta (V. b, 2) = Naratu (II. a, 8) =narrato (Varro wrote narare). Ner (VI. a, 30, &c.) =princeps. Nome (passim) = nomen. No-sve (VI. b, 54) = nisi. Numer (V. a, IJ) = numerus. Numo (V. a, 17) = numus. Orer (VI. a, 26) = oro, ev-^o^ai. Orto (VI. a, 26) = ortus. Ose (VI. a, 26) = ore. Ostendu (VI. a, 20) = ostendo. Oui (VI. b, 43), uve (II. 6, 10) = ovis. Pose (VI. a, 30)=^a^. Pa^r (II. a, 24). Peiko (VI. a, 3) =picus. Peku (VI. a, 30) =^cw5. Pelsana (I. a, 26) = balsamon. Persnimu (I. b, 7) = precator. Pihakler (V. a, 8) piaculum. Pihatu (VI. a, 9) =piato. Pir (I. b, 12) = P ,jire. Plenasio (V. a, < 2)=plenarius. Poplo (passim) = populus. Porka (VII. a, 6) = porca. Post ; postro (VI. b, 5) = postero, i. e. retro. Prokanurent (VI. a, 16) = pro- cinerint. Proseseto (VI. a, 56) = prosecato. Puemune (III. 26) pomona. Puprike (III. 27) =pullice. Pur-tin-sus (I. b, 33) pro-ten^ deris. Pustertiu (I. b, 40) = post-tertio. Rehte (V. a, 24) = recte. Ri (V. a, 6) = res. Ruphra (I. b, 27) = rubra. Sakra (I. b, 29). Salvo, salva, &c. (passim). Seritu (passim) = servato (Miiller, Etrusk^ I. p. 55). Sif(l. a, 7) = sues. 10.] IN THE EUGUBINE TABLES. 101 11. Tio (passim) = te. Uretu (III. 12) = urito. Urnasier (V. a, 2) = urnarius. Umkum (III. 28) = cwm ove. Vas (VI. a, 28) = vas-tus. Vatuva (I. a, 4) =fatua. Veiro (VI. a, 30) = viros. Veru (passim) = fores. Vestra (V. b, 61). Vinu (passim) = mnum. Virseto (VI. a, 28) = visus. Vitlu (II. a, 2l) = vitulus. Voku-kom (VI. b, 43) = cum vel infoco. Vutu (II. b, 39) = vultus. The Todi Inscription contains four words of the same class. Skrehto (VII. b, 3) = scriptus. Sopo (VI. b, 5) = sapone. Stahitu (VI. b, 56) = state. Stmsla (VI. a, 59) = stru-cula, diinin. of strues. Subator (VI. a, 27, &c.) = subactl Suloko (VI. a, 22, &c.) =sul-voco. &ubra (V. a, 20) = supra. Sve (V. a, 24) = Osc. swo?, Lat. si. Seritu (II. b, 24), vide seritu* Sesna (V. b, 9) = cesna, coena. Tafle(ll. a, 12) = tabula. Tases (VI. a, 55) = fac0HS. Tekuries (II. a, 1) = decurias. Termnu-Jco (VI. b, 53) In the year 1835 a bronze figure of a man in armour was discovered near Todi ( Tude,r\ on the borders of Umbria. The inscription, which was detected on the girdle of the breast-plate, has been interpreted from the Greek, Latin, and Hebrew lan- guages by a number of different scholars. It appears to me to contain four words, which may be added to the above list, as they are all explicable from the roots of the Latin language. The inscription runs thus : AHALTRVTITISPVNVMPEPE. The word titis occurs in the Eugubine Tables (I. b, 45), and punum is obviously the accusative of punus, another form of pune, punes, puni, which are known to be Uinbrian words. It is true that the Latin synonym panis and the Eugubine words belong to the i- declension ; but that is no reason why we should not have a by-form of the o- declension, and that this form actually existed in Messapia is well known (Athen. III. p. Ill c.: Travos apros MecrcraViot). These two words being removed from the middle, the extremities remain, namely, ahaltru and pepe. With regard to the first it is to be observed that the lengthening of a syllable, by doubling the vowel and inserting the letter h, is common in Umbrian (see Leps. de Tabb. Eugub. pp. 92, sqq.), and the same practice is often remarked in Latin. 102 THE UMBRIAN LANGUAGE [On. III. Indeed, as we have seen above (p. 82), the elongated form is the more ancient and original. Ahaltru, then, bears the same relation to the Latin alter that ahala bears to ala, nihil to nil, vehemens to vemens, &c. It is true that in the Eugu- bine Tables etre seems to represent the meaning, if not the form of alter; but this is no reason why there should not be the other equally genuine and ancient form alter or ahalter, which is probably the more emphatic word in that language, and corresponds, perhaps, in meaning to the adjective alienus. The signification of the word pepe suggests itself from the context, and is also supported by analogy. It seems to be a reduplication of the root pa (pd-nis, pa-sco, TraadcrOai, Tra-reojuat, &c.), analogous to the reduplication of the root bi (or pi, Tri-vu), &c.) in bi-bo. If the Sabines were a warrior tribe of Umbrians, it is reasonable to conclude that their name for " a warrior" would be Umbrian also ; now we know that the Sabine name for " a warrior" was titus (Fest. p. 366, and above, p. 26), and the warrior tribe at Rome was called the Titienses (Liv. I. 13); accordingly, as the Umbrian Propertius calls these the Titles (EL IV. 1, 31 : Hinc Titles Ramnesque viri Luceresque coloni 1 ), it is not an unfair assumption that titis, pi. tities, was the Umbrian word for " a warrior." We have the same word on an Etruscan monument from Volterra, which re- presents a warrior with sword and spear, and bears the following legend : mi ajiles Tites (Inghirami Mon. Etr. ser. YI. tav. A. Micali Ant. Mon. tav. 51. Muller, Denkmaler, LXII. n. 312). The inscription, then, will run thus : " the warrior eats another's bread ;" the position of ahaltru being justified by the emphasis which naturally falls upon it. Compare Dante, Paradiso, XVII. 58-60 : Tu proverai si come sa di sale Lo pane altrui, et com* e duro calle Lo scendere e '1 salir per 1' altrui scale. This motto, then, either refers to the practice of serving as mercenaries, so common among the Italians, or expresses the prouder feeling of superiority to the mere agriculturist, which was equally characteristic of the oldest Greek warriors. Compare the scolion of Hybrias the Cretan (ap. Athen. XV. 695 F.) : Lucmo in V. 29 is an accurate transcription of the Etruscan Lauchme. 11.] IN THE EUGUBINE TABLES. 103 Trareco TOI/ aSuj/ oa/oj/ OTT* a Sfo-Troras pvcotas KK\rjp.ai. , K. r. X. It is also to be remarked that the Lucumones, or " illustrious nobles," among the Tuscans, seem to have distinguished their plebeians as Aruntes (apovvre^ i. e. mere ploughmen and agri- cultural labourers (Klenze, Phil. Abhandlung. p. 39, note). In general the praenomen Aruns seems to be used in the old mythi- cal history to designate an inferior person (Miiller, Etrusk. I. p. 405). CHAPTER IV. THE SABELLO-OSCAN LANGUAGE. 1. The remains of the Oscan language must be considered as Sabellian also. 2. Alphabetical list of Sabello-Oscan words, with their interpretation. 3. The Bantine Table. 4. Commentary on the Bantine Table. 5. The Cippus Abellanus. 6. The bronze tablet of Agnone. 7. The " Atellanae." 1. The remains of the Oscan language must be considered as Sabellian also. THE Oscan language is more interesting even than the Um- brian, and the remains which have come down to us are much more easily interpreted than the Eugubine Tables. Indeed, as Niebuhr has remarked (I. ad not. 212), " some of the inscrip- tions may be explained word for word, others in part at least, and that too with perfect certainty, and without any violence." This language had a literature of its own, and survived the Roman conquest of southern Italy. It was spoken in Samnium in the year 459 ; J it was one of the languages of Bruttium in the days of Ennius 2 ; the greatest relic of Oscan is the Bantine Table, which was probably engraved about the middle of the seventh century ; and the Oscan was the common idiom at Her- culaneum and Pompeii, when the volcano at once destroyed and preserved those cities. Although, as it has been shown in a previous chapter, the Sabines must be regarded as a branch of the Umbrian stock, who conquered all the Ausonian nations, and though Yarro 3 speaks of the Sabine language as different from the Oscan, yet, as all the remains of the Sabine and Oscan languages belong to a period when the Sabellian conquerors had mixed themselves up with the conquered Ausonians and had learned their language, it seems reasonable that we should not attempt, at this distance of time, 1 Liv. X. 20 : " gnaros llnguce Oscce exploratum mittit." 2 Festus, s. T. bilingues, p. 35 : " Ulingues Bruttates Ennius dixit, quod Brutti et Osce et Greece loqui soliti sint." s L. L. VII. 3, p. 130, Muller. Varro was born at Reate (see p. 301 of Muller's edition), and therefore, perhaps, attached peculiar importance to the provincialisms of the ager Sabinus. 1.] THE SABELLO-OSCAN LANGUAGE. 105 to discriminate between them, but that, recognising generally the original affinity of the Umbrian and Oscan nations, we should consider the Sabine words which have been transmitted to us, as belonging, not so much to the Umbrian idiom, as to the complex Sabello- Oscan language, which prevailed throughout the whole of southern Italy. And this view of the matter is farther justified by the fact, that a great many of these words are quoted, not only as Sabine, but also as Oscan. It is true that some parti- cular words are quoted as Sabine, which are not found in Oscan inscriptions, and not known to be Oscan also ; but we cannot form any general conclusions from such isolated phenomena, espe- cially as a great many of these words are Latin as well. All that it proves is simply this, that there were provincialisms in the Sabine territory properly so called. Still less can we think with Muller (Etrusk. I. p. 42), that the Sabine language is the un-Greek element in the Oscan ; for many of these words have direct connexions with Greek synonyms, as Muller himself has admitted. There are no Sabine inscriptions as such. The Mar- sian inscription, quoted by Lanzi, and which Niebuhr thought unintelligible (I. 105, ad not. 333), is Oscan, if it ought not rather to be called old Latin. In the following observations, then, for the materials of which I am largely indebted to the writings of Professor Klenze (Phi- lologische Abhandlungen, Berlin, 1839,) and of Theodor Momm- sen (Unteritalischen Dialekte, Leipsig, 1850), the Sabine and Oscan will be treated in conjunction with one another. Before proceeding to consider the Oscan inscriptions, it may be as well to give an alphabetical list of those words which are cited by old writers as Sabine, Oscan, or both. $ 2. Alphabetical list of Sabello- Oscan words, with their interpretation. Alpus, Sab. Fest. p. 4, Muller : " Album, quod nos dicimus, a Graeco, quod est d\^), "good" or "fair." As Kupra was the Etruscan Juno, (Strabo, p. 241), this word must have belonged to the Umbrian element common to both languages. Dalivus, Osc. Fest. p. 68 : " Dalivum supinum ait esse Aure- lius, Julius stultum. Oscorum quoque lingua significat insa- num. Santra vero dici putat ipsum, quern Graeci SeiXaiov, i. e. propter cujus fatuitatem quis misereri debeat." Comp. Hesych., ActXi?, fjLwpo^ ; and see Blomf. ad ^Esch. Eumen. 318. Labb. Gloss, daunum, a Kapav-ov ; Gael, earn ; Irish, cair- neach; Sclav, kremeni. Idus, Sab. Yarro, L. L. VI. 28 : " Idus ab eo quod Tusci itus, vel potius quod Sabini idus dicunt." Irpus, Sab. et Samn. Serv. ad JEn. XL 785 : "Nam lupi Sa- binorum lingua hirpi vocantur." Test. p. 106 : " Irpini appellati nomine lupi, quern irpum dicunt Samnites; eum enim ducem secuti agros occupavere." Strabo, V. p. 250 : e^fjs $ eicrlv 'IpTnvoi, Kavrol ^avvlraC TOVVO^U $ ecr^oy aTro TOV yyrjcraiJievov \VKOV r^s awoiKias' 'ipirov yap KCL\OVi>. The proper name Mettius (Fest. p. 158), or Mettus (Liv. I. 23), seems to have been this word Meddix. At least Livy says that Met- tus Fuffetius was made dictator of Alba ; and Festus speaks of Sthennius Mettius as princeps of the Samnites. So, also, we have MEAAEIS OT$ENS (Meddix Ufens) in the inscription given by Castelli di Torremuzza, Sicil. vet. Inscr. V. 45, p. 55: see Muller, Etrusk. II. p. 69, note. Knotel proposes (Zeitschr. f. d. Alterthumsw. 1850, p. 420) to consider Med-dix ^medium- dicens as a compound analogous to ju-dex=jus-dicens, vin- dex vim-dicens, &c. The last word is more truly explained with reference to ven-eo, ven-do, and ven-dico ; and as mediae is properly spelt with one d (see Schomann's Greifswald Pro- gram fur 1840), it would be better to consider med- as the root and x = c-s as a mere formative ending : cf. medicus. In somewhat later times the Sabello-Oscans called their dictator by the name embratur, which is evidently a shortened form of the Latin im-perator^ or indu-perator. Liv. VIII. 39 ; IX. 1 ; X. 29. Oros. V. 15 : " Postquam sibi Samnites Papium Mu- tilum imperatorem praefecerant." Similarly we have coins with the Oscan inscription, G. Paapi G. Mutil Embratur ; which refer to the time of the Social War, when the forces of the confederacy were divided into two armies, each under its own imperator, the Marsi being under the orders of Q. Popce- dius Silo, the Samnites having for their leader this Gains Papius Mutilus, the son of Gains. Of tuticus, see below. Minerva, Sab. s. v. Feronia. Multa, Osc. et Sab. Fest. p. 142 : "Multam Osce dici putant poanam quidam. M. Varro ait poenam esse, sed pecuniariam, de qua subtiliter in Lib. I. quasstionum Epist. I. refert." Cf. p. 144. s. v. Maximam multam. Varro, apud Gell XI. 1 : " Vocabulum autem ipsum multce idem M. Varro uno et vice- simo rerum humanarum non Latinum sed Sabinum esse dicit, idque ad suam memoriam mansisse ait in lingua Samnitium, qui sunt a Sabinis orti." 112 THE SABELLO-OSCAN LANGUAGE. [On. IV. Nar, Sab. Yirg. JEn. VII. 517 : " Sulfurea Nar albus aqua." Ubi Serv. : " Sabini lingua sua nar dicunt sulfur." Ner, nerio, Sab. Suet. Vit. Tiber. I. : " Inter cognomina autem et Neronis adsumpsit, quo significatur lingua Sabina fortis ac strenuus." GelL XIII. 22 : " Nerio a veteribus sic declina- tur, quasi Anio ; nam proinde ut Anienem, sic Nerienem dix- erunt, tertia syllaba producta ; id autem, sive Nerio sive Ne- rienes est, Sabinum verbum est, eoque significatur virtus et fortitude. Itaque ex Claudiis, quos a Sabinis oriundos acce- pimus, qui erat egregia atque praestanti fortitudine Nero appel- latus est. Sed id Sabini accepisse a GrsBcis videntur, qui vin- cula et firmamenta membrorum vevpa dicunt, unde nos quoque nervos appellamus." Lydus, de Mens. IV. 42. Id. de Ma- gistr. I. 23. Compare the Sanscr. nri ; and see above, p. 106, s. v. Cas-nar : cf. p. 97. Novensides, Ops. Sab. s. v. Feronia. Panis= Ceres, Sab. s. v. Lebasius. Panes, Messap. Athen. III. p. Ill c.: iravos apros Mev-, fu), and it is a matter of indifference whether Venus or Ceres comes nearest to the goddess intended. Knotel identifies Evklus with Iphiclus, and of course this is possible ; but the adjunct patri in 1. 25, seems to denote a deity analogous to Liber Pater (cf. Evius). Amma corresponds, as Aufrecht suggests, to the Germ, amme, Sanscr. ambd, " mother." Verehasius, as an 92 132 THE SABELLO-OSCAN LANGUAGE. [Cn. IV. epithet of Jupiter, is explained by the Sanscr. vri, " to grow," whence the Latin virga ; and regator must be rigator, i. e. plu- vius. Patana is Panda or Patella (Gell. XIII. 22. Arnob. IV. 7), who opens the husk of the grain. Teforom answers to the Latin tepidus, and still more nearly to the Etruscan tephral (see above, Chap. II. 11). Akenus is =annus, as in Umbrian (see Au- frecht u. Kirchhoff, Umbr. Sprd. p. 401). Perna is Pales = Pares (v. Festus, p. 222, Miiller; and cf. vetus, veter-nus, lux, luci-na, dies, dia-nus, joy-is, ju-no, &c.). We may compare pistia with pistor, pistum, pisum, &c. 7. The Atellance. It seems scarcely worth while to enumerate the grammatical forms which may be collected from these inscriptions, as they are virtually the same with those which occur in the oldest spe- cimens of Latin, the only important differences being that we have -azum for -arum in the gen. pi. of the 1st decl., that the 3rd declension sometimes preserves the original -ss of the nom. pi., and that this reduplication represents the absorbed m in the ace. pi. of the 2nd and 3rd declensions. It may be desirable, however, before concluding this part of the subject, to make a few remarks on the Fabulce Atellance, the only branch of Oscan literature of which we know any thing. The most important passage respecting the Fabulos Atel- lance, that in which Livy is speaking (VII. 2) of the introduc- tion of the Tuscan ludiones at Rome in the year A.U.C. 390, has often been misunderstood ; and the same has been the fate of a passage in Tacitus (IV. 14), in which the historian mentions the expulsion of the actors from Italy in the year A. u. c. 776. With regard to the latter, Tacitus has caused some confusion by his inaccurate use of the word histrio; but Suetonius has the phrase Atellanarum histrio (Nero, c. 39) ; and the word had either lost its earlier and more limited signification, or the Atel- lanaa were then performed by regular histriones. Livy says that, among other means of appeasing the anger of the gods in the pestilence of 390 A. u. c., scenic games were for the first time introduced at Rome. Hitherto the Romans had had no public sports except those of the circus namely, races and wrestling ; but now this trivial and foreign amusement was introduced. Etruscan ludiones danced gracefully to the sound of the flute without any accompaniment of words, and without $ 7.] THE SABELLO-OSCAN LANGUAGE. 133 any professed mimic action. Afterwards, the Roman youth began to imitate these dances, and accompanied them with unpre- meditated jests, after the manner of the Fescennine verses ; these effusions gave way to the satura, written in verse and set to the flute, which was acted by professed histriones with suitable songs and gestures ; and then, after a lapse of several years, Livius Andronicus ventured to convert the satura into a regular poem, and to make a distinction between the singing (canticum) and the dialogue (diverbia) ; the latter alone being reserved to the histriones, and the former being a monologue, by way of inter- lude with a flute accompaniment *. Upon this, the Roman youth, leaving the regular play to the professed actors, revived the old farces, and acted them as interludes or afterpieces (exodia 2 ) to the regular drama. These farces, he expressly says, were of Oscan origin, and akin to the Fabulce Atellance ; and they had the peculiar advantage of not affecting the civic rights of the actors. In order to understand the ancient respectability of the Atellance, we must bear in mind the opposition which is always recognized between them and the Mime. Hermann has pro- posed the following parallel classification of the Greek and Roman plays (Opusc. V. p. 260, cf. Diomedes, III. p.4SO, Putsch): ARGUMENTUM. ROMANUM ARGUMENTUM. Crepidata (rpay^ia). Prcetextata. Palliata (KW^^IO). Togata, vel trabeata vel taber naria. Satyrica (crdrvpoi). Atellana. Mimus (Mt/uo?). Planipes. 1 Diomed. III. p. 4S9 : "in canticis una tantum debet esse persona, aut, si duse fuerint, ita debent esse, ut ex occulte una audiat, nee collo- quatur, sed secum, si opus fuerit, verba facial." On the canticum see Hermann, Opusc. I. pp. 290, sqq., who has clearly shown that it was not merely a flute voluntary between the acts. 2 As the practice of the Greek and Roman stage involved the per- formance of several dramas on the same day, it matters little whether wo render exodium by " interlude " or " afterpiece." According to the defi- nitions given by Suidas and Hesychius, an eocodium was that which followed an exeunt omnes, whether, which was more common, at the end of a play, or at the end of an act. See the examples given by Meineke on Cratinus, Fr. Incert. CLXX. p. 230, and compare Baumstark's article in Pauly's Real-Encycl. III. p. 360. 134* THE SABELLO-OSCAN LANGUAGE. [On. IV. Adopting this classification, which has at least much to recom- mend it, we shall see that as the Greek satyrical drama was the original form of the entertainment, and, though jocose, was not without its elevating and religious element, so the Atel- lana, as a national drama, was immediately connected with the festive worship of the people in which it took its rise, and therefore retained a respectability which could not be conceded to the performances of foreign histriones. These artists were not allowed to pollute 1 the domestic drama ; and, being free from all contact with the professional actor, the young Roman could appear in the Atellan play without any forfeiture of his social position. Whereas, even in the corrupt days of the later empire, Juvenal saw something especially monstrous in the fact that a noble could appear as a mimus or planipes 2 . With particular reference to the contrast between the mimus and the Atellana, Cicero says to Papirius Psetus, who had introduced some vulgar jokes after a quotation from the CEnomaus of Accius, that he had followed the modern custom of giving a mime for afterpiece instead of adopting the old practice of introducing the Atellan farce after the tragedy 3 . In the same way he says 4 that superfluous imitation, such as obscene gestures, belongs to the domain of those mimi, who caricatured the manners of men. And while Macrobius considers it as an exceptional merit to have introduced mimi without lasciviousness 5 , Valerius Maximus 1 Lir. VII. 2 : " nee ab histrionibus pollui passa est." 2 VIII. 189, sqq. : "populi frons durior hujus, Qui sedet, et spectat triscurria patriciorum, Planipedes audit Fabios, ridere potest qui Mamercorum alapas." 3 Cic. ad Div. IX. 16, 2 : " nunc venio ad jocationes tuas, quum tu secundum CEnomaum Accii, non, ut olim solebat, Atellanam, sed, ut nunc fit, mimum introduxisti." 4 de Oratore, II. 59 : " mimorum est enim etliologorum, si nimia est imi- tatio, sicut obsccenitas." Cf. c. 60, 244. 5 Saturn. II. 7 : " videbimur et adhibendo conyivio mimos vitasse lasciviam" This is the passage referred to by Manutius in his note on Cicero ad Div. IX. 16, 2, where he says in a parenthesis: "itaque Macro- bius Lib. III. Saturn, mi mis lasciviam tribuit." In Smith's Diet, of Anti- quities, Art. Atellance fabulce, Ed. I., this note of Manutius is paraded at full length as a quotation from "Macrobius Satur. Lib. III./' and even 7.] THE SABELLO-OSCAN LANGUAGE. 135 attributes the social respectability of those who performed in the Atellan farces to the old Italian gravity which tempered this entertainment 1 . But besides the moral decency by which the Atellana was distinguished from the mime, it is manifest from the passage in Livy that it derived additional recommendation from the fact that this was a national amusement and was connected with the usages of the country population, who always contributed a varying proportion to the inhabitants of ancient Rome. We infer from the words of the historian that the Roman youth were not satisfied with either the Tuscan or the Greek importations, and that it was their wish to revive something that was not foreign, but national. Of course Livy cannot mean to say that the Oscan farce was not introduced at Rome till after the time of Livius Andronicus Muso, and that it was then imported from Atella. For whereas Muso did not perform at Rome till the second Punic war 2 , Atella shared in the fate of Capua ten years before the battle of Zama, and the inhabitants were compelled to migrate the ut arbitror of the commentator is made to express the opinions of the author quoted. It is evident that the compiler of this Article made no attempt to verify the reference to Macrobius, which he has used without stating that he was indebted for it to Manutius, and which he has care- fully placed at a distance from his reference to Cicero. His blunder is the just Nemesis of his dishonesty. As he quotes from Valerius Maximus, " II. 1," instead of " II. 4," we may presume that in this case also he is using the learning of some commentator. In the new edition of Smith's Dictionary the article Atellance Fabulce is suppressed, and a short account of the subject is included in the article Comcedia, written by another person. The same Nemesis still tracks the dishonest quotation, for there " Macrobius, Satur. III." is quoted for Manutius' statement that the Atellana was divided into five acts. All this may be taken as an example of the false affectation of learning on the part of the compilers, and general incompetence on the part of the editor, which is so frequently conspicuous in Smith's dictionaries. 1 II. 4 : " Atellani autem ab Oscis acciti sunt ; quod genus delecta- tionis Italica severitate temperatum, ideoque vacuum nota est ; nam neque tribu movetur, neque a militaribus stipendiis repellitur." 2 Porcius Licinius, apud Aul. Gell. XVII. 21 : Pcenico hello secundo Muso pinnato gradu Intulit se bellicosam in Romuli gentem feram. See also Hor. II. Epist. I. 162, 136 THE SABELLO-OSCAN LANGUAGE. [On. IV. to Calatia 1 . Now it appears from the coins of this place that its Oscan name was Aderla 2 ; and the Romans always pronounced this as Atella, by a change of the medial into a tenuis, as in Mettus for Meddix, imperator for embratur, fuit for fuid, &c. This shows that the name was in early use at Rome ; and we may suppose that, as an essential element in the population of Rome was Oscan, the Romans had their Oscan farces from a very early period, and that these farces received a great im- provement from the then celebrated city of Aderla in Campania. Ifc is also more than probable that these Oscan farces were common in the country life of the old Romans, both before they were introduced into the city 3 , and after the expulsion of the histriones by Tiberius 4 . For the mask was the peculiar charac- teristic of the Atellanse 5 , and these country farces are always spoken of with especial reference to the masks of the actors. We may be sure that the Oscan language was not used in these farces when that language ceased to be intelligible to the Romans. The language of the fragments which have come down to us is pure Latin 6 , and Tacitus describes the Atellana as " Oscum quondam ludicrumV Probably, till a comparatively late period, i Livy, XXVI. 16, XXII. 61, XXVII. 3. 2 Lepsius ad Inscriptions, p. 111. For the meaning of the word, see above, 5, note. 3 Virgil. Georg. II. 385, sqq. : Nee non Ausonii, Troja gens missa, coloni Versibus incomptis ludunt risuque soluto, Oraque corticibus sumunt horrenda cavatis. Comp. Horat. II. Epist. I. 139, sqq. 4 Juvenal, Sat. III. 172, sqq. : Ipsa dierum Festorum herboso colitur si quando theatro Majestas, tandemque redit ad pulpita notum Exodium, quura personse pallentis hiatum In gremio matris formidat rusticus infans. That the exodium here refers to the Atellana appears from Juv. VI. 71 : " Urbicus exodio risum movet Atellance Gestibus Autonoes." 6 Festus, s. v. personata fabula, p. 217: "per Atellanos qui proprie vocautur personati." The modern representatives of the Atellan charac- ters are still called maschere, and our harlequin always appears with a black mask on the upper part of his face. 6 See Diomed. III. pp. 487, 488, Putsch. 1 Ann. IV. 149. 7.] THE SABELLO-OSCAN LANGUAGE. 137 the Atellana abounded in provincial and rustic expressions 1 ; but at last it retained no trace of its primitive simplicity, for the gross coarseness and obscenity 2 , which seem to have superseded the old-fashioned elegance of the original farce 3 , and brought it into a close resemblance to the mimus, from which it was originally distinguished, must be attributed to the general cor- ruption of manners under the emperors, and perhaps also to.. the fact that from the time of Sulla downwards the Oscan farce was gradually passing from its original form into that of a regular play on the Greek model, so that all the faults of Greek comedy would eventually find a place in the entertainment. The prin- cipal writers of the Latin AtellanaB, after Sulla, who is said to have used his own, that is, the Campanian dialect 4 , were Q. Novius 5 , L. Pomponius Bononiensis 6 , L, Afranius 7 , and C. Mem- mius 8 . The political allusions with which they occasionally abounded, and which in the opinion of Tiberius called for the interference of the senate 9 , were a feature borrowed from the licence of the old Greek comedy ; and to the same source we must refer the names of the personages 10 , which are known to have been adopted by Novius, Afranius, and Pomponius, and which i Varro, L. L. VII. 84, p. 152. 2 Terent. Maur. p. 2436, Putsch ; Quintil. Inst. Or. VI. 3 ; Tertull. De Spectaculis, 18; Schober, uber die Atellan. Schauspiele y pp. 28, sqq. 3 Donat. de Trag. et Com. " Atellanse salibus et jocis compositse, quse in se non habent nisi vetustam elegantiam" * Athenseus, IV. p. 261, C. : e^avi^ovo-t & avrov TO Trept ravra IXapbv ai VTT avrov ypa(^flo~ai SarvpiKal Kco/jLcpdicn rrj Trarpicd (povfj. That tho satyric comedies here referred to must have been Atellance may be in- ferred from Diomedes, III. p. 487, Putsch: "tertia species est fabularum. Latinarum, quse . . . Atellance dictse sunt, arguments dictisque jocularibus similes satyricis fabulis Grsecis." The reference to the Simus in the Atellance (Sueton. Galb. 15) points to a contact with the satyrs. Macro- bius, Saturn. II. 1. 5 Aulus Gellius, N. A. XVII. 2. 6 Macrob. Saturn. VII. 9 ; Fronto ad M. Gees. IV. 3, p. 95, Mai ; Vel- leius, II. 9, 6. 7 Nonius, s. v. ientare. 8 Macrobius, Saturn. I. 10. 9 Tacitus, Annal. IV. 14 : " Oscum quondam ludicrum, levissimse apud vulgus delectationis, eo flagitiorum et virium venisse, ut auctoritate patrum coercendum sit." Cf. Sueton. Nero, c. 39 ; Galba, c. 13 ; Calig. c. 27; where we have special instances of the political allusions in the later Atellance. 10 See Mullcr, Hist. Lit. Or. ch. XXIX. 5. Vol. II. p. 43, note. 138 THE SABELLO-OSCAN LANGUAGE. [On. IV. are either Greek in themselves or translations of Greek words. The old gentleman or pantaloon was called Pappus or Casnar : the former was the Greek ITaTTTros, the latter, as we have seen, was an Oscan term = vetus. The clown or chatterbox was called Bucco, from bucca, and was thus a representative of the Greek TvdOcw. The glutton Macco, Greek MaWw, has left a trace of his name in the Neapolitan Maccaroni ; and Punch or Poli- chinello is derived from the endearing diminutive Pulchellus, which, like the Greek KaXX/ae, was used to denote apes and puppets 1 . The Sannio is the (javvas of Cratinus (Fr. Incert. XXXIII. a. p. 187, Meineke) ; and this buffoon with his patch- work dress is represented by the modern Harlequin, one of whose names is still zanni, Angl. " zany." The modern word harlequin is merely the Italian allecchino, i. e. " gourmand." Menage's dream about the comedian, who was so called in the reign of Henry III. because he frequented the house of M. de Harlai, is only an amusing example of that which was called etymology not many years ago. On the whole we must conclude, that the Atellan farces were ultimately Grecized, like all the literature of ancient Italy, and as the language of the Doric chorus grew more and more identical with that of the Attic dialogue, to which it served as an interlude, so this once Oscan exodium was assimilated in language and character to the histrionic plays, to which it served as an afterpiece, and so gradually lost its national character and social respectability. Thus we find in the destiny of this branch of Oscan literature an example of the absorbing centralization of Rome, which, spreading its metropolitan Latinity over the pro- vinces, eventually annihilated, or incorporated and blended with its civic elements, all the distinctive peculiarities of the allied or subject population. Theatre of the Greeks, Ed. 6, p. [160]. CHAPTER V. THE ETRUSCAN LANGUAGE. 1. Transcriptions of proper names the first clue to an interpretation of the Etruscan language. 2. Names of Etruscan divinities derived and explained. 3. Al- phabetical list of Etruscan words interpreted. 4. Etruscan inscriptions difficulties attending their interpretation. 5. Inscriptions in which the Pelas- gian element predominates. 6. Transition to the inscriptions which contain Scandinavian words The laurel-crowned Apollo Explanation of the words clan, and phleres. 7. Inscriptions containing the words suthi and tree. 8. In- ferences derivable from the words sver, ever, and thur or thaur. 9. Striking coincidence between the Etruscan and Old Norse in the use of the auxiliary verb lata. 10. The great Perugian Inscription critically examined. Its Runic affinities. 11. Harmony between linguistic research and ethnographic tradition in regard to the ancient Etruscans. 12. General remarks on the absorption or evanescence of the old Etruscan language. $ 1. Transcriptions of proper names the first clue to an interpretation of the Etruscan language. IT will not be possible to investigate the remains of the Etrus- can language with any reasonable prospect of complete suc- cess, until some scholar shall have furnished us with a body of inscriptions resting on a critical examination of the originals 1 ; and even then it is doubtful if we should have a sufficiently co- pious collection of materials. The theory, however, that the Etruscan language, as we have it, is in part a Pelasgian idiom, more or less corrupted and deformed by contact with the Um- brian, and in part a relic of the oldest Low-German or Scandi- navian dialects, is amply confirmed by an inspection of those remains which admit of approximate interpretation. The first clue to the understanding of this mysterious lan- guage is furnished by the Etruscan transcriptions of well-known Greek proper names, and by the Etruscan forms of those names which were afterwards adopted by the Romans. This comparison may at least supply some prima-facie evidence of the peculiari- 1 The first impulse to the study of Etruscan antiquities was given by the posthumous publication of Dempster's work de Etruria Regali, which was finished in 1619, and edited by Coke in 1723 4. Bonarota, who furnished the accurate illustrations of this work, insists upon the import- ance of a correct transcription of the existing linguistic materials. 140 THE ETRUSCAN LANGUAGE. [On. V. ties of Tuscan articulation, and of the manner in which the lan- guage tended to corrupt itself. It is well known that the Etruscan alphabet possessed no mediae, as they are called. We are not, therefore, surprised to find, that in their transcriptions of Greek proper names the Etrus- cans have substituted tenues 1 . Thus, the Greek names, ''A$pa- CTTOS, Ti/oeJf, 'OcWoW?, MeXecfy^oos, and IIoXiAJ/cj/s, are written A tresthe, Tute, Utuze, Melakre, . and Pultuke. But the change in the transcription goes a step farther than this ; for, though they actually possessed the tenues, they often convert them into aspiratce. Thus, 'Aya/me/mvcov, "'A^acrro?, Bert?, Hepaeus, HoXvveiKrjs, Tt]\e9, DoXi/ye/K^s 1 , 'O^oe'crr*/?, and of the Latin Minerva, only because the Etruscans did not find it neces- sary to express in writing the articulation-vowels of the liquids. It is interesting to remark that the old poetic dialect of the Icelandic, as distinguished from the modern tongue, exhibits the same pecu- liarity ; thus r is always written for ur, as in northr, vethr, akr, vetr, vitr. There are a few instances of the same brachygraphy in the oldest Greek inscriptions : thus, on Mr. Burgon's vase we have AOHNHGN for 'AO^vrjOev. Bockh (<7. /. No. 33) has wrongly read this inscription, which forms three cretics: TWV 'A0j\vr)9ei> a\Q\(*)i> CJULI. With regard to the form Ercle, for which we have Hercole in Dempster, T. I. tab. VI. ; Lanzi, II. p. 205. tab. XI. n. 1, it is to be remarked that the short u =o before / appears to be a natural stop-gap in old Italian articula- tion. Thus we have ^Esculapius for AtovcX^Tnos. When we remember that 'Hjoa/cX^s- was the tutelary god of the Dorians or Her-mun-duri, who conquered the Peloponnese, we can hardly avoid identifying him with Her-minius. If we pass to the consideration of those proper names which are found in the Latin language, we shall observe peculiarities of precisely the same kind. For instance, the medials in Idus, Tlabonius, Vibius, &c. are represented in Etruscan by the tenues in Itus, Tlapuni, Fipi, &c. ; the tenues in Turius, Velcia, &c. stand for the aspirates in Thura, Felche, &c. ; and the articula- ele-mentum = olementum. See Benaryin the Berl. Jahrb. for August 1841, p. 240. As the Indus, or gladiatorial school was the earliest specimen of a distinct training establishment, and as it has consequently furnished a name to all schools, so its two functions have similarly descended into the vocabulary of education : for rudi-menta, properly the " foil exercises," and ele-menta, properly the " training- food," have become synonymous expressions for early education, just as e-rud-itus, "out of foils," has be- come the term for a completely learned man. 142 THE ETRUSCAN LANGUAGE. [Cii. V. tion-vowels in Licinius, Tanaquil, &c. are omitted before or after the liquids in Lecne, Thanchfil, &c. The transcription Utuxe, for 'O^i/crcreJs, suggests a remark which has been in part anticipated in a former chapter. We see that in this case the Etruscan z corresponds to the Greek -o-cr, just as conversely, in the cases there cited, the Greek - is represented by -ss in Latin. It was formerly supposed that this Etruscan z was equivalent to x = KS, and this supposition was based on a comparison of Utuze with Ulyxes. To say no- thing, however, of the mistake, which was made in assuming that Utuze represented Ulyxes and not 'OSwrcrei/s, it has been shown by Lepsius (De Tabb. Eug. pp. 59, sqq. ; Annali dell' Institute, VIII. p. 168) both that the Etruscans added this z to the guttural K, as in srankzl, &c. and also that, when it was necessary to ex- press the Greek , they did not use the letter z, but formed a representative for it by a combination of K or CH with s, as in Secstinal Sextinia natus, and Elchsntre 'AXef av^pos. Pa- laeographical considerations also indicate that the letter corre- sponded in form, not to f or x, but to the Greek z. We ought, however, to go a step farther than Lepsius has done, and say that the Latin x was, after all, in one of its values, a represen- tative of this Etruscan letter. It is true, indeed, that x does represent also the combination of a guttural and sibilant; but there are cases, on the other hand, in which x is found in Latin words containing roots into which no guttural enters : comp. rixa with epis (epiSos), epifa, &c. In these cases it must be supposed to stand as a representative of the Greek in its sound sh, and also of the Hebrew shin, from which f? has derived its name (see New Crat. 115). With regard to the name Ulysses, Ulyxes, 'O$i/o-creJs, etymology would rather show that the ultimate form of the x, ss, or z, was a softened dental. The Tuscan name of this hero was Nanus, i. e. " the pygmy" (Muller, Etrusk. II. p. 269); and, according to Eustathius (p. 289, 38), 'OAi/crcrevs or 'OAtcroWs was the original form of the Greek name. From these data it has been happily conjectured (by Kenrick, Herod, p. 281) that the name means o-Ai^os, o-At/V = dian-us would become Tina-[s\ or Tinia-[i\. This Tina or Jupiter of the Tuscans was emphatically the god of light and lightning, and with Juno and Minerva formed a group who were joined toge- ther in the special worship of the old Italians. As the Etruscans had no consonant /, the name of Janus must have been pro- nounced by them as Zanus. This god, whose four-faced statue was brought from Falerii to Rome, indicated the sky, or templum, with its four regions. When he appeared as biceps, he repre- sented the main regions of the templum the decumanus and the cardo. And as this augurial reference was intimately con- nected with the arrangement of the gates in a city or in a camp 1 , he became also the god of gates, and his name ultimately signi- fied "a gate" or "archway." Summanus, or Submanus, was the god of nightly thunders. The usual etymology is summus manium; but there is little reason for supposing that it is an ordinary Latin word. As Arnobius considers him identical with Pluto 2 , it seems reasonable to conclude that he was simply the a-va, e-nim, 6-na, &c. Lobeck asserts (Paralipom. p. 121, note) that the v in TI-V-OS is repugnant to all analogy, the literce cliticce of the Greeks being dentals only, as if v were not a dental! The absurdity of Lobeck's remarks here, and in many other passages of his later writings, will serve to show how necessary it is that an etymologer should be acquainted with the principles of comparative philology. There are some observations on this subject in the New Crat. 38, which more particularly refer to Lobeck (Aglaophatn. p. 478, note i.), and to a very inferior man, his pupil Ellendt (Lex. Sophocl. praefat. p. iii.). From what Lobeck said in his Paralipomena (p. 226, note), one felt disposed to hope that his old- fashioned prejudices were beginning to yield to conviction. - In a later work, however (Pathologia, prsef. pp. vii. sqq.)> he reappears in his original character. The caution on which he plumes himself (" ego quoque ssepe vel invitus et ingratis eo adactus sum ut vocabulorum origines abditas conjectura qusererem, cautior fortasse Cratylis nostris, quorum curiositati nihil clausum, nihil impervium est,") is only another name for one-sided obstinacy; and whatever value we may set upon Lobeck's actual per- formances in his own field, we cannot concede to him the right of con- fining all other scholars to the narrow limits of his Hemsterhusian phi- lology. 1 See below, Ch. VII. 6. 2 The Glossar. Labbsei has Summanus, Upo^irfdevs ; and perhaps Pro- metheus, as the stealer of fire from heaven, may have been identified with the god of nightly thunders in some forms of mythology. At Co- $ 2.] THE ETRUSCAN LANGUAGE. 145 Jupiter Infernus ; and as the Dispater of the Tuscans was called Mantus<> and his wife Mania, we may conjecture that Sub-manus was perhaps in Tuscan Zuv-manus or Jupiter-bonus, which is the common euphemism in speaking of the infernal deities. The connexion between the nightly thunders, which the ancients so greatly feared, and the ^Oouiai fipovrai, is obvious. Another gloomy form of the supreme god was Ve-djus or Ve-jovis, who seems to have represented Apollo in his character of the causer of sudden death. The prefix Ve- is a disqualifying negative the name signifies " the bad Jupiter." He was represented as a young man armed with arrows ; his feast was on the nones of March, when an atoning sacrifice was offered up to him ; and he was considered, like Summanus, as another form of Pluto. The second of the great Tuscan deities was Juno (Jovino or Dyuno), who was called Kupra and Thalna in the Etrurian language. Now Kupra signifies " good," as has been shown above; and therefore Dea kupra is Dea bona, the common euphemism for Proserpine. The name Thalna may be analysed with the aid of the principles developed above. The Etruscans had a tendency to employ the aspirates for the tenues, where in other forms, and in Greek especially, the tenues were used. Accordingly, if we articulate between the liquids In, and substi- tute t for th, we shall have, as the name of Juno, the goddess of marriage, the form Tal[a~\na, which at once suggests the root of Talassus, the Roman Hymen, and the Greek raX(s, (Soph. Antig. 629. rctXts* rj vvfjifprj, Zonar. p. 1711. ra\is' rj jueX- Xo'ya/uos 'TrapQevo'S KO.I Ka.Tcovo/uaa'fJLGvr] TIVL ' o\ ce f yvvouKa yafjLeTtji'' ot Se vv/nfyriv, Hesych. x^'Xt^a* OUTGO T^V avvnp- /uoa/u.cvrjv, id. caXicas' TCIS /ue/jn^^crrei/^ei^a?, id. raXt^' 6/oa>$, id.): comp. also yajmoio reXos 1 , Horn. Od. XX. 74, and the epithet ''Hpa reXe/a. The Aramaaan raXiOd (Ffbft, Mark v. 41) is not to be referred to this class. The deity Vulcanus, who in the Etruscan mythology was one of the chief gods, being one of the nine thundering gods, and who in other mythologies appears in the first rank of divinities, always stands in a near relationship to Juno. In the Greek theogony he appears as her son and defender ; he is sometimes lonus, where the infernal deities were especially worshipped, the 6 7rvpKvroto \afipa>6tv O-KOTO) Srvybs KfXaivfjs va(rp.6v, evda Ttp/zievs opKapoTovs fTfvev dfpdiTovs edpas p,e\\()v yiyavras Kanl TLTrjvas irepav. Now Turmus certainly does not differ more from this than Euturpe and Achle from their Greek representatives (Bun- sen, ibid. p. 175). It might seem, then, that Turmus is not the Latin Terminus, but rather the Greek 'E^uijs ; for the Hellenic aspirate being represented in the Pelasgian language, according to rule, by the sibilant, this might pass into T, as in rjiiepa, ffrji^epov, TtffjiCpov ; eVra, reTrra, Hesych. ; ep^is, Tep/j.is, id. &c. The name Ldr, Las, when it signifies " lord" or " noble," has the addition of a pronominal affix -t ; when it signifies " god," it is the simple root : the former is Lars (Larth), gen. Lartis ; the latter Lar, gen. Laris. Precisely the same difference is observable in a comparison between "AmKes, "AVCLKOI, " the Dios- curi," and ai/a/ores, " kings" or " nobles." Similarly the ori- ginal Mar-s seen in the forms Mar-mar } Ma-murius, &c. is 2.] THE ETRUSCAN LANGUAGE. 151 lengthened into Mar-t- t and from names of towns we have deri- vatives with the same insertion of a formative t : e. g. Tuder-t-es, Tibur-t-es, Picen-t-es, Fiden-t-es, Fucen-t-es, Nar-t-es (Corssen, Zeitschr. f. Vergl. Sprf. 1852, pp. 6, 13). Some suppose that the English Lor-d is connected with the same root ; see, how- ever, New Crat. 338 : and as the Lares were connected with the Cabiriac and Curetic worship of the more eastern Pelas- gians, I would rather seek the etymology in the root Act-, Acts-, Aats-, so frequently occurring in the names of places and persons connected with that worship 1 , and expressing the devouring nature of fire. It appears from the word Lar-va that the Lar was represented as a wide-mouthed figure. There are two feminine forms of the name, Lar-unda and Lar-entia. This enumeration of the names of Tuscan divinities shows that, as far as the terms of mythology are concerned (and there are few terms less mutable), the Tuscan language does not abso- lutely escape from the grasp of etymology. If the suggestion thrown out above (Ch. II. 22) respecting the parallelism be- tween Tina and Tor is to be received, the easy analysis of these mythical names is to be explained by the fact that they belonged to the religion of southern Etruria, which was Pelasgian rather than Scandinavian. Many of the common words which have been handed down to us present similar traces of affinity to the languages of the Indo-Germanic family. I will examine them in alphabetical order ; though, unfortunately, they are not so numerous as to assume the form of a comprehensive voca- bulary of the language. 3. Alphabetical List of Etruscan Words interpreted. JEsar, " God." Sueton. Octav. c. 97 : " Responsum est centum solos dies posthac victurum, quern numerum c littera notaret ; futurumque ut inter deos referretur, quod .SESAR, id est, reliqua pars e Csesaris nomine, Etrusca lingua deus vocaretur." Conf. Dio. Cass. LVI. 29 ; Hesych. aiaoi ' Oeoi, VTTO Tvpprjvwv. See Hitter, Vorhalle, pp. 300, 471, who compares the Cabiriac names ^s-mun, JEs-clef, the proper name ^Esyetes, asa the 1 The following are some of the most obvious appearances of this root: Sanscrit, las, "to wish;" Latin, lar-gus ; Greek, Xa-p'a, Xa-/*oy, Xapvy, Xatrfia, &C. A.jp.vos, AT/TOO. 152 THE ETRUSCAN LANGUAGE. [Cn. V. old form of ara, and a great many other words implying "holiness" or "sanctity:" and Grimm, Deutsche Mythol. 2d edit. p. 22. Comp. also alcra. The most important fact is that as or ass, pi. aesir, meaning deus, numen, is " nomen nusquam non occurrens" (Edda Scemund. Vol. I. p. 472) in the old Icelandic. Agalletor, "son." Hesych. dyaXXrJTopa' Tral^a^Tvpprjvoi. This is pure Pelasgian, if not Greek. Thus Sophocles, Antig. 1115, calls Bacchus : Kct$/t/eias vv/JL(pa with yoa<, and the latter with the Hebrew rtyS, mugire, "to low like an ox" (1 Sam. vi. 12, Job vi. 6), and the Latin ceva, which, according to Colu- mella (VI. 24), was the name of the cow at Altinum on the Adriatic. (d) xw " the goose," i. e. " the gaping bird" (^'71' Kexnvns, Athen. p. 519. A). (e) l^t, "the tawny wolf," may be connected with 1HN "yellow" like gold. Perhaps the most remarkable instance of selecting for the name of an object some single attribute, is furnished by the words scudo and " crown," both denoting a large silver coin, and both deriving their origin from a part of the design on the reverse the former from the shield, or coat of arms, the latter from the crown, by which it was surmounted. 156 THE ETRUSCAN LANGUAGE. [On. V. Dea, i.e. bona Dea, "Cybele." Hesych. $eV 'Pea, VTTO Druna, " sovranty." Hesych. Spovva' tj apxv, vwo It is clear that this word can have nothing to do with the Low-Greek Spovyyos, " a body of men," Spovyydpios, " a captain," which are fully explained by Du Cange, Gloss. Med. et Inf. Grcecit. I. pp. 333, 4. We must refer it to the 0. JSTorse, drott = dominus, at drottna = imperare, the dental mutes being absorbed before the n as in ^ei-ros for SeiS-vos, &c. And thus we get another trace of Gothic affinity for the Rasena. Falandum, "the sky." Fest. p. 88: " Falce [0a'Xar oprj, o-KOTTiai, Hesych.] dicta3 ab altitudine, a falando, quod apud Etruscos significat ccelum." This is generally connected with fyaXavQov, blond, &c. Or we might go a step farther, and refer it to (f>a\\w, 0a\o's, &c., which are obviously derived from COLO'S : see Lobeck, Pathol. p. 87. Favissa, " an excavation." Fest. p. 88 : " Favissce locum sic appellabant, in quo erat aqua inclusa circa templa. Sunt autem, qui putant, favissas esse in Capitolio cellis cisternisque similes, ubi reponi erant solita ea, qua3 in templo vetustate erant facta inutilia." From the analogy of favissa, mantissa, and from the circumstance that the Romans seem to have learned to make favissce from the Etruscans, it is inferred that favissa was a Tuscan word : see M tiller, ad Festi locum, and Etrusk. II. p. 239. The word is probably connected with fovea, bauen, &c. We shall see below that lautn was the Rasenic synonym. Februum, " a purification." Angrius, ap. J. Lyd. de Mens. p. 70 : " Februum inferum esse Thuscorum lingua." Also Sabine : see Yarro, L. L. VI. 13. If we compare febris, &c., we shall perhaps connect the root with foveo-torreo, whence favilla, &c., and understand the " torrida cum mica farra," which, according to Ovid (Fast. II. 24), were called by this name. Fentha, according to Lactantius (de Fals. Relig. I. c. 22, $ 9), was the old Italian name of Fatua^ the feminine form of Faunus, " quod mulieribus fata canere consuevisset, ut Faunus viris." The form Finthia seems to occur on an old Tuscan monument (Ann. dell' Instit. VIII. p. 76), and is therefore perhaps a Tuscan word. The analogy of Fentha to Fatua 3.J THE ETRUSCAN LANGUAGE. 157 is the same as that which has been pointed out above in the case of Mantus. The n is a kind of anuswarah very common in Latin : comp. e^is, anguis ; Xe/Trw, linquo ; Xe/^cu, lingo ; Sanscr. tuddmi, tundo ; v$wp. unda ; &c. Floces, "dregs of wine, 31 Aul. Gell. XL 7; "floces audierat prisca voce significare vini fsecein e vinaceis expressarn, sicuti fraces ex oleis." Above s. v. Apluda. Fruntac ; see Haruspex, and Phruntac. Gapus, " a chariot." Hesych. : ydiros o^rj/ma, Tupprjvoi We have here POTTOS, a short Pelasgian form of aTnjvrj. Comp. habena with xafios (Hesych.), pcXip'ii with , Hesych.). Consequently, it expresses on the one hand the amusement afforded by the gesticulations of the ludio (a^/uLori^erai TToi/aAfos- ets yeXwra, Appian, u. s.), and on the other hand indicates the innocent brandishing of weapons by the armed ludio as compared with the use of arms in actual warfare. This latter setffee was preserved by Indus to the last, as it sig- nified the school in which the gladiators played or fenced with wooden foils (rudes) preparatory to the bloody encounters of the arena. That the ludiones were Tuscans even in the clas- sical age, is clear from Plautus, Curculio, I. 2, 60, sqq. : " pessuli, heus, pessuli, vds salutd lubens fite causa mea ludii barbari ; subsilite, dbsecro, et mittite istanc foras," pun- ning on the resemblance of pessuli to the prcesules of these Tuscan dancers (see Non. Marc. c. XII. de Doctorum Inda- gine, p. 783, Gottofr.). Luna, the Tuscan port, probably got its name from the half- moon shape of the harbour. See Pers. VI. 7, 8 ; Strabo, V. p. 222 ; Martial, XIII. 30. The Tuscan spelling was perhaps Losna (=Lus-nd), which is found on a patera (see Muller, Etrusk. I. p. 294). Manus or Manis, "good." Apparently a Tuscan word; at any rate, the manes were Tuscan divinities. Fest. p. 146, s. v. Manuos ; Serv. ad ^n. I. 139, III. 63. So cerus manus, in the Salian song, was creator bonus. Fest. p. 1 22, s. v. Matrem matutam ; comp. Varro, L. L. VII. 26. We may perhaps recognise the same root in a-mcenus, Lithuan. aimesnis. Mantisa, " weighing-meat." Fest. p. 132 : " Mantisa addita- mentum dicitur lingua Tusca, quod ponderi adicitur, sed dete- rius et quod sine ullo usu est. Lucilius : mantisa obsonia vincit" Scaliger and Voss derive it from manu-tensa, " eo quod manu porrigitur." It is more probably connected, like me-n-da, with the root of ^arrfv ; compare frustum with frustra. Nanus, " the pygmy." Lycophr. Alex. 1244 : NaVo? irKavalai ITO.VT epevvqcras jj^v^ov. Ubi Tzetzes : o 'OcWcrei)? Trapd TO?? Tvpcrrjvdis vavos KaAelrcu, crfhovvro's TOV ovo/u.aTo>a, which was its name in Argos and Arcadia. Trutnft=tru-ten-fit : see s.v. Haruspeoc. Vorsus, " one hundred feet square," is quoted as both Tuscan and Umbrian. Fragm. de Limit, ed. GOBS. p. 216: "Primum agri modulum fecerunt quattuor limitibus clausum figurae, quadratae similem, plerumque centum pedum in utraque parte, quod Grseci irXeQpov appellant, Tusci et Umbri vorsum." For the use of 7r\e9pov, see Eurip. Ion. 1137. The fact that vorsus is a Tuscan word confirms the etymologies of Vertum- nus and Nortia. $4-] THE ETRUSCAN LANGUAGE. 165 4. Etruscan Inscriptions Difficulties attending their Interpretation. In passing to our third source of information respecting the Tuscan language the inscriptions which have been preserved we are at once thrown upon difficulties, which at present, per- haps, are not within the reach of a complete solution. We may, indeed, derive from them some fixed results with regard to the structure of the language, and here and there we may find it possible to offer an explanation of a few words of more frequent occurrence. In general, however, we want a more complete collection of these documents ; one, too, in deciphering which the resources of palaeography have been carefully and critically ap- plied. When we shall have obtained this, we shall at least know how far we can hope to penetrate into the hitherto unex- plored arcana of the mysterious Etruscan language. Referring to the theory, that the Etruscan nation consisted of two main ingredients namely, Tyrrheno-Pelasgians, more or less intermixed with Umbrians, and Rzetians or Low Ger- mans 1 , the former prevailing in the South, the latter in the 1 The idea that one ingredient, at least, in the old Etruscan language was allied to the most ancient type of the Low German, as preserved in the Icelandic inscriptions, occurred to me when I was reading the Runic fragments with a different object in 1846. A long series of independent combinations was required before I could bring myself to attach any im- portance to the primd facie resemblances which struck me on the most superficial comparison of documents, apparently so far removed from the possibility of any mutual relations. But I have quite lately discovered that the same first impressions were produced and recorded just one hun- dred years before I communicated my views to the British Association. A folio tract has come into my hands with the following title : Alphdbetum veterum Etruscorum secundis curls inlustratum et auctum a Joh. Chrst Amadutio [Amaduzzi], Rom. 1775, and I find the following statement in p. XLI.: "nemo melius hujusmodi cerebrosa tentamina ridenda suscepit quam anonymus quidam scriptor (qui Hieronymus Zanettius Yenetus a quibusdam habitus est) qui anno 1751 opusculum (Nuova trasfigurazione delle lettere Etrusche) edidit lepidum et festivum satis, in quo .... literas quibus [monumenta Etrusca] instructa sunt Geticas ac Runicas potius . . . statuendas comminiscitur . . . . Id etiam nonnullis Runicis sive Geticis adductis monumentis et cum iis, quee Etrusca censentur, facta comparatione evincere nititur." With more etymological knowledge, but with the same inability to appreciate the importance of the evidence which he 166 THE ETRUSCAN LANGUAGE. [On. Y. north-western part of Etruria, it is obvious that we cannot expect to find one uniform language in the inscriptions, which belong to different epochs and are scattered over the territory- occupied in different proportions by branches of cognate tribes. Accordingly, we must, if possible, discriminate between those fragments which represent the language in its oldest or un-Rasenic form, and those which exhibit scarcely any traces of a Pelasgic character. fi 5. Inscriptions in which the Pelasgian element pre- dominates. Of all the Etruscan cities the least Rasenic perhaps is Caere 1 or Agylla, which stands in so many important connexions with Rome. Its foundation by the Pelasgians is attested by a great number of authorities (Serv. ad ^En. VIII. 478 ; Strabo, V. p. 220 ; Dionys. Hal. III. 58 ; Plin. H. N. III. 8) : its port, Hvpyoi, had a purely Pelasgian or even Greek name, and the Pelasgians had founded there a temple in honour of Ei\q6via (Strabo, V. 226; Diod. XV. 14). In the year 534, B.C., the people of Agylla consulted the oracle at Delphi respecting the removal of a curse; and they observed, in the days of Hero- dotus, the gymnic and equestrian games which the Pythoness prescribed (Herod. I. 167) : moreover, they kept up a con- nexion with Delphi, in the same manner as the cities of Greece, and had a deposit in the bank of the temple (Strabo, V. p. 220). As the Agyllaeans, then, maintained so long a distinct Pe- was adducing, the reviewer of Jakel's superficial book in the Quarterly Review (Vol. XLVI. p. 347) remarks : " It is strange but true that some of the most striking coincidences are between the Latin and the Teutonic dialects of Scandinavia and Friezeland regions which Roman foot never touched. Here are a few of the Scandinavian ones : abstergo, affstryka; dbstrdho, affdraga ; earns, Tcaer ; candela, kindel; clivus, kleif (cliiff) ; &c. In all these cases the word has disappeared, or at least become unusual, in the German. In Friezeland hospes is osb, macula is magi, rete is rhwyd, turtus is turtur, &c." 1 Lepsius (die Tyrrh. Pelasger, p. 28) considers Caere an Umbrian and not a Pelasgian word, -re being a common ending of the names of Um- brian towns; thus we have Tute-re on coins for Tuder. The original name was perhaps Kaiere, which contains a root expressive of antiquity and nobility (above, p. 5). $5.] THE ETRUSCAN LANGUAGE. 167 lasgian character, we might expect to find some characteristics in the inscriptions of Caere, or Cervetri, by which they might be distinguished from the monuments of northern and eastern Etruria. There is at least one very striking justification of this supposition. On an ancient vase, dug up by General Galassi at Cervetri, the following inscription is traced in very clear and legible characters : Mi ni keOuma, mi maOu maram lisiai Qipurenai ; JEOe erai sie epana, mi neOu nastav helefyu. It is obvious that there is an heroic rhythm in these lines ; the punctuation and division into words are of course conjectural. This inscription differs from those which are found in the Umbro- Etruscan or Rasenic districts, and especially from the Perusian cippus, in the much larger proportion of vowels, which are here expressed even before and after liquids, and in the absence of the mutilated terminations in c, I, r, which are so common in the other monuments. The meaning of this couplet seems 'to be as follows : " I am not dust ; I am ruddy wine on burnt ashes : when" (or " if") " there is burning-heat under ground I am water for thirsty lips." Mi is clearly the mutilated e-juu=ecr-/x/. That the substantive verb may be reduced to e-^u, with the first syllable short, is clear from the inscription on the Burgon vase, which Bockh has so strangely misunderstood, (C. I. n. 33), and which obviously consists of three cretics : TU>V 'A6r) \ -vrjQev a- | 6\cov e/A/. ||. Ni is the original negative, which in Latin always appears in a reduplicated or compounded form. The same form appears in Icelandic. KeOuma is the primitive form of y9wv, X&xMa-Xos, x a ^ af 'j humus, &c. ; and may not p-#a/za- be an off- shoot of the Hebrew nDltf , in which the aleph, as in many other cases, represents a stronger guttural? (see above, p. 76). The difference of quantity in the second mi will not prevent us from identifying it with the first, which is lengthened by the ictus. MaOu is the Greek /uedv, Sanscr. madliu. Maram is the epithet agreeing with mathu : it contains the root mar-, found in Mdpwv (the grandson of Bacchus), and in "Icr-jmapos, the site of his vineyards (see Od. IX. 196, sqq.), and probably signifying "ruddy" (/maipw, vaipa, &c.). The fact that Maro was an agri- cultural cognomen at Mantua is an argument in favour of the Etruscan use of the root. Lisiai is the locative of lisis, an old word corresponding to lix, " ashes mingled with water." 168 THE ETRUSCAN LANGUAGE. [Cn. V. Qipurenai is an adjective in concord with lisiai, and probably containing the same root as tepidus, tephral, teforom, &c. (above, pp. 48, 132). EOe is some particle of condition or time. Erai is the locative of epa, " earth." The idea of this second line is conveyed by the sneer of Lucretius, (III. 916, sq. Lachmann): "Tanquam in morte mali cum primis hoc sit eorum, Quod sitis exurat miseros atque arida torres." where Lachmann quotes Cyrill. ctTro/ca^a ustilacio, torres ; and it is probable that epana is synonymous with torres, and that it may be connected with Sairrw, &c., as epulce is with Scnrdvri, daps, SGITTVOV, &c., or ignis with the root dah, " to burn." Sie (pronounced sye) is siet = sit (so ar-sie, ad-sis and si=sit in the Eug. Tables). There can be little doubt that neOu means " water" in the Tuscan language. There is an Etruscan mirror in which the figure of Neptune has superscribed the word Nethuns=Nethu-n-[u]s. The root is ne-, and appears under a slightly different development in the next word, nastav (comp. i/aovuos, vaOfjios, O. H. G. naz)> which is probably a locative in - j is instructive. We might suppose from this that Ari-timi-s, the " virgin of the sea," and 'Ape-Qovcra, " the virgin swiftly flowing," were different types of one and the same goddess (see above, pp. 37, 54). 'A^re/u^s appears to me to be a derivative from ''A precis. The next words probably contain the name and description of the person who made the offering. The name seems to have been Fastia Rufrunia or Rufria. Lanzi and Muller recognise a verb in turce, which is of frequent occurrence on the Etruscan monuments, and translate it by 7roc6c, dedit, aveOrjKe, or the like. Lanzi goes so far as to suggest the etymology [e-]$ft>'p//ce. And perhaps we might make a verb of it, were it not for the context. Its position, however, between the proper name and the word clen, which in all other inscriptions is immediately appended to the name and description of a person, would induce me to seek the verb in ceca (probably a reduplication, like pepe on the Todi statue: compare chu-che, cechaze in the Perugian inscription, and cechase on the Bomarzo sarcophagus, Dennis, I. p. 313), and to suppose 6.] THE ETRUSCAN LANGUAGE. 171 that Turce is the genitive of the proper name Tuscus. The word clen, one of the two to which I have referred, is explained by its contrast to eter, etera, a word clearly expressing the Greek ere^o?, Latin alter (iterum), and Umbrian etre. Thus we have on the same monument ; La . Fenete La . Lethial etera Se . Fenete La . Lethial clan: in which, if etera means, as is most probable, the second in the family, clan must mean the first or head of the family. I would not on this account infer that clan was the ordinal corresponding in every case to primus; but there will be little difficulty in showing etymologically its appropriateness as the designation of the first of a family. The root, which in the Greek and Latin languages signifies head, summit, top, is eel-, cul-, cli-, /co\-, Kop-, or Kpa-. These are in effect the same root, compare glisco, cresco, &c. ; and it is well known, that words denoting height and elevation or head-ship, in fact are employed to signify rank. Now the transition from this to primogeniture the being first in a family is easy and natural : compare the " patrio princeps donarat nomine regem" of Lucretius (I. 88). Therefore, if clen or dens (in Latin clanis or clanius) is con- nected with the root of celsus, cul-men, collis, clivus, KoXofaov, Kopvfprj, Kvpios, Koipavos, Kovpos, AcOjOos, Kupflas, Kpdviov, &c., it may well be used to signify the first in a family. Cf. the Hebrew WVht "de cujuscunque rei initio, principio, origine (velut flumi- nis\ summitate, velut de montium verticibus, &c." (Furst, Cone. s. v.). This etymological analysis will perhaps be complete, if I add that there were two rivers in Italy which bore the name of Clanis or Clanius ; the one running into the Tiber between Tuder and Volsinii, the other joining the sea near the Tuscan colony of Vulturnum. Now the names of rivers in the Pelasgian language seem to have some connexion with roots signifying "height," "hill," or "hill-tower." This has been indicated above in what has been said of the names of the Scythian rivers (Chap. II. $ 10). The Tibe-ris the " Tuscan river," as the Latin poets call it seems to have derived its name from the Pelasgian Teba, " a hill," and the root ri, " to flow" (see above, Chap. IV. 2). And the Clan-is and Clan-ius, which flow down from the Apennines, may well have gained a name of 172 THE ETRUSCAN LANGUAGE. [Cn. V. similar import. If we now pass on to the northern languages, we shall find some curious extensions of these results. For while the root kl- in klif, kliffe, kleyf, signifies altitude and climbing, and while klackr in Icelandic denotes "a rock," we find that, with the affix n, klen or klien in Icelandic, and in Germ, klein, signify " little," but primarily in the sense of " a child" as opposed to "a man;" and it may be a question whether the idea of derivation, which I have just indicated in the river as compared with the mountain, may be at the basis of the ordinary meaning of klen or fcleine. And thus whether the Etruscan clen signifies " the eldest child," or simply " the child," with an implication of primogeniture, as indicating the first contrast with the parents, the Icelandic will help the explanation. The only bilingual inscription, in which I have found clans, seems to imply that, unless otherwise expressed, this word merely denotes sonship. It is (Dennis, II. p. 426) : V. Caszi C. clans C. Cassius C. F. Saturninus. Where C. Clans = C. F., the cognomen Saturninus being an addition in the Latin version. This view is confirmed by the fact that clan sometimes occurs in the same inscription with the matronymic in -al, as in the inscription quoted above ; and while in the bilingual inscriptions this matronymic is rendered by natus, clans, as we have seen, is translated filius, and sometimes filius is added without any corresponding clan in the Etruscan inscriptions. The following examples will show all the different usages of this adjunct : A. Clan or clen used with a genitive case and without any patronymic. a, Phasti Ruphrua Turce den ceca. (Gori, Mm. Etrusc. I. pi. 32). b. V. Caszi C. Clans. (Dennis, II. p. 426). C. Cassius C. F. Saturninus. B. Clan, with a patronymic, and without a genitive : Laris Pumpus Arnthal clan cechase. (Dennis, I. p. 313). And so in the second inscription quoted above. 6.] THE ETRUSCAN LANGUAGE. 173 C. Patronymic without clan, but with filius in the Latin translation, (a). VI. Alphni num. carnal C. AlfiusA.F. Cainnia natus. (Dennis, II. p. 354). (b). Vel. Venzileal Phnalisle C. Vensius C. F. Ccesia natus. (Id. II. p. 371). (c). Cuint. Sen*t. Arntnal Q. Sentius L. F. Arria natus. (Id. II. p. 412). (d). Pup. Velimna Au. Caphatial P. Volumnius A. F. Violens Cafatia natus. (Id. II. p. 475). From this it appears that clan represents the son or daughter as opposed to the father, the mother's name being given in the matronymic. The other of the two words in this inscription, to which I have adverted, is phleres, which clearly means donarium, or something of the kind. This word, as we shall see directly, occurs on a number of small Etruscan objects, which are of the nature of supplicatory gifts. And it would be only fair to con- clude that the word denotes " vow " or " prayer," as included in the donation. Now we know from Festus (p. 230, cf. 77, 109) that ploro and imploro or endoploro in old Latin signified inclamo without any notion of lamentation or weeping. If, then, we compare the Icelandic fleiri^ Suio- Gothic flere with the La,tmplures= pie- ores, we shall easily see how phleres may con- tain the same root as ploro=ple-oro (below, Ch. XII. $ 2), espe- cially since the Latin language recognises a similar change in fleo compared with pluo. The word is then in effect equivalent to the Greek avaOrj^a, as in Cicero (ad Attic. I. 1) : " Her- mathena tua valde me delectat, et posita ita belle est ut totum gymnasium ri\iov avdO^a esse videatur." Thus it means a votive offering, like the votiva tabella of the ancient temples, or the v oto of the modern churches in Italy ; and it is easy to see how the ideas of " vow," " prayer," " invocation," " offering," may be represented by such an object. Accordingly the in- scription of the laurel-crowned Apollo will signify : Sum votivum donarium Apollini atque Artemidi; Fastia Rufria, Tusci filia, faciundum curavit. For if we compare ceca with cechase 174 THE ETRUSCAN LANGUAGE. [Cn. V. or cechase, we may render it with reference to the Icelandic kasa, Danish kokase, "to heap up" or "build." 7. Inscriptions containing the words SUTHI and TRCE. It has been mentioned that the word phleres appears on a number of smaller or moveable objects. In some of these it has appended to it the word tree or three. Thus we have cen phleres tree sansl tenine. (Vermiglioli, p. 31). cen phleres tree. (Micali, Antichi Monumenti, pi. 44. n.2). eca ersce nac achrum phler-thrce. (Dennis, I. p. xc.) The second of these inscriptions is found on the toga of the statue of Aulus Metellus ; the third appears on an amphora found at Vulci, and in connexion with a picture representing the farewell embrace of Admetus and Alcestis. It may be assumed then that the amphora was a farewell offering from a husband to his deceased wife, and that the monument of Metellus was sepul- chral or funereal. If then phleres signifies a votive offering, the additional word tree or three must indicate " mourning" or " sor- row." And here the northern languages at once come to our aid ; for in Suio-Gothic trcega dolere and trcege = dolor ; and in Icelandic at trega = angere or dolere, and tregi - dolor ; and to the same root we may refer the Icelandic ihrek-gravis labor or molestia ; for tregi also means impedimentum. See Specimen Glossarii ad Edd. Scemund. Vol. II. p. 818 : " (at) Trega (A) ' angere/ * dolorem causare/ B. I. 29 : tregr mik that, ' id mihi aegre est,' G, III. 3 : tregrath ydr ' molestum non est vobis/ GH. 2. (B) * dolere 1 ' lugere.' Hinc treginn ' deploratus' 1. * deplorandus' unde foam. pi. tregnar. Priori sensu A. S. tregian. Tregi ' moeror, dolor' (passim), Germ, trauer. Trcege, trege 6 vexatio/ ' indignatio.' Originitus forsan verbotenus : ' onus/ 'moles.' Germ, tracht, Dan. draght, Angl. draught. Cf. tregr ' invitus/ ' segnis/ Germ, trdg, Al. treger. Forsan a draga ' trahere/ ' portare.' Treg-rof * luctuum/ 1. * calamitatum series vel etiam discussio.'" The connexion of this word with traho brings it into still greater affinity with the old languages of Italy, and the evidence from the context is conclusive for the meaning. Many Etruscan inscriptions begin like the three quoted above 7.] THE ETRUSCAN LANGUAGE. 175 with eca, cen, or cehen, which are obviously pronouns or adverbs signifying ' here* or ' this,' in accordance with the root k- which appears in all the Indo- Germanic languages. The Cervetri in- scription has taught us (above p. 168) that era signifies 'earth* (N. H. G. erde, Goth, airtha, Altfr. irthe, Gr. epa). Conse- quently, ersce would naturally denote an earthenware vessel, for -ska is a very common termination in Icelandic names, as bern- ska " childishness," ill-ska " malice," &c. And as cen or cehen is probably an adverb, eca must be the feminine of the prono- minal adjective ecus, eca, ecum, agreeing with ersce. As achrum is clearly the locative of acher which occurs in the great Peru- gian inscription, and which at once suggests the Icelandic akr, Germ, acker, ager, we may fairly conclude that nac is the pre- position which, under the form na, nahe, nach is found in all the Teutonic and Sclavonian languages : and thus the Yulci inscrip- tion will mean : " this earthen vessel in the ground is a votive offering of sorrow." By the side of cen phleres we have, on larger monuments, eca or cehen suthi or suthinesl. Thus we find : eca suthi LartMal Cilnia (Dennis, I. p. 500.) cehen suthi hinthiu times (Vermiglioli, I. p. 64.) eca suthinesl Titnie (Dennis, I. 242, 443.) eca suthi Amcie Titial (Vermiglioli, I. p. 73.) Here again the Icelandic comes to our aid, for sut is dolor, mcestitia, luctus, so completely a synonym of tregi that we have tregnar and sutir in the same stanza of Hamdis-Mal (Edd. Scemund. II. p. 488); and nesla or hnesla=funis, laqueus : so that we may translate eca-suthi, " this is the mourning," and eca suthinesl " this is the sorrowful inscription." Comparisons of individual words in languages not known to be the same are of course eminently precarious. But it is impossible to resist the evidence of affinity furnished by the fact that the words tree and suthi, constantly occurring on Etruscan monuments of a funereal character, are translated at once by the Icelandic synonyms tregi and sut, both signifying " grief" or " sorrow." If we had only this fact we should be induced by it to seek for further resem- blances between the old languages of Northern Europe and the obscure fragments of the old Etruscan. 176 THE ETRUSCAN LANGUAGE. [Cn. V. 8. Inferences derivable from the words SVEB, OVER, and THUR or THAUR. In comparing an unknown with a known language, we derive much help from the collocation of the same or similar words, especially in short sentences. Thus when we find such collocations as the following: phleres zek-sansl ever (Vermiglioli, p. 36), phleres tlen-asies sver (id. p. 39), we can hardly avoid supposing that ever and sver are slightly- different forms of the same word. Now in Icelandic we find the verb thverra = minui, disparere and the adjective thverr = tranversus with its adverb thverz = transversim (vid. JEdd. Scemund. Vol. II. Spec. Gloss, pp. 859, 860). In the cognate languages we find the same change in this word as in the ever and sver of the Etruscans : for while the Icelandic thverr, Engl. thwart, Dan. tver, Germ, zwerch, exhibit the dental more or less assibilated as in sver, the German quer and English queer give us a guttural instead of a sibilant as in ever. The appearance of ever or sver in sepulchral inscriptions (for we have sver in one beginning with eca suthi, Vermiglioli, p. 73), would lead us to suppose that this word or these words must refer to death or prostration, and this is a meaning included in the Icelandic word, whether or not connected with var, "male," " parum." The forms of thverra, when passive, are ek thverr, thvarr, thorinn ; when active, ek thverra, thverda : and thurr, thurt, thyrrinn, signify " aridus," " siccus," like the German durr. Without stopping to ask whether these latter forms are derived in any way from the verb thverr, which is quite possible, it is worthy of remark that in those sepulchral inscriptions, in which the word ever or sver does not occur, we have in corresponding places the word thaure, thurasi (Vermigl. p. 64), thuras, thaura, ihuruni (Inscr. Per. 11. 6, 20, 41). And in one old epitaph (Lanzi, Saggio, II. p. 97, no. 12) we find : mi suthi L. Felthuri thura, where the position of the last word almost leads us to render it : " I am the lamentation for L. Felthurius deceased." The inferences derivable from the appearance of these forms is that connected words significant of decay, prostration, and death, and liable to the same modification, probably existed both in Old 8.] THE ETRUSCAN LANGUAGE. 177 Norse and in Etruscan. The amount of probability depends upon the cumulative effect of the other evidence 1 . 9. Striking coincidence between Etruscan and Old Norse in the use of the auxiliary verb LATA. Whatever may be thought of the verbal resemblances be- tween the Old Norse and the language of the Etruscan fragments, it must be admitted by all sound philologers that we have an indisputable proof of the affinity of these idioms in the gram- matical identity which I communicated to the British Association 2 . Every reader of the Runic inscriptions must have noticed the constant occurrence of the auxiliary or causative verb lata facer e in causa esse, of which the Eddas give us the forms ek Icet, let, latinn. Thus we find : Lithsmother lit hakva stein aufti Julibirn fath, i. e. " Lithsmother let engrave a stone after (in memory of) his father Julibirn." Thorstin lit gera merki stir Suin fathur sin, i. e. " Thorstin let carve marks in memory of his father Sweyn." Ulfktil uk Ku uk Uni thir litu raisa stin iftir Ulf fathur sin, i. e. " Ulfktil and Ku and Uni, they let raise a stone in memory of their father Ulf" (vide Dieterich, Runen-Sprach-Schatz, p. 372). Now we have here, as part of a constantly-recurring phraseology, an auxiliary verb, signifying " to let" or " cause" followed by an infinitive in -a. On reading the first line of the longest Etruscan inscription, that of Perugia, we seem to stumble at once upon this identical phraseology, for we find : eu lat tanna La Rezul amev achr lautn Velthinas. If we had no other reason for supposing that there was some connexion between the Scandinavians and Etruscans, we could not avoid being struck by this apparent identity of construction. As, however, we have every reason to expect resemblances between the two languages, it becomes a matter of importance to inquire whether the grammatical identity can be established, and this amounts to the proof that lat and tanna are both verbs. 1 I may mention in passing that suer actually occurs in Runic inscrip- tions in the sense " father-in-law ;" thus : iftir Kuthrikr suer sin (Die- terich, Runen-Sprsch. p. 265) ; but that I do not regard this as more than an accidental coincidence with the expressions under consideration. 2 Report, 1851, p. 158. 12 178 THE ETRUSCAN LANGUAGE. [Cn.V. Of course there is no primd facie reason to conclude that tanna is a verb. On the contrary, Niebuhr (Kleine Schriften, II. p. 40) thinks that thana is a noun signifying "a lady," and that Tanaquil is only a diminutive of it ; and Passeri, whom he quotes, suggests that Thana is a title of honour, nearly equi- valent in meaning, though not of course in origin, to the modern Italian Donna (from domino). Fortunately, however, about the time when this comparison between the Runic and Etruscan phraseology first occurred to me, Mr J. H. Porteus Oakes re- turned from a tour in Italy, and presented to the Museum at Bury St Edmund's a small patera or saucer, which he had obtained at Chiusi, and which exhibits the following legend : stem tenilaeth nfatia. This at once furnished me with the means of proving that lot tanna in the Perugian Inscription were two verbs, the latter being an infinitive and the former an auxiliary on which it depends. For it is obvious that tenilaeth is the third person of a transitive verb, the nominative being Nfatia, probably the name of a woman (cf. Caphatial = Cafatia natus in Dennis's bilingual inscription, II. p. 475), and the accusative being stem for istam, Umbr. est- (cf. mi with e-mi, &c.). The verb tenilaeth manifestly belongs to the same class of forms as the agglutinate or weak-perfects in Gothic, which are formed by the affix of the causative da, as soki-da, "I did seek" (Gabelentz u. Lobe, Goth. Gramm. $ 127). We have this Gothic formation in the Latin ven-do, pen-do, &c. ; and I have discussed in a subsequent chapter the remarkable causa- tives in -so, -sivi, as arces-so, capes-so, quce*so, &c. It is clear then that lat tanna represents as separate words what tenilaeth exhibits in an agglutinate form. In the latter case the auxiliary is in the present tense, which in Gothic is formed in th; and lat is a strong perfect. There is no difficulty about the meaning of tanna, teni, which are clearly identical with the Icelandic thenia = tendere, O. H. G. danjan, denjan, A. S. dhenjan, N. H. G. dehnen, Gr. Teivto, ravvw, Sanscr. tan-, and therefore signify " to offer," like the Latin porrigo or porricio. If this is the true explanation of the root when it occurs as a verb, we may reasonably apply the same interpretation to its use as a noun. In this use it appears under all the different forms Thana, Thania, Thama, Tania, Tannia, Dana, and Tha (Miiller, Etrusk. II. 303, 315). From the collocation it is clear that the 9.] THE ETRUSCAN LANGUAGE. 179 word is equivalent to phleres, or rather it signifies "an offering" generally, without the implication of a vow or prayer. Thus, while we have in the only urn with an inscription among the Etruscan specimens in the rooms adjoining the Egyptian collec- tion in the British Museum : thana celia cumniza, we find on one of Lanzi's (Saggio, II. 506. no. 15) : mi thana Arntha, which is quite analogous to mi phleres or mi suthi. It is worthy of remark that ten-do, which is an agglutinate form like teni- lata, is synonymous with porrigo ; thus we have in Cicero (de Oratore, I. 40. fi 184) : " presidium clientibus atque opem amicis et prope cunctis civibus lucem ingenii et consilii sui porrigentem atque tendentem;" and we may compare such phrases as duplices tendens ad sidera palmas with porrigit exta manus, and the like. Even the Umbrian has pur-tin-sus = por-rexeris (Eug. Tab. I. b, 33). In ritual phraseology therefore the Latin lan- guage comes sufficiently near the language of this patera, and stem tenilaeth Nfatia bears as close a resemblance to istam tendit (vel porrigit) Nefatia, as we have any right to expect. The Perugian inscription, however, is even nearer to the Runic than this patera legend is to the Latin ; and the evidence fur- nished by the two, taken together, seems to be quite conclusive in proof of the affinity between the Etruscan and Old Norse languages. As lautn and lautnescle occur together on another Etruscan sepulchre, there can be no objection to connect them with the Icelandic laut = lacuna, locus depressus et defossus ; and eu from is is strictly analogous to the Latin ceu from ce, cis ; accordingly, comparing amev with the Icelandic ama = ango, the beginning of the Perugian Inscription will be rendered as naturally and easily as one of the Runes : " Here Lartius the son of Rsesia let offer or give a field of mourning as or for the grave of Velthina." To return to the patera, its companion, now in the possession of Mr Beckford Bevan, bears a legend which is also capable of translation by the help of the Old Norse. The words are : flenim thekinthl thmtflaneth. It is obvious that we have here the name of a man, a transitive verb, and the accusative of the object, which is an open patera or saucer. As therefore in Icelandic flenna = hiatus, chasma, we may explain flenim by an immediate reference to the proper meaning of patera from pateo : cf. patulus; and as in Icelandic tham = egelida obscuritas aeris ; tef - morari ; and lana = mutuum 122 180 THE ETRUSCAN LANGUAGE. [On. V. dare, credere, commodare, Engl. "lend," the compound verb tham-tef-lan-eth will mean " he lendeth for a dark dwelling," and the whole inscription will run thus : Thekinthul dot pateram ad commorandum in tenebris. Verbs compounded of nouns and verbs are not uncommon in Icelandic; thus we have halshoggra, " to behead," brennimerkjd, " to brand," &c. It only remains to remark, that as the Gothic auxiliary -do is found in Latin, so the Norse lata must be recognised in a fainter form in some Latin verbs in ~lo, as well as in the Sclavonic formations in -I, and in the Old Norse diminutives or frequenta- tives in -la, such as rug-la, " to turn upside down," from rugga, " to remove," tog-la, " to let chew," or "chew over again," from tyfffffa* &c - $ 10. The great Perugian Inscription critically examined its Runic affinities. The facility with which the philologist dissects the Etruscan words which have been transmitted to us, either with an inter- pretation, or in such collocation as to render their meaning nearly certain, and the striking and unmistakable coincidences between the most difficult fragments and the remains of the Old Norse language, might well occasion some surprise to those who are told that there exists a large collection of Etruscan inscriptions which cannot be satisfactorily explained. One cause of the un- profitableness of Tuscan inscriptions is to be attributed to the fact, that these inscriptions, being mostly of a sepulchral or dedi- catorial character, are generally made up of proper names and conventional expressions. Consequently they contribute very little to our knowledge of the Tuscan syntax, and furnish us with very few forms of inflexion. So far as I have heard, we have no historical or legal inscriptions. Those which I have in- spected for myself are only monumental epitaphs and the dedica- tions of offerings. These observations might be justified by an examination of all the inscriptions which have been hitherto published. It will be sufficient, however, in this place to show how much or how little can be done by an analysis of the great inscription which was discovered in the neighbourhood of Perugia in the year 1822. This inscription is engraved on two sides of a block of 10. J THE ETRUSCAN LANGUAGE. 181 stone, and consists of forty-five lines in the whole ; being by far the most copious of all the extant monuments of the Tuscan lan- guage. The writing is singularly legible, and the letters were coloured with red paint. The following is an accurate transcript of the facsimiles given by Micali ( Tav. CXX. no. 80) and Vermiglioli (Antiche Iscri- zioni Perugine, ed. 2, p. 85). 25. velthinas. 26. atena . zuk- 27. i . eneski . ip- 28. a . spelane . 29. this .fulumch- 30. va . spel . thi- 31. rene. thi. est. 32. ak . velthina 33. ak . ilune . 34. turunesk . 35. unezea . zuk- 36. i. eneski . ath- 37. umics . afu- 38. nas . penthn- 39. a . ama.velth- 40. ina . afun . 41. thuruni . ein . 42. zeriunak .ch- 43. a. thii. thunch- 44. ulthl . ich .ka. 45. kechazi.chuch- 46. . 1. eu . lat . tanna . la . rezul . 2. amev . achr . lautn . velthinas . e- 3. -st . la . afunas . slel . eth . karu- 4. tezan .fusleri . tesns . teis . 5. rasnes . ipa . ama . hen . naper . 6. xn . velthina . thuras . aras . pe- 7. ras . kemulmleskul . zuki . en- 8. eski . epl . tularu . 9. aulesi . velthinas . arznal . kl- 10. ensi . thii . thils . kuna . kenu . e- 11. plk .felik . larthals . afunes. 12. klen . thunchulthe . 13. falas . chiem .fusle . velthina . 14. hintha . kape . muniklet . masu . 15. naper . srankzl . thii .falsti . v- 16. elthina . hut . naper . penezs . 17. masu . aknina . klel . afuna . vel- 18. thinam . lerzinia . intemam . e- 19. / . fe^ . velthina . zias . atene. 20. tesne . eka . velthina . thuras . th- 21. aura . helu. tesne . rasne . ^* . 22. tesns . m . rasnes . chimth . sp . 23. el. thutas . kuna . afunam . ena . 24. hen . naper . ki . knl . hareutuse . Now, if we go through this inscription, and compare the words of which it is composed, we shall find that out of more than eighty different words there are very few which are not 182 THE ETRUSCAN LANGUAGE. [On. V. obviously proper names, and some of these occur very frequently ; so that this monument, comparatively copious as it is, furnishes, after all, only slender materials for a study of the Tuscan lan- guage. According to the most probable division of the words, the contents of the inscription may be considered as given in the following vocabulary : Achr(2). Afun (40). Afuna (17). Afunam (23). Afunas (3, 37). Afunes (11). Ak (32, 33). Aknina (17). A ma (5, 39). Amev (2). Aras (6). Arznal (9). Atena (26). Atene (19). Athumics (36). Aulesi (9). Gha (42). Chiem (13). Chimth (22). Chuche (45). Einzeriunak (42). Eka (20). Ena (23). Eneski (7, 27). Epi(\l). Eplt (8). Er (18). Est (2, 31). J&A (3). Eu (1). (11). Fulumchva (29). , jfcfort (13, 4). Hareutuze (24). JSfcfo (21). Hen (5, 24). JZwifAa (14). Hut (16). JcA (44). flune (33). Intemam (18). //HI (5, 27). #a (44). .ST^^ (14). Karutezan (4). Kechazi (45). jfti (21). Kemulmleskul (7). (10). i (24). (17). Klen, klensi (9, 12). JT<(19, 24). Kuna (10, 23) [" a wife," Diete- rich, Runen-Sprsch. p. 11 7-] Xa (1, 3). LarthaU (11). Zflrf (1). Lautn (2). Lerzima (18). Jlifam (14, 17). Muniklet (14). (5, 15, 16, 24). (16). Penthna (38). P^TflW (6). Raine, Raines (5, 21, 22). 10.] THE ETRUSCAN LANGUAGE. 183 Thunchulthe (12). Thunehulthl (43). (23). Tularu (8). Turunesk (34). VdtMna, Velthinas, Velthinam (6, 13, 15, 19, 20, 32, 39, 2, 9, 25, 17). (35). (19) [Zia " an aunt" in Mo- Rezul (1). Slel (3). 'Spel, spelane (22, 28, 30). 'Srankzl (15). Tanna (1). Teis (4, 22). Tesne, tesns (5, 20, 21, 22). Thaura (20). 7^, this, thii, thil, thils (29, 31, 10,43). Thuras, thirene, thuruni (6, I dern Tuscan.] 30,41.) I Zuki (7, 26, 35). The first remark to be made respecting this inscription is, that though we have here obviously a different language from that in which the Eugubine Tables are written, still there are many words which in outward form at least resemble the Um- brian phrases. Thus we have eu (v. 1), velthina (passim), est (2), karu- (3), tesns (4), kape (14), muniklet (14), turn- (24), einzeriu- (41), &c., which may be compared with eu> veltu, est, karu, tesenakes, kapi, muneklu, tures, anzeriatu, &c., in the Eugubine Tables, though it does not at all follow that there is any similarity of meaning in addition to the mere assonance. The word naper (5, 15, 16, 24) seems to have the termination -per, so common in Umbrian : we may compare it with the Latin nu-per (pro novo). But although no profitable results can be expected from a comparison between syllables occurring in this inscription and others of similar sound picked at random from the Eugubine Tables, something might be done if we had a large number of smaller inscriptions, written in the same language, derived from the same neighbourhood, and treating in different ways on the same or kindred subjects. To show this I will quote another Perugian inscription, and place side by side in a parallel column the words or phrases of the great inscription which seem to correspond. The text which I have adopted is that of Vermiglioli, (p. 64). The inscription was first copied by Bonarota in his supplement to Dempster, (p. 98) 1 . It was 1 Bonarota describes the inscription as adhuc easterns in antique cedi- ficio ad modum turris lapidibus grandioribus exstructo et vocatur " S. Maiino." Amaduzzi says it comes ex hypogceo Perusino. 184 THE ETRUSCAN LANGUAGE. [Cn. V. also quoted many years ago, with great inaccuracy, by Amaduzzi (Alphabetum Veterum Etruscorum, Rom. 1775. p. Ixi.) : 1.1. cehen . suthi . hinthiu . thues . hintha (14) sains : Etve : thaure . lautnescle . caresri . Aules . lautn (2) Larthia . precu-thurasi. thuras (6) 1.2. Larthial . svle . Cestnal . den . erasi . eth . Phanl . lautn .precus.ipa . murzua cerurum . em 1.3. tunur . cl utiva aras (6) eth (3) lautn (2) ipa (5, 27) ena (23) heczri telur. In another inscription quoted by Vermiglioli (p. 73) we have caratse by the side of carutezan (4), which must be com- pared with hareutuse (24). The starting-point for a profitable comparison between the Perugian Inscription and that just quoted is furnished by an examination of caratse, carutezan, hareutuze, and the word caresri in the document before us. We have seen above (p. 125) that in the Oscan language -tuset or -tuzet occurs as an auxiliary affix to verbs, in the same way as -do and -so = -sino are used in Latin, -do in Gothic, and lota in Old Norse and Etruscan. There is every reason, then, to suppose that the forms cara-tse, caru-tezan, hareu-tu%e, involve the affix tuxet, or that the Etruscan agrees with the Latin, Gothic, and Oscan, in the use of the auxiliary -do. As the Etruscan also agrees with the Old Norse in the use of the auxiliary lata, which probably occurs also in Sclavonian and Latin forms, we may be led to expect a similar coincidence in regard to the auxiliary so -sino. Now it will be shown in the proper place that the isolated form sero, sevi, is only a by-form of sino, sivi, the primary meaning of both being " to put" or "lay down," i.e. as seed in the ground. In Old Norse sero, in the sense " I sow," is represented by soa, which has a peculiar aorist sera, 8 pers. seri. These Old Norse aorists, such as groa " to grow ;" aorist sing. 1. grera, 2. grerir, 10.] THE ETRUSCAN LANGUAGE. 185 3. greri ; pi. 1. grerum, 2. grerut, 3. greru, &c., have been made the subject of special commentaries bj Aufrecht and Knob- lauch (Zeitschr.f. Vergl Sprf. 1851, pp. 471, 573), who agree in identifying the r with the s of eru\|/a and scripsi, and this again with the substantive verb. Whatever opinion may be formed respecting the origin of this r (and the verb pi- rut from pi=fio, shows that it cannot be derived from the contrasted es-se), it is impossible to overlook the fact that seri is, in Old Norse, a past tense of a verb really identical with that^ which constitutes the causative auxiliary in so many Latin forms. So that care-sri would be quite equivalent to care-tuzet. The root is found under the form kar, kra, gra, mostly with a labial auslaut (as in scrib-o, ypafp-co), but sometimes without (as in "O~t, above, p. 147, -^ap-dcrcru)), and sometimes either with or without, as in the Icelandic kira, gera, kiera, kiara, kara, kerva (Dieterich, Runen-Sprsch. p. 134), N. H. G. kerben, A. S. ceorfan, Engl. "carve," to signify any impression made upon a surface by notching, scratching, indenting, painting, or pointing. We may well conclude therefore that care-sri means, " he caused to write or inscribe.'" 1 And as thyr in Icelandic is = serv-us, Greek Oqs, A. S. theov, M. G. thius, and thues is obviously the gen. of a word thu = theov, the beginning of the inscription runs as if it were pure Low- German or some dialect of the Scandinavian. " Here Aulus Lartius let engrave mourning in honour of " (lit. ' after,' hinthiu - hinter, cf. aufti in the Runic Inscription quoted above, p. 177) "his servant Etfus on the sepulchral tomb," i. e. " hier sut hinter theovs seins Etfa thaure lautnescle lat kara Aules Larthia." We should come, however, to a similar conclusion if thu-es were compared with the Pelasgo- Hellenic 0e?os, " an uncle," rather than with 0fa, " a servant." In fact, the two words fall into a remarkable agreement with one another and with the Pelasgic and German words denoting divinity ; cf. (a) thyr, theov, dio, &c. " a servant," (b) 0e7o?, modern Tuscan zio, (Perug. Inscr. zia) "an uncle," (c) Tyr, Tiv, Zio, " God," (Grimm. D. M. p. 175, and above, p. W, s. v. Famel). To say nothing of the possible interchange in the ideas of relationship and servitude which might bring back Oclos and 9qs, to a common origin in the Sanscrit dhava=vir, maritus, pater-familias, the form of the word flelo? in its other mean- ing sufficiently shows that a labial is absorbed, and this would 186 THE ETRUSCAN LANGUAGE. [On. V. account for the identity of Oei-os = #e'Foe, and the Etruscan thu. For the gen. here, cf. Tues in our Tues-day with its original form Tiv = Div-us. The name of a relation is more to be expected here than that of a servant. The preposition hinthiu, with the gen. may be compared with the Gothic use of hindana, e. g. Ulph. Me. III. 8. That this root occurred in the Um- brian we have already seen (above, p. 85). As I and u are con- stantly confused in the transcripts of Tuscan inscriptions, it might be better to write hinthil for hinthiu, and this would come nearer to hinter, hindan, &c. With regard to the form of the pronoun sain, as compared with sein or sin, it may be remarked that in the Runic inscriptions we have sain, san, sian, as well as sin, (Dieterich, p. 289), and that we have stain, as well as sten, stein, stin, (Dieterich, p. 308). I recognise a form like caresri in heczri, the other verb in this inscription, which may obviously be connected with the Runic haka or hakva, " to hew or carve," (above, p. 177), and this being so, it would be a surprising coin- cidence, if it were only a coincidence, that these three lines should contain two of the verbs which appear in the same way in the Runic inscriptions ; as Lithsmother lit hakva stein ; and Thorstin lit gera merki stir Suin fathur sin ; or both together, as, Inkuth lat landtbro kiara ante stain hakva. The last part of the inscription is mutilated : but it seems plain that ipa is a preposition corresponding to our up, Sanscrit upa, Icelandic uppd, Gothic uf t &c. ; and as murzva seems to refer to murus, Icel. mur, a term well applicable to the tower " grandioribus lapidibus exstructa" on which this inscription was found, we may render heczri ipa murzva, " he let carve upon the building." And it is difficult to resist the impression that cerurum is connected with the Old Norse ker=vas, which is used in the JEdda in the sense of vasarium (Scemund. II. p. 528) : " Gudrum hvarf til skemmo, kumbl konunga or kerom valdi," i. e. " Gudruna contulit se ad promptuarium, cristas regias e vasariis delegit." If this com- parison is valid, cerurum is a genitive plural. In some Runic inscriptions ein t which immediately follows, is used as a definite article before an epithet ; as : Sandulf ein suarti, " Sandulf the swarthy" (Worsaae, Danes and Norwegians in England, &c. p. 281). The last word telur, whether or not related to tularu or the Perugian cippus (1. 8), seems to be a verb, not unconnected with the Icelandic at telid, Swed. taeljd, Dutch tellen, Engl. tell, 10.] THE ETRUSCAN LANGUAGE. 187 the inflexion being that of the Icelandic 3 pers. sing., as in brennr, " he burns," from brenna. On an urn in the British Museum, in the same room with the Nineveh sculptures, we find tulati on a mutilated inscription ; and ris-ti or rais-ti, "he erected," on the Runic stones, might justify the assumption that it is a verb ; but it is impossible to form any plausible conjecture as to its signification. If we now turn back from the inscription, which has thus been examined, to the great Perugian cippus, we shall see that some definite conclusions result from the comparison. First of all, as they are obviously written in the same language, the strong resemblances between the phraseology of the shorter legend and that of the Icelandic Runes must confirm our previous conviction respecting the Old Norse affinities of the longer in- scription. Again, as hinthiu and ipa are manifestly prepositions in the former, we may give a similar value to hintha and ipa in the latter. And as ipa is used with the name of a building in the shorter epitaph, ama which follows it on the cippus, and which seems in the first line to refer to mourning or sorrow, must sig- nify an erection for such a purpose, and therefore the amev achr of the first line must mean a field for the erection of a tomb. The word ama also occurs in a very imperfect inscription quoted by Dennis (I. p. 342). Lastly, as we have both lautn and lautnescle in the shorter inscription by the side of lautn in the larger, we may infer that lautnescle is a diminutive form like munusculum, and therefore we may compare kemul-mleskul in the Perugian inscription with kuml, the regular Runic name for a monumental stone (Dieterich, Runen-Sprach-Schatx, p. 124). With regard to the general interpretation of the Perugian inscription, it seems idle to follow in the steps of the Italian scholars, Vermiglioli, Orioli, and Campanari, the last of whom has given us a Latin translation of the whole inscription. Nor can I sympathise in the regret of Dr. C. Von Schmitz, when he complains that he cannot find a publisher for the grammar and dictionary of the Etruscan, which are to explain his forced and unnatural version of this document (Zeitschr. f. d. Alter- thumsw. 1846, Septemb. Beilage). It would, indeed, be easy to found a number of conjectures on the assonances which may be detected in almost every line ; but until a complete collection of all the genuine Etruscan inscriptions shall have furnished us with a sufficiently wide field for our researches, until every extant Tuscan word has been brought within the reach of a 188 THE ETRUSCAN LANGUAGE. [On. V. philological comparison, we must be content to say of this great Perugian inscription, that it appears to be a cippus con- veying some land for funereal purposes 1 . The donor is Larthius, a member of the family of the Reza (Rwsii), who were dis- tinguished people in the neighbourhood of Perusia (see Yermi- glioli, Iscriz. Perug. p. 273), and Rasne, Which occurs thrice in the inscription, seems to be a patronymic of the same family. The relative position of the word, no less than the locality of the inscription, shows that Velthina is the person in whose honour this cippus was erected, and that the word does not re- fer to Felsina, the old name of Bononia (Plin. H. N. Ill, 20. XXXIII, 37. XXXVII. 57. Serv. ad ^En. X. 198). The other personal name, which occurs most frequently in the inscription is Afuna, probably Aponia (Vermiglioli, p. 233) ; and it is worthy of remark, that we have the nom., gen., and accus. of these two proper names in accordance with the regular forms of the first Latin declension, namely, Afuna, A/unas, Afunam, and Vel- thina, Velthinas, Velthinam. The name Velthina may be compared with the well-known name Ccecina, From the pra3- nomen Aulesi in v. 9. it is probably a man's name 2 . If I do not undertake to interpret all that Lartius, the son of Rsesia, has thought fit to inscribe on this cippus, it must not be supposed that this in any way affects the results at which I have arrived respecting the ethnography of the Etruscans. That an inability to interpret Runic monuments may be perfectly consistent with a knowledge of the class of languages to which they belong, is shown, not merely by the known relationship between the lan- guage of the Egyptian hieroglyphics, and the Coptic dialects more recently spoken in that country, but still more strikingly by the fact, that, although we have no doubt as to any of the idioms spoken in ancient Britain, no one has been able as yet to give a certain interpretation of the Runic inscriptions on the i See the commentators on Hor. I. Serm. VIII. 13; and the bon mot of Augustus on Vettius quum monumentum patris exarasset (Macrob. II. Sat. c. 4. p. 232). 2 We have seen above that the termination -I indicates a matronymic ; and I conclude that the Etruscan patronymic ended in -na ; compare in this inscription, Rezul with Rasna, and Cceci-lia, which was the Roman equivalent to the mythical Tanaquil, with the undoubtedly Tuscan form Cceci-na. 1 do not agree with Miiller (Etr. I. p. 453) that the forms in -si) as Aulesi, Clensi, are datives. $10.] THE ETRUSCAN LANGUAGE. J89 pillar at Bewcastle and on the font at Bridekirk, which are both in Cumberland, and which both belong to the same dialect of the Low-German languages, (see Palgrave, History of the Anglo- Saxons, Lond. 1850, pp. 146. sq.). The really important point is to determine the origin of the ancient Etruscans ; and the Peru- gian inscription, so far from throwing any difficulties in the way of the conclusion at which I have arrived, has furnished some of the strongest and most satisfactory confirmations of the Old Norse affinity of the Rasena. J 11. Harmony between linguistic research and ethnographic tradition in regard to the ancient Etruscans. This survey of the Etruscan language, brief and circumscribed as it necessarily is, has enabled us to perceive that there is a perfect harmony and agreement between the results of our lin- guistic researches, so far as the scanty materials have allowed us to carry them, and the ethnographic and historic traditions respecting the ancient Etruscans. We have seen that in the character of their writing, in most of their mythology, in by far the greatest number of those words which have been transmitted to us with an interpretation, and in the oldest inscriptions, espe- cially in those from Caere, there are decisive evidences of an affinity between the inhabitants of Etruria and those Pelasgians who peopled Greece in the earliest times, and who constituted an important element in the inhabitants of Latium. For the residue of the language, and especially in the case of those inscriptions which are found near Clusium and Perugia, we are enabled to recognise an ingredient unmistakably identical with that Scan- dinavian dialect, which Norwegian emigrants conveyed in an ancient form to the inaccessible regions of ultima Thule, where it remained for centuries safe from all risk of corruption or im- provement by an infusion of foreign words or constructions. Now these phenomena, as we have seen, are necessary to reconcile, and do in fact reconcile, all the traditions about the inhabitants of Etruria. The Pelasgian affinities of the old Tyrrhenians are attested by the concurring voice of all antiquity ; and as in Argo- lis, so in Italy, we shall best understand the statement that a more complete civilization was imported directly from Lydia, if we bear in mind that the Lydians referred to in the tradition were Pelasgians, who had appropriated the arts and social culture 190 THE ETRUSCAN LANGUAGE. [On. V. of their Asiatic neighbours. And we shall be able to adopt this universal belief of a connexion between the western coasts of Asia Minor and Italy, without disturbing the well-grounded statement that the Rasena and Raeti were one and the same race, if we infer that these Rasena were a much later ingredient, and one which only established an aristocracy of conquest in the cities of Etruria, without permanently or extensively affecting the great mass of the population. It will be observed that the main obstacle to a general reception of the statement that the Rasena were Rsetians has consisted in the apparent inconsistency between this and the Lydian tradition. The ethnographical inversion, by which Livy makes the Rsetians the fugitive offshoot of a nation which really descended from their own mountains, has not occa- sioned any difficulty. It would be admitted at once that, if the Rastians and Rasena were one and the same people, some foreign interference must have disturbed the continuity of their area in the valley of the Po, and if there was once an unbroken stream of population from the Lech to the Tiber, no ethnographer will doubt that its source must have been in the mainland rather than in the peninsula. But it has not been sufficiently considered, that the bulk of the Pelasgian nation, already settled in Umbria and Etruria, would not lose their original type, merely because they were invaded and conquered by a band of warriors from the north, any more than Anglo-Saxon England was entirely de- prived of its former characteristics by the Norman inroad. The civilization of the Tyrrhenians, their connexion with the commer- cial activity of the Mediterranean 1 , and the advantages which they derived from the arts and social culture of their brethren in Asia Minor, were circumstances long anterior to the invasion from the north ; and as the Rasena would adopt the refinements which they found among the Tyrrhenians, we may make inge- nious comparisons between the tombs of Porsena 2 and Alyattes, without refusing our assent to the well-attested fact that the 1 It is to this that I would attribute the continuance of Hellenic influences, on which Miiller insists (Etrusk. II. 292). 2 It is worthy of remark, that a distinguishing feature in the monu- ment of Porsena, as described by Varro (apudPlin. XXXIV. 13), namely, the bells on the cupolas, is expressly compared with a similar contrivance at the Pelasgian Dodona : " tintinnabula, quse vento agitata longe sonitus referant, ut Dodonce olimfactum" ii.] THE ETRUSCAN LANGUAGE. 191 warriors and city-nobles of historical Etruria derived their origin from the Raetian Alps. With regard to the argument from the remains of the Etruscan language, the philologer will at once admit that, as far as it goes, the evidences of affinity, which have been adduced, are neither precarious nor doubtful. Instead of conjectures founded on a casual agreement of syllables, we have seen that the meaning, which we were led to expect, was at once supplied by the language, which collateral circumstances had in- dicated as the proper source of information ; and not only were ethnical names and common words simply and consistently explained in this way, but we found that some peculiarities of etymology and syntax were at once illustrated by a reference to the same standard of comparison. So that, on the whole, every available resource of grammar and philology tends to confirm and reconcile the otherwise divergent and contradictory statements of ancient history ; and the Etruscans may now without any incon- sistency claim both the Tyrrheno-Lydian and Rsetian affinities, which the classical writers have attributed to them. J 12. General remarks on the absorption or evanescence of the old Etruscan Language. It only remains that I should make a few remarks on the absorption or evanescence of the old Etruscan language. When we see so much that is easily explained ; when, in fact, there is no great difficulty in dealing with any Etruscan word which has come down to us with an interpretation or clue to its meaning ; and when we are puzzled only by inscriptions, which are in themselves mere fragments, made up in a great measure of proper names, and mutilated by, we know not how many, con- ventional abbreviations, it is sufficiently evident that the strik- ing diiferences between the Etruscan and the other ancient dialects of the peninsula were not such as to take the language out of the Indo- Germanic family, and that while these differences affected only an inconsiderable ingredient in the old Etruscan, the main portion of the language must have approximated very closely to the contiguous and surrounding idioms. Otherwise, we should be obliged to ask, where is the bulk of that language which was spoken by the ancestors of Maecenas ? We talk of dead languages ; but this variety of human speech should seem to be not only dead, but buried, and not only buried, but sunk 192 THE ETRUSCAN LANGUAGE. [Cn. V. beneath the earth in some necropolis, into which no Galassi or Campanari can dig his way. The standard Italian of the present day is the offspring of that Latinity which was spoken by the Etrusco-Romans ; but we find no trace of ancient bar- barism in any Tuscan writer. Surely it is a fair inference, that while the Rsetian element, introduced into the northern cities by an aristocracy of conquest, was not permanently influential, but was absorbed, like the Norman French in this country, by the Pelasgo-Umbrian language of the bulk of the population, the latter, which may be termed " the common Etruscan," like the Sabello-Oscan and other dialects, merged in the old Latin, not because the languages were unlike, but because they were sister idioms, and embraced one another as soon as they had discovered their relationship 1 . The only way to escape from all the diffi- culties of this subject is to suppose that the city on the Tiber served as a centre and rallying point for the languages of Italy as well as for the different tribes who spoke them, and that Rome admitted within her walls, with an inferior franchise, which in time completed itself, both the citizens and the vocabularies of the conquered Italian states. If this absorbing centralization could so thoroughly Latinize the Celtic inhabitants of Lombardy, and even the transalpine branch of the Gallic race, much more would it be likely to affect the Etruscans, who extended to the Tiber, and whose language, in its predominant or Pelasgian character, approximated so closely to the cognate idiom of the old Latin tribes. 1 Among many instances of the possibility at least of such a transition, not the least interesting is the derivation of Populonia from Phupluns, the Etruscan Bacchus ; so that this city, the Etruscan name of which was Popluna, is the Dionysopolis of Etruria (see Gerhard in the Mhein. Mus. for 1833, p. 135). Now it is clear that as Nethuns = Nethu-nus, is the god of nethu, so Phupluns = Poplu-nus is the god of poplu. It seems that the ancients planted the poplar chiefly on account of their vines, and the poplar was sacred to Hercules, who has so many points of contact with Bacchus. Have we not, then, in the word phupluns the root of populus, a word quite inexplicable from the Latin language alone? A sort of young, effeminate Hercules, who appears on the coins of Populonia (see Muller, Etrusk. I. p. 331), is probably this Poplunus. The difference in the quantity of the first syllables of Populus and Populonia is not surprising, as the latter is an exotic proper name, and the former a na- turalized common term. CHAPTER VI. THE OLD ROMAN OR LATIN LANGUAGE. 1. Fragments of old Latin not very numerous. 2. Arvalian Litany. 3. Chants preserved by Cato. 4. Fragments of Salian hymns. 5. Old regal laws. 6. Remains of the XII. Tables. 7. Table I. 8. Table II. 9. Table III. 10. Table IV. 11. Table V. 12. Table VI. 13. Table VII. 14. Table VIII. 15. Table IX. 16. Table X. 17- Table XI. 18. Table XII. 19. The Tiburtine Inscription. 20. The epitaphs of the Scipios. 21. The Columna Rostrata. 22. The Silian and Papirian Laws and the edict of the Curule ^Ediles. 23. The Senatus- Consul 'turn de Bacchanalibus. 24. The old Roman law on the Bantine Table. 1. Fragments of Old Latin not very numerous. HAVING in the preceding chapters given specimens of the languages spoken by those nations which contributed in different proportions to the formation of the Roman people, the next step will be to collect the most interesting remains of the old Roman language, considered as the offspring of the Um- brian, Oscan, and Tuscan, such as it was before the predomi- nance of Greek cultivation had begun to work on this rude composite structure. The total loss of the genuine Roman literature l will, of course, leave us but a scanty collection of such documents. Indeed, for the earlier centuries we have only a few brief fragments of religious and legal import. As we approach the Punic wars, the inscriptions become more numerous and com- plete ; but then we are drawing near to a period when the Roman language began to lose its leading characteristics under the pressure of foreign influences, and when it differed little or nothing from that idiom which has become familiar to us from the so-called classical writings of the Augustan age. Polybius, speaking of the ancient treaty between Rome and Carthage (III. 22), remarks that the old Latin language differed so much from that which was spoken in his own time, that the best-informed Romans could not make out some expressions without difficulty, even when they paid the greatest attention : yap n $ia(f)opd yeyove riys ^mXe/crof, Kai Trapd vvv Trpos TYJV ap-^aiav 9 ware TOVS 1 See Macaulay, Lays of Ancimt Rome, pp. 15, sqq. 13 194 THE OLD ROMAN [Cm VI. evict fjioXis e% etricrTacrecos SievKpivelv. The great mass of words must, however, have been susceptible of interpretation ; for he does not shrink from translating into Greek the substance at least of that very ancient treaty. $ 2. Arvalian Litany. Accordingly, we find that the most primitive specimens of Latinity may now-a-days be understood by the scholar, who, after all, possesses greater advantages than Polybius and his con- temporary Romans. This will appear if we examine the song of the Fratres Arvales, which is one of the most important and ancient specimens of the genuine Roman language. The inscrip- tion, in which it is preserved, and which was discovered in the year 1777, is probably not older than A. D. 218; but there is every reason to believe that the cantilena itself was the same which was sung in the earliest ages of Rome, for these litanies very often survive their own significance. The monks read the Latin of their missals without understanding it, and the Parsees of Gujerat cannot interpret their sacred Zend. It appears from the introductory remarks, that this song was confined to the priests, the Publici being excluded : " Deinde subselliis mar- moreis consederunt ; et panes laureates per Publicos partiti sunt; ibi omnes lumemulia cum rapinis acceperunt, et Deas unguenta- verunt, et ^Edes clusa est, omnes foris exierunt : ibi Sacerdotes clusi succincti, libellis acceptis, carmen descindentes tripodaverunt in verba haec : 1. Enos Lases juvate (ter), 2. Neve luaerve Marmar sins incurrere in pleoris (ter) 3. Satur furere (vel fufere) Mars limen salista Berber (ter) 4. Semunis alter nei (vel alternisf) advocapit conc- tos (ter) 5. Enos Marmor (vel Mamor) juvato (ter) 6. Triumpe, triumpe, triumpe, triumpe, triumpe. Post tripodationem, deinde signo dato Publici introiere, et libel- los receperunt." (See Orelli, Inscript. Lat. I. p. 391, no. 2271.) $2.] OR LATIN LANGUAGE. 195 There can be little doubt as to the meaning of any single word in this old hymn, which seems to be written in very rude Saturnian verse, the first half of the verse being alone preserved in some cases ; as in JEnos Lases juvate Enos Mamor juvato. The last line is a series of trochees cum anacrusi, or a still shorter- form of the first half of the Saturnian verse. 1. JEnos is a form of the first person plural, analogous to the German uns. Lases is the old form of Lares (Quinctil. Institut. Orat. I. 4. $ 13 ; see Muller ad Fest. p. 15). 2. Lucerve for luerve-m, according to a custom of dropping the final M, which lasted till Gate's time (see next $). This form bears the same relation to luem that Minerva does to mens. Caterva from catus = acutus (above, p. 106), and its synonym acervus from acus, are derivatives of the same kind 1 . We may also compare bovem, suem, &c. with their older forms, boverem, suerem, &c. Marmar, Marmor, or Mamor, is the Oscan and Tuscan Mamers, i. e. Mars (above p. 146). That Mars, or Mars pater, was addressed as the averter of diseases, bad weather, &c. is clear from Cato, R. R. 141. Sins is sinas: so Tab. Bantin. 1. 19 : Bantins for Bantinus, &c. Pie-ores is the genuine comparative of ple-nus, which bears the same re- lation to 7r\e?os that unus does to olo?. The fullest form would be ple-iores = TrXe-foi/e?. 3. " Mars, having raged to your satisfaction (comp. Hor. I. Carm. II. 37: "longo satiate ludo"), grant that the Sun's light may be warm." Llmen for lumen may be com- 1 Mr. F. W. Newman (Regal Rome, p. 6l) derives caterva from the Welch cad-torva, " battle-troop." I do not know whether this etymology was suggested by the well-known statements in Vegetius, II. 2 : " Galli Celtiberique pluresque barbaricse nationes catervis utebantur in preeliis." Isidor. Grig. IX. 33 : " proprie Macedonum phalanx, Gallorum caterva, nostra legio dicitur." Doderlein, who proposes (Lat. Syn. u. Et. V. 361) to connect caterva with quattuor, properly remarks that these passages do not show that caterva was considered a Gallic word, but only that, as distinguished from the phalanx and legio, it denoted a less com- pletely disciplined body of men. The natural idea of a " heap " of sepa- rable objects is that of a mass piled up to a point, and this is indicated by the roots of ac-er-vus and cat-er-va. The latter therefore, as denoting a body of men, suggests the same arrangement as the cuneus, which is mentioned along with it by Tacitus, Hist. II. 42 : " comminus eminus catervis et cuneis concurrebant." On the form of cat-er-va, see below, Ch. XIII. 5. 132 196 THE OLD ROMAN [Cn. VI. pared with plisima for plurima (Fest. p. 205), scripulum for scrupulum, &c. (see below, $ 5). Sails is the original form of soils : comp. t / ' ' c)e oevrepov cm/cocr/a, TOV ce Tpirov eKarov. Plin. H. N. XXXII. 2, 10, 20 : " Pisceis quei squamosei nee sunt, nei polucetod ; squamosos omneis prceter scarom polu- cetod." Cf. Fest. s. v. Pollucere, p. 253 : " Pollucere merces [quas cuivis deo liceat], sunt far, polenta, vinum, panis fermen- talis, ficus passa, suilla, bubula, agnina, casei, ovilla, alica, sesama, et oleum, pisces quibus est squama, prater scarum: Herculi autem omnia esculenta, poculenta." Id. s. v. Termino, p. 368 : " Denique Numa Pompilius statuit, Eum qui terminum exarasset et ipsum et boves sacros esse." i. e. Qui terminom ecsaraset 9 ipsus et boveis sacrei sunto (See Dirksen, Versuche, p. 334). Id. s. v. Aliuta, p. 6 : " Aliuta antiqui dicebant pro aliter, .... hinc est illud in legibus Numa3 Pompili : Siquisquam aliuta facsit ipsos Jovei sacer estod." 6.] OR LATIN LANGUAGE. 203 6. Remains of the XII. Tables. But of all the legal fragments which exhibit the prisca vetustas verborum (Cic. de Oratore, I. c. 43), the most copious, as well as the most important, are the remains of the Twelve Tables, of which Cicero speaks in such enthusiastic, if not hyperbolical language. These fragments have been more than once collected and explained. In the following extracts I have followed the text of Dirksen ( Uebersicht der bisherigen Versuche zur Kritik und Herstellung des Textes der Zwolf-Tafel-Fragmente). The object, however, of Dirksen's elaborate work is juristic 1 rather than philological ; whereas I have only wished to present these fragments as interesting specimens of old Latinity. It was probably the intention of the decemvirs to comprise their system in six double Tables ; for each successive pair of Tables seems to refer to matters which are naturally classed together. Thus Tab. I. and II. relate to the legis actiones ; Tab. III. and IV. to the mancipium, potestas, and manus, or the rights which might be acquired over insolvent debtors, the right of a father over his son, and of a husband over his wife ; Tab. Y. and VI. to the laws of guardianship, inheritance and property ; Tab. VII. and VIII. to obligationes, delicta, and crimina; Tab. IX. and X. to the jus publicam and jus sacrum ; Tab. XI. and XII. were supplementary to the ten former Tables, both in subject and in date. 7. Tab. I. Fr. 1. (I. 1, 2, Gothofredi) : si . IN . jus . VOCAT . NI . IT . AN- TESTATOR . IGITUR . EM . GAPING . (Porphyrio ad HOT. I. Serm. 9, 65 : " Adversarius molesti illius Horatium consulit, an per- mittat se antestari, injecta manu extracturus ad Praetor em, quod vadimonio non paruerit. De hac autem Lege XII. Tabular um his verbis cautum est : si vis vocationi testamini, igitur en capito antestari. Est ergo antestari, scilicet antequam manum injiciat." Cf. Cic. Legg. II. c. 4 ; Aul. Gell. N. A. XX. 1 ; Auctor ad Herenn. II. c. 13 ; Non. Marcell. de Propr. Serm. c. 1, $ 20, s. v. calvitur. Lucilius, Lib. XVII. : " Si non it, capito, inquit, eum et, si calvitur ergo, Ferto manum"). It seems probable 1 The student will find a general sketch of the old Roman law in Arnold's Rome, I. pp. 256, sqq. 204 THE OLD ROMAN [On. VI. that the original form of the law was, si quis in jus vocatus nee it f antestamino, igitur (i. e. inde, postea, turn, Fest. p. 105) em ( = eum) capita. Cf. Gronov. Lect. Plautin. p. 95. Fr. 2 (I. 3) : si . CALVITUR . PEDEMVE . STRUIT, . MANUM . ENDO . JACITO . (Festus, p. 313). The word calvitur is ex- plained by Gaius, L. 233, pr. D. de Verb. Sign.: u Si calvitur et moretur et frustretur. Inde et calumniatores appellati sunt, quia per fraudem et frustrationem alios vexarent litibus." Pe- dem struere is explained by Festus, 1.1.: "Alii putant signi- ficare retrorsum ire : alii, in aliam partem : alii fugere : alii gradum augere : alii minuere, cum quis vix pedem pedi pra3fert, otiose it, remoratur :" and p. 210 : "pedem struit in xn. signi- ficat fugit, ut ait Ser. Sulpicius." This fragment seems to have followed close upon the previous one : see the passage of Lucilius, quoted above. Fr. 3 (I. 4) : si . MORBUS . AEVITASVE . VITIUM . ESCIT, . QUI . IN . JUS . VOCABIT . JUMENTUM . DATO ; . SI . NOLET . ARCERAM . NE . STERNITO . (Aul. Gell. N. A. XX. 1). Vitium escit means impedimenta erit. Arcera is explained by Nonius Marcellus, de Propr. Serm. I. 270 : "Arcera plaustrum est rusticum, tectum undique quasi area. Hoc vocabulum et apud Varronem et apud M. Tullium invenitur. Hoc autem vehiculi genere senes et cegroti vectari solent. Varro yGpovriSiSaaKaXa): vehebatur cum uxore vehiculo semel aut bis anno cum arcera : si non vellet non sterneret." Fr. 4 (I. 6) : ASSIDUO . VINDEX . ASSIDUUS . ESTO, . PROLE- TARIO . QUOI . QUIS . VOLET . VINDEX . ESTO . (Aul. Gell. 2V. A. XVI. c. 10 ; cf. Cicero, Top. c. 2, who explains assiduus as a synonym of locuples, and derives it, with Julius, ab asse dando ; Nonius, Propr. Serm. c. 1, antepen., who explains proletarius as equivalent to plebeius " qui tantum prolem sufficiat." See Niebuhr, Hist. Rom. I. p. 445, note 1041). Fr. 5 (IX. 2). Festus, p. 348 : " Sanates dicti sunt, qui supra infraque Romam habitaverunt. Quod nomen his fuit, quia cum defecissent a Romanis, brevi post redierunt in amicitiam, quasi sanata mente. Itaque in xn. cautum est, ut ' idem juris esset Sanatibus quod Forctibus? id est bonis (cf. pp. 84, 102), et qui nunquam defecerant a P. R." Whence we may supply, p. 321 : " [Hinc] in xn.: 'NEX[I solutique, ac] FORCTI sANATi[sque idem jus estod'], id est, bonor[um et qui defecerant sociorum]." $7.] OR LATIN LANGUAGE. 205 Where also sanas is explained from Cincius, " [quod Priscus] pra?ter opinio[nem eos debellavis]set, sanavisse[tque ac cum iis pa]cisci potuisset." Dirksen (p. 164) is wrong in referring these extracts to the epitome of Paulus. Fr. 6 (I. 17) : REM . UBI . PAGUNT, . ORATO . (Auctor ad Herenn. II. c. 13). Fr. 7 (I. 8) : NI . PAGUNT . IN . COMITIO . AUT . IN . FORO . ANTE . MERIDIEM . CAUSAM . CONJICITO, . QUOM . PERORANT , AMBO . PRAESENTES . (id. ibid, and Aul. Gell. XVII. 2). The word pagunt is explained by Priscian (X. 5, 32) as a synonym of paciscor ; the common Latin form is pa-n-go, but the medial and tenuis of the gutturals were constantly interchanged after the distinction between them was introduced by Sp. Carvilius (Terent. Scaur, p. 2253, Putsch). Fr. 8 (I. 9) : POST . MERIDIEM . PRAESENTI . STLITEM . ADDI- CITO . (Aul. Gell. XVII. 2). Fr. 9 (I. 10) : SOL . OCCASUS . SUPREMA . TEMPESTAS . ESTO . (id. ibid). The word tempestas is here used for tempus ; the whole afternoon was called tempus occiduum, and the sunset was suprema tempestas (Macrob. Saturn. I. c. 3). Gellius, to whom we owe these fragments, considers the correct reading to be sol, not solis occasus. " Sole occaso" he says, " non insuavi venus- tate (vetustate ;) est, si quis aurem habeat non sordidam nee proculcatam." But Festus (p. 305), Varro (L. L. V. c. 2), and others, consider the phrase to have been solis occasus. There is more probability in the reading of Gellius. Fr. 10 (II. 1). Aul, Gell. N. A. XVI. c. 10 : " Sed enirn quum proletarii, et assidui, et sanates, et vades, et subvades, evanuerint, omnisque ilia xn. Tabularum antiquitas consopita sit," &c, 8. Tab. II. Fr. 1. Gaius, Inst. IV. $ 14 : " Poena autem sacramenti aut quingenaria erat, aut quinquagenaria ; nam de rebus mille a3ris plurisve quingentis assibus, de minoris vero quinquaginta assibus sacramento contendebatur ; nam ita lege xn. Tabularum cautum erat. Sed si de libertate hominis controversia erat, etsi pretiosis- simus homo esset, tamen ut L. assibus sacramento contenderetur eadem lege cautum est favoris causa ne satisdatione onerarentur adsertores." 206 THE OLD ROMAN [Cn. VI. Fr. 2 (II. 2) : (a) MORBUS . SONTICUS (6) STATUS . DIES . CUM . HOSTE (c) SI . QUID . HORUM . FUAT . UNUM, . JUDICI, . ARBITROVE . REOVE, . DIES . DIFFENSUS . ESTO . (a) Aul. Gell. XX. c. 1 : " Morbum vehementiorem, vim graviter nocendi haben- tem, Leg. istar. i. e. xn. Tab. scriptores alio in loco non per se morbum, sed morbum sonticum appellant."" Fest. p. 290 : " Son- ticum morbum in xn. significare ait Julius Stilo certum cum justa causa, quern non nulli putant esse, qui noceat, quod sontes significat nocentes. Nsevius ait : sonticam esse oportet causam, quam ob rem perdas mulierem" (b) Cic. de Off. I. c. 12 : " Hostis enim majores nostros is dicebatur, quern nunc peregri- num dicimus. Indicant xn. Tabulae ut : status dies cum hoste ; iteraque : adversus hostem ceterna auctoritas" Fest. p. 314 : " Status dies [cum hoste] vocatur qui judici causa est constitutus cum peregrine. Ejus enim generis ab antiquis hostes appella- bantur, quod erant pari jure cum populo R., atque hostire pone- batur pro cequare. Plautus in Curculione [I. 1, 5] : si status condictus cum hoste intercedit dies, tamen est eundum, quo im- perant ingrati-s" This passage is neglected by Dirksen, but not by Gronovius, Lectiones Plautince, p. 81. With regard to the original signification of hostis, it is very worthy of remark that the Latin hostis and the Greek fe'yos, starting from opposite points, have interchanged their significations. Hos-tis originally signified " a person entertained by another," " one who has food given to him" (comp. hos-pi-[t-]s, " the master of the feast," hostia, gasts, &c. N. Crat. 474); but at last it came to mean " a stranger, 1 ' " a foreigner/' and even " an enemy" (see Yarro, L. L. p. 2, Miiller). Whereas f evos, originally denoting " a stranger" (extraneus), i. e. " one without"" ([e]f eW), came in the end to signify " an entertainer" and " a friend." I cannot accept Muller'8 derivation of feVos (ad Fest. p. 102). (c) Festus, p. 273 : " Eeus nunc dicitur, qui causam dicit ; et item qui quid promisit spoponditve, ac debet. At Gallus JElius libro 117 Sign. Verb. qu. ad Jus pertinent, ait : Reus est, qui cum altero litem contestatam habet, sive is egit, sive cum eo actum est. Reus stipulando est idem qui stipulator dicitur, quive suo nomine ab altero quid stipulatus est 9 non is qui alteri adstipu- latus est. Reus promittendo est qui suo nomine alteri quid promisit, non qui pro altero quid promisit. At Capito Ateius in eadem quidem opinione est : sed exemplo adjuvat interpreta* 8.] OR LATIN LANGUAGE. 207 tionem. Nam in secunda Tabula secunda lege in qua scriptum est : si quid horum faat unum judici arbitrove reove, eo die diffensus esto, hie uterque, actor reusque, in judicio rei vocantur, iteraque accusator de via citur more vetere et consuetudine anti- qua." Ulpian. L. LXXIV. ad Edict. : " Si quis judicio se sisti promiserit, et valetudine vel tempestate vel vi fluminis prohibitus se sistere non possit, exceptione adjuvatur ; nee immerito : cum enim in tali permissione praesentia opus sit, quemadmodum potuit se sistere qui adversa valetudine impeditus est ? Et ideo etiam Lex xn. Tab. : si judex vel alteruter ex litigatoribus morbo sontico impediatur, jubet diem judicii esse diffensum." I have restored diffensus both in Festus and Ulpian on the authority of Mialler, who has shown (Suppl. Annot. ad Fest. p. 401) that fendo must have been anciently a synonym of ferio and trudo, and consequently that diffensus esto = differatur. Fr. 3 (II. 3) : cui . TESTIMONIUM . DEFUERIT, . is . TERTIIS . DIEBUS . OB . PORTUM . OBVAGULATUM . ITO . (Fest. p. 233 : " Portum in xn. pro domo positum omnes fere consentiunt : si," &c. Id. p. 375 : " Vagulatio in lege xn. [Tab.] significat quces- tionem cum convicio : 5?'," &c.). Fr. 4 (II. 12). " Nam et de furto pacisci lex permittit" (L. 7. 14. D. de Pactis, Ulp. IV. ad Edictum). 9. Tab. TIL Fr. 1 (III. 4) : AERIS . CONFESSI . REBUSQUE . JURE . JUDI- CATIS . TRIGINTA . DIES . JUSTI . SUNTO . (Aul. Gell. XX. C. 1 : " Eosque dies Decemviri justos appellaverunt, velut quoddam justitium, id est juris inter eos quasi interstitionem quandam et cessationem, quibus diebus nihil cum his agi jure posset." XV. c. 13 ; cf. Gaius, Inst. III. 78, &c.). Fr. 2 (III. 5) : POST . DEINDE . MANUS . INJECTIO . ESTO ; . IN . jus . DUCITO . (Aul. Gell. XX. c. 1 ; cf. Gaius, Inst. IV. 21). Fr. 3 (III. 6) : NI . JUDICATUM . FACIT (1. faxsit), . AUT . QUIPS . ENDO . EM . JURE . VINDICIT, . SECUM . DUCITO ; . VIN- CITO, . AUT . NERVO . AUT . COMPEDIBUS, . QUINDECIM . PONDO . NE . MAJORE, . AUT . SI . VOLET . MINORE . VINCITO . (Aul. Gell. XX. c. 1). We should perhaps read faxsit for facit on account of vindicit, for which see Miiller, Suppl. Ann. ad Fest. p. 393. For the form quips see Gronovius ad Gell. I. ; the proper read- 208 THE OLD ROMAN [Cn. VI. ing is ques ; see below, 23. For the meaning of nervus here, comp. Fest. s. v. p. 765. Fr. 4 (III. 7) : si . VOLET, . suo . VIVITO ; . NI . suo . VIVIT, . QUI . EM . VINCTUM . HABEBIT, . LIBRAS . FARRIS . ENDO . DIES . DATO ; . si . VOLET . PLUS . DATO . (Aul. Gell. XX. c. 1; and for the meaning of vivere compare L. 234, 2. D. de Verb. Sign. ; Gams, L. II. ad Leg. xn. Tab. ; Donat. ad Terent. Phorm. II. 1, 20). The student will observe that endo dies = indies. Fr. 5 (III. 8). Aul. Gell. N. A. XX. 1 : Erat autem jus interea paciscendi ; ac nisi pacti forent, habebantur in vinculis, dies LX. ; inter eos dies trims nundinis continuis ad Praetorem in comitium producebantur, quanteeque pecuniaa judicati essent prae- dicabatur." From which Ursinus conjectures : Endoderatim [rather inter utim. Festus, p. Ill] pacio estod. Nei cum eo pacit, LX. dies vinctom habetod. In ieis diebus tertieis nondi- neis continueis indu comitium endo joure im procitato, quan- teique stlis cestumata siet prcedicato. Fr. 6 (III. 9). Aul. Gell. XX. 1 : " Tertiis autem nundinis capite pcenas dabant, aut trans Tiberim peregre venum ibant si plures forent, quibus reus esset judicatus, secare si vellent atque partiri corpus addicti sibi hominis permiserunt verba ipsa Legis dicam : TERTIIS, inquit, NUNDINIS PARTIS SECANTO, si PLUS MINUSVE SECUERUNT, SE FRAUDE ESTO." Cf. Quinctil. Inst. Or. III. c. 6 ; Tertullian. Apol. c. 4. The student will remark that we have here se for sine, as in the compounds se-dulo (- sine dolo), se-paro, se-cludo, se-motus, se-gregatus, &c. (See Festus, p. 336). Se = sed is an ablative form which in later Latin appears only in composition ; sine accords in form with the Sanscrit instrumental, and was used as a preposition to the latest period of the language. Accordingly these two forms may be compared with the Greek m and Kara ; the former being used only as the particle of apodosis or in composition (as Kairerov Find. O. VIII. 38), while the latter retains to the end its regular preposi- tional functions. Fr. 7 (III. 3) : ADVERSUS . HOSTEM . AETERNA . AUCTORITAS . (Cic. de Off. I. c. 12). 10. Tab. IV. Fr. 1 (IV. 1). Cic. de Legg. III. c. 8 : " Deinde quum [Trib. pot. ortus] esset cito legatus \leto datus, Orelli], tarn- $10.] OR LATIN LANGUAGE. 209 quam ex xn. Tabulis insignis ad deformitatem puer." From whence we infer that the xn. Tables authorised the exposure of deformed children. Fr. 2 (IV. 2). From the statement of Dionysius (II. 26, 27), that the decemvirs in their fourth Table continued the jus vendendorum liberorum established in the time of the kings, Ursinus imagines some such passage as this : PATREI . ENDO . FIDIO . VITAE . NECISQUE . POTESTAS . ESTOD, . TERQUE . IN . VENOM . DARIER . jous . ESTOD ; to which he appends the next fragment. Fr. 3 (IV. 3) : si . PATER . FILIUM . TER . VENUM . DUIT, . FILIUS . A . PATRE . LIBER . ESTo. (Ulpian, Fr. Tit. X. 1 ; Gaius, Inst. I. 132 ; IV. 79). Fr. 4 (IV. 4). Aul. Gell. III. 16 : ... " Quoniam Decemviri in decem mensibus gigni hominem, non in undecimo scripsissent ;" whence Gothofredus would restore : si qui ei in x. mensibus proximis postumus natus escit, Justus esto. 11. Tab. V. Fr. 1 . Gaius, Inst. I. $ 145 : " Loquiinur autem exceptis Virginibus Vestalibus, quas etiam veteres in honorem sacerdotii liberas esse voluerunt ; itaque etiam lege xn. Tabularum cautum est." Cf. Plutarch, Vit. Num. c. 10. Fr. 2. Id. II. 47 : " (Item olim) mulieris quae in agnato- rum tutela erat, res mancipi usucapi non poterant, preeterquam si ab ipso tutore (auctore) traditae essent : id ita lege xn. Tabu- larum cautum erat." Fr. 3 (V. 1) : [PATERFAMILIAS] . UTI . LEGASSIT . SUPER . PECUNIA . TUTELAVE . SUAE . REI, . ITA . JUS . ESTO . (Ulpian, Fr. Tit. XL $ 14; Gaius, Inst. II. 224; Cic. de Invent. Rhet. II. c. 50 ; Novell. Justin. XXII. c. 2, &c.). Fr. 4 (V. 2) : si . INTESTATO . MORITUR . cui . suus . HERBS . NEC . SIT, . ADGNATUS . PROXIMUS . FAMILIAM . HABETO. (Ulpian, Fr. Tit. XXVI. 1 ; cf. Gaius, Inst. III. 9, &c.). Fr. 5 (V. 3) ; si . ADGNATUS . NEC . ESCIT, . GENTILIS . FAMI- LIAM . NANXITOR. (Collatio Legg. Mosaic, et Rom. Tit. XVI. $ 4 ; cf. Gaius, Inst, III. 17). I have written nanooitor for nancitor on the authority of Miiller, ad Fest. p. 166 : " nanxitor in xn., nactus erit, prsehenderit ;" where he remarks : " nancitor quomodo futurum exactum esse possit, non intelligo, nisi cor recta 14 210 THE OLD ROMAN [On. VI. una littera. Ab antique verbo nancio fiit. ex. fit nanxo, sicut a capio capso ; idque translatum in pass. form, efficit nanxitur vel nanxitor, ut a turbasso fit turbassitur." Fr. 6 (V. 7). Gaius, Inst. I. 155 : " Quibus testamento quidem tutor datus non sit, iis ex lege xn. agnati sunt tutores ; qui vocantur legitimi." Cf. J 157, where he says that this applied to women also. Fr. 7 (V. 8) : si . FURTOSUS . AUT . PRODIGUS . ESCIT, . AST . EI . CUSTOS . NEC . ESCIT, . ADGNATORUM . GENTILIDMQUE . IN . EO . PEQVUNIAQUE . Ejus . POTESTAS . ESTO. (Cicer. de Invent. Rhet. II. c. 50, gives the bulk of this passage ; aut prodigus is inserted on the authority of Ulpian, 3, i. de Curationibus ; and ast ei custos nee escit is derived from Festus, p. 162 : "Nee conjunctionem grammatici fere dicunt esse disjunctivam, ut nee legit nee scribit, cum si diligentius inspiciatur, ut fecit Sinnius Capito, intelligi possit earn positam esse ab antiquis pro non, ut et in xii. est : ast ei custos nee escit"). For nee see above, Ch. III. 9, and below, Ch. VII. 5. Fr. 8 (V. 4). Ulpian, Frag. Tit. XXIX. 1 ; L. 195, 1. D. de Verb. Sign. : " Civis Romani liberti hereditatem lex xn. Tab. patrono defert, si intestate sine suo herede libertus decesserit Lex : EX EA FAMILIA, inquit, IN EAM FAMILIAM." Gothofredus proposes the following restoration of the law : si libertus intestato moritur cui suus heres nee escit, ast patronus patronive liberi escint, ex ea familia in earn familiam proximo pecunia adduitor. Fr. 9 (V. 5) and 10 (V. 6). From the numerous passages which refer the law de ercti-ciscunda (as the word must have been originally written) familia to the xn. Tables (see Hugo, Gesch. d. Rom. R. I. p. 229), we may perhaps suppose the law to have been : si heredes partem quisque suam habere malint, families ercti-ciscundce tris arbitros sumunto. 12. Tab. VI. Fr. 1 (VI. 1) : CUM . NEXUM . FACIET . MANCIPIUMQUE, . UTI . LINGUA . NUNCUPASSIT, . ITA . JUS . ESTO. (Festus, p. 173 ; Cic. de Off. III. 16, de Orator. i. 57). Nuncupare = nominare : Festus, 1. 1. ; Varro, L. L. VI. 60, p. 95, Muller. Fr. 2 (VI. 2). Cic. de Offic. III. 16 : " Nam cum ex xii. Tabulis satis esset ea prcestari quce essent lingua nuncupata, 12.] OR LATIN LANGUAGE. 211 quce qui infitiatus esset dupli pcenam subiret ; a jureconsultis etiam reticentiaa poena est constituta." Fr. 3 (VI. 5). Cic. Topic, c. 4 : " Quod in re pari valet, valeat in hac, quaa par est ; ut : Quoniam usus auctoritas fundi biennium est, sit etiam cedium : at in lege sedes non appellantur, et sunt ceterarum rerum omnium, quarum annuus est usus." Cf. Cic. pro Ccecina, c. 19 ; Gaius, Instit. II. 42 ; and Boe- thius ad Top. 1. c. p. 509, Orelli. Fr. 4 (VI. 6). Gaius, Inst. I. f 111 : " Usu in manum conveniebat, quse anno continue nupta perseverabat : itaque lege xii. Tab. cautum [erat], si qua nollet eo modo in manum mariti conve[mre, ut quotan]m\9 trinoctio abesset, atque [ita usum] cujusque anni interrumperet." Cf. Aul. Gell. III. 2 ; Macrob. Saturn. I. 3. Fr. 5 (VI. 7) : si . QUI . IN . JURE . MANUM . CONSERUNT . (Aul. Gell. XX. c. 10). Fr. 6 (VI. 8). From Liv. III. 44, Dionys. Hal. XI. c. 30, &c., we may infer a law : prcetor secundum libertatem vindicias dato. Fr. 7 (VI. 9) : TIGNUM . JUNCTUM . AEDIBUS . VINEAEVE, . E . CONCAPITE . NE . soLviTo . (Fest. p. 364). A great number of emendations of this passage have been proposed. The reading which I have adopted is the same as Muller's, except that I prefer concapite to his concape : compare procapis progenies-, " quse ab uno capite procedit" (Fest. p. 225). In the same way as we have capes, capiiis m. = miles ; caput, capitis n. = vertex ; so we have concapis, concapitis f. = continua capitum junctura (comp. Madvig, Beilage zu seiner Latein. Sprachl. p. 33). Fr. 8 (VI. 10). L. 1. pr. D. de tigno juncto, Ulpian, L. XXXVII. ad Edictum: "Quod providenter lex [xii. Tab.] effecit, ne vel sedificia sub hoc praetextu diruantur, vel vinearum cultura turbetur ; sed in eum qui convictus est junxisse, in duplum dat actionem." Where tignum is denned as signifying in the xii. Tables : omnis materia ex qua cedificium constet, vineceque necessaria. Fr. 9 (VI. 11) : QUANDOQUE . SARPTA, . DONEC . DEMPTA . ERUNT . (Fest. p. 384). The word sarpta (which Miiller under- stands of the ipsa sarpta, i. e. sarmenta putata) is explained by Festus, 1. 1. : " sarpiuntur vineae, i. e. putantur," &c. p. 322 : 142 212 THE OLD ROMAN [Cn. VT " [sarpta vinea putata, i.] e. pura [facta ] inde etiam [sarmenta script]ores dici pu[tant ; sarpere enim a]ntiqui pro pur[gare dicebant]." The sentence in the fragment probably ended with vindicare jus esto. 13. Tab. VII. Fr. 1 (VIII. 1). Varro, L. L. V. 22, p. 9 : Ambitus est quod circumeundo teritur, nam ambitus circumitus, ab eoque xn. Tabularum interpretes ambitum parietis circumitum esse de- scribunt." Volusius Msecianus, apud Gronov. de Sestertio, p. 398 : " Sestertius duos asses et semissem. Lex etiam xn. Tabularum argumento est, in qua duo pedes et semis sestertius pes vocatur." Festus, p. 16 (cf. p. 5) : " Ambitus proprie dicitur inter vici- norum sedificia locus duorum pedum et semipedis ad circumeundi facultatem relictus." The law itself, therefore, probably ran thus : inter vicinorum cedificia ambitus parietum sestertius pes esto. Fr. 2 (VIII. 3). Gaius (lib. IV. ad Leg. xn. Tab. L. Jin. D. finium regundorum) refers to a law of Solon, which he quotes in Greek, and describes as in some measure the type of the corresponding law of the xn. Tables, which regulates digging, fencing, and building near the borders of a piece of ground. Fr. 3 (VIII. 6): HORTUS HEREDIUM TUGURIUM . (Plin. H. N. XIX. 4, 1 : "In xn. Tab. leg. nostrar. nusquam nomi- natur villa ; semper in significatione ea hortus, in horti vero Tieredium" Festus, p. 355 : " [ Tugu-~\ria a tecto appellantur [domicilia rusticorum] sordida quo nomine [Messalla in ex- plana]tione xn. ait etiam .... significari"). Properly speaking, the vicus (signifying " several houses joined together") included the villa (- vicula, Doderl. Syn. u. Et. III. 5), which was the residence of the proprietor, and the adjoining tuguria, in which the coloni partiarii lived. All persons living in the same vicus were called vicini ; and the first fragment in this table refers to the ambitus between the houses of those who lived on the same estate. The pasture-land left common to the vicini was called compascuus ager (Festus, p. 40). It is not improbable that the words compescere and impescere occurred in the xn. Tables. See, however, Dirksen, p. 534. Ager is defined as : " locus qui sine villa est" (Ulpian, L. 27. Pr. D. de V. S.). But in a remark- able passage in Fqstus (p. 371), the vicus is similarly described in its opposition to the villa or prcedium. The passage is as 13.] OR LATIN LANGUAGE. 213 follows (see Miiller, Suppl. Ann. p. 413) : " Vici appellari inci- piunt ab agris, [et sunt eorum hominum,] qui ibi villas non habent, ut Marsi aut Peligni, sed ex vicis partim habent rempub- licam, [ubi] et jus dicitur, partim nihil eorum, et tamen ibi nun- dinae aguntur negotii gerendi causa, et magistri vici, item magistri pagi, [in iis] quotannis fiunt. Alter o, cum id genus officiorum [significatur], quae continentia sunt in oppidis, quaeve itineribus regionibusve distributa inter se distant, nominibusque dissimilibus discriminis causa sunt dispartita. Tertio, cum id genus aedifi- ciorum defmitur, quaB in oppido prive, id est in suo quisque loco proprio ita aedificat, ut in eo sedifieio pervium sit, quo itinere habitatores ad suam quisque habitationem habeat accessum : qui non dicuntur vicani, sicut ii, qui aut in oppidi vicis, aut ii, qui in agris sunt, vicani appellantur." Festus here describes (1) the vicus rusticus, (2) a street in a town, as the vicus Cyprius, and (3) a particular kind of insulated house (insula) in the city. Fr. 4 and 5 (VIII. 4, 5). Cicero de Legg. I. c. 21: " Usu- capionem xn. Tabulae intra quinque pedes esse noluerunt." Non. Marcell. de Propr. Serm. c. 5, 34, quotes, as the words of the law : si JURGANT. " Si jurgant, inquit. Benevolorum concer- tatio non Us, ut inimicorum, sed jurgium dicitur." Ursinus supposes the law to have been : si vicini inter se jurgassint, intra v. pedes usucapio ne esto. Fr. 6 (VIII. 10). L. 8. D. de Servit. Freed. Rustic. : " Viae latitude ex lege xn. Tab. in porrectum octo pedes habet ; in an- fractum, id est, ubi flexum est, sedecim." Varro, L. L. VII. 15, p. 124 : " Anfractum est flexum, ab origine duplici dictum, ab ambitu et frangendo ; ab eo leges jubent, in directo pedum vin. esse, in anfracto xvi., id est in flexu." Fr. 7 (VIII. 11). Cicero pro Ccecina, c. 19 : " Si via sit immunita, jubet (lex), qua velit agere jumentum." Cf. Festus, p. 21, s. v. Amsegetes. Miiller and Huschke express their surprise that Dirksen and other learned jurists should have overlooked the passage in Festus, which contains the best materials for the restoration of this law. Festus (s. v. Vice, p. 371) says : " Viae sunt et publicae, per [quas ire, agere, veher]e omnibus licet: privatae quibus [vehiculum immittere non licet] praeter eorum, quorum sunt privatae. [In xn. est : AMSEGETES] VIAS MUNIUNTO, DONICUM LAPIDES ESCUNT I [NI MUNIERINT,] QUA VOLET JUMENTA AGITO." See Miiller, Suppl. Annot. p. 414. 214 THE OLD ROMAN [On. VI. Fr. 8 (VIII. 9). L. 5. D. ne quid in I. publ Paulus, Lib. xvi. ad Sdbinum : " Si per publicum locum rivus aquaeductus private nocebit, erit actio private ex lege xu. Tab. ut noxaa domino caveatur." L. 21. D. de Statuliber. Pompon. L. VII. ex Plautio : si . AQUA . PLUVIA . NOCET. Fr. 9 (VIII. 7). L. 1, 8. D. de Arboribus ccedend. Ulp. L. LXXI. ad Edict. : " Lex xu. Tab. efficere voluit, ut xv. pedes altius rami arboris circumcidantur." From which, and Festus, p. 348, it is proposed to restore the law : si arbor in vicini agrum impendet, altius a terra pedes xv. sublucator. Fr. 10 (VIII. 8). Plin. H. N. XVI. c. 5 : " Cautum est prsBterea lege xu. Tab., ut glandem in alienum fundum prociden- tem liceret colligere." The English law makes a similar provi- sion respecting rabbit-burrows. Fr. 11 (VI. 4). 1, 41, I. de Rer. Divis. : " Venditae vero res et traditaa non aliter emptori adquiruntur, quam si is venditori pretium solverit, vel alio modo satisfecerit, veluti expromissore, aut pignore dato. Quod cavetur quidem et lege xu. Tab., tamen recte dicitur et jure gentium, i. e. jure naturali, effici." Fr. 12 (VI. 3). Ulpian, Fr. tit. 2, $ 4 : " Sub hac condi- tione liber esse jussus, si decem millia heredi dederit, etsi ab herede abalienatus sit, emptori dando pecuniam, ad libertatem perveniet : idque lex xu. Tab. jubet." Cf. Fest. s. v. Statuliber, p. 314. 14. Tab. Fr. 1 (VIII. 8). Cic. de Republ. IV. 10 : " Nostrse xu. Tabulae, quum perpaucas res capite sanxissent, in his hanc quoque sanciendam putaverunt : si quis occentavisset, sive carmen con- didisset, quod infamiam faceret flagitiumve alteri" Festus, p. 181 : " Occentassint antiqui dicebant quod nunc convitium fecerint dicimus, quod id clare, et cum quodam canore fit, ut procul exaudiri possit. Quod turpe habetur, quia non sine causa fieri putatur. Inde cantilenam dici querellam, non cantus jucun- ditatem puto." Plautus, Curcul I. 2, 57 ; Horat. II. Serm. 1, 80 ; II. Epist. 1, 152. Gothofredus would restore the law thus : si quis pipulo (= ploratu, Fest. p. 253 ; cf. p. 212, s. v. pipatio) occentassit, carmenve condidisset, &c. fuste ferito. Fr. 2 (VII. 9) : si MEMBRUM . RUPIT . NI . CUM . EO . PACIT, . TALIO . ESTO . (Fest. p. 363 : " Permittit lex parem vindictam." Aul. Gell. XX. 1; Gaius, Inst. III. 223). $ 14.J OB LATIN LANGUAGE. 215 Fr. 3 (VII. 10). Gaius, Inst. III. 223 : " Propter os vero fractum aut conlisum ccc. assium pcena erat (ex lege xn. Tab.), velut si libero os fractum erat; at si servo, CL." Cf. Aul. Gell. xx. 1. Fr. 4 (VII. 7) : si . INJURIAM . FAXIT . ALTERI, . VIGINTI . QUINQUE . AERIS . POENAE . SUNTO . (Aul. Gell. XX. 1 J cf. Gaius, Inst. III. 223). Fest. p. 371 : " Viginti quinque pcenas in xii. significat viginti quinque asses." Here posnas - poinas is the old form of the genitive singular and nominative plural. Fr. 5 (VII. 2) : RUPITIAS . [QUI . FAXIT] . SARCITO . (Fest. s. vv. pp. 265, 322) i. e. qui damnum dederit prcestato. Fr. 6 (VII. 5). L. 1, pr. D. si Quadrup. Paup. fee. die. Ulp. XVIII. ad Edict. : " Si quadrupes pauperiem fecisse dice- tur, actio ex lege xii. Tab. descendit ; quse lex voluit aut dari id quod nocuit, id est, id animal, quod noxiam commisit, aut sesti- mationem noxia3 offerre." Fr. 7 (VII. 5). L. 14, 3. D. de Prcescr. Verb. : " Si glans ex arbore tua in meum fundum cadat, eamque ego immisso pecore depascam, Aristo scribit non sibi occurrere legitimam actionem, qua experiri possim, nam neque ex lege xii. Tab. de pastu pecoris, quia non in tuo pascitur, neque de pauperie neque de damni injuriffj agi posse" (cf. Tab. VII. Fr. 10). Fr. 8 (VII. 3) : QUI . FRUGES . EXCANTASSIT . (Plin. H. N. XXVIII. C. 2). NEVE . ALIENAM . SEGETEM . PELLEXERIS . (Serv. ad Virg. Eel. VIII. 99). Cf. Seneca, Nat. Qucest. IV. 7, &c. Fr. 9 (VII. 4). Plin. H. N. XVIII. c. 3 : " Frugem quidem aratro qua3sitam furtim noctu pavisse ac secuisse, puberi xii. Tabulis capitale erat, suspensumque Cereri necari jubebant ; gravius quam in homicidio convictum : impubem prsetoris arbi- tratu verberari, noxiamque duplione decerni." Fr. 10 (VII. 6). L. 9. D. de Incend. Ruina Naufr. Gaius, IV. ad xn. Tab. : " Qui cedes acervumve frumenti juxta domum positum combusserit, vinctus verberatus igni necari jubetur, si modo sciens prudensque id commiserit : si vero casu, id est, negligentia, aut noxiam sarcire jubetur, aut si minus idoneus sit, levius castigatur : appellatione autem cedium omnes species sedificii continentur."" Fr. 11 (II. 11). Plin. H. N. XVII. 1 : " Fuit et arborum cura legibus priscis ; cautumque est xn. Tabulis, ut qui injuria cecidisset alienas, lueret in singulas a3ris xxv." 216 THE OLD ROMAN [On. VI. Fr. 12 (II. 4) : si . NOX . FURTUM . FACTUM . SIT, . si . IM . OCCISIT, . JURE . CAESUS . ESTO . (Macrob. Saturn. I. c. 4). Here nox = noctu ; Aul. Gell. VIII. c. 1. Fr. 13 (II. 8). L. 54, 2. D. de furt. Gaius, Lib. XIII. ad Edict. Provinc. : " Furem interdiu deprehensum non aliter occidere lex xn. Tab. permisit, quam si telo se defendat." Fr. 14 (II. 5 7). Aul. Gell. XI. c. 18 : " Ex ceteris autem manifestis furibus liberos verberari addicique jusserunt (decemviri) ei, cui factum furtum esset, si modo id luci fecissent, neque se telo defendissent : servos item furti manifest! prensos verberibus affici et e saxo praecipitari ; sed pueros impuberes prsetoris arbitratu verberari voluerunt, noxamque ab his factam sarciri." Cf. Gaius, III. I 189. For the last part, cf. Fr. 9. Fr. 15 (II. 9). Gaius, Inst. III. 191, 192 : " Concept! et oblati (furti) pcena ex lege xn. Tab. tripli est, praecipit (lex) ut qui quaarere velit, nudus quserat linteo cinctus, lancem habens ; qui si quid invenerit, jubet id lex furtum manifestum esse." Cf. Aul. Gell. XL 18, XVI. 10. Fr. 16 (II. 10) : si . ADORAT . FURTO . QUOD . NEC . MANI- FESTUM . ESCIT . (Fest. p. 162. Gaius, Inst. III. 190 : " Nee manifest! furti per leg. xn. Tab. dupli irrogatur"). For the use' of adoro, see Fest. p. 19 : " Adorare apud antiquos significabat agere, unde et legati oratores dicuntur, quia mandata populi agunt :" add, Fest. s. v. oratores, p. 182 ; Varro, L. L. VI. 76, VII. 41, &c. Fr. 17 (II. 13). Gaius, Inst. II. 45 : " Furtivam (rem) lex xii. Tab. usucapi prohibet." Fr. 18 (III. 2). Cato, E. R. procem. : " Majores nostri sic habuerunt, itaque in legibus posuerunt, furem dupli damnari, foeneratorem quadrupli." Tacit. Annal. VI. 16 : " Nam primo xn. Tabulis sanctum, ne quis unciario foenere amplius exerceret." See Niebuhr, H. R. III. 50, sqq., who has proved that the fcenus unciarium was yL of the principal, i. e. 8^- per cent for the old year of ten months, and therefore 10 per cent for the civil year. Fr. 19 (III. 1). Paulus, Rec. Sent. II. tit. 12, 11: "Ex causa deposit! lege xii. Tab. in duplum actio datur." Fr. 20 (VII. 16). L. L $ 2. D. de suspect. Tutoribus : " Sciendum est suspecti crimen e lege xii. Tab. descendere." L. 55, $ 1. D. de Admin, et Peric. Tutor. : " Sed si ipsi tutores $ 14.] OR LATIN LANGUAGE. 217 rem pupilli furati sunt, videamus, an ea actione, quae proponitur ex lege xn. Tab. adversus tutorem in duplum, singuli in solidum teneantur." Fr. 21 (VII. 17) : PATRONUS . si . CLIENTI . FRAUDEM . FECERIT . SACER . ESTO . (Servius, on Virgil's words, ^Eneid. VI. 609: " pulsatusve parens, et fraus innexa clienti"). I can sup- pose that the original had fraudem frausus siet : see Festus, p. 91, and Gronov. Led. Plant, p. 33, ad Asin. II. 2, 20. Fr. 22 (VII. 11): QUI . SE SIERIT . TESTARIER, . LIBRI- PENSVE . FUERIT, . NT . TESTIMONIUM . FARIATUR(?), . IMPROBUS . INTESTABILISQUE . ESTO . (Aul. Gell. XV. 13). Fr. 23 (VII. 12). Aul. Gell. XX. 1 : "An putas, si non ilia ex xn. Tab. de testimoniis falsis poena abolevisset, et si nunc quoque, ut antea, qui falsum testimonium dixisse convictus esset, e saxo Tarpeio dejiceretur, mentituros fuisse pro testimonio tarn multos quam videmus?" Fr. 24 (VII. 13). Pliny, in the passage quoted in Fr. 9, im- plies that involuntary homicide was but slightly punished. The fine in such a case seems to have been a ram (Serv. ad Virg. Eel. IV. 43) ; and the law has been restored thus (with the help of Cic. de Orat. III. 39, Top. 17) : si quis hominem liberum dolo sciens morti dedit, parricida esto : at si telum manufugit, pro capite occisi et natis ejus arietem subjicito. Fr. 25 (VII. 14). From Plin. H. N. XXVIII. 2, and L. 236, pr. D. de Verb. Sign., the following law has been restored: QUI . MALUM . CARMEN . INCANTASSIT . [CERERI . SACER . ESTO] . [QUI] . MALUM . VENENUM . [FAXIT . DUITVE . PARRICIDA . ESTO]. Fr. 26 (IX. 6). Porcius Latro, Declam. in Catilin. c. 19 : " Priinum xn. Tabulis cautum esse cognoscimus, ne quis in urbe coetus nocturnos agitaret." Which Ursinus restores thus : qui calim endo urbe nox coit, coiverit, capital estod. Fr. 27 (VIII. 2). L. 4. D. de Colleg. et Corporibus : " So- dales sunt, qui ejusdem collegii sunt; quam Graeci eTaipiav vocant. His autem potestatem facit lex, pactionem quam velint sibi ferre, dum ne quid ex publica lege corrumpant." 15. Tab. IX. Fr. 1 (IX. 1). Cicero pro Domo, c. 17 : " Vetant xn. Ta- bulae leges privis hominibus irrogari." Fr. 2 (IX. 4). Cicero de Legibus, III. 19: "Turn leges 218 THE OLD ROMAN [Cn. VI. prsBclarissimse de XH. Tabulis translate duae : quarum . . . altera de capite civis rogari, nisi maximo comitatu, vetat." Cf. Cicero pro Sextio, c. 30. Fr. 3 (IX. 3). Aul. Gell. XX. 1: " Dure autem scriptum esse in istis legibus (sc. xn. Tab.) quid existimari potest ? nisi duram esse legem putas, quse judicem arbitrumve jure datum, qui ob reni dicendam pecuniam accepisse convictus est, capite pcenitur." Cf. Cicero, Verr. Act. II. Lib. II. c. 32. Fr. 4 (IX. 5). L. 2, 23. D. de Orig. Jur. : " Qusestores constituebantur a populo, qui capitalibus rebus prsaessent : hi appellabantur Qucestores parricidii ; quorum etiam meminit lex xn. Tabularum." Cicero de Republ. II. 31 : " Provocationem autem etiam a regibus fuisse declarant pontificii libri, significant nostri etiam augurales ; itemque ab omni judicio poenaque pro- vocari licere, indicant xn. Tabulas compluribus legibus." See above, p. 201. Fr. 5 (IX. 7). L. 3, pr. D. ad Leg. Jul Majestat. : " Lex XH. Tab. jubet eum qui hostem concitaverit, quive hosti civem tradiderit, capite puniri." 16. Tab. X. Fr. 1 (X. 2) : HOMINEM . MORTUUM . IN . URBE . NE . SEPE- LITO . NEVE . URITO . (Cicero de Legibus, II. 23). Fr. 2 (X. 4, 5) : HOC . PLUS . NE . FACITO . ROGUM . ASCIA . NE . POLITO . (id. ibid.). Fr. 3 and 4 (X. 6, 7) : " Extenuate igitur sumtu, tribus riciniis, et vinclis purpuraB, et decem tibicinibus tollit (lex xn. Tab.) etiam lamentationem : MULIERES . GENAS . NE . RADUNTO ; . NEVE . LESSUM . FUNER1S . ERGO . HABENTO." (id. ibid.). For ricinium ^-vestimentum quadratum) see Fest. s. v. p. 274, and for radere genas (=unguibus lacerare malas) id. p. 273. From Servius ad ^En. XII. 606, it would appear that the full frag- ment would be : mulieres genas ne radunto, faciem ne car- punto, &c. Fr. 5 (X. 8) : " Cetera item funebria, quibus luctus augetur, xn. sustulerunt : HOMINI, . inquit, MORTUO . NE . OSSA . LEGITO, . QUO . POST . FUNUS . FAOiAT . Excipit bellicam peregrinamque mortem" (Cic. de Leg. II. 24). Fr. 6 (X. 9, 10) : " HSBC prssterea sunt in legibus de unctura, quibus SERVILIS . UNCTURA . tollitur, omnisque CIRCUMPOTATIO : 16.] OR LATIN LANGUAGE. 219 qusa et recte tolluntur, neque tollerentur nisi fuissent. NE . SUMTUOSA . RESPERSIO ; . NE . LONGAE . CORONAE, . NEC . ACER- RAE . prsetereantur " (Cic. de Legibus, II. 24). For acerra see Fest. p. 18 : "Acerra ara quae ante mortuum poni solebat, in qua odores incendebant. Alii dicunt arculam esse thurariam, scilicet ubi thus reponebant." Festus, s. v. Murrata potione (p. 158), seems also to refer to this law, which, according to Gothofredus ran thus : Servilis unctura omnisque circumpotatio auferitor. Murrata potio mortuo ne inditor. Ne longce coronce, neve acerrce prceferuntor. Fr. 7 (X. 11) : QUI . CORONAM . PARIT . IPSE, . PECUNIAVE . EJUS, . VIRTUTIS . ERGO . DUITOR . El. (Plm. H. N. XXI. 3 ; cf. Cic. de Leg. II. 24). Fr. 8 (X. 12). Cic. de Leg. II. 24 : " Ut uni plura (funera) fierent, lectique plures sternerentur, id quoque ne fieret lege sancitum est." Fr. 9 (X. 13) : NEVE . AURUM . ADDITO . QUOI . AURO . DENTES . VINCTI . ESCUNT, . AST . IM . CUM . ILLO . SEPELIRE . UREREVE . SE . FRAUDS . ESTo . (Cic. de Leg. II. 24). For se = sine, see above, Tab. III. fr. 6. This fragment is interesting, because it shows the antiquity of the dentist's art. Cicero (N. D. III. 22, J 57) raises the first dentist to the rank of an ^Escu- lapius : "^Esculapiorum tertius, Arsippi et Arsinose, qui primus purgationem alvi dentisque evulsionem, ut ferunt, invenit." Fr. 10 (X. 14). Id. ibid. : " Rogum bustumve novum vetat (lex xn. Tab.) propius LX. pedes adici S9deis alienas, invito domino." Fr. 11 (X. 15). Id. ibid. : " Quod autem FORUM, id est vestibulum sepulchri, BUSTUMVE . USUCAPI . vetat (lex xn. Tab.) tuetur jus sepulchrorum." Comp. Festus, s. v. Forum, p. 84. 17. Tab. XL Fr. 1 (XI. 2). Liv. IV. c. 4 : " Hoc ipsum, ne connubium patribus cum plebe esset, non Decemviri tulerunt ?" Cf. Dion. Hal. X. c. 60, XL c. 28. 18. Tab. XII. Fr. 1 (XII. 1). Gaius, Inst. IV. } 28 : " Lege autem in- troducta est pignoris capio, velut lege xn. Tab. adversus eum, qui hostiam emisset, nee pretiuin redderet ; item adversus eum, 220 THE OLD ROMAN [Cn. VI. qui mercedem non redderet pro eo jumento, quod quis ideo locasset, ut inde pecuniam acceptam in dapem, id est in sacri- ficium, inpenderet." Fr. 2 (XII. 4) : "In lege antiqua, si servus sciente domino furtum fecit, vel aliam noxiam commisit, servi nomine actio est noxalis, nee dominus suo nomine tenetur. si . SERVUS . FURTUM . FAXIT, . NOXIAMVE . NOCUIT." (L. II. $ 1. D. de NoXdl. ActlO- nibus). Fr. 3 (XII. 3) : si . VINDICIAM . FALSAM . TULIT, . STLITIS . [ET . VINDICIARUM . PRAE]TOR . ARBITROS . TRES . DATO, . EO- RUM . ARBITRIO . [POSSESSOR sive REUS] . FRUCTUS . DUPLIONE . DAMNUM . DECIDITO . (Festus, s. v. Vindicice, p. 376. I have introduced the corrections and additions of Miiller). Cf. Theodos. Cod. IV. 18, 1. Fr. 4 (XII. 2). L. 3. D. de Litigios. : " Rem, de qua con- troversia est, prohibemur in sacrum dedicare ; alioquin dupli pcenam patimur." Fr. 5 (XI. 1). Liv. VII. 17 : " In xn. Tabulis legem esse, ut, quodcunque postremum populus jussisset, id jus ratumque esset." 19. The Tiburtine Inscription. These remains of the xn. Tables, though referring to an early period of Roman history, are merely quotations, and as such less satisfactory to the philological antiquary than monu- mental relics even of a later date. The oldest, however, of these authentic documents is not earlier than the second Samnite war. It is a senatus-consultum, " which gives to the Tibur tines the assurance that the senate would receive as true and valid their justification in reply to the charges against their fidelity, and that it had given no credit, even before, to these charges" (Niebuhr, H. R. III. p. 310, orig. p. 264, tr.) 1 . The inscription was engraved on a bronze table, which was found at Tivoli in the sixteenth century, near the site of the Temple of Hercules. About a hundred years ago it was in the possession of the Barbe- rini family, but is now lost ; at least, Niebuhr was unable to dis- cover it, though he sought for it in all the Italian collections, i Visconti supposed that this inscription was not older than the Mar- sian war ; but there can be little doubt that Niebuhr's view is correct ; see Beschreibung der Stadt Rom, III. pp. 125, 659. 19.] OR LATIN LANGUAGE. 221 into which the lost treasures of the house of Barberini were likely to have found their way. Niebuhr's transcript (from Gruter, p. 499), compared with Haubold's (Monumenta Legalia, p. 81), is as follows. 1. L. Cornelius Cn. F. Praetor Senatum consuluit a. d. in. Nonas Maias sub aede Kastorus : 2. scr. adf. 1 A. Manlius A. F. Sex. Julius, L. Postu- mius S. 2 F. 3. Quod Teiburtes verba fecerunt, quibusque de rebus vos purgavistis, ea Senatus 4. animum advortit, ita utei aequom fuit : nosque ea ita audiveramus 5. ut vos deixsistis wbeis nontiata esse : ea nos ani- mum nostrum 6. non indoucebamus ita facta esse, propter ea quod scibamus 7. ea vos merito nostrofacere non potuisse ; neque vos dignos esse, 8. quei ea faceretis, neque id vobeis neque rei popli- cae vostrae 9. oitile esse facer e : et postquam vostra verba Senatus audivit, 10. tanto magis animum nostrum indoucimus, ita utei ante 11. arbitrabamur, de eieis rebus af wbeis peccatum non esse. 12. Quonque de eieis rebus Senatuei purgatei estis, credimus, vosque 13. animum nostrum indoucere oportet, item vos populo 14. Romano pur gatos fore. With the exception of a few peculiarities of spelling, as af for ab, quonque for cumque (comp. -cunque), deixsistis for dix- istis, &c., there is nothing in the phraseology of this inscription Scribundo adfuerunt. 2 Niebuhr prefers L. 222 THE OLD ROMAN [On. VI. which is unclassical or obscure. The expressions animum adver- tere, " to observe," animum inducere, " to think," seem to belong to the conventional terminology of those days. After fecerunt in 1. 3 we ought perhaps to add D. E. R. i. c. i. e. " de ea re (patres) ita censuerunt" (cf. Cic. ad Fam. VIII. 8). 20. The Epitaphs of the Scipios. The L. Cornelius, the son of Cnaeus, who is mentioned as praetor in the inscription quoted above, is the same L. Cornelius Scipio Barbatus, whose sarcophagus is one of the most interesting monuments at Rome. The inscription upon that monument ex- pressly states that he had been praetor. All the extant epitaphs of the Scipios have been given by Bunsen (Beschreibung der Stadt Rom, III. pp. 616, sqq.), who does not, however, enter upon any criticism of the text. They are as follows. (a) Epitaph on L. Cornelius Scipio, who was consul in A. u. c. 456. Cornelio' Cn. F. Scipio Cornelius Lucius \ Scipio Barbatus Gnaivod pdtre progndtus | fortis vir sapiensque, Quoms forma virtu \ tei parisuma fuit. Consul censor Aidilis | quifuit apud vos, Taurdsid' Cisauna' \ Sdmnio' cepit, Subigit omne Loucana' \ opsidesque abdoucit 1 . (b) Epitaph on the son of the above, who was aadile in A. u. c. 466 ; consul, 494. L. Cornelio' L. F. Scipio Aidiles . Cosol . Cesor . Hone omo 1 ploirume co sentiont jR[omdni] Duonoro'* optumo' \ fuise viro* Luciom Scipione 1 . | Filios Barbdti Consol, Censor, Aidiles \ hicfuet a\jpud vos~]. Hec cepit Corsica" 1 \ 'Alerid'que urbe', Dedet tempestdtebus \ aide? mereto 2 . 1 See Arnold, History of Rome, II. p. 326. 2 Bunsen, 1. 1. : "In return for the delivery of his fleet in a storm off Corsica he built a temple of which Ovid speaks (Fast. IV. 193) : Te quoque, Tempestas, meritam delubra fatemur, Quum pene est Corsis diruta classis aquis." 20.] OR LATIN LANGUAGE. 223 (c) Epitaph on the Flamen Dialis P. Scipio, son of the elder Africanus, and adoptive father of the younger 1 . Quei dpice 1 , insigne dialis \ fldminis geswtei, Mors perfecit tua ut essent \ omnid brevia, Honos fdma virtusque \ gloria dtque ingenium. Quibus sei in longd licuiset \ tibe utier vita. Facile fdcteis superdses \ gloridm majorum. Qua re lubens te in gremiu\ \ Scipio, recipit terra, Publi, progndtum \ Publio, Cornell. (d) Epitaph on L. Cornelius Scipio, son of Cn. Hispallus, grandson of Calvus, the conqueror of Spain, and nephew of Scipio Nasica : L. Cornelius Cn.f. Cn. n. Scipio. Magna sapientia Multasque virtutes cetate quom parva Posidet hoc saxsum, quoiei vita defecit non The same passage is quoted by Funccius, de Origine et Pueritia L. L. p. 326. 1 As this epitaph seems to deserve a translation, and as no one, so far as I know, has exhibited it hi an English dress, the following attempt may be accepted in the want of a better : The priestly symbol deckt thy brow: But oh ! how brief a share hadst thou Of all this world can give. Honour, and fame, and noble birth, High intellect, and moral worth: Had it been thine to live A lengthened span, endowed with these, Not all the stately memories Of thy time-honoured knightly line Had left a glory like to thine. Hail ! Publius, Publius Scipio's son ! Thy brief but happy course is run. Child of the great Cornelian race, The grave is now thy dwelling-place: And mother earth upon her breast Has lulled thee lovingly to rest. 2 Bunsen, 1. 1. : " Cicero bears testimony to the truth of these noble words in his Cato Mag. 11 : Quam fuit imbecillus African! nlius, is qui te adoptavit ? Quam tenui aut nulla potius valetudine ? Quod ni ita fuisset, altera ille exstitisset lumen civitatis ; ad paternam enim mag- nitudinem animi doctrina uberior accesserat." 224 THE OLD ROMAN [On. VI. Honos. Honore is hie situs quei nunquam Victus est virtutei : annos gnatus XX : is L[aursis] .... datus, ne quairatis honore Quei minus sit mand .... (e) Epitaph on Cn. Cornelius Scipio, brother of the preceding : Cn. Cornelius Cn.f. Scipio Hispanus Pr. Aed. Cur. Q. Tr. mil. II. Xvir si. judik. Xvir sacr.fac. Virtutes generis mieis moribus accumulavi, Progeniem genui, facta patris petiei : Majorum obtenui laudem ut sibei me esse creatum Lcetentur ; stirpem nobilitavit honor. (f) Epitaph on L. Corrfelius Scipio, son of Asiaticus, who was quaestor in 588 : L. Cornell L.f. P. n. Scipio quaist. Tr. mil. annos gnatus XXIII Mortuos. Pater regem Antioco' subegit. (g) Epitaph on a son of the preceding, who died young : Cornelius L.f. L. n. Scipio Asiagenus Comatus annoru* gnatus XVI. (h) Epitaph of uncertain date, but written in very antique characters : Aulla [sic] Cornelia. Cn. f. Hispalli. It will be observed, that in these interesting monuments we have both that anusvdrah, or dropping of the final m, which led to ecthlipsis (e. g. duonoro* for bonorum), and also the visarga, or evanescence of the nominative s (as in Cornelio for Cornelius). The dipththong ai is not always changed into ae, and gnatus has not lost its initial g. We may remark, too, that n seems not to have been pronounced before s : thus we have cosol, cesor, for consul, censor, according to the practice of writing cos. for consul (Diomed. p. 428, Putsch). Epitaph (e) has Xvir si. judik,, i. e. decemvir slitibus judikandis, where we not only observe the initial s of s[]fo'[]s = streit, but also the k before a injudikan- dis. The phraseology, however, does not differ in any important particulars from the Latin language with which we are familiar. The metre in which the three oldest of these inscriptions are composed is deserving of notice. That they are written in 20.] OR LATIN LANGUAGE. 225 Saturnian verse has long been perceived ; Niebuhr, indeed, thinks that they " are nothing else than either complete nenias, or the beginnings of them" (H. R. I. p. 253). It is not, however, so generally agreed how we ought to read and divide the verses. For instance, Niebuhr maintains that patre, in a. 2, is " beyond doubt an interpolation ;" to me it appears necessary to the verse. He thinks that there is no ecthlipsis in apice', c. 1; I cannot scan the line without it. These are only samples of the many differences of opinion, which might arise upon these short inscrip- tions : it will therefore, perhaps, be desirable, that a few general remarks should be made on the Saturnian metre itself, and that these remarks should be applied to the epitaphs before us, which may be placed among the oldest Latin specimens of the Saturnian lay 1 . That the Saturnian metre was either a native of Italy, or naturalised there at a very early period, has been sufficiently shown by Mr. Macaulay (Lays of Ancient Rome, p. 23). It is, perhaps, not too much to say, that this metre, which may be defined in its pure form as a brace of trochaic tripodisD, preceded by an anacrusis, is the most natural and obvious of all rhyth- mical intonations. There is no language which is altogether without it ; though, of course, it varies in elegance and harmony with the particular languages in which it is found, and with the degree of literary advancement possessed by the poets who have written in it. The Umbrians had this verse as well as the Latins ; at least there can be no doubt that the beginning of the vi. Eugubine Table is pervaded by a Saturnian rhythm, though the laws of quantity, which the Latins borrowed from the Greeks, are altogether neglected in it. The following may serve as a sample : *Este persklo aveis a\seriater enetu. Parfd kurnase dersva \ peiqu peica merstu, /-^ Poei dngla dseridto est J eso tremnu serse. These verses are, in fact, more regular than many of the Latin specimens. The only rule which can be laid down for the genuine Latin Saturnian is, that the ictus must occur three times in each member of the verse 2 , and that any thesis, except the 1 Livy's transcript of the inscription of T. Quinctius is confessedly imperfect; the historian says: "his/erwe incisa litteris fuit" (VI. 29). 2 To this necessity for a triple recurrence of the ictus in the genuine 15 226 THE OLD ROMAN [Cn. VI. last, may be omitted (see Muller, Suppl. Annot. ad Fest. p. 396). The anacrusis, at the beginning of the line, is often necessary in languages which, like the Latin and our own, have but a few words which begin with an ictus. When the Greek metres be- came established among the Romans, it would seem that the con- ventional pronunciation of many words was changed to suit the exigencies of the new versification, and no line began with an anacrusis, unless it had that commencement in the Greek model : but this appears not to have been the case in the genuine Roman verses, which begin with an unemphatic thesis whenever the convenience of the writer demands such a prefix. We have seen above ( 2), that the first trochaic tripodia of the Saturnius cum anacrusi, and even an amphibrachys (= trochceus cum anacrusi 1 ). Italian metre I would refer the word tripudium = triplex pulsatio. Pudio meant " to strike with the foot," " to spurn " (comp. re-pudio). The fact is alluded to by Horace, III. Carm. 18, 15: "gaudet invisam pepulisse fossor ter pede terrain." 1 In the common books on metres this would be called a single foot, i. e. an amphibrachys. It appears to me that many of the difficulties, which the student has felt in his first attempts to understand the rules of metre, have been occasioned by the practice of inventing names for the residuary forms of common rhythms. Thus, the last state of the logaosdic verse is called a choriambus; and the student falls into inex- tricable confusion when he endeavours to explain to himself the con- currence of choriambi and dactyls in the commonest measures of Horace's odes. Some commentators would persuade us that we are to scan thus : Mcece\nas atavis \ edite regfbus ; and Sic te \ diva potens \ Cypri. But how can we connect the rhythm of the choriambus with such a termi- nation ? If we examine any of the Glyconics of Sophocles, who was con- sidered a master in this species of verse, we shall observe that his cho- riambi appear in contact with dactyls and trochees, and not with iambi. Take, for instance, (Ed. Col. 510, sqq.: .ev TO T'I TOVTO ras SeiX at lias aTroJpov (j)a\Vto~as '\ ! '& I * f- \ ' aA \yr)Oovos a gvv { o~Tas prj | Trpbs ei>/]aff av\oi$;r)s ras | (raff, TreTTOV, fpy TO rot TroXu | KOI j| /M^Sa/ia | \fjyov Jj XP?7| a) > /I/ '> ! op6bv UK ova-pi d/cjovo-ai. j| Here we see that the rhythm is dactylic or trochaic these two being considered identical in some metrical systems and that the long syllable after the dactyl is occasionally equivalent to the ictus of the trochee. 20.] OR LATIN LANGUAGE. 227 could form a verse. And conversely, if the anacrusis was want- ing, the Saturnius could extend itself to a triplet of tripodiae. We We may apply the same principle to the choriambic metres in Horace, which differ only in the number of imperfect trochees which follow the dactyls in this logacedic rhythm. Thus we have nothing but dactyls in Sic te | diva po]tens Cypri: || we have one imperfect trochee or dactyl in Sic fra|tres Hele|nae |f lucida | sidera; || and two imperfect feet of the same kind in Tu ne | quaesie'ris [j scire nejfas [| quern mihi | quern tibi. || The ere tic bears the same relation to the trochaic dipodia that the cho- riambus does to the dactylic dipodia, or logacedic verse ; and it was in consequence of this reduction of the trochaic dipodia to the cretic that the ancient writers on music were enabled to find a rhythmical identity between the dactyl and the trochaic dipodia (see Miiller, Liter, of Greece, I. p. 228). It appears to me that this view of the question is calculated to settle the dispute between those who reject and those who maintain the termination of a line in the middle of a word. If every compound foot is a sort of conclusion to the rhythm, many rhythms must end in the middle of a word ; and therefore such a csesura cannot be in itself objectionable. We can hardly take any strophe in Pindar without finding some illustration of this. As a specimen, I will subjoin the first strophe of the IX. Olympian ode, with its divisions according to the rhythm : KpoVi KQ)p.a\ovTi (pi Xois 'E | trvv aXXa j vvv cKa]Ta/3oj|Xa>i/ Moi]o-ai/ OTTO Aid Te | 0oii/i|/coo-Tfpdj7rai/ (Tfp,\v6v r TOI oi'crSe jSeXeo-crtv [| AvSos In general, it seems unreasonable to call a number of syllables in which the ictus occurs more than once by the name of " foot " (pes) ; for the foot, so called, is defined by the stamp of the foot which marks the ictus, and therefore, as above suggested, the half-Saturnius would be called tri-pudium, because it consisted of three feet. For instance, if fj-eXos had no ictus except on the first and fourth syllables of ' we might scan it as two dactyls ; but if, as the analogy of -vacv ' would seem to indicate, it had an ictus on the last syllable of we must scan the words as a dactyl + trochee + ictus. This method of considering the Greek metres is exemplified in the Prosody of the Com- plete Greek Grammar. Lorid. 1848. 152 228 THE OLD ROMAN [On. VI. have instances of both practices in the old Latin translation of an epigram, which was written, probably by Leonidas of Tarentum, at the dedication of the spoils taken in the battles of Heraclea and Asculum (B. c. 280, 279), and which should be scanned as follows : Qui dntedhdc invicti | fuvere viri | pater optime Olympi 1 1 Hos ego in pugna vici \ \ V\ctusque sum ab isdem \\ l . Niebuhr suggests (III. note 841) that the first line is an attempt at an hexameter, and the last two an imitation of the shorter verse ; and this remark shows the discernment which is always so remarkable in that great scholar. The author of this translation, which was probably made soon after the original, could not write in hexameter verse, but he represented the hex- ameter of the original by a lengthened form of the Saturnius, and indicated the two penthemimers of the pentameter by writing their meaning in two truncated Saturnians, taking care to indicate by the anacrusis that there was really a break in the rhythm of the original pentameter, although it might be called a single line according to the Greek system of metres. To return, however, to the epitaphs of the Scipios. The scansion of the lines, which I have adopted, is sufficiently indicated by the metrical marks placed over the words. It is only neces- sary to add a few explanatory observations. With the exception of a. 2, 3, b. 3, and c. 7, every line begins with an anacrusis, or unaccentuated thesis ; and it seems to be a matter of indifference whether this is one long or two short syllables. The vowel i is often pronounced like y before a vowel, as in Lucy us (a. 1), Lucyom (b. 3), dyalis (c. 1), brevya (c. 2), ingenyum (c. 3), utyer (c. 4), gremyu (c. 6), Scipyo (ibid.). And u is pronounced like w in c. 2. The rules of synaloepha and ecthlipsis are some- times attended to (as in a. 6), and sometimes neglected (as in b. 5, c. 4). The quantity of fuisse and virc? in b. 2, may be justified on general principles; for fuisse is properly fuvisse, and viro is written veiro in Uinbrian. But there is no consis- tency in the syllabic measurement of the words in these rude The lost original may have been as follows : TOVS Trplv dvuvJTovs, irarep alyXycvTos ' T eKparovv, 01 r' Kparr)(rav 20.] OR LATIN LANGUAGE. 229 lines. Facile, in c. 5, makes a thesis in consequence of that short pronunciation which is indicated by the old form facul (Fest p. 87, Miiller). As all the other verbs in epitaph a. are in the perfect tense, it seems that subigit and abdoucit, in the last line, must be perfect also. Indoucimus is perhaps a perfect in the Tiburtine inscription (1. 10) : " postquam senatus audivit, tanto magis indoucimus ;" and subigit was probably pro- nounced subigit, The beginning of b. seems to have been the conventional phraseology in these monumental nenias. The sepulchre of A. Attilius Calatinus, which stood near those of the Scipios at the Porta Capena (Cic. Tusc. Disp. I. 7, 13), bore an inscription beginning in much the same way : Hone oino ploirume co\sentiont gentes. Populi primarium \fuisse virwn. (Comp. Cic. de Finibus, II. 35, 116 ; Cato M. 17, 61). 21. The Columna Rostrata. The Columna Rostrata, as it is called, was found at the foot of the Capitol in the year 1565. Its partial destruction by lightning is mentioned by Livy (XLII. 20) ; and it was still standing, probably in the existing copy, when Servius wrote (ad Virgil. Georg. III. 29). It refers to the well-known ex- ploits of C. Duilius, who was consul B.C. 260, A.U.C. 494. This inscription, with the supplements of Ciacconi, and a commentary, was published by Funck, in his treatise de Orig. et Puer. L. L. pp. 302, sqq. It is here given with the restorations of Grotefend (Orelli, no. 549). [C. Duilios, M. F. M. N. Consol advorsum Poenos en Siceliad Sicesf]ano[s socios Rom. obsi- dioned crave~\d eocemet leciones r[_efecet dumque Poenei m\aximosque l macistratos l[ecionumque duceis ex n~]ovem castreis exfociunt Macel[am opidom opp]ucnandod cepet enque eodem mac i As it is said that maxumus was the prevalent form before Caesar's time, this more recent spelling may indicate that the inscription is not in its original condition. 230 THE OLD ROMAN [On. VI. [istratod bene r~]em navebos marid consol primos c[eset socios] dasesque navales primos ornavet pa[ravetque~\ cumque els navebos claseis Poenicas om[neis et max~\sumas copias Cartaciniensis praesente\_d sumod] Dictatored ol[pr\om in altod marid pucn[ad meet] xxxque nam\_s cepe]t cum socieis septem\ynilibos quinresm~\osque triresmos- que naveis[xiv. mer set. tone aur]om captom numei DC . . . . [ponded arcen~]tom captom prceda numei ccclooo [pondod crave] captom aes ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo ccclooo .... \is qu]oque navaled praedad poplom [Rom. deitavet atque~] Cartacini[ens~]is \ince~\nuos d\uxet triumpod cum xxx rostr~\eis [clasis] Carta [ciniensis captai quorum erco S. P. Q. R. hanc colomnam eel P.]. 22. The Silian and Papirian Laws, and the Edict of the Curule Festus has preserved two interesting fragments of laws, which are nearly contemporary with the Columna Rostrata. The first of these is the Lex Silia de publicis ponderibus, which was passed in the year B. c. 244, A. u.c. 510. Festus s. v. Publica pondera, p. 246 : " Publica pondera [ad legitimam normam ex- acta fuisse] ex ea causa Junius .... [collegijt quod duo Silii P. et M. Trib. pleb. rogarint his verbis : Ex ponderibus publicis, quibus hac tempestate populus oetier solet, uti coaequetur (l) sedulum, uti quadrantal mni octoginta pondo siet; con- gius mni decem p. siet ; sex sextari congius siet mni; duo de quinquaginta sextari quadranta siet vini; sextarius aequus aequo cum librario siet (s} ; sex dequimque w librari in modio sient. 22.] OR LATIN LANGUAGE. 231 Si quis magistratus adversus hac d. m. pon- dera modiosque vasaque puUica modica, majora, minorave faxit, jusseritve fieri, dolumve adduit quo eafiant, eum quis volet magistratus^ multare, dum minore parti familias taocaf\ liceto; sive quis im (s} sacrum judicare voluerit, liceto" The Latinitj of this fragment requires a few remarks. (1) cocequetur. In the Pompeian Inscription (Orelli, no. 4348) we have : mensuras excequandas. (2) Sedulum. Scaliger sug- gests se dolo m. i. e. sine dolo malo. But sedulo or sedulum itself signifies "sine fraude indiligentiaeve culpa" (Miiller ad l.\ and the law refers to the care and honesty of those who were to test the weights and measures. For sedulus, see Doderl. Syn. u. Et. I. p. 118. (3) "Nihil intelligo nisi librarius qui hie significatur sextarius frumenti erat." Miiller. (4) Sex de- quimque = sex decimque, the qu being written instead of c. (5) The editions have jussit ve re, for which Miiller writes jussitve ; Haubold (Monumenta Legalia) proposes jusseritve, " propter sequens re ;" and I have adopted this reading on account of the word faxit, which precedes. (6) Quis volet magistratus. Cf. Tab. Bantin. Osc. 12. Lat. 7. (7) Dum minore parti fami- lias taxat. Compare the Latin Bantine Inscription, 1. 10 : [dum minoris] partus familias taxsat. Cato, apud Aul. Gell. VII. 3 : " Quse lex est tarn acerba quse dicat, si quis illud facere voluerit, mille nummi dimidium familice multa esto 2" The abl. parti (which occurs in Lucretius) and the genitive partus (com p. Cas- tor us in the Bantine Inscription, ejus, cujus, &c.) depend on multare and multam, which are implied in the sentence. For taxat, see Fest. p. 356. These passages show the origin of the particle dumtaxat, which is used by the classical writers to sig- nify " provided one estimates it," (: estimating it accurately," " only," " at least," " so far as that goes," &C. 1 (8) Im = eum. Fest. p. 103. The Lex Papiria de Sacramento, which is to be referred to the year B.C. 243, A.U.C. 511, is thus cited by Festus s. v. Sacra- 1 It is scarcely necessary to point out the absurdity of the derivation proposed by A. Grotefend (Ausf. Gramm. d. Lat. Spr. 124): " dun- taxat aus dum taceo (cetera) sat (est hoc) I" 232 THE OLD ROMAN [Cn. VI. mentum, p. 344 : " Sacramentum ses significat, quod poanse no- mine penditur, sive eo quis interrogatur, sive contenditur. Id in aliis rebus quinquaginta assium est, in aliis rebus quingentorum inter eos, qui judicio inter se contenderent. Qua de re lege L. Papiri Tr. pi. sanctum est his verbis : Quicunque Praetor post hac factus erit qui inter civesjus dicet, tres viros Capitales populum rogato, hique tres viri [capitales], quicunque [posthac fa]cti erunt, sacramenta ex\igunto~], judicantoque, eodemque jure sunto, uti ex legi- bus plebeique scitis exigere, judicareque, esseque oportet" To these may be added the old Edictum ccdilium curulium de Mancipiis Vendundis, quoted by Gellius, N. A. IV. 2 : Titulus serwrum singulorum utei scriptus sit, ccerato ita, utei intellegi recte possit, quid morbi vitiive quoique sit, quis fugitivus errove sit, nox- ave solutus non sit. 23. The Senatus Consultum de Bacchanalibus. The Senatus Consultum de Bacchanalibus, which is referred to by Livy (XXXIX. 14), and which belongs to the year B.C. 186, A.U.C. 568, was found at Terra de Teriolo in Calabria, in 1 640, and is now at Vienna. A facsimile of the inscription, with the commentary of Matthseus ^Egyptius, will be found in Dra- kenborch's Livy, Vol. VII. pp. 197, sqq. 1. [$] Mar cms L. F, S. Postumius L. F. Cos. Sena- turn consoluerunt N. 1 Octob. apud aedem 2. Duelonai sc* arf? M. Claudi M.F. L. Valeri P. F. Q. Minuci C. F. De Bacanalibus, queifoideratei 3. Esent, ita eocdeicendum censuere. Neiquis eorum Sa- canal* habuise velet; sei ques* 1 Nonis. 2 scribundo. 3 adfuerunt. 4 Bacchanal. 5 ques = qud. See Klenze, Legis Servilice Fr. p. 12, not.2; Fest. p. 261. 23.] OR LATIN LANGUAGE. 233 4. esent, quei sibei deicerent necesus 1 ese Bacanal habere, eeis utei ad pr. urbanum 5. Romam venirent, deque eeis rebus, ubei eorum vtr a* audita esent, utei senatus 6. noster decerneret, dum ne minus senatoribus c. ade- sent [quom e]a res cosoleretur. 7. Bacas* vir ne quis adiese* velet ceivis Romanus, neve nominus Latin\i], neve socium 8. quisquam, nisei pr. urbanum adiesent, isque de sena- tuos sententiad, dum ne 9. minus senatoribus c. adesent, quom ea res cosoleretur, iousisent, censuere. 10. Sacerdos ne quis vir eset, magister neque vir neque mulier quisquam eset, 11. neve pecuniam quisquam eorum comoinem hdbuise velet, neve magistratum 12. neve promagistratud, neque virum neque mulier em quiquam* fecise velet, 13. neve post hac inter sed* conjourase neve comvovise neve conspondise 14. neve conpromesise velet, neve quisquam fidem inter sed dedise velet, 15. sacra in oquoltod^ ne quisquam fecise velet neve in poplicod neve in 16. preivatod, neve exstrad urbem sacra quisquam fecise velet, nisei 17. pr. urbanum adieset, isque de senatuos sententiad, dum ne minus 18. senatoribus c. adesent quom ea res cosoleretur, iousi- sent, censuere. 19. Homines pious v. oinvorsei 8 , virei atque mulier es, sacra ne quisquam 1 necessum. 2 1. utra verba. 3 i. e. Bacchas. 4 adlisse. 6 quisquam. 6 i. e. se as in 1. 14. 7 occulto. 8 universi. 234 THE OLD KOMAN [Cn. VI. 20. fecise velet, neve inter ibei 1 mrei pious duobus, mu- lieribus pious tribus, 21. arfuise velent, nisei de pr. urbani senatuosque sen- tentiad utei suprad 22. scriptum est. Haice utei in coventionid 2 exdeicatis ne minus trinum 23. noundinum, senatuosque sententiam utei scientes ese- tis, eorum 24. sententia itafuit. Sei ques 3 esent quei arvorsum ead fecisent quam suprad 25. scriptum est, eeis rem caputalemfaciendam censuere, atque utei 26. hoce in tabolam ahenam inceideretis. Ita senatus aiquom censuit. 27. Uteique eamfigierjoubeatis ubeifacilumed 4 ' gnoscier potisit 5 , atque 28. utei ea Bacanalia, sei qua sunt exstrad quam sei quid ibei sacri est, 29. ita utei suprad scriptum est, in diebus x quibus vobeis tabelai 6 datai 30. erunt,faciatisuteidismota sient. In agro Teurano* 1 . 24. The Old Roman Law on the Bantine Table. The Roman law on the Bantine Table is probably not older than the middle of the seventh century. The chief reason for introducing it here, is its connexion in locality, if not in import, with the most important fragment of the Oscan language (above, p. 116). Mommsen divides it into six, Klenze into four sections. His transcription and supplements (Rhein. Mus. for 1828, pp. 28, sqq. ; Phil. Abhandl. pp. 7, sqq.), compared with those of Momm- sen (Untevital. Dialekte, pp. 140, sqq.), give the following results : 1 = interea. 2 contione. s ques = quei. 4 facillime. & = potis-sit = possit. 6 =tabellce. 1 in agro Teurano. Strabo, p. 254 c : vrrep fie TWV Qovpiw KOI 17 Tau- pidvr] x&pa 24.] OR LATIN LANGUAGE. 235 CAP. 1. On the degradation of offenders. 1. \n\eque prov [inciam] 2. in sena\tu seiv~\e in poplico joudicio ne sen[tentiam rogato tabellamve nei dato] 3. ... deici]o, neive quis mag. testumonium pop- lice eid[_em defer ri neive deri\ontiari 4. ... [sinito neive joudicem eum neive arbitrum neive recipe~\ratorem dato, neive is in poplico luuci praetextam neive soleas habeto neive quis 5. [mag. prove, mag. prove quo imperio potestateve erit qu]eiquomque comitia conciliumve habebit eum sufragium ferre nei sinito 6. [neive eum censor in senatum legito neive in senatu] relinquito- L. 3. See Quinctil. V. 7, $ 9 : " Duo sunt genera testium, aut voluntariorum aut quibus in judiciis publicis lege denuntiatur" L. 4. luuci, " by day." Plaut. Cas. IV. 2, 7 : " Tandem ut veniamus luci." Cic. Phil XII. 10, 25 : " Quis audeat luci illustrem aggredi ? w CAP. 2. On the punishment of judges and senators who violate the law. 7. \_Seiquis joudex queiquomque ex hace lege~] plebeive scitofactus erit senator ve fecer it gesseritve quo ecc hace lege 8. [ininusfiant quae fieri oportet quaeve fieri oportu] erit oportebitve non fecerit sciens d. m., seive advorsus hance legem fecerit 9. [gesseritve sciens d. m.; ei multa tanta esto HS. . . eamque pequniam] quei volet magistratus exsi- gito. Sei postulabit quei petet pr. recuperatores 10. [quos, quotque dari opor}teat dato jubetoque eum sei ita pariat, condumnari populo, facitoque jou- dicetur. Sei condemnatus 11. [erit, quanti condemnatus erit, prides'] ad q. urb. 236 THE OLD ROMAN [On. VI. det aut bona ejus poplice possideantur facito. Seiquis mag. multam inrogare volet, 12. [ei multam inrogare liceto, dum minoris} partus familias taxsat liceto; eiq. omnium rerum si- remps lex esto, quasei sei is haace lege 13. [multam HS....exegisset.~] 12. dum minoris partus familias taxsat. See above, $ 22, on the Lex Silia. Partus is the genitive case, like Cas- torus, cap. 3, 1. 17. Siremps is explained by Festus, p. 344 : " Siremps ponitur pro eadem, vel, proinde ac ea, quasi similis res ipsa. Cato in dissuadendo legem . . . relicta est : Et prseterea rogas, quemquam adversus ea si populus condempnaverit, uti siremps lex siet, quasi adversus leges fecisset." CAP. 3. On binding the judges and magistrates by an oath to observe the law. 14. [Cos. pr. aid. tr. pi. q. mvir. cap. uimr. a. d. a. qu\ ei nunc est, is in diebus v proxsumeis, quibus queique eorum sciet h. 1. popolum pkbemve 15. [joussisse jouranto utei infra scriptum est. Item die. cos. pr. mag. eq. cens. aid. tr. pi. q. mvir cap. mvir a. d. a. joudex ex h. I. plebive scito 16. [factus queiquomque eorum p]osthac factus erit, eis in diebus v proxsumeis quibus quisque eorum mag. inperiumve inierit, jouranto 17. utei infra scriptum est. Eidem consistunto in ae] de Castorus palam luci in forum vorsus, et eidem in diebus v apud q. jouranto per Jovem deosque 18. [penateis, sese quae ex h. l.facere opart]ebit factu- rum, neque sese advorsum h. I. facturum scien- tem d. m. neque seese facturum neque interce- surum 19. [quo qua? ex h. I. oportet minus fiant. Qu]ei ex h. I. non jouraverit, is magistratum inperiumve nei petito newe gerito neive habeto, neive in senatu 24.] OR LATIN LANGUAGE. 237 20. [si adfuerit sententiam dicere e]um quis sinito neive eum censor in senatum legito. Quei ex h. I. joudicaverit, isfadto apud q. urb. 21. [nomen ejus quei jouraverit sc\riptum siet, quaes- torque ea nomina accipito et eos quei ex h. 1. apud sedjourarintfacito in taboleis 22. \_popliceis scriptos habeaf]. L. 15. i. e. Dictator) consul, prcetor, magister equitum, cen- sor, cedilis, tribunus plebei, qucestor, triumvir capitalis, triumvir agris dandis adsignandis. L. 17. palam luci in forum versus. See Cic. de Offic. III. 24. CAP. 4. On the oath of the senators. 23. [Quei senator est inve senatu sententi]am deixer\in\t post hance leg em rogatam, eis in diebus x prox- sumeis, quibus quisque [eorum sdet h. IJ] 24. [populum plebemve joussisse, j^ouranto apud quaes- torem ad aerarium palam luci per Jovem de [psqu\e penate\is sese quce ex h. L 25. \_facere oportebit facturum, neque see]se advorsum hance legemfacturum esse, neque seese 26. se hoice leegeifi 27. anodni uraver. L. 23. eis = is. L. 24. ad cerarium. See Liv. XXIX. 37. Per Jovem deosque penateis. Comp. Cic. Acad. IV. 20. CAP. 5. 28. e quis magistratus, p. 29. CAP. 6. 30. [u]ti in taboleis popl[iceis'} 31. \tr~\inum nondm\uiri\ 32. is eritun. CHAPTER VII. ANALYSIS OF THE LATIN ALPHABET. 1. Organic classification of the original Latin alphabet. 2. The labials. 3. The gutturals. 4. The dentals. 5. The vowels. 6. The Greek letters used by the Romans. 7 The numeral signs. 1. Organic Classification of the Original Latin Alphabet. THE genuine Latin alphabet, or that set of characters which expressed in writing the sounds of the Roman language be- fore it had borrowed from the Greek a number of words, and the means of exhibiting them to the eye, may be considered as consisting of nineteen letters ; that is, of the representatives of the original Cadmean syllabarium (which consisted of sixteen letters), with an appendix comprising the secondary vowels, or vocalised consonants, i and u, and the secondary sibilant x = sh. If we distribute these nineteen letters according to their natural or organic classification, we shall have the following arrangement : CONSONANTS. Labials. Gutturals. Dentals. Medials . . . B G D Aspirates . . F H R Tenues . . . P Qv T Liquids . . . M L, K Sibilants . . S, X VOWELS. Vowels of Ar-1 Heaviest. Lightest. Medium. ticulations J A E Vocalised 1 Vocalised Labial. Vocalised Guttural, or Dental. Consonants J u I l.] ANALYSIS OF THE LATIN ALPHABET. 239 It will be most convenient, as well as most methodical, to consider these letters according to this classification, which will be justified by the investigation itself. 2. The Labiah. The labials consist of three mutes and the liquid M. The regular changes of the labial mutes, in the principal languages of the Indo-Germanic family, have been thus indicated by James Grimm, to whom we owe the discovery of a most important law (Deutsche Gramm. I. p. 584 J ), which may be stated thus in its application to all three orders of mutes : In Greek, , p ,. In Old High Latin, Sanscrit. hic ' German. Medial corresponds to Tennis and to Aspirate. Aspirate ,, Medial Tennis. Tennis Aspirate Medial. This law, applied to the labials only, may be expressed in the following table : Latin, (Greek, Sanscrit) . B F P Gothic P B F Old High German . . . F P B (V) To take the instances given by Grimm himself, the first column is confirmed, as far as the Latin language is concerned, by the following examples : cannabis (/cctVi/ajSts), Old Norse hanpr, Old High German hanaf; turba (Oopu/3r]\ Goth, thaurp, 0. H. G. dorof; stabulum, 0. N. stopull, O. H. G. staphol To which may be added, labi, Anglo-Saxon slipan, 0. H. G. sliuffan. These instances are confined to the occurrence of the labials in the middle of words ; for there are no German words beginning with P, and no H. G. words beginning with F. The second column is supported as follows : Initials -fag us ((pjT/o's), O- N. beyki, 0. H. D. puocha ; fero ((f>epw), Goth. baira, O. H. G. pirn; fui ((pvio), Ang.-Sax. beon, O. H. G. pirn ; flare, Goth, blasan, O. H. Q.plasan; fra-n-gere (piiyvv/mi), Goth. brikan, 0. H. G. prtchan ; folium (\r], Ke(j)a\ij 9 ypdfatv and such like, the Latin equivalents present b or p ; compare nebula, caput, s-cribere. The reason for this is to be sought in the aversion of the Roman ear from r as a middle sound. The third column rests on the following induction : Initials pes (pedis), Goth, fotus, 0. H. G. vuoz ; piscis, Goth, fisks, O. H. G. vise; pater, Goth.fadrs, O. H. G. vatar ; plenus, Goth. fulls, O. H. G. vol ; pecus, Goth, faihu, 0. H. G. vihu ; palma, Angl.-Sax. folma, 0. H. G. volma ; pellis, Goth, fill, O. H. G. vel ; pullus, Goth, fula, 0. H. G. volo ; primus, Goth, frumists, O. H. G. vromist. Middle sounds sopor, 0. N. svefn, 0. Sax. suelhan; septem, Angl.-Sax. sefon, Goth, sibun; afer, Angl.- Sax. eofor, 0. H. G. ebar ; super, Goth, ufar, 0. N. yfir, 0. H. G. ubar ; rapina, Angl.-Sax. reaf, O. H. G. roub. These may be taken as proofs of the general application of Grimm's rule to the Latin labials. If, however, we examine the use of the separate letters more minutely, we shall find great vacillation even within the limits of the Latin language itself. The medial B seems to have approximated in many cases to the sound of v ; at other times it came more nearly to p. We find in old Latin the forms Duillius, duonus, duellum, &c. by the side ofBillius, bonus, bellum, &c. Now, there is no doubt that the proper abbreviation of these forms would be e. g. donus or vonus, and so on. The labial representative bonus, therefore, shows a sort of indifference between the occasional pronunciation of B and v. This view is confirmed by a comparison of duis, which must have been the original form, with 5/s on the one hand, and bis, bes, vi-ginti on the other. The same appears parti- cularly in the change from Latin to Italian or French, as in haber e= aver e avoir, habebamavevaavois, Aballo=Avalon, Cabellio-Cavaillon, Eburovices=Evreux, &c., or conversely, as in Vesontio ~ Besan$on. The commutation of b and v in the Spanish language gave occasion to Scaliger's epigram : Haud temere antiquas mutat Vasconia voces Cui nihil est aliud vivere quam bibere 1 . 1 Penny Cyd. III. p. 220. See also Scaliger de Cans, L. L. I. c. 14. p. 36. In older Latin we have Fovii by the side of Fabii (Fest. p. 87), Sevini by the side of Sabini (Plin. H. N. III. 12), Stovenses by the side of $2.] ANALYSIS OF THE LATIN ALPHABET. 241 The interchange of B and p may be remarked in burrus, wvp- po$ ; Balantium, Palatium ; bitumen, pitumen (comp. pituita) ; &c. In many Latin words the B stands for a (=P'H) in tho Greek synonym : compare balcena, albus, ambo, nebula, umbi- licus, &c., with (paXaivci) ctX0os, afji ; e-/moXoi>, p.ejuiftXwKa ', /mo^oo?, &c.) : but in the derivative idioms there are many instances of this insertion ; compare numerus, nombre ; camera, chambre ; &c. ; and even when r is substituted for some other liquid, as in hominem, Sp. hombre ; or when a third liquid is retained, as in cumulare, Fr. combler. In classical Latin B is often omitted when flanked by two vowels; this is particularly the case in the dative or ablative plural, as in queis by the side of quibus, filiis by the side of filiabus, &c. ; indeed this omission is regular in the second declension. It is hardly necessary to remark, that the genuine Etruscan element in the Latin language must have been altogether with- out the medial B. As a final, B is found only in the proclitic words ab, ob, sub. When B or v is followed by the vocalised guttural j, we sometimes remark that, in the derived languages, this guttural supersedes the labial, and is pronounced alone, or with an as- similation; so we have cavea (= cavja), cage; cambiare, changer ; debeo, deggio ; Dibio, Dijon ; objectum, oggetto ; rabies, rage; rubere (=rubjere), rougir ; subjectum, sujet ; &c. We see the full development of this change in such words as nager from navigare, while the absolute omission of the labial is justified by ecrire from scribere, in Amiens from Ambiani, and in aimois, which comes from amabam through aimoy=amoue= amava, (Lewis, On the Romance Languages, p. 199). The labial F and the guttural Q V are the most characteristic letters in the Latin alphabet. Of the latter I will speak in its place, merely remarking here that its resemblance to F consists in Stobenses, and in the flexion-forms of the verb -bo, -bam, -bilis, -bundus, by the side of -vi, from/o and/wi (see Corssen, Zeitschr.f. Vergl. Sprf. 1852. p. 17). 16 242 ANALYSIS OF THE LATIN ALPHABET. [Cn. VII. the fact that they are both compound letters, although used from the earliest period as exponents of simple sounds. In considering the Latin F, we must be careful not to confuse it with the Greek on the one hand, or with the modern v on the other. It is true that F corresponds to (p in a number of words, such as fagus, fama, fero, fallo, fari, fastis, frater, frigus,fucus,fugio,fui,fulgeo,fur (Miiller, Etrusk. I. p. 20); but we must consider these words as an approach to a foreign articulation ; for in a great number of words, in which the F has subsequently been commuted for H, we can find no trace of con- nexion with the Greek (p : such are fariolus, fasena, fedus, fircus, folus, fordeum, fostis, fostia, forctis, vefo, trafo (Miiller, Etrusk. I. p. 44). It is generally laid down that F and v are both labio-dental aspirates, and that they differ only as the tenuis differs from the medial ; and one philologer has distinctly asserted their identity, meaning perhaps that in Latin F=the English v, and u=the English w. If, however, we analyse some of the phenomena of comparative philology in which the Latin F appears, and then refer to Quintilian's description of the sound of this letter, we may be disposed to believe that in many cases the English v formed only a part of the sound. Quintilian says (XII. 10, 27, 29) that the Roman language suffered in comparison with the Greek from having only v and F, instead of the Greek v and in the instances cited above, v also appears as a substitute both for and IT. Compare valgus, vallus, veru, virgo, and vitricus, with o\Koy, palus, Tm'pw, jrapOevos, and pater (Buttman, Lexil. s. v. o\Kos). $2] ANALYSIS OF THE LATIN ALPHABET. 243 Quintilian that in his time the Latin F contained, in addition to the labial v, some dental sibilant ; and the sibilant is known to have been the condition in which the guttural passed into the mere aspirate, (c) A comparison of the Greek Qjp with its Latin synonym fera would produce great difficulty, if we could not suppose a coexistence of the sibilant with the labial in the latter ; such a concurrence we have in the Russian synonym svera, Lettish svehrs, Old Prussian svirs. (d) The Sabine words mentioned above (such as fircus), the more modern repre- sentatives of which substitute an aspirate for the r, prove that the F must have contained a guttural aspirate ; for no labial can pass into a guttural, though a compound of labial and guttural may be represented by the guttural only, (e) Those words in the Romance languages which present an aspirate for the F, which their Latin synonyms retained to the last, such as falco, " hawk ;" foris, Fr. " hors ;" facere, formosus, fumus, &c., Sp. " hacer," " hermoso," " humo," &c., prove that, to the last, the Latin F contained some guttural element, in addition to the labial of which it was in part composed. It seems to me that F must have been sv, or, ultimately, HV, and that v must have corresponded to our English w. With regard to the Greek 0, there can be no doubt that it was a distinct p'h, like the middle sound in hop-hazard, shep-herd; reduplications like 7T6(f>vKa (pe-p'huka), and contacts like 2a7r0a> (Sapp'ho), suffi- ciently prove this. The forms of Latin words which seem to substitute F for this must be referred to the Pelasgian element in the Latin language : the Tuscans, as we have seen, were by no means averse from this sound ; and the Romans were obliged to express it by the written representative of a very different articulation. The derivation of Falerii and Falis-ci (cf. Etruria and Etrusci) from a founder Halesus, shows that even among the Tuscans there was an intimate affinity between F and H (see Muller, JEtr. II. p. 273). Of the tenuis p it is not necessary to say much. If we compare the Latin forms with their Greek equivalents, we observe that P, or PP, is used as a substitute for the cf> (P'H) of which I have just spoken. Thus puniceus, caput, prosper, &c., correspond to (poiviKcos, K(pd\rj, 7rpoar lepus, nepos, opera, pauper, recipere, sepelire, sapere, &c., with ouvrir, avril, cheveu, chevetre, chevre, eveque, avoir, genievre, lievre, neveu, ceuvre, pauvre, recevoir, en-sevelir, sa- voir, &C. 1 p is often inserted as a fulcrum to the labial M when a liquid follows : thus we have sumo, sum-p-si, sumptus ; promo, prom- p-si, promptus. Contact with the guttural J will convert p into CH=J or a soft G. Compare rupes, roche ; sapiam, sache ; sapiens, sage, &c. Here in effect the labial is assimilated or absorbed, as in Rochester from Hrof-ceastre. The labial liquid M occasionally takes the place of one or other of the labial mutes, even within the limits of the Latin language itself. It stands by the side of B in glomus, hiems, melior, tumeo, &c., compared with globus, hibernus, bonus (benus, bene, bellus, &c., j3e\T/a>i>, /3eWTTo?, &c.), tuber, &c. We find a substitution of B for M in Bandela, the modern name of Mandela (Orelli ad Hor. III. Carm. 18, 12), and in Lubedon for Laomedon (Scaliger, de Caussis L. L. I. c. 22, p. 54). I am not aware that we have any example of the commutation of M with the labiodental F. With v it is not uncommon: comp. Mulciber, Vulcanus ; pro-mulgare, pro-vulgare, (compare di-vulgare) ; &c. This is still more remarkable if we extend the comparison to cognate languages : thus Mars, mas (maris), may be compared with Fa^s, Fapprjv, vir, virtus, "war," wehren, " warrior," 'Oapicw ; and Minne, " Minion," &c., with Venus, Winnes-jdfte, &c. (Abhandl. Berl Ak. 1826, p. 58). 1 To avoid unnecessary trouble (for independent dictionary-hunting would have led, in most cases, to a repetition of the same results) I have taken several of the commonest comparisons of French and Latin synonyms from the articles on the separate consonants in the Penny Cyclopaedia. It is scarcely worth while to make this reference, for no one acquainted with French and Latin need go to the Penny Cyclopaedia, or any other compilation, in order to learn that ouvrir, avril, &c. are derived from aperire, aprilis, &c. $ 2.] ANALYSIS OF THE LATIN ALPHABET. 245 So also /md-v-Tis may be compared with vates ; at least, Plautus writes mantiscinari for vaticinari. The changes of p into M are generally observable in assimilations such as summus for supimus, supremus : in Greek, and in the passage between Greek and Latin, this change is common enough ; thus we have yuera by the side of TrcSa, and /uioXvfiSos by the side of plumbum. In fact, M and N are more nearly akin to the medials B and D than to the tenues, and a thick articulation will always give the medials for the liquids. At the end of Latin words M is very often omitted in writing, and seems to have been still more frequently neglected in pro- nunciation. With regard to the written omissions, it was the rule to omit in the present tense of active verbs the important M which characterises the first person in many of the other tenses. In fact, the only verbs which retain it in the present tense are su-m and inqua-m : and it is mentioned as a custom of Cato the Censor, that he used also to elide the M at the termination of the futures of verbs in -o and -io (see Ch. VI. ^ 3). The metrical ecthlipsis, which disregards the final -M when a vowel follows, may be explained by supposing a sort of anusvdrah in the Latin language. In the transition to the Romance languages, which make a new nominative of the Latin accusative, the final m is dropt in all but two instances the Italian speme = spem, which extends it by a final vowel, and the French rien = rem, which substitutes the nasal auslaut. 3. The Gutturals. The Roman gutturals are three, the medial G, the aspirate H, and the labio-guttural tenuis Q V . The regular changes of this order of mutes, as far as the Latin language is concerned, are proved by the following examples ; the law itself, as applied to the gutturals, being expressed thus : Latin, (Greek, Sanscrit) . G H C Gothic K G H, G Old High German . . . CH K H, G 1st column. Initials : granum, 0. N. korn, O. H. G. chorn; genus, kuni, chunni ; gena, 0. N. kinn, 0. H. G. cJiinni ; genu, kne, chnio ; gelu, gelidus, Gothic kalds, O. H. G. chalt ; gustare, klusan, chiosan. Middle sounds ; ego, ik } ih (icli) ; ager, akrs, 246 ANALYSIS OF THE LATIN ALPHABET. [On. VII. achar ; magnus, mikils, michil ; jugum^ juk, joch ; mulgere, O. N. miol/ca, O. H. G. melchan. 2d column. Initials : hanser, gans, kans ; heri, hesternus, gistra, kestar ; hortus, yards, karto ; hostis, gasts, kast ; homo, guma, komo. H is of rare occurrence as a middle sound in Latin ; we may, however, compare via, veha, with weg ; veho with Goth, aigan; traho with Anglo-Sax, dragon, &c. 3d column (in which I have substituted c for Q V , because the latter belongs to a different class of comparisons). Initials : claudus, halt, halz ; caput, haubith, houbit ; cor t hairto, herza ; canis, hunths, hund. Middle sounds : lux, liuhad, licht ; tacere, thahan, dagen ; decem, Goth, taihun, Lith. deszimts. Originally the Romans made no distinction between the gut- turals c and G ; the former was the only sign used ; and although Ausonius says (Idyll. XII. de litteris, v. 21) : gammce vice functa prius c (see also Festus, s. vv. prodigia, orcum), thereby implying that c expressed both the medial G and the tenuis K 1 , there is reason to believe that in the older times the Romans pronounced c as a medial, and used Q as their only tenuis gut- tural. This appears from the forms macestratus, leciones, &c., on the Duillian monument, and still more strikingly from the fact that the prsenomens Gains, Gnceus (Fato?, Fewatos), were to the last indicated by the initials C. and Cn. ; for in the case of a proper name the old character would survive the change of application. When, however, the Romans began to distinguish between the pure tenuis K and the labial tenuis Q, they intro- duced a distinction between c and G, which was marked by the addition of a tail to the old character c, the letter thus modified being used to represent the medial, and the old form being trans- ferred from the medials to the tenues. The author of this change was Sp. Carvilius, a freedman and namesake of the cele- brated Sp. Carvilius Ruga, who, in A. u. c. 523, B.C. 231, fur- nished the first example of a divorce. See Plutarch, Qucest. JRom. p. 277 D. : TO K Trpos TO F avyyeveiav e^et irap ai>Tois [the Romans], o\|/e yap e-^ptjcravTo T(p 7rpocrel~upovTos. Id. p. 278 E. : o\|/e , /ecu TT/HOTO? avw% ypa/jLimctToSiSacrKaXelov \\apfil\ios aTreXevOepov Kap/3i\iou TOV irpcoTOv ya/u.eTtjv K/3a- On this confusion in other languages see New Crat. 100. $3.] ANALYSIS OF THE LATIN ALPHABET. 247 XOCTOS-. From the position in the alphabet assigned to this new character, namely, the seventh place, corresponding to that of the Greek z, there is reason to believe that the Roman c stilL retained the hard ^r-sound, while the new character represented the soft sibilant pronunciation of the English J and the Greek z, which is also expressed by the modern Italian gi. It is clear that the Greek K was introduced long before the time of Carvi- lius, and therefore there could have been no need of an additional character except for the expression of an additional sound. And as K was used only in the syllable ka, the additional sound must have been that borne by c and o in modern Italian before the vowels E and i. Before o and u, as we shall see directly, Q was in its original place. The Latin H was a strong guttural aspirate, corresponding in position and in power to the Greek ^. It is true that this cha- racter sometimes indicates a mere spiritus asper ; and in this use it is either dropt or prefixed, according to the articulation. In general, however, it was the strongest and purest of the Roman aspirated gutturals. Graff has remarked (Abhandl. Berl. Ak. 1839, p. 12) that there are three classes of aspirates the guttural (H), i. e. the spiritus ; the labial (w) i. e. the flatus ; and the dental (s), i. e. the sibilatus : and he says that the Latin language entirely wants the first, whereas it possesses the labial aspirate in its Q, and the dental perhaps in its x. This appears to me to be neither a clear nor a correct statement. With regard to H in particular, there can be no doubt that it is a strong guttural, quite as much so as the Greek ^. This is esta- blished by the following comparison. The Latin H answers to ^ in the words hiems (^etVwy), hibernus (xeinepivos), hio (^aivw), humi (^a/xat)? hortus (^0/0x09), &c. It represents the guttural c in trah- o, trac-si, veh-o, vec-si, &c. In a word, it corresponds to the hard Sanscrit A, for which, in the cognate Gothic and Greek words, either g, k, or y, K, ^, are substituted (comp. N. Crat. $ 112). An initial H, or some other guttural, was often omitted in Latin, as in other languages, before another consonant ; thus we have res for hres-hra-is from hir " the hand ;" rus for hrus or eras (karsh = aro), Icena by the side of y\aiva ; ruo by the side of con-gruo, Roma by the side of gruma (above, p. 60), &c. And even before vowels we have frequent instances of the extenuation and omission of an original H. Indeed it is 248 ANALYSIS OF THE LATIN ALPHABET. [On. VII. sometimes a matter of doubt whether the H ought to be retained or dismissed in spelling ; thus some would write Hannibal, others Annibal ; some Etruria, others, more correctly as I think, but less in accordance with authority, Hetruria ; although aut and haud are the same word, and though old MSS. make no distinc- tion between them (Lachmann ad Lucret. III. 330, 632), the former generally omits, while the latter as generally retains the H ; and while hcereo is almost the universally received orthography, we have cesit in Lucret. VI. 1016 (ubi v. Lachm.), in accord- ance with the Tyrrhenian at-cesum, (above, Ch. V. 3. p. 153). With regard to Q or Q V , a character almost peculiar to the Latin alphabet, a longer investigation will be necessary. It has been a common opinion with philologers that there were different classes of the tenuis guttural, varying with the vowel which arti- culated them ; thus, /caV?, kaph, was followed only by a ; H (hetfi) only by e ; ^7 only by i ; Koinra, koph, only by o ; and Q only by u. Lepsius (Zwei Abhandl. pp. 18-31) has given a more rational and systematic form to this opinion, by supposing that there were three fundamental vowels, a, i 9 u ; that i was subsequently split up into i, e, and u into o, u ; that one of the three fundamental vowels was prefixed to each row of mutes in the old organic syllabarium, so that all the inedials were articu- lated with a, all the aspirates with i, and all the tenues with u. This form of the opinion, however, is by no means sufficient to explain the peculiarities of the Roman QV ; and if it were, still it could not be adopted, as it runs counter to the results of a more scientific investigation into the origin of i and u. The difficulty, which has been felt in dealing with the Latin Q, has proceeded chiefly from the supposition that the accompany- ing u or v must be either a distinct vowel or a distinct consonant ; for if it is a vowel, then either it ought to form a diphthong with the accompanying vowel, or a distinct syllable with the Q ; and neither of these cases ever happens : if, on the other hand, it is a consonant, the vowel preceding the Q ought to be long by position ; and this is never the case even in the most ancient writers (see Graff, Abh. Berl. Ak. 1839 : " uber den Buchsta- ben Q (QV)"). It appears to me unnecessary to assume that the accompany- ing u is either a distinct vowel or a distinct consonant. And herein consists the peculiarity of the Roman Q : it cannot be 3.] ANALYSIS OF THE LATIN ALPHABET. 249 articulated without the u, and yet the u has no distinct exist- ence. The true explanation, I conceive, is the following. No attentive student of the Latin authors can have failed to observe how great a tendency there is in this language to introduce sounds consisting of an union of the guttural and labial. Such a sound is the digamma, which may be considered to have been the leading characteristic of the Pelasgian language both in Italy and in Greece. Now there are four states of this sound, besides its original condition, in which both guttural and labial have their full power : the first is when the labial predominates, and this is expressed by the letter F = sv (hv) ; the second is when the guttural predominates, and this is expressed by QV ; the third is when the guttural alone is sounded, and in this state it becomes the strong guttural H or K ; the fourth is when the labial alone is articulated, and from this we have the letter v. The great difference between F and QV consists in this, that in the latter it is necessary to express both the ingredients of the double sound, whereas they are both represented by one charac- ter in the former. Hence it has happened, that, while the guttural element of F has been overlooked by many philologers, they have over-estimated the independent value of the labial which accompanies Q. A sound, bearing the same relation to the medials that QV does to the tenues, is occasionally formed by the addition of v to G. This occurs only after n and r : thus we find tinguo, unguo, urgueo, by the side of tingo, ungo, urgeo. The former were probably the original words, the latter being subsequent modi- fications : compare guerra, " war," guardire, " ward," &c. with the French pronunciation of guerre, guardir, &c. (New Crat. 110). When the labial ingredient of QV is actually vocalised into u, the Q is expressed in classical Latin by the new tenuis c = K ; thus quojus, quoi, the original gen. and dat. of qui, become cujus, cui ; cui rei becomes cur , quom is turned into cum ; sequundus, oquulus. torquular (comp. torqueo), quiris (cf. Qui- rinus), &c., are converted into secundus, oculus, torcular, curis, &c. This is also the case when u is represented by the similar Roman sound of the o. Thus colo must have been originally quolo ; for Q is the initial of quolonia on coins, and in-quilinus is obviously derived from in-colo, which has lost its u, just as 250 ANALYSIS OF THE LATIN ALPHABET. (jDn. VII. quotidie is written cotidie (Schneider, Lat. Gr. I. p. 335). It is known, too, that coquus must have been pronounced quoquus even in Cicero's time ; for he made no difference in pronunciation between the particle quoque and the vocative of coquus : see Quintil. VI. 3, 47 : " Quas Ciceroni aliquando exciderunt, ut dixit, quum is candidatus, qui coqui films habebatur, coram eo suffragium ab alio peteret : ego quoque tibi favebo." The change of qva into cu is particularly remarkable when a syllable is shortened, on account of the heavier form in which it occurs ; as when quatio in composition becomes con-cutio, per-cutio, &c. Perhaps we ought to write acua in those cases in which aqua appears as a trisyllable (Lachmann ad Lucret. VI. 552). The two constituent parts of Q V often exist separately in different forms of the same root : thus we have conniveo, connixi; fio (tj)v(*)),facio,factus; fluo,fluxi; foveo, focus; juvo, jucun- dus ; lavo, lacus ; nix, nivis ; struo, struxi ; vivo, vixi. The last is a double instance ; for there can be no doubt of the con- nexion between " quick" and vivus (for qviqvus) (New Crat. 112, note). Bopp's opinion, therefore (Vergleich. Gramm. pp. 18, 98), that there is some natural connexion between v and k in themselves, is altogether unfounded. In the comparison between Latin and Sanscrit we seldom find that QV is represented by a Sanscrit K, but that it usually stands in cognate words where the Sanscrit has a palatal guttural or sibilant (New Crat. 105, 216) : compare quatuor, Sanscr. chatur ; s-quama, Sanscr. chad, " tegere ;" quumulus, Sanscr. chi, "accumulare;" oc-cultus (ob-quultus), Sanscr. jal, "tegere;" sequor, Sanscr. sajj ; pequus, Sanscr. pa$u ; equus, Sanscr. a$va; &c. When QV stands by the side of a Sanscrit &, it is either when that letter is followed by e or i in which case the gut- tural approximates to the palatal, or when the k stands before u or v. There are some instances in which the QV is represented by the labial P in Greek and Sanscrit ; and this is particularly remarkable in cases where the QV occurs twice in the Latin word : compare the Latin quinque, quoquo (coquo), aqua, loquor, &c., with the Sanscrit and Greek panchan, Tre'/uTre, pach, TreVeo, dp, lap, &c. ; also equus, oquulus, sequor, linquo, &c., with '/TTTTO?, OfJL/Att, 7TO/XCa, Xe/TTO), &C. Quintilian says that the Latin Q is derived from the Greek (I. 4, 9) ; and there can be no doubt that they have a 3.] ANALYSIS OF THE LATIN ALPHABET. 251 common origin. Now this Greek KOTnra, which is of rare oc- currence, is found, where it occurs in Greek inscriptions, only before o. Thus we have popivOoOev (Bockh, C. I. no. 29), opcpov (n. 37), Xi/po^OjO/cas- (n. 166) ; and on coins we have cpopivQos, ^vpaQoviwv, &c. The explanation of this is simple : the letter o before a vowel expressed the sound of w, so far as the mouth of a Greek could convey this sound : compare olcrrpos, pdifiSos, which imitate the whizzing noises of the wings of the gad-fly and the bird ; 6'a, which represents the Persian lamenta- tion wa ! &c. (above, p. 49). Consequently, the syllable /oo$. The last of these words is a mutilation which reminds us of the modern Scotch division of the name Alexander into the two abbreviations Alick and Saunders or Sandy. When the transposition was once effected, the softening of the guttural was obvious and easy : compare cr^erXios, " scathe," schade ; XapM, " s-kirmish," schirm, &c. The Latin s is principally remarkable as standing at the beginning of words, the Greek equivalents of which have only an aspirate : compare sal, sex, septem, sol, sylva, simul, sedere, sequi, somnus, &c., with a\s, ef, CTTTO, 17X10$, vXFty, a/xa, eea-0ai, eiro/uac, VTTVOS, &c. Though in some cases even this aspirate has vanished : as in ai/af , el, eXXo'?, &c., compared with senex, si, sileo, &c. It frequently happens that in the more modern forms of the Roman language an original s has been superseded by the dental sibilant R. Thus Quintilian tells us (I. 4, 13) that Valesius, Fusius, arbos, labos, vapos, clamos, and lases (cf. Fest. s. v.), were the original forms of Valerius, Furius, arbor, labor, vapor, clamor, and lares ; and it is clear that honor, honestus, are only different forms of onus, onustus. It is rather surprising that the Jurist Pomponius (Digg. I. 2, 2, 36) should have attributed to Appius Claudius Csecus (consul I. A.U.C. 447, B.C. 307; consul II. A.U.C. 458, B.C. 296) the inven- 254 ANALYSIS OF THE LATIN ALPHABET. [On. VII. tion of a letter which is the initial of the names Roma and Romulus. He can only mean that Appius was the first to in- troduce the practice of substituting R for s in proper names, a change which he might have made in his censorship. It appears, from what Cicero says, that L. Papirius Crassus, who was consul in A.U.C. 418, B.C. 336, was the first of his name who did not call himself Papisius (ad Famil. IX. 21) : " How came you to suppose," says Cicero, writing to L. Papirius Psetus, " that there never was a Papirius of patrician rank, when it is certain that they were patricii minorum gentium ? To begin with the first of these, I will instance L. Papirius Mugillanus, who, in the year of the city 312, was censor with L. Sempronius Atratinus, who had previously (A.U.C. 310) been his colleague in the consulship. But your family-name at that time was Papisius. After him there were thirteen of your ancestors who were curule magis- trates before L. Papirius Crassus, the first of your family that disused the name Papisius. This Papirius was chosen dictator in A.U.C. 415, with L. Papirius Cursor for his magister equitum, and four years afterwards he was elected consul with K. Duilius." We must conclude, therefore, that Appius Claudius used his cen- sorial authority to sanction a practice, which had already come into vogue, and which was intimately connected with the pecu- liarities of the Roman articulation. In fact, the Romans were to the last remarkable for the same tendency to rhotacism, which is characteristic of the Umbrian, Dorian, and Old Norse dialects. 4. The Dentals. The Romans had five dentals or linguals : the mutes D and T, the liquids L and N, and the secondary letter R, which in most alphabets is considered a liquid, but in the Latin stands for an aspiration or assibilation of the medial D. Grimm's law, as applied to the dentals, stands thus : Latin, (Greek, Sanscrit) . D T Gothic T D Z, TH Old High German . . Z T D The following examples will serve to establish the rule. 1st column. Initials: dingua, lingua, tug go, sunga ; deus, O. N. tyr, 0. H. G. ziu ; dens, dentis, Goth, tunthus, 0. H. G. zand ; domare, tamjan, zemen ; dolus, O. N. tal, %dla ; ducere, 4.] ANALYSIS OF THE LATIN ALPHABET. 255 Goth, tiuhan, 0. H. G. zioJian ; duo, tva, zuei ; dextra, taihsvo, xesawa. Midlde sounds : sedes, sedere, sitan, sizan ; edere, itan, ezan ; videre, vitan, wizan ; odium, hatis, ha% ; u-n-da, vato, wazar ; sudor, sveiti, sweiz ; pedes, fotjus, vuozi. 2d column. The Latin has no 0', and when the R stands for the D, there are generally other coexistent forms in which the medial is found. For the purpose of comparison Grimm has selected some Latin words in which a Latin F stands by the side of the Greek 0. Initials : fores (Ovpa), daur, tor ; fera (6t/p), 0. N. dyr, 0. H. G. tior. Middle sounds : audere, ausus (9appeiv), gadauran, turran ; mathu, Tusc. (Gr. ^e'0i/), Anglo-Sax, medo, O. H. G. metu. 3d column. Initials : tu, Gothic thu ; 0. H. G. du ; tener, 0. N. thunnr, 0. H. G. dunni ; tendere, Goth, thanjan, O. H. G. denen ; tacere, thahan, dagen ; tolerare, thulan, dolen ; tectum, thak, dach. Middle sounds: frater, brothar, pruoder ; rota, O. N. hradhr (" celer"), 0. H. G. hrad (" rota") ; a-l-ter (Umbr. Tusc. etre), anthar, andar ; iterum, vithra, widar. Of the commutations of the dentals one with another in the Latin language alone, the most constant is the interchange of D with L or R. D becomes L in delicare (Fest. pp. 70, 73), impe- limenta, levir, Melica, (Fest. p. 124), olfacit, for dedicare, impedimenta, Sarip, Medica, odefacit ; and is assimilated to L in such words as mala, ralla, scala, sella, from ma-n-do, rado, sca-n-do, sedeo : the converse change is observable in 'O$f doWs, HoXvSevKtjs, SctKpvov (dacrima, Fest. p. 68), $a\l/i\ij$, dingua (Mar. Viet. p. 2547) (0. H. G. zungd), Capitodium, meditari, kadamitas, adauda, &c., the more genuine forms of which are preserved in Ulysses (oXryos), Pol- lux (comp. Seuxes, Hesych. with lux), lacryma (liqueo) 9 lapsilis (XctTrro)), lingua (Xei^eiv), Capitolium, ^eXerav, calamitas, alauda, &c. : $e'co, on the con- trary, is a more ancient form than ligare, (see N. Crat. 155). This change takes place within the limits of the Greek language also : comp. SeiSa) with ^eiXo's, $a? ($&>s) with $aAo's, &c., though in many of these cases there is the residue of an original assimilation, as in /caXo's, root KaS-, cf. jca'^co, &c. The change is also observable in the passage from Latin to the Romance lan- guages ; thus Digentia has become Licenza, the people of Madrid call themselves Madrilenos, and Egidius becomes Giles. The other dentals, T and N, are also sometimes converted into L : as 256 ANALYSIS OF THE LATIN ALPHABET. [Cn. VII. in Thetis, Thelis ; Nympha, Lympha, &c. (See Varro, L. L. VII. 87). In some cases there is a passage from & to X in Greek, as in a^j/, a\is (compare satis) ; and the Greek in 6wpri% is represented by an I in lorica. There is an inter- change of N and R in cereus, ceneus ; in murus, munio ; in Scopov, donum; -jr\ri^,plenus; Londres, London; Havre, Hafen; &c. The ablative or adverbial D has become n in longinquus, pro- pinquus, from longe[d], prope\_d~\ ; compare antiquus, posticus, from antea, postea 9 amicus from amo (amao), &c. In the cor- ruption Catamitus from Ganymedes, both N and D are changed into T, and in caduceus from KYJPVKGIOV we have the converse change from R to D. D is dropt when flanked by two vowels, as es for edis 9 est for edit, esse for edere, item for itidem, &c. So also the dental liquids L and N are liable to excision ; compare vis = volis, and the numberless omissions of the final -nt as. in fuere =fuerunt, regna = regnont. The change from D to R has been often pointed out, in such common instances as au-ris compared with aud-io, apor for apud, meridie for medii die, ar-vocat for ad-vocat, &c. The verb arcesso, which is also written accerso, furnishes a double example of the change : the original form was ad-ced-so = accedere sino , in arcesso the first d is changed into r, and the second assimi- lated to s : in accerso the first d is assimilated to c, and the second changed to r. In the Romance language D is changed into R in the Spanish lampare from lampada, and conversely in the Italian rado from raro, fedire from ferire ; compare the English paddock for parruc, A. S. for par k. As a final letter, D became more and more liable to proscrip- tion. With the exception of the proclitics ad and apud, some- times written et or at and aput, ar and apor ; the conjunction sed, also written set ; and the adverb hand, also written haut and aut (cf. autem) ; we have no D in auslaut in classical Latinity. In the ablative, D was absorbed before the rise of Roman litera- ture, and -ad for -nd or -nt in the neuter plural was finally repre- sented by -d only. N is principally remarkable in Latin from its use as a sort of anusvdrah (see N. Crat. p. 303). In this use it is inserted, gene- rally before the second consonant of the root, as in tu-n-do, root tud- ; fi-n-do, root fid-, &c.; but sometimes after it, as in ster-n-o, root ster-, stra- ; sper-n-o, root sper-, spre- ; si-n-o, root si-, &c. 4.] ANALYSIS OF THE LATIN ALPHABET. 257 Conversely, N becomes evanescent in certain cases, particu- larly before s and v. Thus consul is written cosol (abbreviated into cos) ; and we find cesor, in/as, vicies, vicesimus, for censor, infans, viciens, vicensumus. This omission of N is regular in the Greek participles in -et?, and in other words, e. g. o&ous ; it seems also to have been the rule in Umbrian. In the Romance language the Latin termination -ensis generally loses its N. Thus we have Vaudois by the side of Waldenses, bourgeois for bur- gensis, courtois for cortensis, c. In Italian we have Veronese for Veronensis, marchese for marchensis, paese for pagensis ; and the two last pass into the French marquis and pays. The most important instance of the omission of N before v is furnished by the common word contio, derived from conventio through the form coventio 1 , which is found in old inscriptions. Similarly, convent becomes covent (" Covent-g&rden, &c."), Conflu- entes is turned into Coblenz, amdfiinf into " five." In English the prefix con is shortened into co- before all consonants, in spite of the remonstrances of Bentley. On the contractions of con in Latin, see Lachmann on Lucret. II. 1061. The original preposi- tion is especially disguised in coelebs co-i-lebs coitum linquens. With regard to the changes experienced by the dentals in the passage from Latin to the Romance dialects, the following instances may suffice. D and T are frequently dropt in the French forms of Latin words: (a) D: Andegavi, Fr. Anjou; Ca- durci, Fr. Cahors ; Mediomatrices, Fr. Metz ; Meduana, Fr. Mayenne ; Mediolanum, It. Milano ; Melodunum, Fr. Melun ; cauda (It. coda, Sp. cola), Fr. queue; fides, Fr. foi; media- node, Fr. mi-nuit ; nudus, Fr. nu ; Rhodanus, Fr. Rhone ; vadum, Fr. gue ; videre, Fr. voir 2 . (b) T : acetum, Lomb. aseo ; ad-satis, Fr. as-sez (originally assetz) ; Autura, Fr. Eure ; amatus, Fr. aime ; Bituriges, Fr. Bourges ; Matisco, Fr. Md$on; Rhedones, Fr. Rennes; Rodumna, Fr. Rouanne; Catalauni, Fr. 1 Contio stands related to coventio as nuntius to novi-ven-tius ; comp. nov-i-tius. Domitius, the proper name, seems to signify "the horne- goer ;" so propitius, as the antecedent of praesens, when said of a deity. Ilithyia (old fern, of el\ci6a>s) might be rendered Propitia. 2 The French sometimes drop the D before a guttural in words of German extraction, as in Huguenot for Eidgenossen, or Eid-genoten, i. e. " conspirators." 17 258 ANALYSIS OF THE LATIN ALPHABET. [Cn. VII. Chalons ; pater, Fr. pere ; Rutheni, Fr. Rodez ; vita, Fr. vie. There is a double abbreviation in Arras from Atrebates. So also we have, Mayence from Moguntiacum, page from paeda- gogium (N. Crat. 225), and Rich-borough from Rutupium, where we have also the change from pi to ch (above, p. 244). In Grenoble from Gratianopolis the first three syllables are contracted, just as in gre from gratia, in malgre, c. On the con- trary, D intrudes or is revived in certain prepositions when com- pounded with verbs beginning with a vowel ; thus we have prod- est but pro-sunt, red-eo, but re-verto, and as we have re-fero, it may be doubtful whether re-tuli or ret-tuli is for red-tuli or re-tetuli. Relligio, relliquice, &c. favour the former supposition. In the Romance languages this letter is sometimes inserted as a fulcrum between the liquids n and r, as in cendre, Dordogne, gendre, tendre, from ciner-is, Duranius, gener, tener ; viendr-ai, tiendr-ai, for venir-ai (venire habeo), tener-ai (tenere habeo), &c. ; vendredi for Veneris-die, &c. This will remind the classical student of the similar insertion in the Greek av-$-po$, &c. ; and both the Greeks and the Romans apply the same principle to the labials also. The combination TI is almost always represented by a soft G in French words derived from the Latin ; as age, etage, mariage from cetatium, statio, maritatio. In these cases it is matter of indifference whether we suppose a softening of the whole combination (N. Crat. $ 112) or an omission of the dental and substitution of the i = /, as in the labial forms men- tioned above (p. 244). The indistinctness with which the French pronounce N at the end of a word has given rise to some etymological, or rather orthographical, inconsistencies in that language. Not the least remarkable of these is the appearance of s instead of M or N in the first person of many verb-forms. If we compare suis with the Italian sono on the one hand, and the Spanish soy on the other, and remember that the first and third persons of the present tense in the Romance verbs do not exhibit a final s in the oldest examples of the language, we may conclude that the s in this and other French forms is an arbitrary orthographic appendage. The termination -ois=ensis shows that soy is not an inadequate representative of sono. L, N, R, are frequently interchanged as the Latin passes into 4.] , ANALYSIS OF THE LATIN ALPHABET. 259 the Komance idiom. L passes into R 1 in apotre, epitre, Orne, rossignol, titre, &c., from apostolus, epistola, Olina, lusciniola, titulus, &c. ; N into L in alma, Barcelona, Bologna, Lebrixa, from anima, Barcino, Bononia, Nebrissa ; N into R in diacre from diaconus, in sero, sevi by the side of sino, sivi, and in Langres from Lingones, Never from Noviodunum. In Old Latin r passes into I, as in Cceles Vivenna from Cceres (above, p. 26) ; but I passes into r in cceruleus from cceluleus. We seem to have a change of I into r, or vice versa, in Us, litis, from stlit, compared with the German streit. L is a representation of D in Giles from JEgidius, in ellera for edera, and in Versiglia for Vesidia. The Italians vocalise L into i when it follows certain conso- nants : compare clamare, clarus, clavis, flos, Florentia, fluctus, flumen, obliquus, Placentia, planus, plenus, &c., with chiamare, cliiaro, chiave, fiore, Fiorenze (Firenze), fiotto, fiume, bieco (Fr. biais, Engl. " bias 2 "), Piacenza, piano, pieno, &e. The French vocalise the Latin L into u, which seems to have been in the first instance only an affection of the previous vowel, into which the L was subsequently absorbed. Thus alter was first written aultre, and then autre. This affection of a preceding vowel by the liquid which follows is not uncommon in other languages. The Greeks in some of their dialects pro- nounced the vowel broad before or after p : comp. (ppaai with (fipeai, &c. : and the common people in Dorsetshire pronounce o like a when it is followed by r and another consonant ; thus George is pronounced Gearge, storm, starm, &c. The French absorption of the L is almost universal: it is regular in the dative of the article au=a le, aux-a les ; in the plurals of nouns in I, as animales, animaux ; canales, canaux, &c. But it is also found in a number of other words, in which the vowel 1 Ad-Mare seems to be an instance of the converse change from R to L : for this compound is from ad and ula = ovpd, and refers, like the Greek traiveiv (= o-tietv, " to shake or wag "), to the dog blandishing, fawning, and wagging his tail. The older etymologers connect it with ad-oro ; but this admits of a different interpretation. 2 It is probable that the word " bias " came from France with the game of bowls, and as denoting that one-sided weight which makes the sphere run obliquely, it is connected in meaning as well as origin with biais = bieco = obieco = obliquus. 172 260 ANALYSIS OF THE LATIN ALPHABET. [On. VII. preceding I is not a ; even when it is u : compare aliquis unus, altare t eXerujLoa-vvrj, Bulgare, felix (like o /maKapirr]^ used in speaking of the dead), ulna, &c., with the French aucun, autel, aumone, bougre, feu (anciently written feux sm&feulx), aune, &c. 5. The Vowels. The philological student must always bear in mind that there are two distinct classes of vowels ; the one containing the vowels of articulation, A, E, o ; the other comprising the vocalised conso- nants i and u. In other words, there are only three distinct vowels, A, i, u ; for E and o differ from A in weight only. The original alphabet is a syllabarium consisting of breathings and Consonants, which are articulated by the sound A. Now the character A, in its original application, denotes the lightest of the breathings, the character E the heaviest of them, and the cha- racter o a breathing which is intermediate in weight. Conse- quently, on the principle that the lightest vowel always co- exists with the heaviest form (see N. Crat. 101, 222, &c.), when these breathings were no longer indicated by distinct characters, A would represent the heaviest articulation-vowel, E the lightest, and o that which stands between them in point of weight. That this is actually the order of the articulation-vowels, considered in respect to the weight of the combinations in which they are found, is clearly established by an examination of the existing forms in the most perfect of the Indo- Germanic languages. The vowels i and u result from the vocalisation, not of breathings, as is the case with A, E, o, but of mutes. The former is the ultimate state of the softened or assibilated gut- turals and dentals, the latter is the residuum of the labials (^V. Crat. 108). But, though they are of different origin from A and its subordinates, they must be considered, especially in the Latin language, as occasionally approximating in sound to the vowels derived from breathings, and as representing them in certain cases, where forms of an intermediate weight require an intermediate weight of vowels. This will be best shown by examples, from which it will appear that the vowels i and u have shades of value, or rather that they admit of subdivision into other vowels, differing from them in weight, as E and o differ from A, but not expressed jn different characters, at least in the existing written remains of the Latin language. 5.] ANALYSIS OF THE LATIN ALPHABET. 261 It has been remarked that the a of the root-syllable is changed into i or e in secondary formations according to a fixed rule: namely, the a becomes i when the root-syllable in the longer form remains otherwise unchanged ; but the a is turned into e when the root-syllable is followed immediately by an adsci- titious consonant, or when the consonant following the root- vowel is thrown back upon the vowel by some consonantal vowel, like i, or e-y (see Bopp, Vergleich. Gramm. p. 5 ; Rosen, Journal of Education, VIII. p. 344; N. Crat. 222 '). The following examples may suffice to establish this : A I E amicus . . . in-imicus . . . "enmity." arma in-ermis. ars in-ers. barba . im-berbis. {oc-ciput . . . (bi-ceps. prin-cipium . < prce-ceps. sin-ciput . . [prin-ceps. , fce-cidi. \stilli-cidium. cano facio . (ce-cini \tubi-cinis icon-ficio \pro-fici ciscor con-centus. tubi-cen. con-fectus. pro-fectus. factum . pro-fecto. fallo fe-felli. fastus pro-festus. gradior re-gredior. jacio . . . ab-jicio . . . ab-jectus. taceo . . . con-ticesco. tango . . . con-ting o. The cause of the change from i to E is farther shown by the change back again from E to i when the root is not followed by two consonants : thus, bi-ceps, &c., become bi-cipitis, &c. in the genitive ; and similarly tubi-cen[s] makes tubi-cinis. Another change from i to E is to be remarked in the transformation of 1 Similar to this is the case of qametz c hatwph in Hebrew, for here the long d becomes 8 in consequence of the consonant in auslaut being thrown back on the vowel of articulation. 262 ANALYSIS OF THE LATIN ALPHABET. [On. VII. the diphthongs AI, 01 into AE and OE. It was also a peculiarity of the Latin writers from the earliest times to use E as a repre- sentative of EI, for which also they occasionally substituted i. Thus, while'' HVe ipos becomes Epirus ; Dei, Di; Deis, Dis; &c.; we have naves by the side of naveis=navis, and both tris and tres by the side of treis. Schwartze (alte JEgypten, I. p. 605) distinguishes three main periods of Latin orthography in regard to the pronunciation of i and E. The peculiarity of the first and oldest period consisted in the employment of E with a dull i sound, which Schwartze terms the E pinguis. The second period, which immediately preceded the classical, wrote i instead of this E pinguis. The third or classical period in a considerable number of forms introduced an E, which formally corresponded to the old E pinguis, but was materially different from it, and this, as it possessed the true sound of E, he calls the phonetic E. The next comparison, in point of weight, which suggests it- self, is that between the secondary vowels i and u ; and in order to make this comparison satisfactorily, it will be well to consider first their subdivisions. It appears, then, that there are three distinct uses of each of these vowels : i is (1 ) a very long vowel, the representative of the diphthong AI=AE ; (2) a vowel of medium length, frequently, as we have seen above, the representative of a, the first part of that diphthong ; (3) a very short vowel ap- proximating to the sound of the shortest u, and used chiefly before R. Similarly, u is (1) a very long vowel, the represen- tative of the diphthong OI=OE; (2) a vowel of medium length, generally answering to o, the first part of that diphthong ; (3) a very short vowel, approximating to the sound of the shortest i, and used chiefly before L. The old Italians had separate cha- racters for I 3 and u 3 , which differed from the other characters by the addition of certain marks : i 3 was written F, like a mutilated F, and u 3 was written T. It is remarkable that the emperor Claudius, when he introduced his new letters into the Roman alphabet to express the consonant v, the Greek \[/, and the modi- fication I 3 , while he inverted the digamma (thus d) to express the first, and joined two sigmas (thus X ) to express the second, which was consequently called antisigma (Priscian, p. 545, Putsch ; I. p. 40, Krehl), was contented to borrow the third from the old alphabet of the Oscans. The following examples will justify the subdivision which I have made of the vowels i and u. $ 5.] ANALYSIS OF THE LATIN ALPHABET. 263 l^ In composition we find this long vowel in the root- syllable of words which contain the diphthong ai = ae. Thus, from ces-timo we have ex-istimo ; from cequus we have in-iquus; from ccedo, con-cido, oc-cido ; from queer '0, in-quiro ; &c. This long i, as we have seen, also represents the diphthong Ei, and it is used as a contraction for n, especially in the genitives of nouns in -ius. When employed for either of these purposes, it is expressed in the inscriptions by an exaggeration of form ; thus we have D!S, AL!, OB!T, for Deis, alii, obiit. Conversely, we sometimes find that a doubled vowel is written to represent one long vowel ; thus we have (Orelli, no. 1287): LEEGEALBAANA for lege Albana. I 2 . This is the commonest power of the Roman i. It is, however, a representative of A in other cases besides those given above : thus, inter stands for the old antar, ille represents the Sanscrit anya, old Latin ollus, &c. From the examples quoted by Schwartze, das alte jtEgypten, I. pp. 543, sqq., there need be no doubt that the older Romans used E as a representative of r 2 . I 3 . The sound of this letter is indicated by a passage in Velius Longus (p. 2235, Putsch): " Unde fit, ut seepe aliud scribamus, aliud enuntiemus, sicut supra (p. 2219) locutus sum de v iro et virtute, ubi i scribitur et paene v enuntiatur ; unde Ti. Claudius novam quandam litteram excogitavit, similem ei notae, quam pro aspiratione Grseci ponunt, per quam scriberentur eae voces, quaa neque secundum exilitatem litteras i, neque secun- dum pinguitudinem litteraa v sonant, ut in viro et virtute, neque rursus secundum latum litterse sonum enuntiarentur, ut in eo quod est legere, scribere" From this passage we learn that i before R was pronounced somewhat like u, as is the case with us ; and we also draw the important inference that legere and scribere must have been pronounced lire and scrire. In augur and the proper name Spurius this pronunciation seems to be ex- pressed by the vowel u. The latter is a derivative from super, and is equivalent in meaning to Superbus (above, p. 26) ; the former is a derivative from avi-gero, as may be proved by a curious analogy between the derivatives of avis, " a bird," and ce-s, " a weight or burden." For as cedi-ti-mus means a person who is conversant with a temple (Fest. p. 13 = cedis intimus), so avitimus would mean " conversant with birds," ces-timus, 264 ANALYSIS OF THE LATIN ALPHABET. [On. VII. " conversant with weights ;" hence, as augury and weighing were the two most usual means of forming a judgment, both au-tumo and ces-tumo signified " to judge." Comp. the use of con-templor, con-sidero. Again, as ce-ger signifies " bearing a burden," or "burdened," and ne-ger, "not able to bear," or " weak" (Fest. p. 165, s. v. ne-gritu[do]\ so augur would mean " bearing a bird," or " dealing with birds" (belli-ger, &c.) : com p. au-spex, &c. On the proper orthography of Virgilius or Vergilius the student will find the principal authorities in Wag- ner's Virgil, Vol. V. p. 479. The existence of such a short vowel as i 3 is necessary for the explanation of those forms in which i appears to be lighter than E. Thus, from lego, rego, teneo, we have col-ligo, di-rigo, re-tineo ; and the i thus introduced is so short, that it is omitted altogether in some compounds of rego, as per[r]-go, sur\r]-go. In the rustic pronunciation of the Italians i was frequently drop- ped (as in ame, from animus), and the E, on the other hand, was lengthened improperly ; see Cic. de Orat. III. 12, $ 46 : " Quare Cotta noster, cujus tu ilia lata, Sulpici, nonnumquam imitaris, ut iota litteram tollas, et E plenissimum dicas, non mihi oratores antiques, sed messores videtur imitari." U r The interchange of the diphthong oi = oe with this value of u is of constant occurrence. Thus we have oinos, unus; moenus, munus ; &c. ; and in Boeotian Greek e^v for e/uo/ (Apol- lon. de Pronom. p. 364). The observation of some of these changes leads to interesting etymologies ; as, for instance, in the case of the word prcelium, formerly written proilium (see Mure- tus, Var. Lect. VI. 4). The Greeks, like the Highlanders of Scotland, placed their best-armed soldiers in the first line, and by these the battle was begun and generally decided. Hence these or oVXtrat were called Tr^uXe'es, which is interpreted (see Hermann, Opusc. IV. p. 289 ; Miiller, Dor. III. 12, ^ 10), and is undoubtedly another form of 7rpoi\ees ; and hence the skirmish or battle between the van of the two armies was termed 7rpo-i\iov or prodium. This etymology is confirmed by the obvious derivation of milites. The Greek language ex- pressed large numbers in terms derived from common objects : thus, x/Xtot, " a thousand," is connected with ^t\o?, " a heap of fodder," from ^/o>, "to scatter abroad;" and vvpioi, "ten thou- sand," with nvpa), " to pour forth water." Similarly, the Latin $ 6.] ANALYSIS OF THE LATIN ALPHABET. 265 m-ile, " a thousand," means only " a large number," " a crowd" (ofjL-i\ia) ; and m-il-ites are " those who march in a large body" (compare pari-etes, " those which go round," scil. the house), i. e* " the common soldiers" (cf. above, p. 25). So that we have three classes of warriors : (1) the Tr^jAe'e?, i. e. Tr^o-iXe'e? or tjpcocs, "the choice troops, who fought in the van ;" (2) the [ha]m-ilites, or, " common soldiers, who marched in a body ;" (3) the equ-ites, or " cavalry, who went on horseback." The rorarii seem to have derived their name from the idea of spreading out or pouring forth, which is conveyed by ^iXioi and /uivpioi, and not from the fanciful resemblance of slight drops before a heavy shower. In the same way as the diphthong AI becomes i v the diph- thong AU becomes U L : comp. causa, ac-cuso ; claudo, in-cludo ; &c. The same is the case with the Greek diphthong ov, Oov- KviiStfs 9 ThucydideS) &c. ; and even with its Latin equivalent ou 9 thus we have indouco for induco on the bronze table of Tivoli (above, Chap. VI. 19). The diphthong AU is sometimes represented by 6 = au, as in Sanscrit : comp. plaudo, ex-plodo ; Claudius^ Clodius ; &c. In ob-oedio, from audio, AU is repre- sented by the lighter diphthong 01 ; and it is a further proof of the tendency to interchange u x and i lf that the diphthong 01 = OE, which is so often represented by u 1? also appears as 1^1 thus, oiconomus is written iconomus, oSotSoKos appears as hodido- cus 9 QivoiJiaos as Inomaus, Koi^riTripiov as cimeterium, &c. Sometimes, on the contrary, OE is represented by the first vowel only, as in diocesis, poema, &c., from Sioucijffi?, Troika, &c. (see Gifanius, in Mureti Opp. I. p. 550, Ruhnken.). With regard to Troiew, the omission of the t was common enough in Greek (see Person, Tracts, p. 63 ; Dindorf, ad Arist. Nub. 1448, Acharn. 410). The pronunciation of yi = w, as in Ilithyia = EiXeiOuia, is best explained on the hypothesis that the y - v became eva- nescent, just as the a in ai and au is omitted in the derived forms, for yi = vi is certainly pronounced with a single utterance. That ui may be shortened to i is clear from the forms posit for posuit (Orelli, C. I. nos. 71, 1732, 1475, 3087, 4139), tis for tuis (Id. no. 4847), sis for suis (Lucr. III. 1038 ; Y. 1076. Fest. s. v. sos). In the same way uu is shortened into u (Orelli, nos. 1108, 3488) and ii into i (Gruter, p. DLXXIIL, and cf. all the genitives of nouns in ius). U 2 . This is the common short u of the Romans. It corre- 266 ANALYSIS OF THE LATIN ALPHABET. [Cn. VII. spends generally to the short o of the Greeks ; and nouns of the o-declension always exhibit this u in Latin : comp. Xv/co?, lupus; '/TTTTOS, equus ; &c. It is probably a remnant of the Etruscan u. In the older Latin inscriptions we have seen o used for this value of u. Thus we have consol for consul, Luciom for Lucium, &c. U 3 . This letter, like I 3 , must be considered as a point of contact between i and u. Indeed, it may be doubtful in some cases whether u 3 has not been written for i 3 . The passage of this u 3 into an approximate i is of the following nature : First, a short o is changed into u 2 . The genitive of the Greek im- parisyllabic declension ends in -o? : for this the oldest Latin substitutes -us, as in Castorus, nominus, &c. compared with Senatuos, &c. Some of these old genitives remained to the end of the language, as alms, ejus, hujus, illius, &c. Again, the 1st pers. plur. of the Greek verb ended in -O/JLCV = -o^e? : for this the old Romans wrote -umus, a form still preserved in sumus and volumus. Again, in old Latin the vowel of the crude form is preserved in the inflexions, as in arcu-bus, optu-mus, pontu- fex, &c. But in all three cases the later Latin exhibits an i : thus we have Castoris, nominis, &c. ; dicimus, scribimus, &c. ; arcibus, optimus, pontifex, &c. In these cases we observe that u = o passes into a simple i. But there are other instances in which the transition seems to go still farther. As the reduplica- tion-syllable is generally shorter than the root-syllable in the preterite of verbs, we should expect that the u or o in the first syllable of cu-curri, mo-mordi, pu-pugi, tu-tudi, would be an approximation to Ug. 1 Then, again, in cultus, culmen, &c. from colo, columen, &c., the u is clearly less significant than o, though the u here may have been partly occasioned by that affinity between u and I of which the French furnishes so many ex- amples, and which we also see in the transition from the Greek 'AovcXifTTtos, 'H|oa/cXf;s to the Latin ^Esculapius, Hercules. But there are some cases in which we conclude that the u, which is written, has less weight even than i. This might be inferred from con-culco, the secondary form of cako, which, according to 1 The older writers wrote memordi,peposci,pepugi, spepondi, according to Gellius, N. A. VII. 9, who, however, says of the common spelling, " ita mine omnes ferme doctiores hujusmodi verbis utuntur." 5.] ANALYSIS OF THE LATIN ALPHABET. 267 the above table, should be either con-cilco or con-celco; and also from difficultas> sepultus, derived from difficilis and sepelio. The fact seems to be, that what would be i before R, becomes u 3 before L; so that u 3 , I 3 , are both ultimate forms of their re- spective vowels, and as such are in a state of convergence. Accordingly, if we should seek to arrange the Latin vowels in regard to their comparative weight, we should, as the result of this inquiry, have the following order : A (as in musd, &c.) ; U 1? l t ; A; 0, U 8 , I 2 ; E ; U 3 , I 3 . f 6. The Greek Letters used by the Romans. The Greek letters subsequently employed by the Romans were z, K, and Y. The period at which the first of these was introduced is doubtful ; for while, on the one hand, we are told that z is found in the Salian songs (Velius Longus, p. 2217 : " Mihi videtur nee aliena sermoni fuisse z littera, cum inveniatur in carmine Saliari"), on the other hand, we find that, even in words borrowed from the Greek, this letter is represented by di, as in Sabadius for 2e/3ao9 (Apulei Met. VIII. 170), judai- diareim judaizare (Commodian, Instruct, adv. Gent. c. XXXVII. 634), trapedia for trapeza (Auctor. Rei. Agrar. p. 248), schidia for schiza, oridia for oriza, &c. (vide Schneid. Elementarl. I. p. 386 ; and Lobeck, Aglaoph. p. 296, note /.) The fact seems to be, that the Romans had two diiferent characters to express the two different values of the Greek z, which was a dental, either assibilated (as 0-$), or softened (as Sy). Now, in its latter use it becomes equivalent to the softened guttural ; for the dental and guttural, when combined with y, which is the ultimate vocalisation of the gutturals, converge in the sound of our j or sh (New Crat. $ 112, 216). When, therefore, the Greek z more nearly approximates to the sound cr, either this is pre- served in the Latin transcriptions, as in Mesdentius, Sdepherus, for Mezentius, Zephyrus (Max. Victor, p. 1945) ; or the $ is assimilated to the , K(t)fj.d, paXaJCt0, &c. ; or else one or other of the two component parts is omitted, as in Saguntus for Zakyn- thus, or Medentius for Mezentius. In this case, too, we may consider that the letter x occasionally steps in, as in rixa by the side of epi[$\s. When, however, the Greek z is a softened 268 ANALYSIS OF THE LATIN ALPHABET. [On. VII. $, and therefore equivalent to a softened guttural, we find that it is represented either by the full combination di, as in the cases quoted above, or else by the vocalised guttural (j) only. Of this latter substitution there are numberless instances : such as, Ju-piter, Zei)s irarrip ; jugum, fyuyos ; &c. Of these the most important are the cases connected with the first-quoted example, Ju-piter Dies-pater ; and I must take this oppor- tunity of returning to one etymology belonging to this class, which has always appeared to me to open the way to a chain of the most interesting associations. It has been shown elsewhere (N. Crat. 116) how the Greek H, originally the mark of aspiration, came to be used as a sign for the -long e. Out of that investigation it appeared (1) that a short vowel aspirated may be equivalent to an un- aspirated long vowel ; (2) that the vocalised consonants i and u may change their place ; (3) that these vocalised consonants may be absorbed into or represented by the long vowel only. To the instances given there, I will now add the iota subscriptum of the Greek dative, and the Ionic Greek absorption of v after w, as in OwvjuLa, ewvrov, &C. 1 These principles explain the con- nexion between qirap, jecur (Sanscr. yakrit); jjfjuav, Sidnecros, dimidius ; and between j^epa = Sidnepos, and dies 2 (comp. diu- turnus, juturna ; Diana, Janus, &c.). Now, besides q/mepa, we have an adjective ^e^oos, " civilised," " cultivated," &c., the re- gular antithesis of aypw, and it has been suggested (ibid. 150), that this word was originally applied to a country through which there was a road or passage, a country divided by a road (Staimepos) ; just as aypios was properly applied to a rude, open country, with nothing but aypot*. This is sufficiently 1 In many editions of Herodotus we have these words written eovroO, &c. ; but the accentuation of tiavpa sufficiently proves that it is a dissyllable j and even if we had not this evidence, it would be contrary to all analogy to infer a resolution of a diphthong in a crasis, the sole object of which is to shorten the word. Why should rdvr6 be written, if it were a word of as many syllables as TO avro ? 2 In the name of the city 'Ipepa (another form of jpepa, see Bb'ckh's note on Pindar, O. XII. 13-21, p. 210), the preposition did is represented by the aspirated i. In the words anti-quus, posti-cus, from antea, postea, we have l = ed = eai. 3 Hence xupos with its old synonym xP* ( New ^ rat - 28 ) ma y be considered as an adjective agreeing with the suppressed word aypos, just $ 6.] ANALYSIS OF THE LATIN ALPHABET. 269 proved by ^Esch. Eumen. 13, 14: KeXevOoiroioi Trainee ' TOV, ^Oova dvri/uLepov TiOevTes tj/jLepcojuLevrjv. Find. Isthm. III. 76 (IV. 97) : pdimXiaurl re TropOfjiOv d/mepcocraTo. Herod. I. 126 : evOavTo. o Ki^os (YIV "V 01 ) \***ps ctKavQworis ) TOVTOV a(pi TOV J^topOV 7TjOOei7T6 J~r]IULp(x)(TCll GV rjfJLpa. IV. 118 I TOfJ CU6* e/uLTToScov yivoiuiGi'ovs qimepovTcu Trai/ra?. In all of these passages the verb rjfiepow implies making a clear passage or road ; and in Plato (Legg. p. 761 A.) the adjective rj^po^ is used as a predi- cate of o^os 1 : o^tov TG G7rL(j.G\ov[jLevovs, OTTO)? tos y/uepwTaTai eKacrrai yiyvwvTai 1 . That the Greeks connected road-making with civilisation in general, and with the peaceful commerce of man with man, appears from many passages (Aristotle, Trepl Oav/macricov o,Kovcrf^a.Ta)v, c. 85, p. 837, Bekk.; Thucydides, I. 2, compared with I. 13, &c.) ; and this is generally implied in all the legends relating to Hercules and Theseus. But it has not been sufficiently remarked that this road-making was also in- timately connected with the cultivation of land. It may, how- ever, be shown, that as the Greek aypos becomes ramepos when divided by a road, by a similar process the Latin ager becomes jugerum di-ager-um. Whenever a piece of unemployed ground of ager, so called was to be taken into use, whether for cultivation, or for the site of a city or a camp, the rules of the ancient limi- tatio were immediately applied. Now this very word limi- tatio signifies, the dividing of a certain piece of ground into main-roads (vice) and cross-roads (limites); and the same pri- mary notion is conveyed by tern-plum, so obviously derived from tern-no, Gr. TCL^-VW, comp. re'/u^os 1 * &c. For in all limitation the first thing done was to observe the templum, i. e. as we should say, to take the bearing by the compass 2 . Suppose the as x<*pa refers to the suppressed word y^ : and thus x<*P os signifies " land not built on" cither the open space in a town, or fields in the country (Herod. II. 154: di'Soxri x^povs eVoi/^o-ai), and x^P a rather signifies "a region," " a territory," in the wider sense. 1 The word rj7rcipos = ^ Siairepav x^P a > furnishes another instance of the substitution of 77 for did: comp. the epithet StaTrpvo-to?, Find. N. IV. 51, where see the note. 2 Most ancient nations seem to have connected the regiones coeli with the regiones marwm. Thus in old English " the milky way " was called " Watling-street," which was the name of one of the four great roads in this country ; see Grimm, Deutsche Myth. p. 330, 2d ed. 270 ANALYSIS OF THE LATIN ALPHABET. [On. VII. augur stood with his back to the north, then the line from north to south would be called the cardo, as corresponding to the axis of the globe ; and that from east to west, which cut the cardo at right angles, would be called the decumanus, or " tenth line." For both these lines repeated themselves according to the number of separate allotments into which the land was divided, or the number of separate streets in the city or camp 1 . Now the Roman actus or fundus [120 feet] was the unit of sub- division; two of these fundi made a jugerum = di-ager-um, and two jugera constituted the heredium of a Roman patrician : con- sequently, 200 jugera made up the ager limitatus of a century of the old Roman populus (Fest. s. v. Centuriatus, p. 53). If this ager limitatus, then, were arranged as a square, we have, of course, for each side 20 x 120 feet. Supposing, then, a road between each two of the fundi, which there must have been, as every two fundi made a di-ager-um, the cardo which passed between the tenth and eleventh fundus would be properly called the decumanus, and it would consequently be the main road, and would be terminated by the main gate (porta decu- mand). The point at which the decumanus crossed the cardo was called groma or gruma ; and here, in a city or camp, the two cross-roads seem to have spread themselves out into a kind of forum. There is as much probability in the supposition that the immortal name of Rome was derived from this ancient word, as there is in any of the numerous etymologies suggested by Festus (p. 266). From this it appears, that among the Romans it was the same thing to speak of a territory as divided by roads, and to call it cultivated, occupied, or built upon ; and the jugerum, or divided ager, implied both. To the same principle 1 It would seem that the word sicilicus (from seco) was properly and originally applied to this apportionment of land. In the Bantine Table (1. 25) we have nep him pruhipid mais zicolois x nesimois ; which I have translated (above, p. 127) : ne in hoc prcehibeat (i.e. prcebeai) plus sicilicis x contiguis. According to Klenze (Abhandl. p. 60) x nesimois = decimis; but I cannot understand why we should have an ordinal here. The root of ne-simus appears in nahe, near, next, &c. ; and I would understand it of so many adjoining allotments. The sicilicus was 600 square feet, i. e. ^ of the jugerum, or ^ of the actus. Consequently, the 30 contiguous sicilici mentioned in 1. 17 would be f- of the jugerum, or f of the actus; and the ten contiguous sicilici would, therefore be ^ of the former and n of the latter. 6.] ANALYSIS OF THE LATIN ALPHABET. 271 we may refer the importance attached by the ancients to straight ploughing l ; for the furrow was the first element of the road ; and the urbs itself was only that space round which the plough had been formally and solemnly drawn. The Romans were very sparing in their use of the Greek letter K. It was occasionally employed to form the syllable ka, as in kalumnia, kandidatus, kaput, Karthago, Kastor, evoka- tus, judikandus, Parkarum ; but in these instances it was con- sidered quite superfluous ; and Quintilian thinks (I. 4, 9, and 7, 10) that its use ought to be restricted to those cases in which it serves as the conventional mark of an abbrevation, as in K. = Kceso, and K. or Kal. = Kalendce. Isidor (Origg. 1, 4) and Petrus Diaconus (p. 1582, Putsch) tell us that the letter K was added to the Roman alphabet by the ludi-magister Sallustius, in order to mark a distinction between K and Q. It occurs in the oldest Latin inscription which has come down to us (above, p. 220) in the Greek word Kastorus, and was probably suggested by an increased intercourse with the Greek colonies of southern Italy long before Sp. Carvilius introduced the distinction between c and G. The letter Y was never used by the Romans except as the transcription of v in words derived either from or through the Greek ; and it seems to have been a representative of those sounds which have been designated above by the characters v t and u 3 , both of which involve an approximation to the sound of i. Hence, in the French alphabet it is not improperly called " the Greek i" (i grec). In many words, rather connected with the Greek than derived from it, the v is represented by i, as in cliens, in-clitus (K\VW), clipeus (KPVTTTW), silva (vXFrj), &c. ; while in others the v has become E, as in socer (cfcvpoi), remulco (pvfjiov\Kea)), polenta (TraXvvTij), &c. The Roman U 2 sometimes represents the common v of the Greeks, as in lupus (Xv/cos), nunc (yvi), fui (0i/co), &c. ; sometimes the Greek o, as in all nouns of the o-declension. i See Hesiod. Op. et D. 443 : 05 K epyov fjLcXerwv Welav avXait fj.r]KTi irairraivav jicff 6}irj\iK.as. Luke ix. 62 ; and corap. the tropical use of delirare. 272 ANALYSIS OF THE LATIN ALPHABET. [On. VII. 7. The Numeral Signs. This examination of the Latin alphabet will not be complete without some remarks on the signs which were used by the Romans to denote the numeral adjectives. Priscian, in his usual school-boy way, has endeavoured to establish the connexion between the numeral signs as we have them, and the ordinary Roman capitals. Thus, quinque, he tells us, is represented by V, because this is the fifth vowel ; quinquaginta is L, because, etymologically, L and N may be interchanged, and N is TTCVT^- Kovra in Greek ; quingenti is D, because this is the next letter to C ! and so forth (Priscian, II. p. 388, ed. Krehl). Now there can be no doubt that the Roman numeral signs are derived from the Tuscans ; though in certain cases a Roman capital has been substituted for an Etruscan character which does not correspond to it in value, and though in these instances the figures are either inclined or reversed. The Etruscan cha- racters are as follows: I, II, III, IlII, A, AI, All, AIII, IX, X, &c. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10. XX, XXX, XXXX, or XT T TX, &a 20, 30, 40, 50, 60, , 8, ]j)> <0> &c. 100, 1000, 5000, 10000. It is sufficiently obvious that the first ten of these characters are identical with the Roman figures, the A, &c. being reversed ; and as T> is often written T, and as vb X, frequently occur on Roman family coins, we may recognise in this character the original of the Roman L, and therefore identify the Etruscan and Roman ciphers from 1 to 99. The Roman C and the Etruscan do not appear to be connected ; but the Etruscan 8, or, as it is also written Q, is clearly the same as the Roman I-H, Q), and do, for which M was subsequently written ; and the same remark applies to the still higher numbers. If, then, the Roman ciphers were derived from the Tuscans it is obvious that we must seek in the Tuscan language for an interpretation. Now it cannot be doubted that the Tuscan numeral signs are either letters of the alphabet slightly changed, 7.] ANALYSIS OF THE LATIN ALPHABET. 273 or combinations of such characters made according to fixed rules. Thus, A is the inverted V = u; T or T is an inverted ^ = ch l ; and 8 =/ Since, therefore, the position of these letters in the organic alphabet does not correspond to their value as numeral signs, we must conclude that they represent the initials of the numerals in the Etruscan, just as M afterwards denoted mille in the Latin language. We do not know any Etruscan numeral, and therefore cannot pretend to any certainty on this subject ; but this is the most probable inference. The manner in which the elementary signs are combined to form the intermediate numerals is more easily and safely investigated. The character denoting unity is perhaps selected from its simplicity ; it is the natural and obvious score in every country. This character is combined with itself to form the next three digits, though four is sometimes expressed as 51, according to the principle of sub- traction so common among the Romans (comp. duodeviginti, &c.). The same plan is adopted to form the numerals between 5 and 10. The number 10 is represented by a combination of two V's thus, X ; and this figure enclosed in a circle indicates the multiplication of 10 by itself, or 100. The letter 8, or 0, being assumed as the representative of 1000, its half,- or D, would indicate 500 ; and as multiplication by ten was indicated by a circle in the case of 100, on the same principle gjj) would be 10,000, and its half or ]j) would represent 5000. These rules for the formation of one numeral from another are more obvious than the origin of the elementary numeral signs. But where certainty is not within our reach, we must be contented with a solution of those difficulties which may be sub- mitted with safety to a philological analysis. 1 It is possible that this character may be the half of that which denotes 100, according to the principle stated below. 18 CHAPTER VIII. THE LATIN CASE-SYSTEM. 1. Fulness and deficiencies of the Latin case-system. 2. General scheme of the case-endings. 3. Differences of crude form. 4. Hypothetical forms of the nominative and accusative plural. 5. Existing forms the genitive. 6. The dative and locative. 7- The accusative singular. 8. The ablative. 9. The neuter forms. 10. The vocative. 11. Adverbs considered as cases of nouns. 12. Adverbial expression for the day of the month. 1. Fulness and deficiencies of the Latin case- system. rjlHE system of cases, with which the Latin noun is furnished, JL presents a greater abundance and variety of forms than that of the Greek declension. The Greek noun has no distinct ablative case ; its accusative has frequently lost its characteristic termina- tion ; the genitive includes the ablative meaning ; and the loca- tive is almost obsolete. The greater number and variety of the Latin cases is due to the more ancient state or condition of the language, and perhaps also to its composite structure. As the language degenerates into the so-called Romance idioms, we find that its cases are gradually lost, and their place taken by a number of prefixes, which add indeed to the syntactical distinct- ness of the language, but purchase this advantage by sacrificing the etymological development. The student of Latin, however, very soon discovers that the variety of case-forms is the very reverse of an advantage. For idiomatic usage has introduced so much confusion into the use of the genitive, dative, "and ablative, that the two latter derive all their distinctions from the preposi- tions attached to the ablative, while the genitive, in many cases, differs from the ablative only as an arbitrary form, and without any reference to a distinction of meaning. If we revert to the Greek language, which still retains the more a'ccurate distinctions of case, we shall see that the genitive, or case of ablation, denotes the origin of motion or action ; the dative, or case of accession, denotes juxta-position, immediate proximity, rest and presence ; the accusative, or case of transition, denotes the end of motion or action, the object to which something is proceeding. Now the Latin, in most instances, is unable to express this simple relation of unde, ubi, and quo by the mere case-endings. If we except certain adverbs derived from nouns, certain agglutinate .!.] THE LATIN CASE-SYSTEM. 275 forms, such as meridie, postridie, &c., some few nouns, as rus, domus, humus, bellum, militia, and the proper names of cities, we have no locative in Latin, and no case for the simple expres- sion of departure or approach, and are obliged to use prepositions, such as in, ab, ad, to convey these meanings. And even with regard to the forms which are still used as locatives, differences of declension produce endless confusions, which all the old and some modern grammarians have enhanced by making arbitrary rules for differences of case in the syntax of different declensions. Thus because nouns in -a, -us, of the first and second declension, had a locative in -a-i = ce, and in -o-i = i, we are told that mili- tia, Romce, domi, Cypri are genitive cases ; whereas ruri, Carthagine, Athenis are ablatives, because the locative approxi- mates or corresponds to the mutilated ablative in the consonantal declension. These labourers in the work of making the Latin language unlearnable, except by the parrot use of the memory, could not perceive that as dies is masculine when it means " a day,"' ho-die and postrl-die must belong to the same forms, and that if the former is from ho-i-die, the latter must be from postero-i-die. The fact is that the locative originally ended in -in or -im, and this was corrupted in every form with the ex- ception of such words as partim, enim, &c. ; hence, to restore the original ending, we must write, with different amounts of alteration or addition, militia-im (-in), Roma-im (-in), domo-im (-in), Cypro-im (-in), rur-im (-in), Carthagin-im (-in), Athenis- im (-in). 2. General scheme of the case-endings. In treating of the Latin cases, our attention is directed to three different aspects under which they may be considered. We may regard them either according to a general scheme de- rived from all the declensions, or as modified by those varieties in the termination of the crude form which constitute differences of declension ; or we may take both of these together, and add to them those additional phenomena which are furnished by the adverb. A supplementary source of information respecting the cases may be derived from those nouns, whether substantive or adjective, which are obviously formed from the oblique cases of other nouns. Thus, we know that the original Greek genitive ended in -cno (Sanscr. sya) from the form of the possessive ad- 182 276 THE LATIN CASE-SYSTEM. [On. VIII. jective Swxoo-co? (Bopp, Vergl Gramm. p. 294, note). Simi- larly, a case in -ine, analogous to the Sanscrit instrumental, may be inferred both from the particle sine and from the derivative forms urbdnus (= urbainus), &c., and officma (= officTma), &c. If we confine ourselves to the forms of the noun, we get the following general scheme of the case-endings. SING. PLUR. TVT~~ (sometimes absorbed, assimilated. r -\ IMOm. 5 V or dropt by V i sarga h) [$J 6S (variously modified) Gen. is, JUS, Sis (originally -siom) [r~]um (originally siom-s) Dat. i or bi Accus. m Abl. a[d~\ (the d is found only in old Latin) [b] US IS Loc. i[m~\ or i\n\ is- [im~\ or is- \in\. $ 3. Differences of crude form. By taking the different crude forms according to the usual classification, we shall at once see how this scheme is modified and applied. The declensions will be fully discussed in a sepa- rate chapter, and it will be sufficient in this place to show how the different cases attach themselves to the different characteristics. CONSONANT-NOUNS. SING. PLUR. Norn. lapi[d'\s lapid-[s]-es (= es) Gen. lapid-is lapid-e-rum l Dat. lapid-i-\bf\ (= t) lapid-i-bus Accus. lapid-e-m lapid-e[m\s (- es) Abl. lapid-e[d~\ lapid-i-bus Loc. lapid-im? lapid-is-im? VOWEL-NOUNS. A SING. PLUR. Nom. familiars'] familia-[ses] (= ai, ce) Gen. familia-is (= as, ai, ce) familia-rum Dat. familia-[l>]i (= ce) familia-bus (- is) 2 1 Charisius, I. 40. 2 For the form in -bus comp. Orelli, Inscr. nos. 1628, 1629, 4601, &c.; and K. L. Schneider, Formenlehre, I. pp. 25, sqq. $3.] THE LATIN CASE-SYSTEM. 277 Accus. familia-m Abl. familia-[d] (= a Loc. familia-i (= ce) familia-i-bus familia-is-im ? E = A-I SING. Norn, die-s - dia-is Gen. die-i[s] 1 Dat. dic-[6]t Accus. efo'e-[m] Abl. die-Id] Loc. die = To-i is dative and instrumental in ti-bi 9 vo-bis, but simply local in u-bi, i-bi, &c. Commonly the Latin locative ends in -i, agreeing in this with the Sanscrit. But when the characteristic of the noun is a consonant, it is generally shortened into e, especially if the word is of more than two syl- lables. The locative of rus is ruri. In the plural the dative 6.] THE LATIN CASE-SYSTEM. 283 and locative are always confused with the ablative ; and instances occur even in classical Latin where the dative of an ordinary noun, with the sense of limitation, appears in the form of the ablative in e. In some phrases this is rather the rule than the exception; such are pignore dare, for pignori; lllviri auro argento cere flando feriundo, for ceri; jure dicundo for juri ; qui dant quique accipiunt foenore, farfoenori ; &c. (see Schneider, Lat. Gr. II. pp. 200, sqq. ; Muller, ad Varro. L. L. V. p. 16). If there is any reason for using the term dativus in reference to the case of a noun, it must surely be applicable to morte in the epitaph of Plautus, quoted by Gellius (N. A. I. 84) : Postquam est morte datus Plautus, Comoedia luget, for here the form in -e actually follows a verb of giving. Thus we see that ore is not the ablative but the dative in (Virgil, Georg. I. 430) : si virgmeum suffuderit ore ruborem; and that it is a locative in (Georg. III. 439) : linguis micat ore trisulcis. 7. The Accusative Singular. The m, which marks the accusative singular in Latin and Sanscrit, is only a weaker form of the dental v, which appears in Greek. This dental is the residuum of the third pronominal element, and denotes distance and objectivity. We are not to suppose that partem and partim are the same word, or generally that the accusative and locative are the same form. The i which appears in the latter, with or without the accusative affix, con- stitutes the essential difference between the two cases. Belonging to the second pronominal element, this i is in itself an expression of proximity ; and thus, while parte-m denotes that " the part" is an object to be approached or acted on, part-i-m indicates that not only is the part an object, but also that it is close at hand for use or superposition. It is true that the temporal particles quum, turn, nun-c, jam, &c., are not less locative in meaning than olim, and that the causal nam, though accusative in form, coincides in signification with the locative enim. But we must remember that quod, quod si, quippe = quia-pe, on, ore, are, &c. are used as general expressions of objectivity ; and we must not allow syntactical equivalences to interfere with our etymological discrimination. 284 THE LATIN CASE-SYSTEM. [Cn. VIII. 8. The Ablative. In ordinary Latin the ablative is used as the case of instru- mentality in both numbers; and in the plural there is no dis- tinction between it and the dative. The specimens of old Latin in Chapter VI. have sufficiently shown that the termination of the ablative was d, or, perhaps, at one period of the language, t. The instrumental ending in Sanscrit is, as we have seen, -ina ; and the Sanscrit ablative ended, like the Latin, in -d. The tendency of the instrumental and ablative the case of proximity and the case of derivation, to interchange their significations, is a phe- nomenon, in which the philosophical grammarian finds no difficulty. The fact that sine and sed are so nearly synonymous is an obvious exemplification of this tendency. It is a more serious imperfection of the Latin case-system that the ablative, though distinguished in form from the genitive, should sometimes agree with it in meaning, and sometimes coincide in sense with its direct opposite the dative. With regard to the singular number, which has an ablative properly so called, there can be no doubt that in Latin and Sanscrit, as well as in Greek, the genitive and ablative are traceable to a common origin. The full, original, and proper form of the genitive singular was -sion, and this in Greek often appeared as -9ev : cf. 0eos = cnoV In Sanscrit the ablative vrik&t bears the same relation to the genitive vrikasya that the genitive 7roAeo>s does to a more ancient TroAto'crioy, or the adverb KO\WS to an original KaXo-Oev, or the common TVTTT( 19 to the inevitably assumed Tvirre-cri. It is well known that the Latin adverbs in ~tus correspond to the Greek in -Qev\ thus cceli-tus = ovpavo-9ev ; and the Greek termination $- in -5i/s, &c. involves this ending -0ev (New Crat. 263). There is there- fore every reason to believe that the Latin ablative in -d or -t is an apocopated form of a case in -dus or -tus, which is resolvable to an ultimate identity with the genitive. 9. The Neuter Forms. The neuter accusative, which serves also as a nominative (see New Crat. 236), ends, like the usual accusative, in -ra in all nouns of the vowel-declensions. There is no doubt, however, that this m may be traced back through the dental liquid n, which represents it in Greek, to the dental mute ~d or -t. Thus 9.] THE LATIN CASE-SYSTEM. 285 we have i-d, illu-d, quo-d, &c. to the latest period of the lan- guage ; we have also met, tet, set, or med, ted, sed ; ego-met, me-met, ted-ipsum, inter sed (Senat. Consult, de Bacch. 11. 13, 14) ; and we shall see in the next chapter that the final s or r, in nouns like corpu-s, robo-r, genu-s^ &c., is a softening of an original t or d. We must take care not to confuse this t or d with the same letter appearing as the affix of the ablative. The long vowel, which precedes the dental in that case, shows that there is apocope or absorption of something more than a mere consonant, and abundant reason has been given for the inference that this d has passed through ih from an original sibilant repre- senting the second pronominal element. On the contrary, the accusative m, n, d or t is merely the residuum of the third pro- nominal element, denoting simple objectivity. The forms of the neuter-plural show, a fortiori, that the dental affix in the singular was a mere letter, and not a syllable, as in the case of the ablative. For all neuter nouns, to whatever declension they belong, form their plural nominative-accusative in a in the Zend and in the old European languages of this family. Now the Greek language shows us that n, when it stands by itself at the end of a word, or precedes a dental mute, may be changed into a, and this vowel may even represent the combination -i/r. Thus we have Trarepa for irarepv, rerJ^arai for Terv(pvrai, GW- fyiaTo for aco^oivTo, 7rd0os for 7r6vOosi and even SGKO, for $e/ce/>T, and crcofjia for GCO^GVT. There is therefore no objection, a priori, to the hypothesis, but rather a presumption, that the plural -a represents an original -vr\ and it seems quite reasonable to assume that v\a ~ ^uXevr ; for if the objective v or r of the singular had to be extended into a plural, we should not in this case append the personal or subjective s, as in the case of mas- culine and feminine nouns, but should rather repeat the objective affix. Now it is known that the neuter plural in Latin originally ended in -d ; thus we find in the Senatus Consult, de Bacch. 1. 24 : quei advorsum ea-d fecisent. Again, we find in Sanscrit that neuter plurals end in -ni ; thus madhu = peGu makes mad- hu-ni=/ue9v-a', and the final i must be a vocalization of a second n, just as conversely nn is substituted for ni in %ewos = eVo9 = fe?i>o?. Lastly, while the Erse plural of the third personal pronoun is sidd for swiad, the Welsh form of the plural is hwynt for swynt. Putting all these facts together, we must 286 THE LATIN CASE-SYSTEM. [Cn. VIII. come to the conclusion that the neuter accusative singular ended in -m -n = -t or -d, and that the plural a represents an original -nd = -nt = -nn or -mm. The pronominal neuters in ae, as quce, Jicec, &c., are ex- plained in a subsequent chapter. $10. The Vocative. The vocative, i. e. the case of allocution, exhortation, or ex- clamation, is not distinguished from the nominative except in nouns of the second declension, and in certain Greek words adopted by the classical writers. When a noun in -us has to be used in the vocative, the crude form is employed with the lightest substitution for the characteristic vowel. Thus dominus makes domine. If i precedes the characteristic, the vocative e is ab- sorbed, and filius makes fill -file. The same is the case with meus which has for its vocative mi = mee. As the regular nomina- tive plural of deus is di, the Romans, to avoid confusion, did not use a vocative dee = di. This rule does not apply to adjectives, as Cynthte from Cynthius, SpercJue from Spercliius. The vo- cative Cat exposes the common error of pronouncing the dactyl Cams as a trochee ; for if this had been true the vocative must have been Cai-e. In point of fact, Caius is scanned regularly in three syllables ; thus we have (Martial, IX. Ep. 93) : v. 4. Pervigil in pluma Coins, ecce, jacet. v. 7. Quod debes, Cat, redde, inquit Phoebus. v. 10. Caius et mallet verbera mille pati. v. 12. Non mavis quam ter Cams esse tuus. Although the vocative, as a distinct case, is thus limited to a few forms in the language, the Latin writers give it occa- sionally a very remarkable extension of use. Thus it is made to agree with the nominative tu : as Stemmate quod Tusco ramum millesime ducis, Censoremne tuum vel quod trabeate salutas. (Pers. III. 27, 28). This is regularly the case in the idiomatic use of macte magis aucte (i. e. frugibus et mold) ; thus we have : macte virtute esto, "be increased in virtue" (Hor. I. Serm. II. 31); macte nova virtute puer, " be increased in your young valour" (Virg. ^En. IX. 641). And even in an oblique sentence, as: juberem \te~] macte virtute esse (Liv. II. 12). 11.] THE LATIN CASE-SYSTEM. 287 11. Adverbs considered as Cases of Nouns. If now we add to the observations derived from the actual cases of nouns, the additional phenomena furnished by the ad- verbs, the subject of this chapter will have received all the examination of which it is capable. Adverbs are, properly speaking, certain cases of pronouns and nouns, and under particular circumstances they are deduced from the participles or supines of verbs. Their syntactical use is as secondary predicates, inasmuch as they convey predication only through the verb of the sentence. The Greeks employ their adjectives and participles for this purpose without any additional inflexion ; but the Roman adverbs are always cases, and some- times, if one may use the expression, double or superimposed cases of nominal or pronominal forms. Pronominal adverbs are secondary predicates either of place or of time. The former indicate (a) " locality," in which case they generally exhibit the locative endings -bi and -im or the accusative -m : thus, from the demonstrative is and the relative qui, we have i-bi and u-bi, originally cubi, comp. ali-cubi, &c. ; from iste we have istim, &c. ; and the ending -m appears in us-quam or uspiam, &c. ; (6) " motion towards," in which case they end in -o : as ul-tro, " to a place beyond" (see Doderlein, Syn. u. Etym. III. pp. 105, sqq.); quo, "whither ;" eo, "thither;" &c. ; sometimes -c is appended : thus we have illuc, istuc, by the side of illo, isto ; (c) " motion from," in which case the ending is -nde, or -nee, -nque : thus we have i-nde from is, [c]u-nde from qui, aliu-nde from alius, hi-nc from hi-c, illi-nc from ille, utri-nque from uter ; (d) "the way," in which case we have a feminine ablative in -d agreeing with via understood, as qua, ed, &c. The forms of class (c) deserve some special remark. The comparison of turn with tune shows that the n would have been written m, if the c had not been appended. And the same remark applies to exin-de, hin-c, illin-c, istin-c : for exim occurs in Lucretius, (see Lachmann on III. 161), and Ritschl has claimed illim and istim for the text of Plautus (Rhein. Mus. 1850. pp. 472, sqq.). But this does not interfere with the inference that the accusative and locative m is the re- presentative of an original dental. There can be no doubt that the termination -de is identical with that of the ablative, and, as we have seen, with the termination -tus. Bopp, who was aware 288 THE LATIN CASE-SYSTEM. [On. VIII. of this (Vergl. Gramm. p. 610), proposes to consider the same letter as included in liinc, illinc, istinc, which he regards as cor- ruptions of hindc, illindc, istindc. I should not desire any other proof of the importance of the distinction which I first in- troduced into the analysis of the pronominal elements (New Crat. 130). According to the principle which regulates all combi- nations of these elements, n -f c denotes motion " from the there to the here," and therefore expresses ablation or removal quite as naturally as the affix -de = -tus, which is in fact ultimately referable to the same source (N. Crat. 262). Pronominal adverbs of time generally end in -m, as turn, quum ; in ~nc, -nque, as tu-nc, cu-nque ; or in -ndo, -nquam, as qua-ndoy nu-nquam. Adverbs derived from nouns adjective and substantive either end in e, o, or ter ; or else they are merely adjectives in the neuter objective case. (a) Adverbs in e or o, anciently ending in -ed, or -od, are, in fact, ablative cases of adjectives : thus valde, originally vali- dod ; bene, originally bonod ; cito, originally citod; certe or certo, originally certod, &c., are the ablative cases of validus, bonus, citus, certus, &c. respectively. The Greeks had a large class of adverbs of the same kind ; but in these the final -d of the ablative has been softened down, according to the laws of Hellenism, into an -9 : thus, ovrcos, KaXws, &c. represent the old forms of the ablative, ovroS, /ca\o$, &c. (see N. Crat. $ 249). There are two cases where this $- seems still to exist, '/$->? and 'AtfrpoS-iTrj (Sanscr. Abhrdd-ita) ; and there is one instance in which the metre of Homer will not allow its modern represen- tative to stand, namely, in those passages where with the Latin in -mus). There is yet a third form in which it appears, namely, -tim, which is the termi- nation of a most interesting class of participial adverbs ; for I cannot consent to consider any of them as strictly formed from nouns ; and though the verbs in all cases are not forth- coming, the adverbs themselves prove that they must have existed in part at least. Instances of this class of adverbs are caterva-tim, carp-tim, grada-tim, priva-tim, punc-tim, separa- tim, vica-tim. Compare with these the German participial forms in -ingen, and the Greek participial adverbs in -vSa, -vStjV) -$r)v (N. Crat. 263). The most striking result from a proper appreciation of the origin of adverbs in -tim, is the explanation which it supplies for those adverbs in -ter which are derived from active participles. The termination of the supine is already -tu ; the adverb, therefore, is a locative case of the supine ; for caterva-tim stands to caterva-tus in precisely the same relation as par-tim to pars (par[f]s). Similarly, aman-ter, sapien-ter, &c. are cases of the participles amans, sapiens, &c. ; for the crude forms of these participles already contain the t. Now, if I am right in concluding that these terminations, -6ev, -dhas, -ter, -tus, -tim, &c. are lengthened forms of that dental affix which marks the ablative of the noun, most interesting conclusions 19 290 THE LATIN CASE-SYSTEM. [On. VIII. may be drawn from this respecting the origin of the participle and of the passive person-endings of the Latin verb : for if the dental, which must be added to the noun to form the ablative case or adverb, is already included in the participle, it follows that the crude form of the participle is already an ablative formation. That there is no essential distinction between the terminations -tim and -ter, and that the former is not restricted to participles of the passive formation, is clear from such forms as pede-tentim, &c. In fact, while the -d or -t alone are sufficient to express the ablative and participial relation (as in cupi-dus=cupiens ; the terminations -$ov, -Srjv, by the side of -v$ov, -vStjv ; the participle reri;^)or[-w9] by the side of TVTTTOVT- ; and the adverbs in -tus by the side of those in -nde, both signifying "motion from"= " ablation "), yet we must admit that the strengthened form of the active participle, which contains the liquid as well as the mute dental, is no less ablative than those forms in which the mute appears alone ; for there is no less opposition between i-bi and i-nde from i-s 9 than between avro-Oi and avro-Qev from ai/To-s 1 . The participle, therefore, is an ablative or adverbial formation from a verbal root, expressing that which comes out of the action of a verb, i. e. the manner of it ; and differs only from these adverbs, and from the persons of the verb, in the circum- stance, that it is not an immoveable form, but one which is capable of regular flexion through the whole system of cases (N. Crat. 300, 415). Adverbs, used as conjunctions, are such as jam (from is), enim (Sanscr. ena), ideo, tamen, igitur, &c. These are, in fact, cases of different pronouns. Most of them are of obvious origin : 1 In the text I have merely put together some of the analogies suggested in my former work. The late Mr. Garnett, who was one of the soundest, and, at the same time, most original philologers in this country, hajd arrived at some results which were calculated to confirm and extend these views. In a letter to me (dated 3d May, 1842) he said : " I flatter myself that I can make it appear from a pretty copious induc- tion that the Indo-Germanic present participle is formed upon the abla- tive case of the verbal noun [Sanscrit tupat], in much the same way as the pronoun possessive in Latin, German, &c , is formed upon the geni- tive of the personal. If I am not mistaken, this is calculated to throw an important light upon the organization of the Indo-Germanic and many other languages." 11.] THE LATIN CASE-SYSTEM. 291 idea (comp. adeo) is equivalent to the Greek eTr/r^^es (= eirl Ta^ecrir, Buttmann), and from it is derived idoneus - ideoneus = Gr. 67riTt)$eios. Igitur is either the case in -tur (- tus, -9ev) from a pronoun which is found in Oscan, under the form of esa, the soft Latin g representing the sound of s or #, or it is the locative of the third pronoun strengthened by a prefix equivalent to the combination e-ho, e-go, which is found with similar adjuncts, especially in the case of i-s-te, the first syllable of which includes the same elements as i-gi, and e-ho. In old Latin its signification was i-nde, " out of that" (Festus, p. 105 ; above, Chapter. VI. $ 7), which is the usual force of the termination -tus = Oev, or " thereupon/' which agrees with the other analysis of igitur, with the use of -tur in the third person passive, and with the obvious meaning of e-s-te in Umbrian. Some adverbs are merely cases of common nouns, which usage has made indeclinable. These appear sometimes as con- junctions, and sometimes as prepositions. Instar, gratia, and ergo, may be compared with SiKrjv, xdpiv, and eW/ca (see New Crat. 271, sqq.). Prope[d] (cf. propin-quus) is the ablative of an old adjective, and prop-ter is its case in -ter = tus = 6ev. Penes and tenus are forms of the same kind as instar, and contain the roots of pen-dere, ten-dere. Clam and palam are locatives of the same nature as partim, &c. The former, which was also written calim (Fest. p. 47), contains the root of celo, K\7rTw, KaXvTTTw, &c. Palam is the same case of an adjective connected with palatum, irv\tj, &c. That it is a noun appears farther from the fact, that it is used also with the preposition in (in palam = aperte, Gloss. Isid.), like in-cassum ; comp. pro- palam. The same is the case with cor am = co^oram (/car' o/x/ua) 5 comp. co'minus, e'minus (e/c x ei p^' Sometimes the adverb is merely the crude form of the noun. We have examples of this in simul, procul (from similis, procilis) ; and the ancients wrote facul (Fest. p. 87) and perfacul (id. p. 214) for faculter or facile, and perfacile. Again, the full form of the noun is occa- sionally used as an adverb : in the xn. Tables we have nox for noctu (above, p. 216); and Virgil (JEn. I. 215 ; VII. 624) and other writers use pars for partim. There is an approximation to this usage in the indeclinable Greek Oejuis (Buttmann, Ausf. Sprachl. I. p. 227). 192 292 THE LATIN CASE-SYSTEM. [On. VIII. $ 12. Adverbial expression for the day of the month. To these instances of the adverbial use of nouns may, perhaps, be added the phrase by which the Romans designated the day of the month. Here a locative of the day is inserted between the preposition and the word which denotes the standard of reckoning. Thus, "on the fourth day before the Nones of April," is expressed by, ante (die quarto) Nonas Apriles = quarto die ante Nonas Apriles. And this whole expression is regarded as one word, which may be dependent on a preposition : thus we may say, ex ante die iii. Non. Jun. usque ad pridie Kal. Septembres, or differre aliquid in ante xv. Kal. Novembres. If the inserted date was ever written or pronounced in the accusative case, according to the ordinary practice among modern Latinists, it is obvious that this must have originated in an attraction, or in a mistaken usage. The well-known employ- ment of the locative pridie to indicate the day immediately before the Calends, Nones, or Ides, shows that the other days must have been expressed in the same case. CHAPTER IX. DECLENSIONS OF THE LATIN NOUN. 1. The usual arrangement is erroneous. 2. General rules for the classification of Latin nouns. 3. First or -a declension. 4. Second or -o declension. 5. Third declension or consonantal nouns. 6. A. First class or purely consonantal nouns. 7. B. Second class or semi-consonantal nouns. $ 1. The usual Arrangement is erroneous. THE arrangement of Latin nouns in different declensions (K\I~ ow) or forms of inflexion has been managed by grammarians without any regard either to the internal organization of the word or to the real convenience of the learner. Among the ancient grammarians, Varro proposed a simple convention namely, to distinguish the declensions of nouns according to the vowel of the ablative singular (L. L. X. 62, p. 257, Muller) : "nam ejus cassuis literarum discriminibus facilius reliquorum varietatem discernere poterit, quod ei habent exitus, aut in A, ut hac terra ; aut in E ut hac lance ; aut in I, ut hac levi ; aut in O, ut hoc ccelo ; aut in U, ut hoc versu. Igitur ad demonstrandas declinationes vice prima lisdc." Diomedes distinguished seven declensions, dividing the nouns in -ius, -mm from those in -us, -urn, and the neuters in -u from the feminines in -us (see Zeitschr. f. d. Wiss. d. Spr. III. 315). The favourite and oldest method in this country has been to consider the noun according to five distinct declensions. The a and o declensions stand in their proper place at the head of the list. Then follows the conso- nantal declension considered as one. And the nouns in -u and -e are treated as two distinct schemes of case-formations. One of the objects, which I proposed to myself in writing a new Latin Grammar, was to correct this vicious and faulty exhibition of the different forms of the noun ; but I was unable in that elementary treatise 1 to explain and justify every feature in the new system which I adopted. That and other developments were reserved for the present work ; and I shall now proceed to show that the arrangement, which appears in the Latin Grammar, is the only classification which is consistent with the results of scientific phi- A complete Latin Grammar for the use of learners. London, 1852. 294 DECLENSIONS OF THE LATIN NOUN. [On. IX. lology ; while I know by experience that it is at least as easy to the learner. fi 2. General rules for the classification of Latin Nouns. The true classification of the crude or uninflected forms of the Latin noun is obviously that of the letters which constitute the distinctive characteristics. At first sight, all these forms fall into two great divisions, according as they terminate in vowels or consonants. But while, on the one hand, the vowels them- selves are distinguished by their structure and origin as vowels of articulation and vocalized consonants, so that the latter belong to the consonant class when considered according to the genesis of the crude form, on the other hand, the consonants are not less distinguished among themselves, according to the organ by which they are uttered, and according to the difference between mutes and liquids, than they are discriminated from the pure vowels. The scientific or methodical order of the declensions must be one which enables us most easily to fall back on the root of the noun, and on the original form of those pronominal affixes by which it is extended or developed, before it becomes the vehicle of the case-endings. And if the vocalized consonants i and u may be traced to an ultimate identity with guttural or labial mutes, it is clear that the nouns of which they are the characteristics ought to be ranged among the consonant declen- sions. In this way, we shall have two main classes of nouns those whose characteristic is one of the pure vowels a or o, and these may be considered as subdivided into two declensions ; and those whose characteristic is a consonant, whether mute, or liquid, or one of the semi-consonants i and u, considered as a representative of some mute, and these may be regarded as constituting one declension. While this scheme of the declensions is the only arrangement, which can be justified on the grounds of scientific etymology, it is at least as convenient as any other to the mere learner : for we cannot give any practical rule to a beginner more simple than that which results from this arrange- ment namely, that the vowel-nouns invariably form their geni- tive plural in -a-rum or -o-rum, which is rarely contracted into -um; that they form their dative and ablative plural in -is, which rarely appears under the uncontracted form -bus; that the accusative singular is always -am or -urn, the accusative 2.] DECLENSIONS OF THE LATIN NOUN. 295 plural -os or -as, and the ablative singular always -a or -6 ; and, on the other hand, that the consonant nouns generally form their genitive plural in -um, which is rarely preceded by the characte- ristic r; that, conversely, they form their dative and ablative plural in -bus, which rarely, if ever, loses its characteristic b ; that the ablative singular is always e or i; and the accusative plural always -es, except when the characteristic is u. These general distinctions do not apply to the nominative-accusative plural of neuter nouns, which are uniformly terminated by -a in all declen- sions. If then the classification, which I am about to explain, is not only true, but most convenient to the student, there can be no reason why it should not supersede the old-fashioned method even in elementary grammars. $ 3. First or -a Declension. The Latin -a declension, as compared with the Greek, pre- sents one remarkable contrast. In pure Latin nouns, the termi- nation is invariably -a, whereas in corresponding forms the Greek declension exhibits -a, -a, -as, -77, -179. Thus we have not only cella by the side of a'juuAXa, but amicitia, scriba, area, nota, ho- micida, by the side of (pi\id, Ta/uids, T-5 by the side of quantus, or Ta'joas = Tctjoai/r-s by the side of Tarentum no inference can be drawn. But as -d- is generally, if not always, a shortened form of the articulation which appears as the second personal pronoun and the second numeral, and as we have verbal forms in -dus (as cupidus, &c.) by the side of verbals in -re'os, -TVS, -TI?, it is not unreasonable to conclude that if orien-t-s = oriu-n-dus, the former is an abridgment of orien-tis analogous to sementis, &c., and this explains the genitive plural in -ium. Although there are some nouns in -i- which re- tain their characteristic throughout the cases as sitis, Tiberis, febris, puppis, &c., it not unfrequently happens that the shorter 302 DECLENSIONS OF THE LATIN NOUN, [Cn. IX. vowel e is substituted in the nom., ace. and abl. sing., and this is always the rule in the nom. and ace. pi. So that, generally, the criterion of a noun in i is furnished by the form of the gen. pi. Thus, although we have nubes, nubem, nube, nubes > we have always nub-i-um. The peculiar nouns in -es = -a-is, in which this characteristic i is appended to a crude form in -a, sometimes ap- pearing as a distinct noun of the first declension (cf. mater-ia, " the mother-stuff," or " materials," v\rj, with materies - mate- ria-is), always retain this e = ai, and consequently exhibit the full or proper form of the gen. pi. in -rum. For, according to the rule, s = r is not usually elided except between two short vowels, and the contraction e-ai produces the same result as the contractions a = a-e and 6 == o-e in the first and second declension, so that we have arum - a-erum, orum - o-erum and erum = a-irum. As canis, juvenis and vates form the gen. pi. in -urn, we infer from this simple fact that they are as improperly included in the -i- declension as other nouns are excluded from it. If we compare canis with KVWV = /cJoi/-s, we shall see that the i is merely an unorganic insertion after the liquid, and the same is the case with juvenis ; whereas vates must be explained on the same principle as the Greek compounds in -qs from neuter nouns in -os, which exhibit the lengthened form only in the nom. and accus. (New Crat. 228). The neuter nouns in -e, which are shown by their abl. sing, in -i, their nom. accus. pi. in -ia, and their gen. pi. in -ium, to belong to the class of -i nouns, are really the neuter forms of adjectives in -is. Compare, for example, maenia with com-munis, mare and mille with acris, agilis, rete with restis and irretire, animal, for animale t with cequalis, &c. One of the strongest proofs that the additional -i is an indication of the adjectival inflexion is furnished by the fact that while the immoveable vetus, veteris, forms its gen. pi. in -urn, and while celer, denoting " a horseman," has no gen. pi. but celerum, the regularly inflected adjective celer, celeris, celere, has a gen. pi. celer-ium. With regard to the nouns in I and r in particular, we must consider that the extensions in -Us and -ris are the basis of further extension in -leus and -rius, such as nuc-leus, prceto- rius, &c., which in Greek would sometimes appear as -Xt-Acos, and for this there is an occasional parallel in Latin, as in fame-li-cus. The following classification will show how far the whole group of i nouns has retained or lost the original characteristic. 7.] DECLENSIONS OF THE LATIN NOUN. 303 pupp'is nube[=i~\s urb[i~]s serpen[ti]s di[=a-i]s mar\t=]e animal\ji\ pupp-is .... .... .... .... mar-is animal-is kbl.pupp-i .... .... .... .... mar-i animal-i pupp-im] (N.A. v ........ .... ....-< , mar-ia anmial-ia, or em j { pi. pi. pupp-ium nub-ium urb~ium serpen-t-ium di\j=a-i\-r-um mar-ium animal-ium (fi) There can be little doubt that nouns in u either included or were ultimately identical with the nouns in -i, which have just been discussed. Thus in Greek -i/-s was originally -Fts or -w?, and the Oscan Ke-us stands by the side of the Latin ci-vis (above, p. 125). In most existing instances, however, this i has been lost, and we have either a noun in v, declined like the purely consonant nouns, or a form in which the u is retained throughout, just as the i alone keeps its place in the most regular of the i nouns. Of the former class, we have only two remaining : bos, for bov-s (Greek fiovs), gen. bov-is, and Jus for Jov-s (Greek Zeus), gen. Jov-is. The nominative of this latter noun is always connected with pater under the form Ju-piter, corresponding more nearly to the Greek vocative. Thus Catullus (LXIV. [LXVL], 48) translates the line of Callimachus word for word as follows : ZeC TrcLTfp eos "KaXvftav rrav aTroXoiro yevof. Ju-piter ut Chalybon omne genus pereat. The analogy between the nouns in i and u will be seen from the following comparison. N. pupp-is trib-us N. A. ret[i=~]e corn-u G. pupp-is trib-us G. ret-is corn-us D. pupp-i tribu-i or tribu D. ret-i corn-u A. pupp-im trib-um N. pi. ret-ia corn-ua Abl. pupp-i trib-u G. pi. ret-ium corn-uum G.pl. pupp-ium tribu-um There are two nouns of the i declension, which deserve es- pecial consideration, not only on their own account, but also on account of some remarkable assonances in the cognate languages, which might lead to misconception or confusion : these are res, "a thing or object," and mare, "the sea." I have shown, in another work, that res = h-ra-is is a derivative from hir = ^eip (Varro, L. L. IY. 26), and that it must therefore be compared with the Greek ^eos, x/oe/a, -^prj^a, to which it bears the same relation as Icena, luridus, &c. do to yXaiva, ^Xwjoos, &c. Con- sequently, res is " that which is handled," and means an object of thought in accordance with that practical tendency of the 304 DECLENSIONS OF THE LATIN NOUN. [On. IX. Roman mind which made them regard all realities as necessarily palpable l , whereas the Greeks were contented with the evidence of the eyes. Thus while a Greek declared his certainty by the predicates evapyi}s a t ejULcpavys, craV evapy^s ravpos; which is opposed to dvfyeia KVTCI ftoinrpapos or the partial assumption of the bovine form. Just in the same way we find in Shakspere (K. John, I. 2) : Mine eye hath well examined his parts, And finds them perfect Richard. And Milton says (Parad. Reg. I. 82) : I saw A. perfect dove descend; i. e. evapyrjs ircpiarTepd. Aristotle (Eih. Nicom. I. 1, 3) uses evapyrjs and avcpbs as synonymous expressions for that which falls within the reach of our ordinary experience. 7.] DECLENSIONS OF THE LATIN NOUN. 305 the hand," is also used to signify " the surface of the sea" (see Pmd.Isthm.lll.74:). But these are merely accidental coincidences: for, as we have seen above (p. 75), ma-re and the Sclavonian mo-re must be referred to the Semitic D?D, the second syllable being that which appears in the Greek pew, the Etruscan ril, &c. Besides, mare does not signify " the surface of the sea," but the mass of water, as opposed to dry-land. The surface of the water is denoted by pelagus, directly borrowed from the Greek 7reXa7os, which is connected with 7rAa', and means " an extended sheet of water;" hence TreXayos signifies "the high-sea," and TreAcryio? means " out at sea" (New Crat. 280). If a river had burst its banks and covered a large expanse of country, it would be called a mare, or " flood," and might in that case ex- hibit a pelagus or " wide surface of water." Thus Virgil says of the mouth of the Po (^En. 1. 246) : It mare proruptum, et pelago premit arva sonanti. " It rushes forth in a flood, and covers the lands with a roaring sheet of water" This view of the origin and signification of ma-re is important with reference to its form as a noun in i. We see this i in other words involving the root re, as ri-vus, ri-l, &c. ; and considering the general meaning of adjectives in -is, we must come to the conclusion that ma-r-e is the neuter of an adjective ma-re-is =ma-r-is=vc) pop poos. To return to res=hra-is, the ter- mination seems to indicate it as a doing, rather than as a thing done as a " hand-ling" (handlung) rather than as a work, as a xpijcris rather than as a xpw a - Practically, however, res means a mere object of thought, a thing which is or may be handled ; and this appears still more clearly from the use of re-or, " I think," i. e. " I propose a res to my mind," and its derivative ra-tio (from ra-tus) 9 which implies the action of the verb, and denotes the mode or act of thinking. Still, it may be seen, by a little care in the examination, that the fixed or passive meaning of res is quite consistent with its original use as a noun of action. As we shall see, when we come to the gerundia and gerundiva, 'the difference between active and passive becomes evanescent when we descend to the infinitive or abstract use of a word. When we are speaking of the " winding-up of a business," " the closing of a shop," &c., it is obvious that we direct attention to the thing done, rather than to the act of doing it. Just so with res as opposed to ratio. Between these two the substantive reus 20 306 DECLENSIONS OF THE LATIN NOUN. [Cn. IX. and the verb reor may be presumed to intervene. If res means a " handling," or " action," reus will denote the person impli- cated in the action; and as res, in a legal sense, denotes the cause and object of the controversy, in the same technical appli- cation reus will denote a person accused or impeached cujus res agitur. And as ratio has no existence save through the verb reor, it must mean something more than the mere bodily handling implied by res. It must denote a mental operation consequent upon this contact. And, in point of fact, ratio always implies some intellectual process, or the plan and system which emanate from it. While res or res familiaris is the property, ratio is the account kept ; res publica is the state or object of government, ratio is the mode of governing ; res is the outer world, as in natura rerum, &c,, ratio is the inner reason, which deals with it theoretically. And this opposition is even carried so far that, while verborum ratio is the arrangement of words, or the style (Cic. de Oratore, II. 15, $ 64), we have rerum ratio ( 63) for " history," or the arrangement of facts and actions. The neuters in e of this declension are interesting as examples of the form which appears by the side of all masculine and feminine adjectives in -is, as tristis, neut. triste. Of course this theory assures us that the original ending of their neuter must have been -id, just as ante was originally antid. And this inference is confirmed by an obsolete neuter in -is, which bears the same relation to -id that corpus, opus, &c., do to the original corpud, opud, &c. This neuter is found in potis, satis, by the side of pote and sat (for sate) ; thus, Lucret. I. 452 : Conjunctum est id, quod nunquam sine perniciali Discidio potis est sejungi seque gregari. V. 716 : Corpus enira licet esse aliud, quod fertur, et una Labitur oranimodis occursans efficiensque, Nee potis est cerni, quia cassum lumine fertur. Terent. Adelph. IV. 1, 5 : " ita fiat et istoc, si quid potis est rectius." Catull. LXXV. 24 : " quod non potis est." LXXI. 7 : " qui potis est." Corn. Nep. Epam. 4 : " abstinently erit hoc satis testimonium : " cf. Hannib. 6. These passages are quoted by Schwartze, das alte ^gypten, I. p. 637. The same expla- nation applies to necessus for necessum or necesse, in the Senatus Consultum de Bacchanalibus. CHAPTER X. PRONOUNS AND PRONOMINAL WORDS. 1. General definitions. 2. Personal Pronouns. 3. Indicative Pronouns. 4. Distinctive Pronouns. 5. Relative, interrogative, and indefinite Pronouns. 6. Numerals and degrees of comparison. J. Prepositions. 8. Negative Particles. 1. General Definitions. THE term pronoun, in accordance with its original meaning, (pronomen, avr^vv^ia), ought to denote only those words which are used as substitutes for nouns. But according to that which appears to me to be the only scientific classification, all words fall into two great divisions, pronouns, or words which indicate space or position; and words containing roots, which express the positional relations of general attributes. The former do not allow any admixture with the other element of language : the latter require the addition of at least one pronominal suffix to make them words. I have therefore proposed 1 to call the pronouns, or positional words, the organizing, constituent, or formative element of inflected language, and the roots I would designate as the material element of human speech. With this extension of meaning the term pronoun will include not only the personal, demonstrative, and relative words, which it generally denotes, but also the prepositions, the conjunctions, and those adverbs which are not merely cases of nouns. ^ 2. Personal Pronouns. Although the verb has three persons, the Latin language does not use more than two personal pronouns or general indi- cations of the nominative case. For although ego and tu may be used with the first and second persons of the verb, which, as we shall see, are not consistently expressed by the inflexions; with the third person, which always ends in -t or -tur, the nominative is either omitted or expressed by a noun substantive. When, however, in the objective construction it is necessary to introduce a pronoun referring to the nominative of the verb, we employ the reciprocal or reflexive se. Thus, although diceba-t is a suf- 1 New Crat. 128. 202 308 PRONOUNS AND PRONOMINAL WORDS. [On. X. ficient expression of " he said, or used to say," we must introduce se before an infinitive expressing the assertion ; as : diceba-t SE esse bonum virum, " he said that he (the person, in question, who said) was a good man ;" and as we should write ego diceba-m ME esse, or tu diceba-s TE esse, we may infer an ori- ginal pronoun of the third person beginning with s- and corre- sponding to the Greek 6 or '/, just as e corresponds to se. But this form occurs only in the oblique cases, sui, sibi, se, and in the particles si-c, si-ne, si, and se-d. The original inflexions of the two personal pronouns were as follows : SING. 1ST. e-go or ego-met tu or tw-te G. 'mis ti-s D. mi-hi (for mi-fi or mi-bi) ti-bi A. me-he te-he Abl. me-d. te^d. For the plural, or rather the collective form, of the personal pronouns, we have two different roots corresponding to vwi and acftwi, which are used as the dual in Greek; and from these roots we have the nom., ac., voc. no-s, vo-s; dat., abl. no-bi-s, vo-bi-s. According to the analogy of vwiv, cr^auV, we ought also to have genitives no-urn or no-sum, and vo-um or vo-sum. But these are not found. Indeed, although the singular genitives mis, tis, which may have been originally forms in -jus, like 7iu-jus } e-jus, &c., retained their use as late as Plautus, these also became obsolete in classical Latinity, and the genitive forms for the singular and plural were derived from the possessive adjec- tives meus, tuus, nos-ter, ves-ter. The connexion between the genitive and the epithet is well known (New Crat. 298), and in all languages the possessive may take the place of the genitive of a pronoun. But in Latin and Greek we have not only a possessive in direct adjectival agreement with its noun, but, by a singular attraction, we have the genitive of the pos- sessive used as if it were the genitive of the pronoun itself. I call this an attraction, for I think it must be explained by a transition from those idiomatic collocations, in which a dependent genitive stands by the side of the possessive. Thus we may say not only mea scripta, " my writings," for " the writings of me," but even mea scripta recitare timentis (Hor. I. Serm. 4, 23), 2.] PRONOUNS AND PRONOMINAL WORDS. 309 J the writings of me fearing to recite ;" and not only jperipa epis, " our contention/' for " the contention of us," but even ayaOwv epts ^erepa (JEschyl. Eum. 975), " the contention of us good persons." We see then how easy the transition may be from such phrases as mea unius opera respublica est salva, or vestris paucorum respondet laudibus, to 6am unius tui studio me assequi posse confido, or vestrum omnium voluntati paruit. Hence we find that ultimately mei and tui were the only geni- tives of ego and tu, and nostri or nostrum, and vestri or vestrum t the only genitives of nos and vos. The same applies to the very defective pronoun of the third person, the reciprocal se, which has lost its nominative, and has only the genitive sui, the dative sibi, and the accusative or ablative se 9 for all genders and numbers. We must also consider the Greek ejuoiJ, or /uoD, anciently /jieov (N. Crat. 134), and crow, as properly belonging to the possessive. The hypothesis of an attraction, which I have proposed, is the only way of explaining the difference in the usage of nostri, nostrum, and of vestri, vestrum. That nostrum, vestrum are genitives plural, is clear from the fact that they were anciently used in the full forms nostrorum, vestrorum ; thus in Plautus (Mostell.I.3,123) we have : verum illud est, maximaque pars vostrorum intelligit. As genitives they can only be explained by an attraction into the case of some plural genitive expressed or understood. In general, we do not find the genitive except when the personality is emphatically expressed; as in Ovid, Her oid. XIII. 166 : Si tibi cur a mei, sit tibi cur a tui. Cic. Catil. IV. 9 : habetis ducem memorem vestri, oblitum sui. And here it may stand by the side of an inflected possessive, as in Cic. ad Fam. XII. 17 : grata mihi vehementer est memoria nostri tua ; or even be opposed to one, as in Ovid, Her oid. VII. 134 : parsque tui lateat corpore clausa meo. But whereas nostri, vestri, are used only when we speak of the persons as a whole; as: memoria nostri tua, "your recollection of us," as a single object of thought ; nostrum, vestrum are employed when we speak of the persons as a collection of separate or separable elements. Accordingly, the latter is the form adopted after such a word as pars (in the passage quoted above from Plautus), and by the side of omnium, as in Cic. Cat. I. 7 : patria est com- munis omnium nostrum par ens, "our native land is the common parent of all of us," many and separable as we are. But that it is really in this case an attraction from the inflected possessive, is 310 PRONOUNS AND PRONOMINAL WORDS. [On. X. clear from such passages as Cic. Cat. IV. 2 : hi ad vestram omnium ccedem Romce restiterunt. We have a genitive by the side of the possessive in the construction of the impersonal verbs, or rather phrases, re-fert =rei fert, "it contributes to the in- terest," and interest, "it is concerned about the business," where rei is understood in the sense in which the Latin verb has become an English substantive 1 . In these phrases we have either a gen. of the person or persons interested, or the pos- sessive pronouns, mea, tua, sua, nostra, vestra, agreeing with the dative rei, expressed in re-fert, and understood in interest. Thus we have : faciundum aliquid, quod illorum magis, quam sua re-tulisse videretur, " he must do something which might seem to have been more for the interest of those others than for his own;" Ccesar dicere solebat non tarn sua quam reipub- licce inter esse, ut salvus esset, " Caesar used to say that it was not so much for his interest as for that of the state that he should be safe." That re for rei is the dative, and consequently that mea, sua, &c., here stand for mece, suce, &c., is proved by the competent testimony of Yerrius (Festus, p. 282, ed. Miiller): re-fert quum dicimus, errare nos ait Verrius. Esse enim rectum REI FERT, dativo scilicet, non ablativo casu. In Cato, R. R. c. 3, we have: et rei et virtuti et glorice erit. That fero may be used absolutely without any accusative is clear from such phrases as : dum tempus ad earn rem tulit (Ter. Andr. I. 2, 17), dum cetas tulit (id. ibid. II. 6, 12), nunc ita tempus fert, ut cupiam (Heaut. IV. 1, 54), scilicet ita tempus fert (Adelph. V. 3, 5). And it is unnecessary to show that fero, like XvviTeXeid, may govern the dativus commodi. The change of ce into a is found also in post-hac, inter-ea, &c., which will be explained immediately. $ 3. Indicative Pronouns. The three pronouns, hie, iste, ille are called indicative, be- cause they indicate, as objects, the three personal pronouns, which, in the cases already considered, are expressed as subjects of the verb. Hie, " this," " the person or thing here," indicates the speaker and all close to him ; iste, " that of yours," indicates the person addressed and those in his proximity; ille, "that 1 For re = rei in this sense cf. Plaut. Trinumm. III. 2, 9 = 635 : tua re consulere cupio. 3.] PRONOUNS AND PRONOMINAL WORDS. 311 other," indicates all distant persons and objects. This distinction was well known to the oldest grammarians, and is fully borne out by the consistent usage of the best writers. Priscian's dis- tinction is rather vague : he says (XVII. 9. 58, Vol. II. p. 39, Krehl) : " Demonstrativa [sunt] hie, iste, et ille. Sed interest quod ille spatio longiore intelligitur, iste vero propinquiore ; hie autem non solum de pra3sente, verum etiam de absente possumus dicere, ad intellectum referentes demonstrationem, ut, hoc regnum dea gentibus esse Virgilius ad absentem Carthaginem retulit demonstrationem." But Laurentius Valla has given the personal reference of the three pronouns with the greatest accuracy (Elegant. II. c. iv. p. 39. ed. Aldina 1536) : " de me loquens dicere debeo hoc caput, hcec manus, hcec civitas. De te vero istud caput, ista manus, ista civitas. De tertia autem per- sona illud caput, ilia manus, ilia civitas. Cicero in Antonium (Phil. II. 25) : tu istis faucibus, Sec., h. e. istis tuis faucibus, &c. Unde nascuntur adverbia istic, istinc, istac, istuc, istorsum, isto. Ut idem ad Valerium juris consultum : qui istinc veniunt aiunt te superbiorem esse factum, i. e. qui ab ista provincia in qua agis, hue in Italiam Romamque veniunt." Practically we find that hie and iste are opposed as I and you, and hie and ille as near and distant. Thus we find (Cic. Acad. IV. 33) : " iisdem hie sapiens, de quo loquor, oculis, quibus iste vester terram, mare, intuebitur ;" and (pro Rabirio II.) : " si illos, quos jam videre non possumuS) negligis, ne his quidem, quos vides, consuli putas oportere." And thus in reference to circumstances previously mentioned, ille denotes the former or more distant, hie the latter or nearer particular ; as in Propert. III. 14, 17 : Qualis et Eurotse Pollux et Castor arenis, Hie victor pugnis, ille futurus equis. The same distinctions are observable in certain peculiar usages. Thus Terence has (Andr. IT. 1, 10): " tu si hie sis, aliter sentias," " if you were in my place, you would think otherwise." In lawsuits iste, " the man before you," i. e. the judices, is the defendant : hence, we find this pronoun used with a certain ex- pression of contempt to indicate a person who has been brought unfavourably before the notice of those whom we are addressing ; whereas ille, " that other," as indicating a person so striking as to attract our attention in spite of his remoteness, is often used to denote a well-known or eminent individual, as : " magnus ille 312 PRONOUNS AND PRONOMINAL WORDS. [On. X. Alexander," or " Medea ilia" In all these usages the triad hie, iste, ille, correspond to the Greek oe, OVTOS, e/ceti/o?. This is especially seen in the employment of o$e and OVTOS to designate the first and second persons respectively. Thus CEdipus is made to say of himself: OVTL JULTJ Xa^wat TOVOC cv/uLna^ov ((Ed. C. 450) ; but he is addressed by the subterraneous voice (ibid. 1627): M OUTO?, OVTOS Oto/Trovs, TL yue'XXojuey ; The speaker in a law-court designates himself, his client, and his affairs, by oSe; but the defendant is OVTOS = iste, " the man before you" (the judges). In continuous narrative rdSe are the things which I am about to say, which are before me, but not yet before my readers ; whereas Tavra are the things just said, and which have been submitted to them. This shows that the true reading in uEschylus, Suppl. 313, must be : XO. TSrj\ov StTraifta irarepa rovf? e'/zov Trarpos. BA. TO ndv tradeoff vvv ovop.a TOVTOV pot (fopdo'ov. For the Chorus having spoken of their father as present by them (rovSe), the King, in his reply, would designate him as by their side (TOVTOV). With regard to the etymology of the indicative pronouns, there can be no doubt that the first part of hi-c corresponds to the Greek '/, which appears as the nominative of the reflexive e'o = ov, of, e. It is therefore a subsidiary form of o = ;, to both of which it partly corresponds in meaning, and to the latter of which it is directly related. In the classical use of ipse, on the contrary, the first part, or the is, remains uninflected, while the second syllable is regularly declined ; thus : i-psus, i-psa, i-psum, gen. i-psius, &c. There are two ways of explaining this phenomenon. We may either suppose that the ps- represents an inversion of the reci- procal : and thus the in- flexion of the second part only will correspond to the Greek forms e/uavTov, eavrov, &c., where the first part is immoveable. This is Bopp's theory. But it may with justice be objected that ipse corresponds to aurog, and that we have the combina- tions me ipsum, se ipsum, &c. Besides, we find in the older writers that the included is is regularly declined, while the affix -pse remains as an immutable appendage, just like the -dem of i-dem ; thus we have eam-pse (Plaut. CistelL I. 3, 22 ; Aul. V. 7), ea-pse ilia (Curcul. IV. 3, 2), eo-pse illo (ibid. 5) : and especially in the combination re ea-pse, or reapse (Festus, p. 278, Muller). Since therefore we find another affix -pte also appended not only to the declined forms of is, as in eo-pte (Festus, p. 110, cf. ipsippe = ipsipte, p. 105), but also to vos, mihi, meo, suo, &c. as vo-pte, mihi-pte, meo-pte, suo-pte, &c., as this cannot be re- ferred to an inversion of sv, but may bear the same relation to- 4.] PRONOUNS AND PRONOMINAL WORDS. 317 -pse that the original supines in -turn do to the secondary forms in -sum, I fall back on the other explanation, and consider -pte an indeclinable affix analogous to Trore, which has been softened into -pse, perhaps from an original assimilation in is-pte (cf. ctu/cos for o//c-(7/co9, \s are identical. The guttural anlaut of the Latin relative and interrogative 21 322 PRONOUNS AND PRONOMINAL WORDS. [Cn. X. is lost in ubi, unde (cf. ali-cubi, ali-cunde), un-quam (cf. -cunque), uter (cf. /corey>os), &c. Extensions of the relative or interrogative form indefinite or indefinite-relative pronouns, which are accurately distinguished by the best writers. Thus ali-quis = alius-quis or ille-quis, quis-piam, and qui-dam, denote " some one in particular," though the object is not named; quis-que means, "every one;" quis- quis and qui-cunque "whosoever;" qui~vis and qui-libet, "any you please ; " quis-quam and its adjective ullus = unulus, " any at all." Hence the words in the first group are obscurely defi- nite ; quisque, quisquis, and quicunque include all persons or things referred to ; quivis and quilibet allow an unlimited range of choice ; and quisquam and ullus exclude all the objects speci- fied. The first syllables of ali-quis have been discussed above, and there is no difficulty in understanding the compound as sig- nificant of separative uncertainty " that other some one." As quis-piam and qui-dam very nearly correspond in meaning, their etymological analysis ought to lead to similar results. With regard to the former there can be no doubt that quis-piam = quis-pe-iam. Now pe is obviously equivalent to que and re : cf. nem-pe, nam- que. Consequently quis-pe-iam = quis-que-jam = osrt? re Srj, " some one whoever it may be." The correspondence of pe and re in this case is confirmed by the exact agreement of quippe = quia pe and are (to which $q is sometimes added) in the sense " inasmuch as : " for quia is the old neuter plural of quis. In many of its usages jam corresponds in meaning to the Greek ^, as in the cases just now compared. But in form there is a much closer affinity between Stj and the affix -dam or -dem. Thus qui-dam is exactly os 77, and qui-dem is ye Stj, To the same class belongs demum, which Ebel (Zeitschr. f. Vergl Sprachf. 1851, p. 308,) would explain as a superlative from the preposition de, on the analogy of primum from prce. The forms tan-dem and pri-dem show that this explanation is untenable ; and the latter at all events proves that dem and pri are not contradictory designations of time. The true explanation is suggested by deni-que and its by-forms done-c and doni-cum. Greek particles expressing time end either in Ka - /cei/, as avri-Ka, rjvi-Ka, or in re, as 6-re, ro-re, Tro-re, CV-TC, e/cacrro- TF, &c. It is clear that these endings are ultimately identical ; but it may be concluded, that, while the latter gives rather a 5.] PRONOUNS AND PRONOMINAL WORDS. 323 degree of precision to the term, the former, which more immediately corresponds to the well-known particle of the apodosis, comes nearer in meaning to the Latin cnn-que^-wo-re, and our -soever. The Latin -que corresponds in some cases to -/ca or ai>, in others to -re. Thus, while -cun-que is Tro-re, there can be no doubt as to the equivalence of ubi-que and oirov av, of Trjvi-Ka and deni- que (New Crat. J 196). The substitution of the tenuis for the medial in the Greek forms is not universal, for we have ore 5iJ by the side of quan- do, and when this apparent difference is removed, we have no difficulty in seeing the exact correspondence between Trj/mos, as opposed to was, and demum, for which, according to Festus (p. 70, Miiller), Livius Andronicus wrote demus. As the element dem is placed indifferently before or after the particle which it qualifies (cf. deni-que with tan-dem, pri-dem) we shall understand the correspondence between qui-dam, osrts S^, and the synony- mous Stj Ti? =*nescio quis (Heindorf ad Plat. Phcedon. p. 107 d). Jam is related to dam, Sqv, as Janus to Dianus, &c., and thus quispiam os r*s re Stj or 6's TIS $q TTOTC falls into a near re- semblance to qui-dam = o? $y or Stj rz?. The difference between aliquis and quispiam consists in the shade of definiteness con- veyed to the former by its prefix ali-, so that while aliquis means " some one in particular," quispiam means generally " some one" or " any one." Thus in Cicero (de Orat. II. c. 9. J 38), we have : " si de rebus rusticis agricola quispiam^ aut etiam, id quod multi, medicus de morbis, aut de pingendo pictor aliquis diserte dixerit aut scripserit, non idcirco artis illius pu- tanda sit eloquentia." The addition of the id quod multi shows that quispiam is more general than aliquis : " if any person versed in agriculture shall have written or spoken with eloquence on rural affairs, or even any physician on diseases, as many have done, or some painter on painting, &c." That there is much the same distinction between aliquis and quispiam as between aliquis and quis, is proved by the existence and usage of the compound aliquispiam or aliquipiam (see Cic. Tusc. Disp. III. 9). In the case of aliquis itself a stronger signification of se- paration or definiteness may be conveyed by writing at length alius quis or quis alius (see the passages quoted by Draken- borch, ad Liv. V. 13. 4. p. 59). The parallelism between quippe - quia-pe and are might lead us to conclude that ut-pote, 21 2 324 PRONOUNS AND PRONOMINAL WORDS. [On. X. which is all but a synonym of quippe, is merely a compound of ut and a form involving -pe and equivalent to the termination -pte discussed above. As however there is no Latin word -pote equivalent to the Greek TTOTC, and as the phrase ut pote = ut potest actually occurs in Yarro (apud Non. c. 2. n. 876 : viget, veget, ut pote, plurimum), we may fairly conclude that we have here a phrase like scilicet, duntaxat, and not a mere combina- tion of pronominal elements, so that ut pote means " as is pos- sible." The suggestion of Doderlein that it stands for ut puta does not deserve a moment's consideration. That quilibet involves the impersonal libet is obvious on the slightest examination; and notwithstanding the difficulty occasioned by the particle -ve, we must conclude that the 2nd pers. sing, of volo is the affix of quivis. This is not only deducible from the analogy of quilibet, but is shown by a passage in Cato (R. R. c. 52) where a noun is interposed between qui and vis : " hoc modo quod genus vis propagabis." What has been already said of cun-que = cum-que = TTO-TC applies to other uses of the affix -que, as quis-que, uter-que, undi-que, utrin-que, ubi-que, us-que, quo-que. There is much general truth in Schmidt's definition of quisque (de pronom. Gr. et Lat. p. 100) : " pronornen indefinitum rem mente conceptam et e rerum ejusdem generis cumulo ac serie exemtam significat. Que autem particula si ad pronomen additur, pronominis vis ex- tcnditur, idque ad omnem rem, in quam cadere possit sententia, transferri significatur. Itaque quis, particula que adjuncta, non hominum incertum quondam, sed omnem, ad quern pertinere pos- sit sententia, notat. Ab omnis igitur ita differt, ut hoc quidem cunctos simul significet, quisque autera distributionem quandam exprimat." Referring to the comparison made above between the Roman affix, and the Greek --a, KCV, or av appended to re- latives in general expressions, it is clear that the only principle, which will explain all the facts, is that which lies at the basis of the true theory respecting these Greek particles. Now it appears that av and Kev are connected with the second pronominal element, and therefore claim the same pedigree as the relative pronouns. But they are not only immediately attached to the relative word in the hypothesis or protasis, as in orav, edv, o? ai/, &c., but also appear as antecedents or correlatives in the apodosis of a condition. In the latter case they can only be $ 5.] PRONOUNS AND PRONOMINAL WORDS. 325 considered as hints suggestive of the hypothetical or general nature of the whole sentence ; for if I say Xeyoi/j.' av, even with- out any condition expressed, the hearer feels that a condition is implied, which would not be the case if I had said Ae'fco. Such being the fact in regard to the apodosis, it is still more evident that the addition of a relative particle in the protasis, which is already a relative sentence, must add to the generality or com- prehensiveness of the reference. And so we constantly find that the multiplication of relative or indefinite elements makes the range of supposition wider ; and if quis means " any one," quis- que, quis-quis, qui-cun-que will mean " any any" or " every possible" individual. This view is confirmed by the Semitic usages : for we not only find pronominal repetitions, such as J1DW D = HD1 TO = quid et quid, but even repetitions of general terms, as tt^NI WN vir et vir = quis-que. Tn comparing qnis- que with qui-cun-que we observe, besides the constant distinction between quis and qui, that the latter is strengthened by the in- sertion of the temporal particle cum ; and it is worthy of notice that not only is cunque used by itself as an expression of time ; as in Hor. I. Carm. 32, 15 : " mihi cunque salve rite vocanti," where cunque quoque tempore ; but we even find it after cum, as in Lucretius, II. 113: " contemplator enim, cum solis lumina cunque inserti fundunt radii per opaca domorum." Us-que for cus-que (cf. us-piam, us-quam) is only a different inflexion of the same elements as cun-que, for us-que and un-quam both refer to time, (see Schmidt, 1. 1. p. 96) ; and quo-que " too," " still," " con- tinuing that state of things," must also be regarded as a particle of time, like its synonym etiam = et jam l . As the latter part of the words quis-que, quis-quis, qui-cun- que is manifestly of relative import no less than the affix of quis-quam, it is clear that the absolute difference in meaning between these words, and between us-que and un-quam, us-quam, cannot depend upon the etymology of the suffix. If we compare tarn, quam with turn, quum, we shall see that while the former pair refer to manner, the latter imply time. As dies signifying a particular day is always masculine, and as we have a number of adverbs counting time by days, as pridie, hodie, nudius tertius, 1 For the parallelism and difference of quoque and etiam see Plaut. Trin. IV. 3, 42 : "illis quoque abrogant etiam fidem." S26 PRONOUNS AND PRONOMINAL WORDS. [On. X. diu, interdiu, &c., it is fair to conclude that turn, quum mean " on the particular day," " on which day ;" and the same expla- nation will apply to olim, "on that day." Similarly, as the Greek adverbs in -rj are properly explained by an ellipse of o&3 signifying " way," " process," " manner," and as we have the adverbs obviam, perviam signifying directions or modes of motion, it may be inferred that there is an ellipse of mam in tarn, quam, which would at once explain their meaning. If we apply the same explanation to quis-quam, we shall see that it means " any one in any way," i. e. " any one at all," which is always its distinctive meaning ; for quisquam can only be used in a negative or conditional sentence, where all are excluded, or where the range of choice is circumscribed between the narrowest possible limits. Hence in Terence (Eunuch, prol. 1) we have : " si quisquam est in his poeta his nomen profitetur suum" " if there is any person at all, if there is any one person in all the world" where the number is especially limited. Hence unus is often appended to quisquam (cf. Liv. XXVIII. 37, where quisquam unus is opposed to alii omnes, and II. 9, where quisquam unus is opposed to universus senatus). Hence also ullus=unulus, "a little one," " a mere one," serves as the adjective of quisquam, which, as we have seen, has no feminine or plural forms, though it occurs occasionally with feminine nouns. The exclusive force of unus and ullus is well shown by the modern French aucun = aliquis unus, which performs all the functions of quispiam, although the first word belongs to the most definite of these general pronouns. Thus non vidi quenquam might be rendered je rfai vu personne, or aucune personne. And in English we sometimes use the word " single" for the purpose of excluding all of the kind as, " I have not a single shilling." Opposed as quisquam is to quis- quis, it is very strange that no editor should have observed its intrusion into the place of the latter in a passage of Ovid (Fast. Y. 21) : Ssepe aliquis solio, quod tu, Saturne, tenebas, Ausus de media plebe sedere deus; Et latus Oceano quisquam deus adrena junxit: Tethys et extreme seepe recepta loco est. It is obvious that quisquam is inadmissible, and that we must read quisquis, with the punctuation : et latus Oceano, quisquis deus advena, junxit, i.e. " whatever god happened to come 5] PRONOUNS AND PRONOMINAL WORDS. 327 up." Cf. Plaut. Amph. I. 1, 156 : quisquis homo hue venerit, pugnos edet. j 6. Numerals and Degrees of Comparison. In regard to the general discussion of this part of the subject, I have nothing to add to the full investigation which it has re- ceived in the New Crat. Book II. ch. 2. For the sake of method, however, it will be desirable to mention a few facts referring more particularly to the Latin language. While unus, more anciently osnus or oinos, corresponds in origin to the Greek efs, ei;-, Goth, aina, Celtic aenn, the Sanscrit eka is represented only by the adjective cequus. We have eV, with $ instead of tho aspirate, in sem-el, sim-plex, sem-per, and sin-gulus. The ordinal primus is derived from the preposition prce, just as the Greek irpwTos comes from irpo. All the ordinals end in -mus (which is perhaps contained in octavus for octau-mus, nonus for novimus), with the exception of secundus, "following," which is merely the participle of sequor 9 and of tertius, quartus 9 quintus, sextus, which represent the Greek -ro?. In tertius this ending is length- ened by the qualitative or possessive -ius, so that ter-t-ius is a derivative of ter-tus, and the same is the case in the Sanscrit dvi-tiyas, tritiyas, and in the Sclavonic tretii, fern, tretiza. The Sclavonic relative kotoroia exhibits a similar extension of a form corresponding to Korepos. By the side of duo we have ambo, which is nearly synonymous with uterque. The distinction of these words is well known. While duo merely denotes an ag- gregate of two individuals the number "two" ambo signifies " both together" and uterque, "both the one and the other." This is clear from such passages as the following ; Ter. Adelph. I. 2, 50: Curemus sequam uterque partem; tu alterum, Ego alterum: -nam ambos curare propemodum Reposcere ilium est, quern dedisti. " Let both the one and the other of us look to his own : for to concern yourself with both together is almost to demand back again the boy whom you gave me." Auson. Ep. 91 : " vis ambas ut amem ? si diligit utraque vellem." " Do you wish me to love both together ? If both the one and the other loves me, I should be glad to do so." Hence it is clear that, as Doderlein says (Lat. Et. u. Syn. IV. 349), ambo regards the two as two halves, 328 PRONOUNS AND PRONOMINAL WORDS. [Cn. X. but uterque as two integral unities : and the former corresponds to a/u0w, the latter to e/care/jo?, and both in different cases to an(f>oTepos. The separability of the two constituent units in uterque is farther shown by the fact that this word may have either a singular or plural verb, whereas ambo always takes the plural. The formation of the degrees of comparison in adjectives and adverbs is intimately connected with that of the numerals. For all ordinals are of the nature of superlatives, and the most genuine form of the comparative in the Indo-Germanic languages is the combination of pronominal elements, which forms the third numeral, considered as indicating something beyond two. Although the Latin language is almost the only idiom which exhibits the full development of the separate usage of the form ter=ta-ra (New Crat. 157), for it has not only the numeral under the forms tres, ter t ter-nio, ter-tius, but also a noun ter- minus, and a regular preposition trans, it does not use -ter as a comparative suffix except in the case of pronominal forms. For all common words we have instead of -ter t -repos, -taras, which are so common in cognate languages, either the merely relative adjective in -ius, corresponding to the Sanscrit -tyas, Greek -to?, or a derivative from this in -ior, corresponding to the Sanscrit -fydns, Greek -KDV -LOV-S. Thus we have both al-ter and al-ius, and from the same root ul-tra, ul-tro. Many prepositions have a fixed or adverbial form in -tra, which is extended by the addition of -ior into an inflected comparative. Thus we have ci-tra, ci-ter-ior, ex-tra, ex-ter-ior, in-tra, in-ter-ior, ul-tra, ul-ter-ior, &c. The forms an-ter-ior, de-ter-ior, pos-ter-ior, show that there must have been originally derivatives like an- tra, de-tra, pos-tra> as well as the existing an-te, de,pos-t[e\; and we have seen that pos-tro is still extant in Umbrian. In some words the original affix was -ra only, as in inf-ra, sup-ra, whence inferior, superior. Some prepositions have no interme- diate adverb in -tra or -ra, but merely add the termination -ior, as prior from prce, propior from prope; and to this class we must add pejor for pes-ior, from per. All regular adjectives form their comparative in this way namely, by adding -ior to the crude form of the positive, as dur-us, dur-ior, facil-is, facil-ior, or, if the adjective involves a verbal root, from the crude form of the participle ; thus, the comparative of maledicus 6.] PRONOUNS AND PRONOMINAL WORDS. 329 is not maledicior, but maledicent-ior. There is no doubt that al-ius and med-ius are comparative words. The regular com- parative in -ior, gen. -wris, is formed from the genitive of these forms, as appears from the Sanscrit -tyans, Gr. -IMV=-IOV-S (New Crat. 165). As the ordinal admits of two forms in -tus and in -mus, and as the superlative is of the nature of an ordinal, we should expect that it would be indicated by one or both of these terminations. And this is the case. We have -mus alone in pri-mus, extre-mus, postre-mus, inji-mus or imus, and sum-mus for supi-mus. We have -ti-mus in ul-timus, in op-timus, " uppermost," from ob, in in-timus, " most inward," from in, in pes-simus (for pes-timus) "most down," from per (cf. pessum- do with per-do, and per-eo). The termination -timus is univer- sally assimilated in the superlatives of ordinary adjectives. For these superlatives are formed, like the comparatives in -tra, -Tepos, from an adverbial form, and not from the crude form of the adjective, like the comparatives in -ior (see New Crat. 165 ; Gr. Gr. Art. 269, sqq.). The adverb derived from the adjectives in -us or -er, which ended in e or o in ordinary Latin, originally terminated in -ed; and as the supines in -turn of dental verbs generally changed their t into s, or, in combi- nation with the characteristic, into -ss, we are not at a loss to account for the similar phenomenon in the superlatives : for ces- sum-ced-tum from cedo, and sessum^sed-tum from sedeo, fully correspond to dur-i-ssimus from dured-timus, and moll-i-ssimus from mollid-timus. The change of e into i in the former case is in accordance with the usual practice ; cf. teneo, con-tineo, sedeo, assideo, &c. When the crude form of the adjective ends in I or r, the t of -timus is assimilated to this letter : thus from celer we have celer-rimus for celer-timus, from facilis we have facil-limus for facil-timus. The junction between the crude form of the adjective and an affix properly appended to a derived adverb is due to the fact that adjectives of this kind may use their neuter and even their crude form as adverbs ; thus we have not only faciliter, but facile, and even facul (Festus, p. 87, Mulier). ' $ 7. Prepositions. The most important of the pronominal adverbs, which are used as the basis of degrees of comparison, are the prepositions. 330 PRONOUNS AND PRONOMINAL WORDS. [On. X. One of these, trans, is merely an extension of the affix of the comparative, and they are all employed more or less in qualifying those expressions of case, on which the mutual relations of words so much depend. We have seen that, according to the proper and original distinctions of the oblique cases, the genitive or ablative (for they were originally identical) denotes motion from a place, or, generally, separation ; the dative or locative implies rest in a place, or, generally, conjunction ; and the accusative signifies motion to a place, or, generally, approach with a view to conjunction ; but that these primitive uses of the oblique inflex- ions have become obsolete in Latin, with the exception of a few general nouns and the proper names of cities. In other instances, motion from and to, and rest in a place, together with the other mutual relations of words, are expressed by some preposition ; and in this use of the prepositions, the genitive, as distinct from the ablative, and the dative, whether identified with the locative or distinguished from it, are utterly excluded. The ablative alone is used with those prepositions which signify separation, and takes the place of the dative or locative with those which imply rest or conjunction, while the accusative properly accom- panies those which denote approach or motion. It will be convenient to class the Latin prepositions under three heads, corresponding to the three primitive distinctions of the oblique cases namely, separation or motion from, rest in, and approach or motion to. To each of these may be appended the derived or compounded prepositions, which introduce some new modification of meaning. The three simplest auxiliaries of the primitive relations of case are ab (shortened in a, and extended into abs, absque) for the expression of separation or motion from, with the ablative ; in for the expression of rest in or on, with the ablative, as the usurper of the place of the dative or locative ; and ad for the expression of approach or motion to with the accusative. There is no doubt as to the origin and linguistic affinities of these prepositions. Ab or abs corresponds in etymology and meaning to the Greek, awo or a\|/, which was originally av-iros, or va-7ros (New Crat. 169), and, as such, denoted motion from a distant object to the subject, according to the principle which I have stated and elucidated elsewhere (New Crat. 130, 169; Gr. Gr. Art. 77). Practically ab and airo denote motion from the 7.] PRONOUNS AND PRONOMINAL WORDS. 331 surface of an object, and are so distinguished from ex (e), e'f (e/c), which imply that we pass through intermediate proximity ; in corresponds in use to the Greek eV and e!s = ew, and in origin not only to these prepositions, but also to avd. In with the abla- tive and e v with the dative express the simplest and most elementary notion of locality the being in a place. With the accusative, in signifies into or unto a place, deriving the expres- sion of motion from the case with which it is connected. When ev is connected with the accusative in this sense, it is always ex- panded to eis = evg, except in some of the lyric poets, such as Pindar, who, like the Romans, use ev to express both location with the dative and motion with the accusative. There is no doubt that >, elv, eiv'i, avd, iva, are ultimately identical, the original form having been Fa-i/a, which expresses motion through the nearer to the more distant object. Practically, in represents all the uses of ev, e\j 7r6\i-=in urbe, et? rtjv 7ro\w = in urbem, dv-iipi6juios = in- numerus. The preposition ad is obviously another form of the conjunctions at = " still," and et = "too," "and." The late Pro- fessor Hunter showed 1 that there was the same relation between the Greek $e, which signifies " too," " in the second place," and the affix -$e, as in otoVSe, "to-home," implying motion to a place. We learn from the other form el-ra (New Crat. 193) that e-n is compounded of the second element Fa, and the third ; conse- quently it corresponds in etymology, as it does pretty nearly in meaning, to the Greek ets = evs, and to in used with the accusative. In its use with the ablative of the agent, ab corresponds rather to the Greek V7ro 9 than to dwo. Thus : mundus a deo creatus est would be rendered o xoajtos VTTO (not dwo) TOV 9eov eKTiffOfj. But we aro not to conclude from this that VTTO, dir6 9 are different forms of the same word. The u is found in all the cognate words VTTO, sub, virep, super, subter, uf, ufar, upa, upari; and it is clear that while CL-TTO = VCL-ITO, is com- pounded of the third and first, V-TTO = Fa-iro is made up of the second and first pronominal elements, and so denotes a passage 1 A Grammatical Essay on the nature^ import, and effect, of certain Con- junctions ; particularly the Greek 8e : read June 21, 1784. Trans, of the Royal Society of Edinburgh, Vol. I. pp. 113 34. 332 PRONOUNS AND PRONOMINAL WORDS. [Cn. X. to the subject from that which is proximate or under the feet. As the act of separation implies nearness at the moment of sepa- ration, we find that idiomatically ab is used to express relative positions, as : a fronte, " in front," a tergo, " behind," libertus a manu, " a freedman at hand," i. e. an amanuensis. But this meaning is more fully expressed by ap-ud, compounded of ab and ad, and combining the meaning of these two prepositions ; for apud signifies " being by the side of but not part of an object," and this implies both juxta-position and separation. It is used with the accusative, because this is the case of the latter preposition of the two, and because the passage from ab to ad implies motion. The Greek napd, which answers exactly to apud) takes different cases according to the meaning implied by the special reference (Gr. Gr. Art. 485). In low Latin we have the compound ab-ante from which comes the French a-vant, and even de-ab-ante from whence comes devant (see Pott, Zeitschr. /. d. Vergl. Sprf. I. p. 311). The preposition in has also the comparative forms in-ter and in-tra, or in-fra, which imply motion, and are consequently joined to the accusative. The same is the case with an-te, which retains the a found in an-ter, Sanscr. an-tar, Gr. a-Tep for av-rep (New Crat. 204). In meaning an-te corresponds to the Greek O.V-TL only so far as the latter signifies " in front of," which is the primitive signification of the Latin particle. The Greek Trpo, from whence comes Trio's, or Trpori, claims a com- mon origin with pro ; and there can be no doubt as to the connexion between napd, whence Trapai, and prce ; but there are many shades of meaning in which the Latin and Greek terms by no means coincide. Prce-ter, which is a comparative of prce, and prop-ter, which is similarly formed from pro-pe, an extension of pro (above, 5), express exactly certain meanings of wapd : thus -n-apd $6%av = prceter opinionem, and irapd ravra propter ista. Per exactly answers to Trapd, in its negative or depreciating sense, in compounds such as pe-jero forper-juro = TrapopKew : cf. pejor for perior. Although per and Trepl are identical words, there* are only some few cases in which their significations strictly correspond (see New Crat. 177, 8). It is perhaps still more difficult to show the exact relation in meaning between the Greek and Latin affix -nep, -per: cf. airep, oacnrep, &c. with paullisper, nuper, &c. In many of its employments the Latin 7.] PRONOUNS AND PRONOMINAL WORDS. 333 per coincides exactly with the Greek Sid, which, with the geni- tive, and, in the older poets, with the accusative also, signifies " through," and which, with the accusative in ordinary Greek, corresponds to the use of irapd, propter, to which I have just adverted. Etymologically there can be no doubt that Sid finds a representative in the Latin de, which implies descent and derivation, and is of course used with the ablative. It has been remarked already, that ab differs from ex, the other preposition most directly connected with the meaning of the ablative, by refer- ring to the surface of the object from which the separation takes place, whereas ex denotes a removal from or out of the interior of the object or objects. Now de also presumes that the thing removed was a part of the object from which it is removed. Thus while we have no ab-imo from emo, we have both ex-imo, " to take out/' and demo, " to take away a part " (as partem solido demere de die), to say nothing of sumo, " to take up," promo, " to take forth," which imply approximation to the same idea of partition. This signification of partition brings us back very closely to the primitive meaning of Sid, &, Svo ; and we have absolute division in such phrases as dedi de meo. From the same idea of partition we may get the sense of derivation and descent implied in these and other compounds of de. And here de comes into close contact with the affixes -0ei>, -tus, which un- doubtedly belong to the same original element (see New Crat. 263) ; thus de ccdo is exactly equivalent to cceli-tus. While Sid corresponds to per in its sense of " through," and to de in its meaning of division into parts, we find that de conversely coincides with Trepl in the sense of " about," " concerning," as denoting the subject from which the action or writing is derived, i. e. the source of agency or the subject-matter (v\r{). Thus scripsit de republica means " he took the subject of his writing from the general theme of the commonwealth ;" for which a Greek would have said : eypa^e ircpl Ttjs TroAtre/as, i. e. " his writing was about or derived from the republic/' The connexion of de and Sid is seen still more plainly in the form di or dis which the former bears in composition. As de, though connected with Sid, thus corresponds to one of the uses of irepi, while Sid in its general meaning coincides with per, so we find that ob, which is etymologically identical with djUL(pi, a synonym of Trepi, agrees in one of its uses with 334 PRONOUNS AND PRONOMINAL WORDS. [Cii. X. propter, and so with Sid when used with the accusative. The fact, that ob may be traced to a common origin with CTTI and afjiffti, has been elsewhere established (New Crat. 172, 3), by the following proofs. There can be no doubt as to the identity of eirl with the Sanscrit api and abhi. Now abhi is related to afji(bi 9 as abhra is to '6ju./3po$, abhau to a/i0o>, ambo, &c. And the analogy of O.TTO for av-Tro, shows that CTTI must originally have been ev-irl or aVwi = d/m-cpi. Moreover eiri and afj.(pi con- cur not only in their ordinary meanings, but especially in that sense of interchange or reciprocity which I have claimed for CTTI (New Crat. 174). Now 06, which resembles the Sanscrit abhi in its auslaut, shows by its vowel the last trace of a lost nasal ; comp. obba, umbo, a/u/3/f . And its usage, in other senses than that of propter, indicates a close connexion in meaning with ewl and ajULiKa\v7rTU) in oc-culo ; if ob-edio suggests 67ra/coyo>, ob-esus (bassus) refers us to aV0'Xa0>ys, ob-erro to 7repi7r\ava)/u.ai, and ob-liquus to a/jL(i). The preposition circum (circa, circiter), which is limited to the local or temporal meaning of irepi, is a case of the substantive circus, which may be connected with cis (citra), a form of the pronominal element -ce ; and ci-tra, citro are opposed to ul-tra, ul-tro, as ce = " here" is opposed to ul- (al- 9 an-, ^7-, in-) =" there," and there is no doubt that the preposition in is ultimately identical with the pronoun ul-, al- (cf. Sanscr. any a, Greek KCIVOS, &c.). The pronominal root ce obtains another prepositional extension in cum = %vv, and this again has its comparative in con-tra, " against," implying extension from and in front of that which is here. The first element po- combined with the second -s and the third -n gives in po[s~\ne a sense of extension "backwards" and "behind," i.e. through all three posi- tions ; and this is also the meaning of pos-t, which bears the same relation to po-ne that se-d or se-t does to si-ne. The latter, which is really po-s-ne without the first syllable, expresses the idea of simple separation. The compound post, or even the syllable po alone, is used as a preposition almost equivalent to trans, as in po-mcerum or post-moerium, "the space beyond the wall," post-liminium, " the space beyond the threshhold, within which a resumption of civic rights is possible." Trans, involving the elements of the comparative suffix, with a new affix, differs little from ul-tra, for it includes nearly the same elements in a dif- ferent order. As cir-cus is probably connected with cis, so ter- minus undoubtedly contains the root of tr-ans. A finis or ter- minus strictly excludes the citra as well as the ultra, and the circus, as a line, is neither the space, which it encloses, nor that, which it shuts out. Erga, which bears the same relation to ergo that ultra does to ultra, must be explained by the corre- spondence of ergo and igitur. The latter, as we have seen, is an extension in -tur = -tus of i-gi = es-gi ; and erg-a = esg-a is only a different form of the same word ; for the ending of igi-tur is -tur = -tim, and while circa stands by circi-ter we shall see directly that juxta presumes a juxtatim. It has been shown (in Chapter VIII.) that clam, coram, penes and tenus are adverbs derived from nominal or verbal roots; and juxta-jug-sta is a compound of the root jug- in jungo, jugum, jugis, and the crude form of sto. Like con-tinuo 3-36 PRONOUNS AND PRONOMINAL WORDS. [On. X. it expresses contiguity. Some consonantal affix, equivalent to a case-ending, is involved in the last syllable. The old gram- marians remark that " statim pro firmiter primam producit ; pro illico corripit ;" and such forms as stdtio, &c., prove that the contraction is not always exhibited. But the analogy of ai/a- /u.i'y-orjv, avct-fJiiy-ca, ai'a-fjuya, ava-/miq (Greek Grammar, Art. 265), shows that some affix was to be expected, and that it might be extenuated into a mere vocal auslaut. From the almost synonymous tenus and e%fjs, compared with the ablatives in a for ad, and with erga by the side of igi-tur, we can easily infer the nature of the appendage which has been rubbed off from the prepositional adverb jug-sta = jug-sta-tim. It may be worth while to add that prepositions compounded with verbs are liable to certain changes from assimilation or absorption, which perhaps typify a similar change in the separate use of these proclitic words. A f ab, abs may appear as au, and we have seen it assume the form af in old Latin (above, p. 221). Ad may change d into the first letter of the word with which it is compounded ; thus it may become ac, af, ag, al, an, op, ar, as, at; and we have seen that the last of these represents one of its separate usages ; compare also et, and the Greek en. Ante sometimes appears as antid, which may have been its original form (see above, p. 306). Circum may lose its final m or change it into n. Cum appears as com, co, col, con, or cor. De either remains unaltered, or assumes the form des before t ; it is found also with a different, but cognate signification, as dis- 9 di-, dif- and dir-. E, ex, enters into compounds either in its separate form, or assi- milated to /-, as in ef-fero. In is im before labials, i before g, il and ir before the liquids I and r, but otherwise unchanged; in old writers or their imitators we have endo or indu. Inter is not changed, except before t, when it becomes intelr. Ob becomes obs before dentals, it is assimilated to labials and gutturals, and is shortened into o before m; sometimes it resumes its original m : thus we have amb, shortened into am, or an before c, as in an-ceps. Per is sometimes, but not always, assimilated to I. 7.] PRONOUNS AND PRONOMINAL WORDS. 337 Post) or pone, becomes po, in pomoerium, pomeridianus. Pro is written prod before a vowel, as in prod-est ; it suffers metathesis in pol-liceo, por-rigo, where it approaches to the cognate per, if it is not identical with it. The inseparable re, really a form of in=avd, is written red be- fore a vowel, or the dentals d, t ; compare red-eo, red-do, ret-tuli. Sine, or sed, appears only as se. Sub may change b to the following letter, and sometimes as- sumes s before t, as in subs-traho. Trans may be shortened into tra. Ve, or vehe, is not a preposition, but a particle containing the same root as via=veha, velio, weg, &c. $ 8. Negative Particles. Negative particles fall into two main classes essentially dif- ferent in signification ; for they denote either denial, which is categorical negation, or prohibition, which is hypothetical nega- tion ; in the former case, we negative an affirmation, i. e. affirm that the case is not so ; in the latter, we negative a supposition, i. e. prohibit or forbid an assumed or possible event. As these differences are absolute in logic or syntax, it is necessary that they should be expressed by the forms of the words ; and the three classical languages have sufficient, but by no means iden- tical, methods of conveying these distinctions. The Greek lan- guage expresses categorical negation by the particle ov or OV-K, amounting to d-i/a-Fa-/c, which denotes distance and separation, but takes for the expression of a prohibition or negative hypo- thesis the particle fj.tj, which is connected with the first personal pronoun, and is therefore opposed to OUK as subject is to object (New Crat. 189). The Hebrew language has the same root $, which is ultimately identical with the Indo-Germanic na or a-na, to express both negation and prohibition ; but while the categorical negative tib conveys this idea by a lengthened stress on the vowel which follows the liquid, the hypothetical b$ denotes the prohibition of an act present or intended by an initial breathing which throws the emphasis on the anlaut (Maskil le-Sopher, p. 15). The Latin language, like the Hebrew, contents itself with one pronominal element, namely, n\ signifying " distance " and " separation," for both negation and 22 338 PRONOUNS AND PRONOMINAL WORDS. [Cii. X. prohibition, but distinguishes these in form by adopting a com- pound or lengthened word for the categorical negative, while the hypothetical word appears without any such strengthening addition. Thus, while the common expression for the cate- gorical negative is non for nenu or noenu, which is obviously ne cenum or ne unum with the ecthlipsis of the final m, we find merely ne in the prohibitive sense, in ordinary Latin. There are traces in single words and in the older authors of a strengthening affix c in this latter use (above, p. 98), corresponding to the affix which appears in OV-K or ov-%i. We must distinguish this affix from the conjunction -que, which appears in the disjunction ne-que (Muller, Suppl. Ann. ad Fest. p. 387). If, then, we compare oi)-/c=a-j/-Fa-: with ne-c, we shall see that they differ only in the inserted element Fa, and there is no reason to suppose that the categorical rfon differs from the hypothetical ne, other- wise than by the strengthening word unum, which is also in- volved in nullus = n'unu-lus. On the other hand, we see from the categorical use of n'unquam, riusquam, ne~quidem and ne- que, that the negative ne may always be used in a denial of facts, if it is only sufficiently strengthened. The identity of d-i>a-[Fa]-/c and ne-c is farther shown by the use of the negative as a prefix in Latin. Of this we have three forms ; the simple ne or rii as in ne-fas, ne-scio, ni-hil, ni-si, &c. ; the same with i = Fa pre- fixed, as in in^-iquus, in-numerus, im-mensus, i-gnavus, &c. ; with c affixed, as in nec-opinus, neg^otium, neg-ligo or nec-ligo. As it is quite clear that in these instances the element n is that which gives the negative force, and as this element is common to n'on and ne, it follows that the Romans did not distinguish between the form of the prohibition and categorical negation otherwise than by strengthening the latter. And this extenuation of the negative emphasis in subordinate expressions is also shown by the fact, that, in conditional and final sentences, the mere dimi- nution of assertion expressed by minus took the place of the shorter negative; thus we have si minus for sin, and quominus for quin. It is a question whether the shorter form ne can appear without some strengthening affix, as -dum, -que, or quidem, in the categorical negation. Of the passages quoted some are manifestly corrupt, and it seems that ne is not used categorically, except when it stands for ne-quidem, " not even " (see Drakenborch, ad Liv. VIII. 4 ; XXXIII. 49). It may be S.] PRONOUNS AND PRONOMINAL WORDS. 339 doubted in these cases whether there is not a concealed prohi- bition, as in the Greek rf on. On the other hand, when non appears, as it occasionally does, in a final sentence, there is always some reason for the employment of this more emphatical par- ticle. Thus ne plura dicam, or ut ne plura dicam, means merely "not to say more," but ut plura non dicam neque alio- rum exempli* confirmem (Cic. pro lege Manil. 15, 44) implies a more deliberate abstinence from irrelevant details. The dif- ference between ne-quidem and non-quidem or nec-quidem con- sists in the greater degree of emphasis conveyed by the former, which is much the more usual combination ; for ne-quidem means " not even ;" but 'non (or nee) -quidem denotes merely a qualifi- cation of opposed terms, so that quidem is simply the Greek yue'r; : this appears from Quintilian's rendering (IX. 3, f 55) of Demosthenes (de Corona, p. 288) : OVK elirov p*v Tavra, OVK eypa\}sa Se' ovc? eypa^a jue^, OVK eTrpevfievcra Se m ovo' eTrpea- fievaa nev, OVK eTreiaa ce 9>;/3a/oi/?, " non enim dixi quidem, sed non scripsi ; nee scripsi quidem, sed non obii legationem ; nee obii quidem, sed non persuasi Thebanis :" (see Wagner on Virg. Georg. I. 126). This distinction in emphasis regulates the employment of the negative particles in interrogations, and we observe the same relation between the Greek and Latin particles in this use also that is, we employ nonne in Latin, where we write ap ov in Greek ; num, which bears the same relation to ne that ipsus does to ipse or necessum to necesse, corresponds to the Greek use of /my or /XT} ovv=fjiv > and the enclitic -ne is used when no nega- tion appears in Greek ; thus we have : ap' OVK ecrnv a . . < je lus 2 m , - . . . some other time ., . , The Pluperfect . . .anteriority } I /own* CM*. 1 pendant que vous ecriviez. 2 apres que vous eutes fini d'ecrire. 3 avant que vous eussiez e"crit. 3.] THE THEORY OF THE LATIN VERB. 345 of composite tenses which is peculiar to the Latin language, and still more so by the irregular use of the affix -s to express derived or indefinite tenses. $ 4. The Substantive Verbs. Before I proceed to examine the tense-system of the Romans, as it appears in all the complications of an ordinary verb, it will be as well to analyse, in the first instance, the substantive verb which enters so largely into all temporal relations. The Latin language has two verbs signifying " to be :" one contains the root es-, Sanscr. as-, Greek eo--, Lith. es- ; the other, the root/if-, Saner. 6M-, Gr. u- represents the ultimate form of the constituent labial. In the ordinary forms of the Greek, the transitive (puoo, u 0aF, and fu =fac for faf, "to bring to light," or "cause to be." TRANSITIVE. Pres. utf~oTs fceminus = fui-minus (cf. fcemind) fuius =Ji'lius. The omission of iya in e with $IKVVJULV. It will be seen at once that the Latin verb is much more complete than the Greek : and besides these forms, which admit of direct compa- rison, the Latin neuter verb has a present subjunctive fuam = fu-iam, a pluperfect indicative fu-eram =fuesam, a perfect sub- junctive fuerim (or fuero) fuve-sim, and a corresponding plu- perfect fuissem = fuve-se-sim. The s = r, which appears in the last three of these forms, is best explained by a comparative analysis. of iretpvKa and fui fufui. As i is the regular ex- ponent of guttural vocalization, as the guttural, before it subsides into i, is generally softened into s and h, and as we find A;, s, h in the perfect and aorist of Greek verbs, we see that 7repev, 8co f^coKev av = si quid habuisset, dedisset. The distinction between cases (3) and (4) is also observed in the expres- sion of a wish : thus, utlnam salvus sis ! pronounces no opinion respect- ing the health of the party addressed; but utinam salvus esses! implies that he is no longer in good health. 358 THE THEORY OF THE LATIN VERB. [Cn. XL case: si hoc nunc vociferari velim, me dies, vox, later a deficiant ; where we should have in Greek : et TOUTO ev no TrapavriKa yeywve7v eOcXoijui, we pa? av IJLOL Kctl (t)i>fjs /ecu aQevovs evSeriveiGv. In the fourth case : (a) si scirem, dicerem = el rfirLa-TafjLrjv, eXeyov av. (b) si voluissem plura, non ne- gasses = ei 7r\eovwv eTre^J^ui/cra, OVK av ypvrjaw. And this confusion becomes greater still, when, by a rhetorical figure, the impossible is supposed possible ; as in Ter. Andr. II. 1, 10 : tu si hie sis, aliter sentias. For in this instance the only differ- ence between the two cases, which is one of tense, is overlooked. In the apodosis of case 4, b, the Romans sometimes used the plusquam-perfectum of the indicative, as in Seneca, de Ira, I. 11 : perierat imperium, si Fabius tantum ausus esset^ quantum ira suadebat; and Horace, II. Carm. 17, 27: me truncus illapsus cerebro sustulerat, nisi Faunus ictum dextra levasset. Some- times the perfect was used in this apodosis, as in Juvenal, X. 123 : Antoni gladios potuit contemnere, si sic omnia dixisset ; or even the imperfect, as in Tacitus, Annal. XII. 39 : nee ideo fugam sistebat, ni legiones pugnam excepissent. Again, particles of time, like donee, require the subjunctive when future time is spoken of; as in Hor. I. Epist. 20, 10 : earns eris Romce, donee te deserat cetas. But this becomes a past tense of the indicative when past time is referred to ; as in Hor. I. Epist. 10, 36 : cervus equum pellebat donee [equus] imploravit opes hominis frcenumque recepit. The confusion between the Latin indicative and subjunctive is also shown by the use of the subjunctive pre- sent as a future indicative (a phenomenon equally remarkable in Greek, New Crat. 393), and conversely by the employment of the periphrastic future (which is, after all, the same kind of form as the ordinary composite form of the future indicative) as an equivalent for a tense of the subjunctive mood. Thus Cicero uses dicam and dicere instituo in the same construction ; Phil. 1. 1 : " antequam de republica dicam ea, quae dicenda hoc tempore arbitror, exponam breviter consilium profectionis mese." Pro Murena, 1 : "antequam pro L. Murena dicere instituo, pro me ipso pauca dicam." And we have always the indica- tive in apodosis to the subjunctive when the future in -rus is used : e. g. Liv. XXXVIII. 47 : " si tribuni prohiberent, testes citaturus fui" (for " citarem") ; and Cic. Verr. III. 52: "illi ipsi aratores, qui remanserant, relicturi omnes agros erant" $ 11.] THE THEORY OF THE LATIN VERB. 359 (for "reliquissent"), "nisi ad eos Metellus Roma literas mi- sisset." The Romans also used the perfect subjunctive exactly as the Greeks used their perfect indicative with Kal Stj in sup- positions. On the whole, it must be confessed that the Latin sub- junctive, meaning by that term the set of tenses which are formed by the insertion of -i-, differs modally from the indicative only in this, that it is uniformly employed in dependent clauses where the idiom of the language repudiates the indicative ; and it is not a little remarkable, that in almost all these cases in all, except when final particles are used, or when an indirect question follows a past tense the indicative is expressly required in Greek syntax. The title subjunctive, therefore, does but partially characterise the Latin tenses in -i-; and their right to a separate modal classification is scarcely less doubtful than that of the Greek optative as distinguished from the conjunctive. The differences between the indicative, imperative, and infi- nitive equally exist between the two latter and the subjunctive. The indicative and subjunctive alone possess a complete appa- ratus of person-endings ; the imperative being sometimes merely the crude form of the verb, and the infinitive being strictly impersonal. $ 12. Forms of the Infinitive and Participle- how con- nected in derivation and meaning. He who would investigate accurately the forms of the Latin language must always regard the infinitive as standing in intimate connexion with the participles. There are, in fact, three distinct forms of the Latin infinitive : (a) the residuum of an abstractum verbale in -sis, which remains uninflected ; (6) a similar verbal in -tus, of which two cases are employed; (c) the participial word in -ndus, which is used both as three cases of the infinitive governing the object of the verb, and also as an adjective in concord with the object. There are also three forms of the participle : (a) one in -ns= -nts, sometimes lengthened into -ndus; (ft) another in -tus ; and a third (7) in -turns. The participle in -ns is always active ; its by-form in -ndus is properly active, though it often seems to be passive. The participle in -tus is always passive, except when derived from a deponent verb, 360 THE THEORY OF THE LATIN VERB. [On. XL in which case it corresponds in meaning to the Greek aorist middle. The participle in -turns is always active and future. It is, in fact, an extension of the noun of agency in -tor ; com- pare prcetor, prcetura ; scriptor, scriptura, &c. with the corre- sponding future in -turns of prceo, scribo, &c. (see New Crat. 267). The Greek future participle is sometimes used as a mere expression of agency ; thus we have in Soph. Antig. 261 : oi>$ o KU}\V(T(*)V Trapijv. Aristot. Eth. NIC. II. 1, ^ 7 : ovcev av e$ei rou $i<)a%oi>Tos where we should use the mere nouns of agency " the make-peace " " the teacher." Now it is impossible to take an instructive view of these forms without considering them together. The participle in -turns (7) is a derivative from the verbal in -tus (b) ; and it would be difficult to avoid identifying the participle in -ndus and the corresponding gerundial infinitive. In the following remarks, therefore, I shall presume, what has been proved elsewhere (New Crat. 416), the original identity of the infinitive and the par- ticiple. That the verbal (a), which acts as the ordinary infinitive in re-se, is derived from the crude form of the verb by the addi- tion of a pronominal ending si- or sy-, is clear, no less from the analogy of the JSolic Greek forms in -19, where the i is trans- posed (com p. N. Crat. 410, (3)), than from the original form of the passive, which is -rier~-syer, and not merely -rer. This infinitive, therefore, is the indeclinable state of a derivative precisely similar to the Greek nouns in -0-19 (7rpai?, jo^-cri?, &c.), which express the action of the verb. This Greek ending in -o-*9 appears to have been the same in effect as another ending in -rJs, which, however, is of less frequent occurrence (eV^-Ti/s, iiq-rvst opxriff-Tik) &c.), but which may be compared with the Latin infinitive (b) in -turn, -tu, (the supine, as it is called), and with the Sanscrit gerund in -tvd. The verbal in -tus, which is assumed as the origin of these supines, must be carefully distin- guished from the passive participle (/3) in -tus. For it appears, from forms like venum, &c., and the Oscan infinitives moltaum, &c., that the t of the supine is not organic, but that the infinitive (b) is formed like the infinitive (a) by a suffix belonging to the second pronominal element, so that the labial (u = v) is an essential part of the ending. On the other hand, the participle (/3) has merely a dental suffix derived from the third pronomi- 12.] THE THEORY OF THE LATIN VERB. 361 nal element, and corresponding to the Greek endings in -ros, -1/09, and the Latin -tus--nus. In fact, the suffix of infinitive (6) is tv = Fa or va, while that of participle (/3) is t- only. 13. The GERUNDIUM and GERUNDIVUM shown to be active and present. The infinitive (c) and the participle (a) are, in fact, different, or apparently different, applications of one and the same form. In its infinitive use this verbal in -ndus is called by two names the gerundium when it governs the object of the verb, and the gerundivum when it agrees with the object. Thus, in " con- silium capiendi urbem," we have a gerundium ; in " consilium urbis capiendce" a gerundivum. As participles, the ordinary grammatical nomenclature most incorrectly distinguishes the form in -ndus as " a future passive/* from the form -n[_t]s con- sidered as " a present active." The form in -ndus is never a future, and it bears no resemblance to the passive inform. The real difficulty is to explain to the student the seeming alternation of an active and passive meaning in these forms. Perhaps there is no better way of doing this than by directing attention to the fact, that the difference between active and passive really be- comes evanescent in the infinitive use of a verb. " He is a man to love"="he is a man to be loved ;" " I give you this to eat" = "I give you this to be eaten," &C. 1 The Greek active infini- tives in -/xei'm, -rat, are really passive forms in their inflected use 2 ; and that the Latin forms in -ndus, which seem to be 1 Wo observe the same fact in the use of the participles in English and German. Thus, in Herefordshire, "a good-leapt horse" means "a good- leaping horse;" and in German there is no perceptible difference between kam ge.itten and Team reitend. See Mr. Lewis's Glossary of Provincial Words used in Herefordshire, p. 58; and Grimm, D. Or. IV. p. 129. 2 Conversely, the forms in -VT-, which are always active when used in concord with a noun, are occasionally employed in that infinitive sense in which the differences of voice seem to be neglected. Thus we have, Soph. Aj. 579 : dprjvelv eVroSas npos TOjj,a>vTi nijpaTi (" ad vulnus quod secturam desideret" s. " secandum sit"). (Ed. Col. 1219 : orav TIS es jr\eov TTfa-rj TOV 6e\ovTos (" quando quis cupiendi satietatem expleverit" s. "id quod cupiebat plene consecutus fuerit"). Thucyd. I. 36: yi>tora> TO /** dedibs avTov roi/y Ivavriovs /zaXAoi/ ^o^a-ov {" sciat timer e illud suum majorem adversaries metum incussurum esse"). 362 THE THEORY OF THE LATIN VERB. [Cn. XI. passive in their use as gerundiva, are really only secondary forms of the participle in -n[f]s, appears not only from etymo- logical considerations (New Crat. 415), but also from their use both as active infinitives and active participles. When the gerundivum is apparently passive, it seems to attach to itself the sense of duty or obligation. Thus, we should translate delenda est Carthago, " Carthage is to be destroyed" =" we ought to destroy Carthage ;" and no one has taken the trouble to inquire whether this oportet is really contained in the gerundivum. If it is, all attempts at explanation must be unavailing. But since it is not necessary to seek in the participial form this notion, which may be conveyed by the substantive verb (e. g. sapientis est seipsum nosse), it is surely better to connect the gerundivum with the gerundium, and to reconcile the use of the one with the ordinary force of the other. Supposing, therefore, that da-ndus is a secondary form of da-n[t~\s, and synonymous with it, on the analogy of Acraga[nt]s, Agrige-ntum ; orie-n[f]s, oriu-ndus; &c. ; how do we get the phrase da-nda est occasio, " an oppor- tunity is to be given," from d-a-ndus=dan\f\s, "giving?" Simply from the gerundial or infinitive use of the participle. Thus, (A) da-ndus =da-n[f]s signifies " giving ;" (B) this, used as an infinitive, still retains its active signification, for ad dandum opes means "for giving riches"="to give riches;" (c) when this is attracted into the case of the object, the sense is not altered, for ad opes dandas is precisely equivalent to ad dandum opes ; (D) when, however, this attraction appears in the nomi- native case, the error at once takes root, and no one is willing to see that it is still merely an attraction from the infinitive or indeclinable use of the participle. Even here, however, the intransitive verb enables us to bring back the student to a con- sideration of the real principle. For one can hardly fail to see that vivendum est=vwere est i. q. oportet vivere ; and that there may be no doubt as to the identity of the uninflected with the inflected gerund in this case, Horace has put them together in the same sentence : " nunc est bibendum, nunc pede libero pul- sanda tellus," where it is obvious that tellus pulsanda est is no less equivalent to " oportet pulsare tellurem," than " bibendum est " is to " oportet bibere." At all events, his Greek original expressed both notions by the infinitive with 13.] THE THEORY OF THE LATIN VERB. 363 vvv xpi) fj,e6v(r6^v KO.I TWO. Trpbs fBiav Trivrjv, 7rei6/} KarOave MupcriXoy. (Alcseus, Fr. 20. p. 575, Bergk.) The strongest proof, that the involved meaning of the gerun- divum is strictly that of the active verb, is furnished by the well-known fact that the attracted form is regularly preferred to the gerund in -di, -do, +dum governing the case, when the verb of the gerund requires an accusative case; thus we have: ad tolerandos rather than ad tolerandum, labores; consuetudo homi- num immolandorum rather than homines immolandi; triumviri reipublicce constituendce rather than constituendo rempublicam. Indeed this is rarely departed from, except when two gerunds of a different construction occur in the same sentence, as in Sail. Cat. 4 : " neque vero agrum colendo aut venando, servi- libus officiis, intentum sotatem agere," because venando has nothing to do with agrum. The student might be led to suppose at first sight that the phrase: lex de pecuniis repetundis, "a law about extortion," literally denoted " a law concerning money to be refunded," and that therefore the gerundivum was passive in signification. But this gerundivum is used only in the genitive and ablative plural, to agree with pecuniarum and pecuniis, and we happen to have a passage of Tacitus (Annal. XIII. 33) which proves that the verbal is transitive : for the words : a quo Lycii repetebant are immediately followed by : lege repetundarum dam- natus est ; and thus we see that lex de pecuniis repetundis does not mean "a law concerning money to be refunded," but, "a law which provides for the redemanding of money illegally exacted." This view of the case appears to me to remove most of the difficulties and confusions by which the subject of the gerund has hitherto been encumbered. There are three supplementary considerations which deserve to be adduced. The first is, that in the particular case where the gerundivum appears to be most emphatically passive namely, when it implies that a thing is given out or commissioned to be done it is found by the side of the active infinitive : thus, while we have such phrases as: " Anti- gonus Eumenem mortmain propinquis sepeliendum tradidit" (Corn. Nep. Eum. 13), we have by their side such as: "tristi- tiam et metus tradam protervis in mare Creticum portare ventis" (Hor. I. Carm. 26, 1). That the gerund in this case is really present, as well as active, appears from its opposition to the use 364 THE THEORY OF THE LATIN VERB. [Cn. XI. of the past participle ; thus : hoc faciundum curabo means " I will provide for the doing of this:" hoc factum volo means "I wish it were already done." The second point to be noticed is that deponent verbs, which have no passive voice, employ the gerundivum in the attributive use, which, we are told, cannot easily be wrested to an active signification ; as : prcelia conju- gibus loquenda, " battles for wives to speak of." The third case is this ; that the supines, which are only different cases of one and the same verbal, appear as active infinitives when the accusative is used (-turn), and as passive when the ablative is em- ployed (-tu). Now, this seemingly passive use of the supine in -tu arises from the fact, that it appears only by the side of adjectives, in which case the active and passive forms of the infinitive are often used indifferently, and some adjectives take the supine in -tu when they expressly require an active infinitive, as in : " difficile est dictu (-dicere), quanto opere conciliet homines comitas affa- bilitasque sermonis " (Cic. Off. II. 14). Now this supine, which is thus identical with the infinitive active, frequently alternates with the gerund ; compare, for instance: quid est tamjucundum auditu (Cic. de Or. I. 8), with: verba ad audiendum jucunda (id. ibid. I. 49). The active sense of the verbal in -tus = -sus is equally apparent in the dative case : thus we find such phrases as (Sal- lust, Jugurth. 24) : " quoniam eo natus sum ut Jugurthai sce- lerum ostentui essem," i. e. " since I have been born to serve as an exhibition of (=to exhibit) the wickedness of Jugurtha." But the form in -ndus is not only active in voice, but also, as has been mentioned, present in tense. Thus, if we take a depo- nent verb, we often find a form in -ndus acting as a collateral to the common form in -w[tf]s, and opposed with it to the form in -tus. For instance, secundus and sequen[f]s both signify " fol- lowing," but secutus " having followed." The same is the distinction between morien[f]s, moriundus ; orien[t]s, oriundus; irascen[t]s, ira[s]cundus ; &c., on the one hand, and mortuus, ortus, iratus, &c., on the other. This cannot be remarked in active verbs, because the Latin language has no active past par- ticiple. If, however, we turn to the gerund ial use of the form in -ndus, we may observe a distinction of tense between it and the participle in -tus even in the case of active verbs. Thus volvendus is really a present tense in Virgil, .sEneid. IX. 7 : volvenda dies, en, attulit ultro ; comp. Ennius (apud Varro. L. L. 13.] THE THEORY OF THE LATIN VERB. 365 VII. 104, p. 160, Miiller), and Lucretius, V. 1275 ; because, in its inflected form, it is equivalent in meaning to volvendo; and the following passages show that the gerund is equivalent to the present participle : Virgil, Georg. II. 225 : " multa virum volvens durando saecula vincit;" Lucret. I. 203: " multaque vivendo vitalia vincere saecla;" and id. III. 961: "omnia si pergas vi- vendo vincere ssecla." And the words of Livy (prcef. ad Hist.) : " quse ante conditam condendamve urbem traduntur," can only mean "traditions derived from a period when the city was nei- ther built nor building" 14. The Participle in -turns. The participle (y) in -rus or -urus, which always bears a future signification, is supported by an analogy in the Latin lan- guage which has no parallel either in Greek or Sanscrit. The Greek desiderative is formed from the ordinary future by the insertion of the element i- : thus pa-w, fut. Spa-eta, desiderative $pa-&ei(t>. This desiderative is the common future in Sanscrit; though the Vedas have a future, like the Greek, formed by the element s- only, without the addition of i- 1 . Now the regular future of scribo would be scrip-so, indicated by the aorist scripsi; but the desiderative is scripturio. We may infer, then, that in the loss of the regular future of the Latin verb, the desiderative and future participle have been formed by the addition of the future r = s and the desiderative ri = si, not to the crude form of the verb, but to the verbal in -tus, so that the desiderative is deduced immediately from the future participle in -tur-us or from the noun of agency in -tor (above, p. 360). 15. The Perfect Subjunctive. We have seen above ( 4) that the form fuerim =fuesim is really a subjunctive tense of the usual kind derived from the perfect indicative fui =fuesa. As, however, the first person is occasionally written fuero, just as sim = esim or erim is short- ened into ero, it has been common among grammarians to ima- gine two tenses as distinct as ero and sim. But this view is represented under two different forms : for while the older gram- i See Rosen, on the Rig-V&da Sarihita, p. iv. 366 THE THEORY OF THE LATIN VERB. [On. XI. mars make/wm'm andfuero two tenses of the subjunctive mood, the former being perfect, and the latter future, the more modern writers on the subject increase the confusion by referring the latter, as a futurum exactum, to the indicative mood, while the former retains its place as perfect subjunctive. Those, who have had any thing to do with the business of teaching the Latin language, need not be told that a young and thoughtful student will not derive much edification from the doctrine that fuerit is both indicative and subjunctive, both past &nd future. And those who are conversant with the higher kind of philology, know that, while fuero and fuerim are merely euphonic distinctions, all the other persons, having only one set of meanings, are necessarily inflexions of the same form. "With regard to the signification of this perfect subjunctive, it is clear that, as it is formed from the perfect indicative just as the present subjunctive is formed from the present indicative, it must exhibit the same modification of meaning. Now dicam = dic-yam means "there is a proba- bility of my speaking;'* consequently dixero = dic-se-rim must mean, " there is a probability of my having spoken ;" and in proportion as the former approximates to the predication, " I shall speak," in the same proportion does the latter express, " I shall have spoken." In strictness that which is called & futurum exactum, or paulo-post-futurum, can only exist in forms derived from the perfects of intransitive verbs. These forms exist in Greek both with the active and with the middle inflexions ; thus from OvjffKo), " I am dying," reOi'rjKct, " I am dead," we have TeOv^oiJ-ai or reO^co, " I shall have died," i. e. " I shall be found in the state of death ;" from ypd(pco, " I am writing," we have yeypa(f)a, " I have written," yeypaimimai, " I have been written," i.e. " I stand or remain written," yeypa^o/mai, " I shall have been written," i. e. "I shall stand and remain written." Now it has been observed even by the old grammarians, that the Romans did not use these futures of the intransitive or passive perfect. Thus Priscian says (Let. VIII. c. 8. p. 388, Krehl) : " quamvis Grseci futurum quoque diviserunt in quibusdam verbis, in futurum infinitum, ut ri5\|/o^at, et paulo post futurum, ut Terv^ojULdi, melius tamen Romani considerata futuri ratione, quae omnino incerta est, simplici in eo voce utuntur, nee finiunt spatium futuri." But if the Romans had no futurum exactum of the passive form, still less would they have one with active 15.] THE THEORY OF THE LATIN VERB. 367 inflexions. The question of moods, as we have seen above, is not one of forms, but one of syntactical usage. And if we wish to inquire whether there is any justification for those who place fuero in the indicative mood, we have only to ascertain whether there is really any difference in syntactical usage between this form and fuerim, and generally, whether the tense, which we call perfect subjunctive, is ever used as an indicative, that is, as a categorical predication, without any reference to a protasis, expressed or plainly implied. The confusion, into which some modern grammarians have fallen in regard to this tense, has arisen entirely from the use of the Latin subjunctive in the apodosis, without a qualifying particle of reference like the Greek av. Hence the imperfect grammarian is extremely liable to confuse between a categorical and a consequential assertion, where the protasis is omitted; and while the Greek optative, with av 9 is rendered by the future indicative, without any risk of a misunderstanding as to the logical intention of the phrase, the perfect subjunctive in Latin has been supposed to be merely a future indicative referring to completed action. The following comparison -will show that there is no use of the tense now under consideration, which may not be referred to some parallel em- ployment of the Greek conjunctive or optative aorist. . , ,/ a / . i fhabeas 1 , , . a. eav TI e^j?s, vwcreis = si quid < , , , . > , dabis. b. eav TI &xf]S> Sweets = si quid habueris, dabis. c. ei TI e^ot?, $i$oi rj$ av = si quid habeas, des. d. el TL Gxpiris, $oi w " = si quid habueris, dederis. If in the second and fourth cases habueris and dederis are subjunctive or potential, the same explanation must apply to the following : a. si plane occidimus, ego omnibus meis exitio fuero, " if we have altogether fallen, I shall have been (i. e. I shall prove in the result, yevoi^rjv av) a destruction to all my friends." b. si pergis, abiero, "if you go on, I shall have departed (i. e. I shall go at once, aTreXOoifj.' av)" c. tu invita mulieres ; ego accivero pueros, " do you invite the ladies ; after that, when you have done so, I shall be found to have sent for the boys (, &c., as Grimm has shown in a recent paper on the subject (Abh. Ak. Berl. 1848 : " uber schenken und geben "). The other verb, which appears to belong to the -a conjugation, but has a reduplicated perfect, is sto, which makes A. III. st&ti. This verb does, not give the same indications as do of a mere articulation-vowel ; for even the compounds retain the long a, which appears in stdbat, &c. But we have a by-form, si-sto, to which steti may be referred, just as our transitive "stay," intransitive "stand," are represented by the German present stehe, perf. stand, both of which are intransitive. And I am inclined to explain the long a in sto, as resulting from a contraction of staho = steyo, Germ, stehen, which is still found in the Umbrian stahito = stato (above, p. 82). So that sto can- not be considered as a verb, of which the characteristic or for- mative adjunct is -a, but, like do, owes its contraction to the con- tact of the root-syllable with the termination. With these two ex- ceptions, all -a verbs form their perfect in -ui or -vi. Although the Greek vowel-verbs particularly affect the aorist in -, which have a short a before the - which are also written -uo (New Crat. 432, y\ so there are many special reasons for inferring the presence of this auxi- liary in the Latin verbs in -eo. Perhaps the most important of these special reasons is suggested by the phenomenon that many active verbs in Latin, either (a) uncontracted, or (b) contracted in -a, have a neuter or passive verb from the same root distin- guished by the formative characteristic e; thus we have (a) 3.] THE LATIN CONJUGATIONS. 379 active jacere, passive jacere; active pandZre, passive pater e ; active pendere, passive pendere ; active scandere, passive scatere; (b) active liquare, passive liquere ; active pardre, parere, pas- sive parere; active seddre, passive sedere. Now it is well known that the insertion of ya between the root and the ending forms the passive voice in Sanscrit (New Crat. 379), and I have shown (ibid. 381) that a similar explanation is applicable to the Greek passive aorists in -Orjv and -tjv; and as one of these aorists is e Sclav, mjeshu, Polish mieszam, Bohe- \L s*~ mian misyti, Russian s-mjeshat\ Persian JLsr^ 0. H. G. misc-jan, Lith. maiszyti, Gael, measgaim, Sanscr. mif-ra, &c. From the extreme antiquity and universal prevalence of this compound root, and from the formative affix with which it appears as a verb in most of the Indo- Germanic languages, it is fair to conclude that its origin is to be sought in a pronominal combination analogous in meaning and form to the Irish measg, " among," " between,'* Welsh ym-musk, Greek /ue-ra, /uL-cr(f>a, /uLe-^pi 9 fjieacros, Lat. me-dius, Hebrew Ip/T^l, which would serve as a sufficient basis for such a causative verb. It has been mentioned above (p. 76), in a general way, that deb-eo is con- nected with the important Semitic and Sclavonian root liD, dhob, and dob, signifying " good." But it will be necessary in this place to justify this comparison with especial reference to the formative syllable of the conjugation. In its impersonal use, oportet corresponds to the personal and impersonal use of debeo, and as the former is clearly connected with opus, so the latter expresses, as Forcellini says, rationem officii, convenire, oportere, obstrictum esse ad aliquid faciendum. In both, the ideas of interest and duty are mixed up, and in general, when we say that it is good for us to do anything, we combine in one notion the thought of a moral fitness or propriety and that of an advantage to be gained. We feel that we owe it to ourselves, when we feel that we owe it to our principles or to our fellow- men. Hence, being in debt, which is the reverse of a good thing, is expressed by an application of the verb, which conveys the idea of justice or moral obligation, just as officium, " duty," belongs to the same family with officit, or obest, "it harms." In English we have only one word for what we " owe" and what we " ought to do ;" and the German sollen, " to be in duty bound" (connected with our "shall," and "should"), be- longs to the same root as schuld, "a debt." The Greek phrase &Kaio9 eijuLi TOUTO TroieTi/, " I am in justice bound to do this"= " I ought to do it," shows how the two ideas run into one another. But the most decisive illustration of the etymology of deb-eo is fur- nished by the affinity between the Greek o-0e'XXw, " to increase," "enlarge," "benefit," "aggrandize." o-0eXos, "advantage," " help," "profit,'* fc)-0eXeo>, " to be of service" (all from the root 382 THE LATIN CONJUGATIONS. [On. XII. phel-, "to swell," and all showing the ordinary meaning of yi and dob), and their derivatives o0Xt-<7/c-a-i>a>, " to incur an obligation," and o-$e/Xo>=o-, "to owe," the impersonal use of which ofaiXei, " it is fitting," reverts to the meaning of the other class of words and of the Latin oportet and opus est. As then o-0e/Xft)=o0eX-yw, with the same pronominal adjunct ya, forms the expression of duty from that of advantage, so deb-eo by the same machinery passes to the same extension of the primitive dob, " a fitting time," dob-ro, "good, useful," &c. $ 4. The third or -i Conjugation. The best general rule for distinguishing between the verbs in -io, which belong to the vowel-conjugation, and those which have for their characteristic the letter i considered as a semi- consonant, or vocalization of a guttural, has been already given ( 1). With regard to their origin and analysis, we must con- sider the former as an extension of the -e conjugation, and while the vowel-verbs in -io will thus represent a set of derivatives in which a crude form in -i is strengthened by the affix -ya, in which case there will always be a contraction, the semi-conso- nantal verbs, which outwardly resemble them, merely strengthen the present and its immediate offspring with a vocalized guttural, to which the person-endings are attached without any inter- mediate agency. Thus, as we shall see in the next chapter, all verbs of the third conjugation are derived from nouns actually existing in -i, or which may be inferred from the inflexions of existing nouns, while the semi-consonant verbs have no such primitives. We see the manner in which the second conjugation is included in the third, from a verb of the second conjugation, of which the root happens to end in the vowel -i, and which, there- fore, is liable to the double contraction observable in all genuine i verbs. From the root ci- (Greek K/-OJ) we have, with an entire correspondence of meaning, two forms ci-eo and ci-o, and as the perfect is always ctvi, we must consider the latter as a condensation of the former. The great peculiarity of this verb is that its participle (E. III.) is indifferently citus or citus, the latter being found not only in compounds like concttus, incitus, percitus, but also in the simple form citus, both when it is used as a participle, as in Virgil (^Eneid. VIII. 642) : Haud procul inde citce Metium in diyersa quadrigse Distulerant, 4-1 THE LATIN CONJUGATIONS. 383 where we must take citce with in diver sa, " chariots moved in different directions ;" and also when it appears as a simple adjective signifying "swift." The short penultima is contrary to all rule; for the participle of ci-eo must be ci-ttus=citus ; and we can only explain it as a result of Roman abbreviation. But the existence of the forms cieo and do is quite sufficient to prove the fact, for which I contend, that true verbs in -i include the formative in -e. And in the next chapter I shall show that, as I have mentioned above (^ 1), the same remark applies also to the a verbs. To this rule, respecting the i verbs, there are only two exceptions the verb eo (root i) and the verb queo (root quen- or /con-). These two verbs are distinguished from the regular verbs in i by their omission of the e in the imperfect ibam, quibam, and by the adoption of the agglutinate form in the futures i-bo, qui-bo. With regard to the former point, although we have occasional exceptions in the poets, as lenibat, polibant, &c., we generally find that the imperfect of the i verb ends in -iebam, as audi-e-bam ; and in this particular it is imi- tated by the semi-consonant verb in i, which gives capiebam, faciebam, fugiebam, &c. With regard to the future, we rarely, if ever, find an -i verb which follows the analogy of ibo, quibo ; but in almost every case we have the subjunctive form in -am (-es, -et, c.), which is invariably adopted by the consonant verbs. The substitution of e for i in the verb eo, which does not involve the formative element of the second conjugation, leads to some momentary confusion with the e- verb, in those instances in which eo is used as an agglutinate auxiliary to express the passive of certain compounds of do and facio, just as the -eo verb stands as the corresponding intransitive to verbs merely differing from it in conjugation. Thus we have inter-eo, " I go between," i. e. vanish, by the side of inter-ficio, " I cause to go between," i. e. make away with ; per-eo, " I go through," i. e. disappear, by the side of per-do, " I put through," i. e. anni- hilate ; and similarly, pessum-do (cf. TrcpOw); ven-eo (=venum eo), " I go for sale," i. e. " I am sold," by the side of ven-do (=venum-do), " I put up for sale," and ven-dico or vin-dico (=venum-dico\ " I declare for sale." But the confusion is only instantaneous, for the first comparison shows that these verbs are distinguished from the neuter verbs mentioned above (as pateo, pendeo, sedeo) both by the conjugation of the present 384 THE LATIN CONJUGATIONS. [On. XII. (in -eo, -es, -et, &c., not -eo, -is, -it, &c.) and by the form of the perfect (which is never in -ivi). On the other hand, we must distinguish the causative verbs in -do, Greek -Ow, from the aorist formations in -Qrjv, -rjv, which involve^the element ya, and have precisely the converse meaning. Of these latter forms enough has been said elsewhere (New Crat. 379, sqq.). I will only remark in passing, that the explanation of these forms will not justify the monstrosity eyptjyopOacri, in which all the gram- marians have acquiesced. As this word rests only on a single passage (Horn. //. X. 419) and as the context shows (cf. II. VII. 371 ; XVIII. 299) that the true reading is: ot ' eypyyopdai re = 7r/0-ya>, whence Trr-a^o's and the Italian pit-occo. Now if the primary meaning of this root is " to fall down " and "make an inclination," like the Hebrew Tfil, "to make a reaching towards another," so that the root will be contained in pe[d]-s, TTf-TTT-o), 7re$-ov,fotus, "foot," the present must have required the strengthening observed in TreiOco = TriO-yco, and presumed in peto-pet-yo. It is also clear that rudo is only another form of rugio, which has passed into rudio ; compare the Gothic rauhts = " fremitus," with the Greek pdOos, poOciv, pvfyiv, ypvfyiv, &c. Several of the consonant verbs strengthen the root in the present tense and its derivatives by a nasal insertion analogous to the Sanscrit anusvdra : but this insertion is never retained in the perfect, if this tense is or was formed by reduplication ; thus we have pu-n-go, pupugi, ru-m-po, riipi, fra-n-go, fregi, tu-n-do, tutudi, sci-n-do, sc*idi, &c. The same rule applies to n, when it is appended to the root, for in this case also it appears to be inconsistent with reduplication, not only in the Greek and Latin, but also in their elder sister the Sanscrit, and in the Sclavonian, which furnished the Pelasgian element to both of them. Thus we have da-ddmi, but ap-nomi ; $i$co/ju, TiOrjfu, iaTrjfjii, but <(evy-vvfjii 9 Gap.-VY]^i, iK-veo/mai ; TT'I-TTTCO for TTt-Trero), but TTIT-VW ; bibo, but TTI-VU) ; and, as we shall see, sper-no, but spre-vi, contem-no, but contemp-si. In Scla- vonian there is a particular class of verbs, which the grammarians call semel-factive, and in which this nu is the distinctive mark. As then the reduplication clearly denotes iterative or continuous action, we must conclude that n is in these cases the pronominal element denoting separation and distance, which is opposed to the idea of abiding presence connected with that of continuance. Whereas in those cases in which the perfect formation retains the 25 386 THE LATIN CONJUGATIONS. [On. XII. -n, as in jungo, junxi, funyor, functus sum, &c., we may infer that the n is merely euphonic, or intended to express, in con- junction with the guttural, the sound of the Semitic y (See Report of the British Association for 1851, p. 148). Most of the Greek verbs in -ro> exhibit the r- as a pronominal adjunct of the same kind with the -v- which has just been mentioned : compare TVTT-TCD, TIK-TCO with re/m-i/co, SciK-va), &c. We may come to the same conclusion with regard to the Latin verbs in -to, as flee-to from the root flac- in flaccidus, &c. As n is opposed to the continuous or iterative meaning of the verb, it may seem surprising that the most common Latin frequen- tatives end in -ito ; but these, as we shall see in the next chapter, are derivatives of a very different kind. Of the Latin verbs in -to, -tis, &c., the most instructive is ver-to. The ideas of turning, changing, and beginning to be, have a common source, and refer themselves to one conception in the mind. It is difficult to say which is the primary modification of the thought. Perhaps the word vertumnus, which has long been recognised as a participial form from verto, will lead us most easily to the primary meaning of the root. It is usual to consider the Etruscan deity Vertum- nus as the god of the autumn or of the ripe fruits (so Creuzer, Symb. III. 665) ; but the co-existence of the word auctumnus shows that this cannot be the correct view of the matter. As the husband of Pomona, the summer-goddess, Vertumnus begets Cceculus, the darkening time of the year, and must therefore, in himself, be a personification of the spring, ver, which is actually included in his name. For ver=ver-t (feap-r) is the period when the germs of the fruits first come into being (compare wes-en with wer-deri), and this, as the beginning of new life, is a change from the previous state of decay and non-existence. We may say that Vertumnus (or Vertunnus, cf. Neptunus for Nep- tumnus) is the year when " it changes itself," or puts on a new dress ; and as the aura Favom, in the language of Lucretius, is not only reserata, or released from its former bondage in the dungeons of winter, but also genitabilis, or the cause of birth, we. may see that Vertumnus, the god of change (Ovid. Fast. VI. 410 ; Prop. IV. 2, 10 ; Horat II. Serm. 7, 14), is also the representative of the generation or birth of the fruits, which lie fecundating under the care of Pomona, until they spring up into the Auctumnus =Auctomenos or growing year. Thus the Hebrew 5.] THE LATIN CONJUGATIONS. S87 h, which denotes the autumn, is used as an expression for maturity, as in Job XXIX. 4 ; and if the same root indicates also a falling away, decadence, and consequent reproach, we only come to the idea suggested by Cceculus, another expression for the Autumn, as the child of Vertumnus and Pomona. The Umbrian Propertius (IV. 2, 46) expressly tells us that the name of Vertumnus was explicable in the Etruscan language ; for he says : At mihi, quod formas unus vertebar in omnes, Nomen ab eventu patria lingua dedit, and that this patria lingua must be Etruscan (i. e. in this case Pelasgian) is clear from the beginning of the Elegy (v. 3) : Tuscus ego, et Tuscis orior: nee poenitet inter Proelia Volsinios deseruisse focos. And Varro expressly tells us that he was a chief divinity with those Etruscans who came with Crelius Vibenna (L. L. V. 46, p. 18, Mliller) : " ab iis dictus Vicus Tuscus, et ideo ibi Ver- tumnum stare, quod is Deus Etrurise princeps." From this we learn that the Pelasgian religion was peculiarly distinguished by its elementary character (above, p. 36), and that ver-to, and consequently auc-to, were Pelasgo-Tyrrhenian words. In its middle sense, vertor often appears in the compound re-vertor, " I turn myself back or return." The verb rego, which, as we have seen (above, p. 76), has important affinities with the Greek, Sclavonian, and even the Semitic languages, is never used as a deponent to signify motion in a straight line, like the Greek e-p^-oiJiai, nor is it used as a neuter verb like r-pe^ao, and yet the term regio or regio viarum expressly denotes the straight course or direction, like the avopevwv /BrjuaTcov o-pey/ua of ^Eschylus (Choeph. 799). The uncompounded verb lego has the perfect Icgi, which is undoubtedly a remnant of reduplication ; but in the derivative forms, such as intel-ligo, " I make a dis- crimination," i. e. I understand, diligo, " I make a choice," i. e. I prefer or love, neg-ligo, " I make no option," i. e. I leave behind neglected, we have only the aorist in -si, as intellexi, dilexi, neglexl. But we have also intellegi, neglegi, and conversely collexi, in the older writers (see Lachmann, ad Lucret. VI. 17). This aorist revives the lost guttural of the present tense in fluo, fluxi, in struo, struxi, in vivo, vixi, and mfruor,fructus sum; and strengthens an ultimate guttural in traho, traxi, and veho, vexi. 252 388 THE LATIN CONJUGATIONS. [On. XII. 6. B. Liquid Verbs. Some of the verbs, which have I for their characteristic, double this letter in the present tense, but not in the perfect, thus we have pello, pepuli, pulsus, &c. The analogy of ille 9 alius, &c., would lead us to infer that these verbs belong strictly to the semi-consonant class, and the singular participle tlatus or latus from tollo, tetuli, coupled with the Greek form r\dw, would almost suggest the idea that there was once a collateral verb in -a. There are only two n verbs, the reduplicated gigno, root gen-, perfect genui, and cano, perfect cecini. But the known relationship between ille, alius and dvd, together with the meanings of alo, al-mus, al-u-mnus, which imply " bringing up" suggest the possibility that this verb may have belonged ori- ginally to the same form of the liquid characteristic. We have seen above that I and n are both dentals, and that they are frequently interchanged. Although s is by its origin a result of the gutturals, it often passes into the dental r ; and there can be little doubt that most of the verbs in r and s must be placed in the same category. Indeed it has been suggested that sero, serui is merely a reduplication for seso. While the other liquids are all capable of some connexion with the dental articulation, the labial m stands apart from any interchange with the other letters of this class, except in the case of an assimilation, as in pressi from premo (cf. jubeo, jussi). The most important and remarkable of the m verbs is emo, which is worthy of special examination, not only on its own account, but also on account of its numerous compounds. The primary meaning of emo is, " I take up or select," and thus it comes very near in signification to lego. This idea of selection lies at the root of the ordinary meaning of emo, " I buy ;" for this presumes a selection from a variety of objects offered for sale. In our own colloquial English, " I will take this," is the usual phrase for expressing an intention to purchase some particular article. The Greek Trpiaf/iai ap- pears as the middle of TrnrpdaKw, " I cause to pass over ;" and the two together express the changing of hands (nepav) which always attends a sale. And as ct7ro/oyuai means, " I give away for my own benefit," i. e. " I part with a thing on advantageous terms," so oWo/mai (from the same root as ov-ivrjiu) declares the fact that the purchaser finds his benefit in the transaction. A 6.] THE LATIN CONJUGATIONS. 389 recent theological writer has remarked that "the verb emo, which signifies literally 'to select for use' (whence amor and its derivative an[a]0, cf. diligo), is employed in its compounds promo and sumo to denote the use made of the selected articles, or of the money which is their representative ; these must be in promptu before they can be in sumptu, they must be KT^UTU before they can be ^jo^/xara. Hence promptus is the primary as well as the secondary synonym of eroT/xos." When we re- collect that the compounds ad-imo, ex-imo, inter-imo, give us the i, which presumes an a in the weaker form (as in con-ficio, fromfacio, &c., above, p. 261), we are entitled to suppose that emo represents a primary amo, amis, and a secondary em-io ; (comp. ten-eo, con-tin-eo, with raw, TCL-VVW, &c.). We shall see in the next chapter that amor presumes an original am-ior, and that am[a]o suggests a form am- a = am-ya which is included in amor=am-ior, formed from the genitive case of such a noun. It is usual to connect amor with the Sanscrit kdma, which corresponds to it in meaning. But as the analysis now before us shows that " love " is a secondary meaning, derived from that of " selection," we may leave out of the question any results arising from this immediate comparison ; and as the Greek 7rpi-afj.at 9 TTI- Trpd-o-KU), are manifestly connected with the pronominal combi- nation Tre-pu-v or tra-pd, signifying a transit, we may compare a-ma with a-yua, sa-ma, cu-m, which express union or conjunction, and hence appropriation {New Crat. 181), and bring us ultimately to the most probable origin of the Sanscrit kdma. It is worth noticing that the Greek d-a-Trd^oiuai, " I draw to myself," really includes in its prefix this pronominal combination (New Crat. 213), and the same is the case with am-plector and com-plector. No difficulty will be created by the fact that we have a compound co-emo, in the secondary sense, " I buy up." It would be paying too great a compliment to the etymological knowledge of the Eomans to suppose that they dreamt of an affinity between the preposition cum, and the root of emo; and even if this had been so, the repetition of the same elements under different forms would have been in accordance with the oldest examples of pronominal agglutination. The perfect of emo, is emi, and this form is retained by the compounds, except when the prepositional prefix coalesces with the first syllable of the verb : thus we have ademi, exemi, intercmi, but demo- S90 THE LATIN CONJUGATIONS. [On. XII. de-emo makes dem-p-si, promo = pro-emo makes prom-p-si, sumo-su-emo makes sum-p-si ; and while co-emo, "I buy up," makes co-emi, co-emptus, the same verb in the older sense, " I take and put together," i. e. the hair, makes como, com-p-si, com-p-tus. 7. C. Semi-consonantal Verbs. It has been already mentioned that the vowel- verbs in -i differ from the semi-consonantal forms, which they so nearly resemble, both in the origin and in the extent of the pronominal adjunct by which they are qualified. For while the vowel i- verb in- volves not only a crude form in -i, but a repetition of the same pronominal element, the semi-consonantal i- verb uses this adjunct merely to strengthen the present tense and its immediate deriva- tives, and loses all traces of it in those formations in which a contraction is most conspicuous, namely, in the second person singular of A. I., and in the present infinitive. Thus, while we have, from the crude form of ves-ti-s, vesti-o = vesti-yo, vestl-s = vesti-is, and vestire = vesti-yere, the mere root fac- gives us fac-io -fac-yo, fac-is and fac-ere. As cupio has a perfect cupivi and derivatives like cup'ido, we may perhaps be inclined to consider cupere as a degenerate form, and to refer this verb to the vowel-conjugation; and this opinion might be confirmed by its relation to capio. For, according to a principle pointed out elsewhere (New Crat. 53), capio and cupio are related by the association of contrast ; and the shorter vowel u shows that the latter is a longer form than capio ; but this implies that cupio = capi-yo, which is in accordance with the theory respecting the i- verbs. In all other verbs, however, which form the present in -io and the infinitive in -ere, it is plain that there is only one affection of the root with a formative appendage, and the nature of this adjunct is clearly seen in the case of fug-io. For there can be no doubt that we have here the root fug-, and that the same root is found in (pevyw, aor. e-(pvy-ov, where it is strengthened by guna (New Crat. 442), and in (fiu-y-yd-vw, where it is not only strengthened by anusvdra, but supported by an additional nasal (ibid. 435). To the same class as , and transferred to the root-syllable in (pOeipc*}, root (pOap-, (paivw, root (pa-, Kplvco root tcpl- (New Crat. 432). With re- gard to the u- verbs, the known derivation of many of them, and the termination of the participle (E. III.) in -titus or -uitus, shows that they are abridgments or degenerate forms of e- verbs. Thus it is clear that metu-o comes from metu-s, tribu-o from tribu-s, &c. ; and as the verbs are thus connected with crude forms of the semi- consonantal declensions, they require in addition another pro- nominal adjunct, and thus stand in the same relation to the genuine semi-consonant verbs in -u, such as ruo, ruere, rutus, that the vowel i- verbs bear to the semi-consonantal verbs in i. As the i is after all a representative of some guttural, those apparently u- verbs, which exhibit their guttural characteristic in the perfect, as struo, struoci, structus, do not essentially differ from those, which, like metuo, have absorbed the element ya. 8. Irregular Verbs. A. Additions to the Present Tense. From the formations, which we have just discussed, and in which the second element, under the modification i = ya, plays so prominent a part, there is an immediate transition to the first class of the so-called irregular verbs, which strengthen the present by the addition of one or more actual consonants. As far as the epithet " irregular" is concerned, we have seen that there are deviations from perfect uniformity even in those conjugations which we take as the type of the Latin verb ; and it is only in consequence of an excess in the degree of deviation that we are induced to place the verbs with a consonantal accretion in a class by themselves. The additions, by which the present is strength- ened in these verbs, are the liquid N, which in a solitary instance appears also as R, and the combination sc. The former of these adjuncts may or may not be the same with the inserted anusvdra, which we find mjungo, root jug-, fungor root fug-. It is possible that such a nasal may have resulted from euphony ; on the other hand, the manner, in which the adjuncts -vi, -w are melted down so as to combine themselves with the root, e. g. in (paivw - (pd-vyw (root (>a-), eKavvw = i\a-vvw, (root 6\a-), renders it possible that the addition may be pronominal or formative. And this view is confirmed by the fact (noticed 392 THE LATIN CONJUGATIONS. [Cn. XII. above, p. 385), that the inserted nasal seems, like the added n, to be inconsistent with reduplication (cf. ru-m-po, rupi, &c.). We do not find, in Latin as in Greek, that the adjunct n coexists with the inserted n, as in Tu-y-^a-va), Xa-ju.-/3a-i;a>, &c., or with the appended sc, as in o(f>\t-crK-a-i>w, &c. Many of the Latin forms in n have corresponding verbs in Greek ; thus we have cer-no by the side of Kpivw = Kpi-vyuAd, and in the iterative or inchoative tenses in -, is a pronominal affix, springing from a repetition of the idea of proximity (New Crat. J J 386, 7). Whether we say at once that s+c is a junction of two forms of the same element, 8.] THE LATIN CONJUGATIONS. 397 like the common endings n + t, t + n, or identify it with the affix sy found in the Sanscrit future, and in the Greek and Latin desideratives, the result will be the same, for s=i=k come to an ultimate agreement as forms of the second pronominal element. As pronominal elements and their combinations appear also as verb-roots (as e. g. ^ev- in juevw, Oa- in riOrj/uu, &c.), we shall have no difficulty in recognising the reduplication sc, with its inchoative and iterative meaning, in " a large class of words of which the general idea is that of the inequality of the limbs" (Kenrick, Herod, p. 24), or rather which denote progression by successive steps ; such as aneXo^, sca-ndo, &c. $ 9. B. Abbreviated forms. Most of the abbreviated forms, or the verbs which are liable to syncope in certain of their inflexions, have received sufficient notice already. Possum for potis-sum or pot'sum is merely an assimila- tion. The perfect pot-ui may be referred to the same class as the other agglutinate perfects. The omission of d in certain inflexions of edo belongs to an analogy which is particularly observable in the Romance languages (above, pp. 256,7). The same may be said of vis for volis, malo for mage'volo, &c. There are, however, some etymological peculiarities about fero, which deserve a special examination, independently of the fact that it borrows its perfect tuli for letuli, and its participle latus for tlatus or toltus, from the root of tollo, tolyo or tlao. No difficulty is suggested by an immediate comparison of fer-o with the Greek -ro, hinthan, can-is, "hand," "hound," ^ai/$aVa>, pre-hendo (New Crat. 162, 281), we see at once the pos- sibility of a community of origin in fero and fendo. And as we cannot explain the $ or th in either case as a mere adjunct to the root, we must not be led by the actual change of r into n, in some of these forms, to the conclusion that this change has taken place in hendo and fendo. As in the case of ^a-v-^a-vta, ^ is more in accordance with scientific reasoning to suppose that the n is here an anusvdra or euphonic nasal ; and the insertion of this sound would naturally introduce the medial d before r, as in dv-S-pos, ven-d-re-di, &c. But, as we have seen, the Latin r has a natural tendency to commutation with d. Conse- quently, its absorption or assimilation in -hend-o, fend-o, would follow as a matter of course. And thus fer-o, fen-d-o, smdfer-io 9 establish their claim to be considered as members of the same fer-ii\Q stock. 10.] THE LATIN CONJUGATIONS. 399 $ 10. Defective Verbs. The epithet "defective" is applied to verbs with a very restricted signification. Properly speaking, all impersonal verbs are defective in the 1st and 2nd persons, and all neuter and de- ponent verbs are defective in voice, except when the former are defective in person. But it is customary to restrict the term defective to those verbs which are specially incomplete in the machinery of their conjugation. Some of these are really only irregular appendages of existing verbs. Thus ccepi is the usual perfect of in-cipio, memini of reminiscor ; ausim and faxim are obsolete tenses of audeo and facio, and the former of these, with gaudeo, jfido, and soleo, has no perfect of the active form ; quceso, qucesumus are the original articulations of qucero, quceri- mus ; forem and/ore are used with sum and/m*. Some few verbs are employed in a sort of interjectional sense in the imperative only, as apage, cedo, &c. ;. others, as vale, which are thus used, appear also as regular verbs. Odi, " I hate," " I have conceived a dislike," is the intransitive perfect of a lost deponent, corre- sponding to the Greek oSva-ffofiai (cf. oXwXa from oXXv/im, &c.) ; this deponent form exists in the compound participles exosus and perosus. We can have no difficulty in understanding the parenthetical use which gradually reduced the oldest verbs of "speaking," aio, inquam, andjfaH, to a few of their commonest inflexions. We have the same result in the Greek rj $ 05, and in our " quoth," which, as has been remarked above (p. 112), exists as an independent verb only in the compound " be-queath," and which contains the same root as in-quam. The forms of the im- perfect and future (in-quiebam, in-quies\ and the diphthong in the derivative quce-ro = quai-sino, show that the root in-quam must have contained something more than a mere vowel of articulation, and that it was probably strengthened by the semi-vowel i. It therefore stands on a different footing from sum, the only other verb which retains the first person-ending in the present ; for here the u is a mere sh'va like that in Hercules (above, p. 266) : cf. as-mi and ea-/Jii In the by-form in-fit we havey= qv, which is not uncommon. CHAPTER XIII. DERIVATION AND COMPOSITION. 1. A. Derivation. General principles. 2. Derivation is merely extended, or ulterior inflexion. 3. (I.) Derivative nouns. 4. (a) Forms with the first pro- nominal element only. 5. (b) Forms with the second pronominal element only. 6. (c) Forms with the third pronominal element only. 7- (a) Terminations compounded of the first and other pronominal elements. 8. (/3) Terminations compounded of the second and other pronominal elements. 9. (y) The third pronominal element compounded with others and reduplicated. 10. (II.) Derived verbs. 11. B. Discrimination of compound words. 12. Classifi- cation of Latin compounds. 1. A. Derivation. General principles. THE term derivation was once used to denote the process of guess-work by which the etymology of a word was ascer- tained, and it was formerly thought that the most satisfactory derivation of a Latin wctfd was that which consisted in its direct deduction from some Greek word of similar sound 1 . The student of scientific or comparative philology does not need to be told that, although the Greek and Latin languages have a common element, or are traceable, in part at least, to a common source, their mutual relationship is collateral, and not in the direct line of descent, and that in these and other old languages of the Indo-Germanic family " derivation is, strictly speaking, inapplicable, farther than as pointing out the manner in which certain constant syllables, belonging to the pronominal or formative element of inflected languages, may be prefixed or subjoined to a given form for the expression of some secondary or dependent relation" (New Crat. Pref. 1st Ed.). According to this view, derivation includes a de- partment of what is called word-building ( Wort-Uldung), so far as this is distinguished from mere inflexion. The modifications of the noun and verb, by which inflected language is characterised, belong indifferently to all forms, whether primary or derived, whether simple or compound. And after considering these for- mations, the grammarian naturally passes on to an investigation 1 Doderlein is perhaps the last representative of this school, and some of his derivations (e. g, fraus from ^evSos !) are equal to the worst attempts of his predecessors. $ !] DERIVATION AND COMPOSITION. 401 of the cognate but subsequent procedure by virtue of which, (1) an existing noun or verb developes itself into a secondary form of the same kind, or (2) two or more distinct words are combined in one, and furnished with a single set of inflexions. This procedure is called word-building, and might be designated as derivation in reference to the materials, and composition in reference to the machinery. Practically, however, we confine the term derivation to the former department ; namely, to the deve- lopement of secondary words containing only a simple root ; while composition is used to denote the subordination of two or more crude forms under the influence of some set of formative appen- dages and inflexions. J 2. Derivation is merely extended or ulterior inflexion. In considering the distinction between derivation and in- flexion, we must bear in mind, that the former process is really nothing more than an extension of the latter. In forming a word, in the first instance, by the addition of cases or person-endings, we derive our formative materials from the same limited and classified stock of pronominal elements, which furnishes us with the machinery of derivation. Indeed, the new crude form, which becomes the vehicle of the inflexion, is very often neither more nor less than the oblique case of some existing word, and it is probable that this process has been repeated in successive de- rivations. This remark applies only to derivative nouns, for the new forms of verbs cannot rest upon the inflexions, i.e. person- endings, of their primitives. In general, we observe that there is much greater variety in the secondary formations of nouns than in those of verbs. For the person -endings of the latter antici- pate the distinctive use of the three pronominal elements in their most prominent and important application, whereas the cases of the noun are connected only with a special developement of the second element, signifying proximity, and transition of agency or the point of motion, and of the third, denoting position and dis- tance. In the derivative forms we find the converse phenomenon : for while the verbs are contented with extensions of their crude form, by pronominal additions limited to that special develope- ment of the second and third elements, which is found in the cases of the noun, and which does not exhibit any direct reference to the primary distinctions of position; in the nouns all three prono- 26 402 DERIVATION AND COMPOSITION. [On. XIII. initial elements are used, in their distinctive senses and in combi- nation with one another, to form nominal derivatives, which may- be extended by successive accretions to a considerable length of after-growth. A verb in the finite moods must always be distin-i guished by person-endings, which cannot become the vehicle of ulterior formations ; and, for the same reason, all pronominal elements, which might be mistaken for person-endings by re- taining the original distinctions, are excluded, in the verb, from the function of extending the crude form, which they exercise in the derivative nouns, both when they are and when they are not identical with the case-affixes of the primitive words. 3. (I.) Derived Nouns. It is not always possible to assign a definite meaning to all the elements or combinations of elements, which contribute to the extension of the crude form in Latin nouns ; but so far as we can arrive at the signification of the affix, we can see that the distinctive use of the pronouns is preserved in this application ; namely, that the first pronominal element expresses that the thing proceeds from, or immediately belongs to, the subject ; the second, that it has a relation to the subject ; the third, that it is a mere object, or something removed from the proximity of the subject. We also observe that the combinations of these elements are regulated by the same principle as that which explains their use in prepositions and other independent words; namely, "that if any one of the elements of position is combined with -ra, an ultimate form of the third element, it indicates motion and continuation in a direction of which the element in question represents the point nearest to the subject; and that by sub- joining any one of the pronominal elements to any other of them, we denote a motion or continuation from the position signified by the first element towards that indicated by the second, and so on, if the combination involves more than two." (New Crat. 169). 4. (a) Forms with the first Pronominal Element only. There are comparatively few Latin nouns in -ma or -mus, which express an action as immediately proceeding from the subject : such are fa-ma, " a speaking " (root fa-), flam-ma, "a burning" (root flag-), tra-ma, "a drawing" (root trah-), $4.] DERIVATION AND COMPOSITION. 403 ani-mus, " a blowing," ar-mus, " a joining," re-mus (root ret- or rot-) " a turning round " (in the water), i. e. " a rowing thing," al-mus, " a nourisher," pri-mus, " the first of a series beginning with the subject," &c. $ 5. (b) Forms with the second Pronominal Element only. The second element, under one or other of its various modi- fications, contributes most largely to the formation of derivative nouns. A great number of these are abstract or qualitative terms, and they differ from those in -ma and -mus by their more general and relative predication. For all those formed by the first element only may be translated as expressing the sub- ject of action, and some of them, as re-mus, al-mus, cannot be regarded as mere abstractions. Whereas the nouns, which exhibit the second element as their termination, always depart from the idea of a subject or agent, and express only an agency or quality, like the English words in -ness, -hood, -y, &c. Some- times the second element appears under a guttural form, as in vo-c-s (vox), " a voice"' or " speaking" (Sanscr. hve, cf. fioq, r\-yr\, &c.); and to this class belongs the copious list of adjectives in -cus, -1-cus, ac-s (=ax), &c., denoting quality or disposition, as civi-cus from civis, ami-cus from amo, loqu-a-x from loquor, &c. But by far the most common form of the second element, in its use as an affix, is that in which the guttural is vocalized to i. Besides the numerous words in -ia, -ius, -ea, -eus, -ium, -is, as grat-ia from grat-us, mod-ius from mod-us, pic-ea from pix, calc-eus from calc-s, consil-ium from consul, febr-is from ferv-eo, nubes=nube-is from nubo, materies=mater-ia-is from mater, &c., it seems reasonable to infer that the masculine nouns in a, together with some feminines, involve vocalized gutturals ; for we cannot otherwise account for the formation of such words as scrib-a, nota, agri-col- a, Sec., as compared with the Greek Kpirris, Tifj.tj, avKea, and raMxs, than by supposing an omission of the extenuated i=y : thus scrib-a^ scrib-y as will be legitimately formed from scribo, nota=not-ya=no-tia, will properly correspond to T^J &c. in Greek, and to amici-tia, &c. in Latin. We may also compare ad-vena^ad-ven-ya-s with ad-venio. That such an extenuation is possible is shown by the transference of ^covtj, Sec. into zona, &c. (above, p. 295). We have also seen that the affix i lies more or less hid in some nouns of the third declension, 262 404 DERIVATION AND COMPOSITION. [On. XIII. and especially in participles and adjectives (above, p. 301). This is particularly the case with the forms in nt-s or nti-s, and we may compare the affix -tis or -tus, mpes-tis, "a destroying," ves-tis, "a covering," po-tus, "a drinking," spiri-tus, "a breathing," with the Greek nouns in -cris, -rts, and -rus, as 7rpaK-(ris, "a doing " = Tr/oa/c-rJ?, , and tem-p-lum with re/xe-i/o?, whether as re/ueyos alOepos (^Esch. Pers. 357), or as the portion of land cut off and set apart for divine uses. I 1. (a) Terminations compounded of the first and other Pronominal Elements. Many nouns exhibit in their affix a combination of the first element with the third, under the form m + n, which is often 406 DERIVATION AND COMPOSITION. [On. XIII. strengthened by a repetition of the objective affix under the form t, so that the whole affix is m + nt. Of nouns in -mon corresponding to the Greek nouns in -(JLCDV, we have only three, ser-mo[n~], pul-m6[n\, te-mo\n\, which may be compared with yvco-fjiuov, TrAev-fiwi/, &c. ; we have also a limited number of nouns in -mnus, corresponding to the Greek passive participle in -yuei/os, such as auctu-mnus, da-mnum, vertu-mnus, alu-mnus, colu-mna, ceru-mna, &c. A comparison of alu-mnus, "the person nourished," with al-mus, "the nourisher," shows that the combination m + n completes the agency and carries it on to the object acted on. As in Greek we have -/movrj^-iJio-v-ia by the side of -JULODV, so in Latin we find an extended termination -monia, in such words as acri-mon-ia, cegri-mon-ia, ali-mon-ia, cere-mon-ia, casti-mon-ia, parsi-mon-ia, sancti-mon-ia, all of which express a quality or abstraction inferred from an act done. The force of these words is best shown by a comparison between these and the nouns in m+nt, which have a repetition of the third element instead of an addition of the second. These words, which agree with the Greek neuters in -par = -^eyr, either omit the final t, as in car-men, cri-men, legu-men, stra-men (above, p. 299), or, which is more common, exhibit the lengthened form -mentum, as in ali-mentum, ar-mentum, arma-mentum, aug-mentum, orna-mentum, pul-mentum. Now all these words express an action proceeding from the subject (m), but become objective (n), and exhibited in its results (t). Thus car-men\t\ v ai " \ = Sanscrit kar-man means " a thing made," with (jtxei/T-J especial reference to the maker. But cere-monia, which con- tains the same root (cere-, ere-, kri-) 9 calls attention by its affix to the doing or process. Similarly, al-mus is " a nourisher," alumnus, "a person nourished," ali-mentum, "a thing for nourishing," but ali-monia, " the process of nourishing." $ 8. (/3) Terminations compounded of the second and other Pronominal Elements. Of these combinations the most usual and important are the forms in which the second element, vocalized into i, is prefixed to the third element with or without further extensions. A very large class of nouns end in -io[w], and express, if one may say so, a fixed or consolidated abstraction. These nouns, which are 8.] DERIVATION AND COMPOSITION. 407 always of verbal origin when the noun is feminine, are formed either by affixing -io\n\ to the simple stem, as in leg-io, opin-io, reg-io, rellig-io, and this is always the form in the masculine nouns, as cur-io, centur-io, scip-io, &c. : or by adopting the t=s of the supine, as in man-sio, ses-sio, con-fu-sio, quces-tWj col-lec-tio, dis-trac-tio, dubita-tio, pulsa-tio, con-jura-tio, posi- tio, ad-moni-tio, erudi-tio, &c. The masculine nouns generally denote a person or thing belonging to that from which the noun is derived ; thus cur-io\n\ is the man of the cur-ia, centur-io[n~\ the man of the centur-ia, &c. And as the genitive ended origi- nally in -ion = -sion, we must consider these nouns in -idn-ion-s as extensions of the genitive case. The same explanation will apply to the nouns in -o[w], as epul-o\ri] from epul-a; for there is reason to believe (above, p. 295) that these forms have lost or absorbed an i. As the termination -ia, -is, -sis, -tis is parti- cularly appropriated to verbal nouns expressing the action of the verb, we must conclude that the verbal nouns in -io, -sio, -tio, are also derived from the genitive of nouns in -ia, -sis, &c. And this will lead us to the meaning already suggested, namely, that these words denote the result of an abstraction which has become fixed and objective. The important word relligio will furnish a good exemplifi- cation of my meaning. There have been two different opinions with regard to the etymology of this word. For while most mo- dern scholars adopt the suggestion of Servius (ad Virg. ^En. VIII. 349), Lactantius (IV. 28), and Augustin (Retract. 1. 13), namely, that the word comes from religare, supporting this view with the quotation from Lucretius (I. 931, IV. 7) : " relligionum nodis animos exsolvere;" Cicero makes religere the main verb, and gives the following explanation (de Nat. Deor. II. 28. fin.) : " qui omnia, quse ad cultum deorum pertinerent, diligenter re- tractarent et tamquam relegerent, sunt dicti religiosi ex rele- gendo, ut elegantes ex eligendo, tanquam a diligendo, diligentes, ex intelligendo, intelligentes : his enim in verbis omnibus vis le- gendi eadem, quae in religioso;" and similarly, in another part of the same work, he says (ibid. II. 3, $ 8) : " relligio est qus9 superioris cujusdam nature (quam divinam vocant) cur am ca3ri- moniamque affert." This etymology is in accordance with the verse quoted by Aul. Gell. IV. 9 : " religentem esse oportet, rel- ligioswn nefas." And there can be no doubt that it is perfectly 408 DERIVATION AND COMPOSITION. [On. XIIL true. It is clear from the use of the word that relligio is not derived from religare, " to bind back," but from religere, " to gather over and over again," " to think perpetually and carefully on the same subject," " to dwell with anxious thought on some idea or recollection:" so that re-ligens is nearly a synonym of di-ligens, and an opposite of neg-ligens. The word expressing the abstraction of the verb should end in -ia, but this, as in most of these words in -io\n\, is lost, and we have only the derivation from the genitive case expressing the result of the abstraction the realized ideal. Hence, practically, relligio sig- nifies, (1) " religious worship," considered as scrupulous obedience to the exactions of the conscience, and with especial reference to the act of worship ; as (Cic. Verr. II. 4, $ 49) : " qui sacris anniversariis ac summa religione coleretur ; " or to the religious sanctity of an object ; as (id. ibid. 46) : " fanuni est Junonis antiquum, quod tanta religione semper fuit, ut semper inviolatum sanctumque fuerit;" (2) "religious scruple" or "superstitious fear," considered as something objective and real; as (Caas. Bell. Civ. III. 72, 4) : " non recordabantur quam parvula3 ssepe causse vel falsas suspicionis vel terroris repentini vel objectce rel- ligionis magna detrimenta intulissent ; " and especially in the plural, as (Lucret. I. 109) : " relligionibus atque mineis obsis- tere vatum ; " (3) by substituting the cause for the effect, " guilt causing religious scruple or fear," and " the divine curse and consequent remorse or oppression of the conscience caused by a sense of violated religious scruples." In the second and third sense it is used in a curious connexion with violare and expiare in three passages of Cicero, which have never, so far as 1 know, been compared by any lexicographer or commentator : (a) ad Atticum, I. 17, ^ 16 : " quare et ilia, quse violata, expia- buntur ; et base nostra, quae sunt sanctissime conservata, suam religionem obtinebunt." Here, it should seem, religio means "scrupulous observance;" and the maintenance of uninterrupted intimacy between Cicero and Atticus is opposed to the atone- ment necessary to restore the violated harmony between Quintus and his brother-in-law, (b) Tusc. Disput. I. 12, 27 : " id quum multis aliis rebus, turn e pontificio jure et caeremoniis sepul- crorum intelligi licet; quas maximis ingeniis prasditi nee tanta cura coluissent nee violatas tarn inewpiabili religione sanxissent, nisi hsBsisset in eorum mentibus mortem non interitum esse omnia 8.] DERIVATION AND COMPOSITION. 409 tollentem atque delentem." It is clear from the collocation in this passage, compared with that of the former, that religio means not only the scrupulous observance of religious obligations, but the lasting curse or remorse, which, as a punishment, waits on those who violate the sanctity of divine worship. This feeling may, as the former passage shows, be expiated, atoned, or re- moved by the performance of suitable rites, or the guilt may be so heinous that no reconciliation can take place between the offender and- his conscience; and thus we find in the third passage to which I have adverted, (c) Philipp. I. 6, 13 : "an me censetis, P. C., decreturum fuisse, ut parentalia cum supplica- tionibus miscerentur, ut inexpiabiles religiones in rempublicam inducerentur ? " that a state would be involved in an 0709, or pollution, which no KaOapfjLol could wash away, if funereal sacri- fices in honour of the departed were mixed up and confused with public thanksgivings to the immortal gods. From all this it appears that the formation in -o [n] brings the mere abstract noun, from which it is derived, into a more concrete reference, so that the meaning is rather the result of the verb's action than the action itself. This is the signification also of Greek nouns in -wi/, -wvos, many of which, as Xeijuwi', denote some object or thing. The most important, and perhaps the least understood of these Greek nouns is aiwv, which denotes not only an unlimited extension in time, which is one meaning of de /, but also present existence, or existence for the time being, as in o KpctTwv aei, " whoever happens to be in power" (^Esch. Prom. 973) ; thus aiwv may signify not only an age or eternity, but also the present life, as opposed to the future, which is sometimes its meaning in the New Testament, and the existing generation of a family for the time being, as opposed to the series of yeveai, which make up the whole succession or con- tinuance of a race (see the note on the Antig. 580, p. 179). From religio we have the adjective religiosus ; and the occurrence of the same form in derivatives from nouns in -ia as cerumn-osas from cerumna, glori-osus from gloria, luxuri- osus from luxuria, c., tends to confirm the supposition that the noun in -io is an extension of the noun in -ia. We find adjectives in -osus from other crude forms, as dol-osus, libidin-osus, and we may conclude that in these cases also the intermediate form is the genitive in -ion. The forms in ~tivus, mentioned above 410 DERIVATION AND COMPOSITION. [On. XIII. (p. 404), show that the origin of the abstract nouns in -is, -sis, -tis, &c., may be traced back to the supine in -turn and the infinitive in -se (cf. New Crat. $ 410, (3)). That in all the nouns in -o[n~\ the i of the genitive -ion is absorbed, may be proved by an examination of the abstract nouns in -or, such as amor, favor, honor, &c. For no one will doubt that the Latin comparative ending -ior=ion-s is equivalent to the Greek -HDV=IOV- and -a-rium, as arbor-e-tum, ov-i-le, gran-d-rium ; for if arbor-eus is formed by an adjunct of the second element under the form ya, arbor-e-tum must extend the same form by an addition of the third element, and a similar explanation will be required by the long i = ii and a = ea of ov-i-le and gran-d-rium, to which the I and r terminations are appended. We see then that all nouns expressing agency, or the place, means, and occasion of agency, are formed by adding a combi- nation of the second and third pronominal elements and this is what we should a priori expect for the idea of agency is that something, i. e. a doing, proceeds from the subject, who by the nature of the case is presumed to be near, and passes on to an object, which by the nature of the case is presumed to be relatively more distant. But we observe that the same sort of endings are used to form ordinary adjectives derived from nouns and not from verbs ; thus from rex we have reg-a-lis = reg-ya-lis, from Roma we have Rom-anus = Roma-ya-nus, from consul we have consul-a-ris = consul-ya-ris, from civis we have civl-lis = civi-ya-lis, from asinus we have asin-i-nus = asin-ya- nus, &c., which fully correspond to the forms ov-i-le, gran-d-rium, &c. ; and there is also a class of diminutives in -cu-lus, which exhibit the same termination as the verbal nouns veh-iculum, &c. It will be easy to show that the combination of elements in these cases is as consistent with their primitive signification as in the class previously examined. To begin with the dimi- nutives. As there are objective nouns in -turn, -lum, -rum, as well as nouns combining this affix with one belonging to the second element, so there are diminutives in -lus and -leus, as well as those exhibiting the compound termination now under consi- deration. Thus we have libel-lus - liber-u-lus, filio-lus, &c., as well as pisc-i-culus, homun-culus, &c. The origin of the diminutive expression, or vTroKopiajuLa, is to be sought in the tendency to speak of a darling object, as, at the same time, little. Whether this has or has not any connexion with a 8.] DERIVATION AND COMPOSITION. 415 mother's fondness for a child is doubtful. But it is a universal practice to speak of a petted object as a glycerium, yXvKepiov, or " dear little thing." In classical Latin the diminutive puella -pueru-la is invariably used instead of the original word. Now in these terms the feeling of personality becomes evanescent, and that of mere objectivity takes its place. With a view to the expression of this idea it seems to be a matter of indifference whether we merely append the objective ending -lus, Greek -v, or connect this with the main verb by some possessive affix de- rived from the second element in Latin -c, Greek -i. For example, we may form the secondary noun juven-cus from juve- nis without any change of meaning ; and pul-lus, catu-lus, &c., will be just as good diminutives as juven-cu-lus. The other derivatives, mentioned above, must be regarded as extensions of the case in -i-na or i-n (p. 276). Thus Romdn-us-Roma-in-us is a man who lives " at Rome," Roma-i\n}. This is shown still farther by the relation between these nouns and their extensions in -en-sis. These derivatives are either formed directly from their primitives, as praten-sis, " that which belongs to or grows in the meadow" (prato-en=prat-in, in a heavier form prat-en), or else they involve some noun already formed upon the locative, as Roman-ien-sis from Romanus. " In genere," says Ruhnken (ad Suet. Ccesar. 37, p. 58), " adjectiva, qus9 in -ensis exeunt, designant res hominesque, qui sunt in aliqua regione, sed aliunde originem habent. Romanus, qui Romas natus est ; Roman\i\ensis, qui Rom ae degit: Siculus, qui in Sicilia ortus est ; Siciliensis qui incolit Sicilian!, aliunde ortus : v. Fest. v. Corinthiensis et Intt. ad Vellei. Pater c. II. 51. Idem discrimen apud Grsecos in 'IraXos et 'IraXiwr^?, St/ceXos et 2t/ceXtfor>79, &c. : v. Ammo- nium in his vocibus et ibi Cl. Valckenar." This is a correct statement of the fact, but it does not explain the formation of the secondary nouns in -ensis. As TraXiwr^?, &c., are formed from nouns in -ia (New Crat. $ 259), so we always find that, if there are co-existing derivatives in -nus and -en-sis, there is an intervening form in -ia. Thus from Hispanus we have His- pania, and from this again Hispaniensis as from the locative Hispania-in. Accordingly, we may infer that Romaniensis, which is the true form, comes from an intervening Romania as the country of the Romani. The permanence of this rule of secondary derivation is shown by the practice of our bishops, 416 DERIVATION AND COMPOSITION. [On. XIII. who call themselves Cantuariensis,Dunelmensis,8cc., to show that they are temporary incumbents, rather than hereditary peers. A comparison of these nouns with the equivalent Greek forms in -HD-TW, -iri-rrjs, -i-rrjs, teaches us that the termination -sis, attached to the locative -en and belonging to the second element, is identical with the similarly derived -Tr)<$. We shall therefore not be surprised to find it also under the forms -tis and -ter. This is the fact when the locative, to which it is attached, is plural, as in the case of those nouns, which express an extended region rather than a definite locality. As we say, in agris, in campis, in sylvis, in terris, rather than in agro> in campo, &c., it is natural that we should find, as we do, agres- tis, campes-ter, sylves-ter, terres-ter, rather than agr en-sis t &c., which do not occur. At first sight we might feel disposed to refer eques-ter and pedes-ter, rather to the substantives eques, pedes, than to the locatives equis, pedibus. But the omission of b in queis for quibus, &c., shows us how pedeis might be a loca- tive, and we have a passage in Virgil, which actually places the locative equis on a parallel footing with the derived pedes = ped-it-s; ^Eneid. VII. 624 : Pars pedes [i. e. pedibus iens] ire parat campis : pars arduus altis Pulverulentus equis furit : omnes anna requirunt. The noun seques-ter does not belong to this class. As de- noting a functionary, it connects itself at once with magis-ter and minis-ter, and as these involve adverbs, which are of the nature of locatives, we must derive seques-ter, not from sequor with the old grammarians (for then we ought to have secu-tor), but from secus=sequis (cf. 8equior)*=Ka$, and thus sequester t which means a mediator, umpire, or other indifferent party, will naturally imply one who stands apart from both the litigants ; for quod secus est is opposed to quod interest (Plautus, Trin.1. 2, 93). The patronymics in -ilius must not be referred to the same class with the nouns in -He, -inus, -arius, &c. As it is known that in this case li-di (compare Acilius, Epilius, Hostilius, Petilius, Pompilius, Popilius^ Venilius, with their original forms Acidius, Epidius, Hostidius, Petidius 9 Pompedius, Umbr. Pumperius, Popidius, Venidius), we must refer these words to the same class with the Greek patronymics in -<^?, where the second pronominal element appears under the form of an approximate dental sibilant (New Crat. 262). 9.] DERIVATION AND COMPOSITION. 417 9. (y) The third Pronominal Element compounded with others and reduplicated. The most common extension of the third pronominal element is its reduplication under the forms t+n, or n+t, the latter com- bination being by far the most usual. With regard to other forms into which the pronoun enters under the type t, it is not always easy to say whether this is a corruption of ty, or merely the expression of the objective word. Thus we have seen that in t+r there is something more than the third element extended by the addition of r. It is probable, however, that in such affixes as -ti- mus and -ti-nus we have merely the third element in the first syllable; compare the Sanscrit punya-ta-mas, hya-ta-nas and nu-tnas, in which the dental appears unaffected by any foreign element, with ex-ti-mus, legi-ti-mus, cras-ti-nus, hes-te-r-nus ; and taci-tu-r-nus with the passive participle taci-tus. We come to a similar conclusion by comparing the older spelling of the affix, as in op-tumus, with the change in TUTTT-o-jmev = TV7rT-o-iu.es, vol- u-mus, dic-i-mus, whence it appears that the u is not a vocalized consonant, but a mere change of articulation for an original o - a. In this inversion, it really matters very little, so far as the mean- ing of the affix is concerned, whether the dental syllable is re- ferred to the second element or the third. This has been shown in the analysis of the third numeral, which admits of a similar explanation, whether we consider it as made up of ta+ra, or re- gard it as a corruption of an original tva-ra (New Crat. 1 57). Be this as it may, there can be no doubt that the combination n + t, which plays so important a part in Latin derivatives, is a reduplicated form of the third pronominal element, expressing ob- jectivity in its vaguest signification. Hence we find this combina- tion (resolved into a 1 ) as the neuter plural of all nouns; and either unresolved, or in various forms of assimilation, in the third person plural of verbs, in the active participles, and as a further affix to nouns corresponding in meaning and often in origin to the per- fect passive participle of the Greek verb and to obsolete Latin participles. In all these usages it denotes collective or vague 1 A curious collateral proof of this resolution is furnished by $\ido-ios for 3>\i,ovo-ios=3>\iovvTios from Xio[>T> : vide Steph. Byz. s. v.', and from this we may see that di7r\a9, or -airs = -a?, generally appear in Latin under the form -entum ; thus Acragas, Crumoeis, Maloeis, Pyxus, Taras, be- come Agrigentum, Grumentum, Maleventum, Buxentum, Ta- rentum. Similarly, we have ar-mentum, orna-mentum, &c., by the side of aw-^a[r\ = o-oj-^ei/r, &c. These extensions have occasioned some difficulties in Latin etymology ; it will be suf- ficient here to take the two interesting examples supplied by fundus and pondus. The former is obviously, on the principle just mentioned, an extension of fun\f\s or fon[f]s, the participle of fuo, " to pour out," which is involved in the agglutinate form fu-n-do (cf. per-do, cre-do, &c.), and in the frequentative fu-to. The nouns fon[f]s, "a fountain," i. e. "that which pours forth water," and fundus, properly " the bottom of a vessel for pour- ing out," hence the lowest part or basis of any thing, the solid part or foundation of a man's property, his estate or TO virapxov, exhibit the formation under discussion, without any additional elements. But pondus, gen. ponder-is, leads us to the same class of words as opus, operis, and these, as we have seen (above, p. 299), are terminated by the softened dental, as an additional mark of objectivity. The ablative pondo, however, shows that there must have been a word pondus, pondi, corresponding to fundus, fundi, and the synonymous ablative sponte, " by the weight or inclination," proves that the participial noun pons, pontis (in old Latin abbreviated into pos, Varro, L. L. V. I. p. 3, Miiller) originally referred to a weight laid down, or poured forth, such, for example, as an embankment, a mass of earth- work, or separate stones thrown into the water (ye-(pvpa), which was the primary notion of a bridge, as the means of crossing a stream : for we need not go far to prove the antiquity of stepping- 9.] DERIVATION AND COMPOSITION. 419 stones. While we have the d inpendo, pendeo, &c,, the t of sponte is retained in ponti-fex, as describing the functions of the priest, who settled the atonement for a specific fault by the imposition of a fine, on payment of which he pronounced the offender free from guilt, so that he stands opposed to the carni-fex, who ex- acted satisfaction on the body of the delinquent, without incur- ring the guilt or the danger of Shylock. We have a similar idea in the Hebrew JPS (see Prcelect. Phil in Deborce Can- ticum, Cantabr. 1848, p. 10). The connexion of the root fo = svo = hvo = xeF or ^Fe with po or spo, is farther shown by the community of meaning between ^o)/xa and pons, between oWi/Aw audfundo. And we may also compare fons with pontus, which properly indicates the depth of the sea (whence TTOVT'I( we have e-sur-io, &c. The variation in quantity between the desiderative verb and the noun or participle, with which it is so intimately connected, may be explained by the lengthened form of the verb, and illustrated by minis-tr-i compared with minis-ter-ium, &c. In some of the desiderative verbs the unorganic t is absorbed or assimilated, as in the supines or infinitives, like ven-um, molta-um, &c. (above, p. 360). Thus, from scat-eo we have scat-urio, from lingo, lig- urio, &c. It is scarcely necessary to observe after what has been said, that the verbs of the first conjugation in -ico must be referred to adjectives in -icus, whether they still exist or are only contained in these verbs : thus, alb-i-co presumes an alb-icus as well as albus ; compare aXcpos with Aev/ro's, &c. The same remark applies to the verbs in -ulo, which must have proceeded from nouns in -ulus ; compare modulo with modulus, &c. In speaking of derivative verbs we must bear in mind that, although a verb may furnish the basis of a series of derivative nouns, it may still have some parent stock among the older names of things. For example, although rog-atio, preca-tio, &c., are derivatives from rogo, precor, the fact that these verbs belong to the a conjugation shows that they are themselves derived from some primitive noun like p-rec-es. The following tables will help the student to determine when, in a given case, the substantive is formed from the verb, or vice versa. In general he will see that this depends on the appearance of a derivative pronominal adjunct in either case. 422 DERIVATION AND COMPOSITION. [On. XIII. I. NOUNS DERIVED FROM VERBS. Nouns in E = A-I are derived from consonant- verbs. facere .... faci-es Jldere .... fid-es con-spicere . . . speci-es Nouns in u or su from TU (compare ven-um with fal-sum and moni-tum) are derived from consonant-verbs. currere . . . currus^cur-sus discedere . . . discessus gradi (aggredere, &c.) . gradus ludere .... lusus=lud-sus vertere .... versus Consonant-nouns are derived from consonant-verbs. ducere .... dux legere . . . .lex munus capere . . muni-ceps pct-n-gere . . . pax regere .... rex Here the final -s of the noun must involve the syllable -us in the last-mentioned class. II. VERBS DERIVED FROM NOUNS. Verbs in A=aya l are derived (a) from nouns in A=ya. curare .... cura fugare .... fuga morari .... mora prcedari .... prceda (6) from nouns in i, in a causative sense. celebrare .... Celebris ditare .... ditis gravare .... gravis levare . . levis 1 The fact that the a- verb really includes the element i = ya is con- clusively shown by the form nego ^-neg-[a\o = nec-aio, " I say no " (above, p. 98). 10.] DERIVATION AND COMPOSITION. 423 Here the i of the crude form coalesces with the A, as mfunalis forfuni-alis, navalis for navi-alis, &c. A noun of the i declension occasionally forms a verb in A without any absorption of the i ; thus we have ab-brevi-are from brevis, and al-levi-are, as well as levare, from levis. (c) from nouns in o. bellare .... bellum donare .... donum numerare . . . numerus populare . . . populus probare . . . probus regnare . . . regnum sanare .... sanus (d) from consonant-nouns. fraudare , . . fraus generare . . . genus laborare . . . labor laudare . . . laus nominare . . . nomen onerare . . . onus orare . . . . os vocare .... vox This is particularly the case in compounds, as in belligerare from belliger, which is formed from bellum and gerere. And we must not overlook the fact, that nouns in A.-ya are formed in the same manner from consonant-verbs, not only in compounds, like agri-cola, homi-cida, &c., from colere, ccedo, &c., but also in simple forms, as ala, " that which raises," from alere ; lingua, " that which licks," from lingu-ere ; toga, " that which covers," from tegere, &c. ; so that we may always assume an intervening a- noun. Verbs in E are generally secondary extensions of simple roots. Some, like lucere, are derived from consonant-nouns. Not a few, like ardere, favere, fulgere, pavere, coexist with nouns in -or=yor. The same, however, may be remarked of verbs in A ; compare amare by the side of amor=am-yor, or ama-yon-s. For in-dulg-eo we must go back to an assumed dulgus, cf. the Greek ^oXi^os, ei/^eXe^s, &c. (above, p. 76). And fceteo must be derived from fce-dus (originally foetus, " by- 424 DERIVATION AND COMPOSITION. [Cii. XIII. gone = stale," cf. ef-fce-tus), which signifies "nasty," referring, in the first instance, to the smell, and, by a natural transition, to whatever is disagreeable : thus we speak of "a nasty accident," &c. Verbs in i are derived from nouns of the i declension. Thus we have audire .... auris=auvis finire .... finis lenire .... lenis mollire .... mollis vestire .... vestis When we seem to have an exception to this rule, we can always find, on looking into the question, that the crude form of the noun, from which the verb in i is derived, does involve this letter. Thus we have sepire from sepe, which is really an i noun; punire is from pcena, but the Greek 7roivq=iroi-vy-a 9 and the adjective impuni-s, show that the form ends in i; moliri comes from moles = mole-is ; sortiri from sors=sor-ti-s, gen. pi. sorti-um ; and blandiri is referred to blandus, which is really the participle of bl\a\o**fl\a\o t "to breathe or blow gently" (cf. fjidXa-Kos, fMaX-Oa-Kos, &c.) ; such phrases as blandus prece vel hostia, " soothing with prayer," or "sacrifice" (Hor. Ep. XL 1, 135. Carm. III. 23, 18), whence we have blandce preces (id. Carm. IV. 1, 8. A. P. 395), still retain the participial meaning; and this is presumed in the adverb blanditer (Plaut. Asin. I. 3, 69), so that the true form is blan-ti-s, whence bland-i-ri. Verbs in u, when this amounts to u-ya, are derived from nouns in u. Thus we have acuere .... acu metuere .... metus tribuere .... tribus This may be regarded as a singular case ; for no contraction is possible in a derivative verb of this kind. $ 11. B. Composition. Discrimination of Compound Words. The proper distinction between a compound word and the apparently compounded form consists in the fact, that the former is an union of two or more words, of which the last only is inflected, so that the preceding crude forms remain in a con- struct or subordinate state, whereas the mere juxta-position, or 11.] DERIVATION AND COMPOSITION. 425 apparently compounded term, is made up of separable elements, the inflexions of which are retained. Thus in such words as mag- nanimus, cedifico, we have entirely new compounds ; for the for- mer is an adjective made up from the ablative of quality, so that magnanimus = is qui magno animo est; and the latter is a deri- vative from a compound adjective cedificus, which involves the whole predication cedemfacio. On the other hand, the compo- sition is only apparent in res-publica, " the commonwealth," jus- jurandum, "an oath," juris-peritus, " a lawyer," animadverto animum adverto, " to pay attention to," " to take strict notice of," "to punish," &c. That these are not compounds, but merely juxta-positions of separable elements, is clear from the fact that, in those which are in direct agreement, both parts may be inflected throughout, as rei-publicce, jure-jurando, and all may be separated by particles, as in res vero publica, juris legumque peritus. There is no doubt, however, that these parathetic structures may pass into regular compounds, in the course of long usage. Thus from the phrase sesque for as semisque, " one and a half," we have the compound sesquipes, "a foot and a half," and its derivative adjective sesquipedalis. Again, when the first part of a real compound is an indeclinable word, it may be separated by a tmesis from the inflected part of the compound; thus we have inque salutatus for insalutatusque, and per mihi mirum videtur for permirum. In such forms as nihilo-minus, dum-taxat, ut-pote, vide-sis, sodessi-audes, sci- licet, &c., the two words are merely written in continuity to show their hasty utterance in the flow of conversation. Some- times it requires a careful analysis to prove that the word is really a compound. Thus annus or anus seems at first sight to be necessarily a simple word ; but it is proved by philological dissection (p. 163) to be a shortened form of aei-vos = del i/eo- ^uei>os (cf. oupa-vos and wKea-vos, according to the old notion of a wide superincumbent firmament, and a swift stream flowing round the earth), and the idea attached to the word is that which is expressed in Virgil's lines (Georg. II. 401) : Redit agricolis labor actus in orbem, Atque in se sua per vestigia volvitur annus. Then again it is an etymological discovery that prces, custos, opu-lentus, vio-lentus, &c., are not merely derivative forms, but real compounds (above, pp. 298, 393) ; and the same remark 426 DERIVATION AND COMPOSITION. [Cn.XIII. applies to the verbals in -bilis and -lundus, which involve the verb of becoming (fio), and are not to be explained, like the derivatives in -bulum, as vena-bulum, by a mere reference to the pronominal formations. $ 12. Classification of Latin Compounds. If we consider the Latin language only, we may conve- niently distribute all the compound words into four classes. (a) Determinative compounds are when the first part of the word defines the second ; such are the prepositional compounds : cognomen, dedecus, interrex, semideus, injuria, nefas, consul, collega, pronepos, &c., where the prefix qualifies the meaning of the whole word. In some cases the meaning is defined by an involved epithet, as in: cav-mdium, lati-clavis, lati-fundium, quatri-duum, &c. (b) Syntactical compounds are when the first word is governed by the second, whether the regimen is that of a noun dependent on another noun, as in galli-cinium, " the crowing of a cock," opu-lentus, "loaded with wealth," stilli-cidium, "a falling of drops ;" or, what is much more common, that of an accusative case governed by a verb, as in: agri-cola -quiagrum colit, brevi-loquens-breviter loquens, male-dicus = qui maledicit, signi-fer = qui signumfert; and in the verbs derived from such compounds, whether the intervening noun is still extant or not ; as : cequi-paro = cequum paro, castigo-castum ago, pur go = purum ago, &c. To the same class belongs aurigo from auriga or aureax = qui aureas agit, according to Festus (p. 8) : " aureax, auriga. Aureas enim dicebant frenum quod ad aures equorum religabatur ; oreas quo ora coercebantur" (cf. pp. 27, 4, 182, 23). If this interpretation is not sufficient, we must con- sider the aures or avrvyes of the chariot as referred to in the compound ; for as this term is applied to the side-pieces of the plough, which Virgil terms a currus 1 (Georg. I. 174), it may have been also a designation for something corresponding to 1 Modern editors read cursus, but it is difficult to see why cursus should be applicable to a plough, when the same word with an assimi- lation is considered inapplicable. It appears to me that the secondary word is more suitable to the metaphor than the direct verbal. Besides, it is clear from the verb torqueo that the plough itself, not its motion, is here alluded to in " currus a tergo torqueat imos." $12.] DERIVATION AND COMPOSITION. 427 these side-pieces in the wheeled vehicle. I may remark, in passing, that the oriel window, in Gothic architecture, was un- doubtedly so called from its projecting like the human ear from the side of a building. The old spelling shows this. Thus we find in an ancient MS. : " The Lords always eat in Gothick Halls, at the high table or oreille (which is a little room at the upper end of the hall where stands a table,) with the folks at the side tables ;" in accordance with which we find in Matthew of Paris (ap. Ducang. s. v.) : " ut non in infirmaria, sed seorsim in oriolo, monachi infirmi carnem comederent." Now it is well known that oreille is a representative of auriculus. So that the oriolum or "oriel" is the "ear-window" or projecting chamber used for privacy and retirement. (c) Auxiliary compounds are when two verbs come together, and the second helps the former either in a predication of time, or by introducing a modification of meaning or reference ; thus we have : ama-vi = amare-fui, ven-do=venum do, ven-eo = venum eo, arcesso - accedere sino, treme-facio = tremere facio, &c. ; and to the same class belong all the tenses in -bam and -bo, -vi and -veram. (d) Possessive compounds are when the first part denotes the manner, in which the thing, denoted by the last word, is pos- sessed by the subject, to which the whole compound is referred either as predicate or epithet; thus we have: aheno-barbus, alti-sonus, crassi-pes, magn-animus, in which the first part is a declinable word; and affinis, concors, nefastus, benefaus, inermis, bimaris, elinguis, in which the first part is an unin- flected particle : in both cases the possessive adjective, consti- tuted by the whole compound, involves a determinative com- pound, which is made rnoveable, so as to agree with different substantives. Among these nouns, we must take care to dis- tinguish between those in -ceps from caput, as bi-ceps t gen. 5i- cipit-is, and the syntactical compounds involving -ceps from capio, as muni-cep-s, gen. muni-cip-is, &c. Although this classification of the compounds is sufficient for all practical purposes, so far as the Latin language alone is con- cerned 1 , it is convenient, with a view to comparative philology, i Livy remarks incidentally (XXVII. 11) that the Latin language was inferior to the Greek in the power of forming compound words. 428 DERIVATION AND COMPOSITION. [On. XIII. to inquire how far these composite formations admit of arrange- ment in accordance with the system of the Sanscrit grammarians. As I have compared the six classes of the samdsa with the Greek compounds (New Crat. $ 309), and as Bopp has subse- quently adapted this arrangement to his more general purposes (Vergl Gramm. pp. 1427, foil. VI. Abtheil. 1852), it may be as well to place the Latin formations under these heads. The six classes of the Sanscrit samdsa are designated by names some of which describe and others exemplify the nature of their construction ; and they are arranged by Vopadeva in the following order : (1) The first are described by the term dvan- dva, i. e. " two and two," " pair," or " doubling," and consist of mere aggregations of words which might be written separately and joined by a copulative conjunction, as agnt-$6mdu, " Agni and Soma," in the dual; brdhmana-kshatriya-vit'-^udras, the four Indian castes, in the plural; &c. ; (2) the second are exemplified and named by the compound bahu-vrihi, " that which has much rice," and therefore consist of compound epithets ; (3) the third are called karma-dhdraya, " that which comprehends (dhdrayati) the object (karma)" and include such words as mahd-rdjah, " a great king," where a substantive is defined by an uninflected epithet prefixed; (4) the fourth, exemplified by tat-purusha, " the man of him," comprises compounds formed of two or more nouns, the first set being in some oblique case governed by the last, which may be a substantive, adjective, or participle in -ta, as rdja-purushah, " the king's man ;" (5) the fifth, called dvigu from dvi, " two," contains compounds of which the first part is a numeral and the second a noun, as chatur-yuga-m, " the four ages of the world ;" (6) the last class is called avyayi-bhava, or " adverbial," and is made up of indeclinable words, the first part being some particle, and the last a noun in the neuter gender, as a-san$aya-m t " without doubt," ati-mdtra-m, " over the mea- sure." It appears from this enumeration that classes (3) and (5) are determinative, class (4) is syntactical, class (2) is pos^ sessive, and class (1) is merely an aggregation of terms. The following examples will suffice, so far as the Latin language is concerned. (1) There are no Latin dvandva, unless we recognise such a form in su-ovi-taurilia = suile+ovile + taurile. But the Latin language, especially in its oldest form, abounds in examples of 12.] DERIVATION AND COMPOSITION. 429 nouns aggregated together so as to form one notion, and without any copulative conjunction ; thus we have populus Romanus Quirites for the united people of Romans and Sabines (Niebuhr, H. R. I. p. 294) ; Patres Conscripti, for the combination of two elements, the original and the elected deputies, in the senate ; sarta tecta for sarta et tecta, " sound in wall and roof " (Festus, p. 322), &c. Notwithstanding this old Roman usage of com- bining related words by mere juxta-position, we find that in later times the language became pedantically accurate in the employ- ment of copulative conjunctions ; two epithets to the same word required the intervention of one of these particles ; and the best writers made a consistent distinction between et-ad=GTi the particle of addition, -que the particle of combination and paral- lelism, and at-que (shortened into ac), which is compounded of the other two, and implies that there is not only an addition, but also an intimate connexion between the things coupled together. (2) Of bahu-vrihi compounds there is a long list in Latin. In addition to the possessives mentioned above, we have com- pounds made up of substantives and their epithets, as versi- color, multi-caulis, acu-pedius ; of numerals and substantives, as quadru-pes, bi-dens, quinque-folius ; of prepositions and substan- tives, as corn-modus, corn-munis, ex-cors, &c. ; of verb-roots preceded by particles, as male-dicus, bene-ficus, &c. To this class belong the opposites, pro-sper or pro-sperus^ " in accordance with our hopes" (Non. 171, 25: sperem veteres pro spem dice- bant, unde et prospere dicimus, h. e. pro-spe) and a-sper, " con- trary to our hopes " (i. e. a spe), as in Sallust, Cat. c. 26 : " aspera foedaque evenerant," compared with Jug. c. 63 : "cuncta prospera eventura." It is more usual to compare prosper with 7rpoopos. (3) Karmadhdraya compounds in Latin are such as pcen- insula, neg-otium, pro-nepos, ab-avus, in-imicus, &c. (4) We have tat-purusha compounds in Latin words like tibl-cen, for tibii-cen, auri-fodina, opi-fex for operi-fex, lapi- cidina for lapidi-cidina, mus-cipula, imbri-citor, &c. (5) The Latin determinatives include many dvigu com- pounds as a subordinate class ; such are bi-noctuum, quinqu- ertium, bi-ennium, quadri-vium, &c. (6) Adverbial compounds or avyayi-bhava are in fact cases 430 DERIVATION AND COMPOSITION. [On. XIII. of nouns with or without epithets or prepositions ; as : obviam, a/atim, admodum, multi-modis, imprimis, &c. To this class we must refer the correlatives se-dulo = se-dolo, " without feeling any weariness," and se-fraude, " without incurring any loss." The epithet mains, technically applied to dolus in the old laws, proves that it does not of itself imply " deceit" or " guile " (see Festus, p. 69), and the verbs dol[a]o, " to belabour," doleo, " to labour," whence dolor, "labouring," show that the primary meaning of the word is " pain" as connected with exertion. The root is that of tol-lo, tolero, rXdw, a-0\ios, &c., and Doderlein (Syn. u. Et. I. p. 118) has well compared sedulo with d-Trof^s = hand gravate in Soph. (Ed. C. 293. In the same way, it may be shown that frau[d]s-fra-va[d~\s (above, pp. 122, 298) sig- nified deprivation as an effect, before it indicated dishonesty as the cause. All these examples refer only to nouns, whether substantives or adjectives, and adverbs, considered as cases of nouns. Strictly speaking there are no synthetic or organic compounds of verbs ; those, which have a preposition or adverb by way of prefix, are merely parathetic combinations, and, with the exception of an. occasional assimilation, the two parts of the word are not really fused into one, and a tmesis or separation is still possible. When a verb contains two or more distinct roots, so melted down into one whole as to be incapable of divulsion, we also find that the verb is a derivation from some compound noun. Thus while bene-facio, male-dico, com-pono, per-lego, and the like, are shown by the unaltered conjugation of the verb to be mere juxta-positions of separable elements, lceti-jfic[a]o, belli-ger[a]o are manifestly not merely parathetic combinations of Icetumfacio and bellum gero, but verbs derived from the adjectives Iceti-ficus, belli- ger, pro- bably through a noun of action in -a = ya. As verbals in -us, like beneficus, Icetificus, maledicus, &c. are equivalent in mean- ing to the present participles of the parathetic verbs which they represent, and as their comparatives are actually formed from the participles (e. g. maledicus, maledicentior), we may conclude that the termination is the mutilated form of some pronominal affix, like that of the Greek participles in -w^vas or via (New Crat. 414). When the first part of a genuine compound is an inflected word and the second begins with a consonant, the vowel of con- 12.] DERIVATION AND COMPOSITION. 431 nexion is generally ?, as in causi-dicus, corni-ger^ cedi-Jico. The vowel of connexion is sometimes omitted, as in nau-fragus for navi-fragus, mus-cipula for muri-cipula, puer-pera for pueri-pera. Sometimes a consonantal affix is also dropt, as in homi-cida for homini-cida. And in a few cases the connecting vowel is not ?, but o or u ; thus we have aheno-barbus, opu- lentus, turbu-lentus, Troju-gena, vio-lentus. It is possible that the articulation may be affected here by the letters n and/, which precede, or by the liquid I which follows the vowel. In tib'icen tibi-i-cen we have a contracted i, but tubt-cen follows the general rule. CHAPTER XIV. CONSTITUTION AND PATHOLOGY OF THE LATIN LANGUAGE. 1. Genius of the Latin language. 2. Abbreviations observable in the written forms. 3. Ancient testimonies to the difference between the spoken and the written language. 4. The poetry of the Augustan age does not represent the genuine Latin pronunciation ; 5. which is rather to be derived from an exami- nation of the comic metres. 6. The French language is the best modern repre- sentative of the spoken Latin. 7. The modern Italian not equally so ; and why. 8. Different dialects of the French language. 9. But all these dialects were closely related to the Latin. 10. Leading distinctions between the Roman and Romance idioms. 11. Importance and value of the Latin language. J 1. Genius of the Latin Language. EVERY language may be considered as an organic body pos- sessing within itself a principle of vitality, but also capable of disintegration and decay. We may therefore, without strain- ing the metaphor, speak of its constitution, or power of con- tinuing in a healthy state ; and also of its pathology *, or of the symptoms of that disease to which it is by its very nature more peculiarly liable. Accordingly, if it were necessary to describe in one sen- tence the genius and constitution of the Latin language, one could not do this better than by defining it as a language which is always yearning after contraction. Whether this tendency is indicated in the written remains by the usual processes of syni- zesis, assimilation, and apocope ; whether it appears in the slur- ring-over of syllables, by which the scansion of the comic metres is effected ; or whether we perceive it in the systematic abbre- viations which mark the transition from the Roman to the Ro- 1 Lobeck, who has called one of his works Pathologies Sermonis Greed Prolegomena, gives the following explanation of this term as applied to language : " Cui nomen Pathologice imponere non nefas duxi, fretus auc- toritate et exemplo Theodoreti, qui, similitudinem a re medica transferens, librorum suorum elegantissimos 7ra6rj^,ar(i!>v 'EXXqi/tKeSz/ depaTrevriK^v in- Bcripsit. Videlicet, vocabula quoque affectiones suas habent, non homines solum, et eas similes humanis, pleonasmos, ellipses, tropasque rarias, ad quas et cognoscendas diagnosi opus est et ad corrigendos therapia ; nam et hoc nomen usu ceperunt grammatici" (Prcef. pp. v. vi.). $ 1.] CONSTITUTION AND PATHOLOGY, &c. 433 mance languages, it is still one and the same, it is the type of the language, in its infancy, its maturity, and its decay. The most distinct and vivid picture of the Latin language is, therefore, to be derived from a consideration of this peculiarity, as developed I. In the written language of ancient Rome. II. In the spoken language of ancient Rome, so far as we can discern it in the remains of the comedians. III. In the modern languages (and particularly in the French) which are derived from the Latin. ^ 2. Abbreviations observable in the written forms. I. With regard to the written forms in which the Latin language has been handed down to us, it would not, perhaps, be too large an assertion, if we said that every etymological diffi- culty arises more or less from this systematic abbreviation. It is true that all languages are more or less liable to this dimi- nution of the forms of speech, and it is the more observable in proportion as the syntax militates against the permanence of the etymological structures. But the distinctive peculiarity of the Latin appears in the fact that this abridgment coexists with a perfect maintenance of the word-forms, as far as the inflexions are concerned, and does not spring from the superabundance of syntactical substitutes. It is in fact a result of the haste and impatience of the Roman lords of the world, and is quite inde- pendent of the inherent principles of the language. If we look to other idioms, we shall see that, although the Sanscrit f loka runs the words into one another, and so affects the terminations, there is no appearance of abbreviation in the middle of the words. The Hebrew and other Semitic dialects have broken down all the for- mative machinery, but the triliteral root maintains its consonants, except where assimilation becomes inevitable. To the latest period of Hellenistic Greek the spoken and written language tolerated the syllabic articulation of the longest compounds. High-German still revels in the manufacture of polysyllables. And even the Scla- vonic idioms, which have so many points of contact with the Latin, are not led, even by the concourse of consonants, to abridge their composite forms ; and in the haste of polite conversation we may 28 434 CONSTITUTION AND PATHOLOGY [Cn. XIV hear the most sesquipedalian utterances at St Petersburg 1 . Ifc is only the Latin language and its daughters, in which we observe this systematic shortening, first of spoken, and afterwards of written words, and therefore we may both attribute it to the habits of the people, and describe it as the characteristic feature of the Roman and Romance form of speech. There are two ways in which this tendency manifests itself in the loss of the termination, and in the coalition of syllables in the middle of tbe word. When dipt or mutilated words are common in any language, the cause is to be sought in the strength and prominence of the single accent 2 , which is generally thrown forward as far as pos- sible, and in the impatience with which practical and busy men hurry through that part of their work which consists in talking. The rules of the Latin metrical system might have prepared us for something of the kind. It has been shown in a former chap- ter (above, p. 225), that the triple recurrence of the ictus was the essential feature of the Saturnian verse, the thesis being ob- served or neglected at the pleasure of the composer. Similarly, the accentuated syllable of a word, or that on which the emphasis of pronunciation was allowed to fall, was supposed to represent the significance of the term, just as the weight of a body is con- sidered to be collected at its centre of gravity ; and the other syllables were slurred over or cast aside as superfluous and un- necessary incumbrances. As instances of this, one might adduce a number of syncopised forms of common words. We have ac for atque, amavere for amaverunt, amare for amaris, ccel for ccelo, do for domo, dein for deinde, gau for gaudio, nee for neque> neu for neve, ni for nisi, pa for parte, po for populo> seu for siv e 3 , &c. ; and, not to speak of the visdrgah, by which a final s, though written, was not pronounced (New Crat. 242), we have a number of words in which the termination -is or -us was re- 1 E.g. the common Russian for "present my compliments to your father" is zasvidyetel'stvuete moe pochtenie vashemu batyushkye i. e. testi- faaminor meam venerationem vestro patri, where the conventional verb is as long as an Aristophanic compound. 2 See Dietrich, Zur Gescli. d. Accents im Lateinischen, Zeitschr. f. d. Vergl. Sprf. I. pp. 543, sqq. 3 See other instances in Columna's Ennius, p. 137. 2.] OF THE LATIN LANGUAGE. 435 gularly abridged to -e : such as, ille, ipse, mage, &c., for ollus t ipsus, magis, &c. The contemptuous familiarity with which the master addressed his slaves gave rise to a number of abbrevi- ations of the Greek names of the latter. Thus Artemidorus was called Artemas (Varro, L. L. VIII. 21), EpapJiroditus became the Epaphras of St. Paul, and Demodorus shrunk into Demas or Dama (Hor. II. Serm. 5, 101 ; ibid. 6, 54). But the hasty pronunciation of the Romans, so far as it was exhibited in the written forms of the language, appears chiefly in the omission of letters or syllables in the middle of words. If the hurried talker has time to pronounce more than one syllable, he would rather preserve the termination than any of the middle sounds. Indeed, the accent sometimes stands over the ruins of a number of syllables, which it has fused into one compound articulation. The following instances, selected from a very large number, may serve to illustrate this : Ala for Axilla (Cic. Orat. c. 45, 153), aula (plla) for auxilla, bruma (scil. dies), " the shortest day," from brevimus, career from co-arceo, contami- nare, the derivative verb from contagimen, contio for conventio, cunce for cubince, dixti for dixisti, exilis for exigilis (from egeo, cf. exiguus), imus for iiifimus, jusso for jussero, lapicidince for lapidicidince, mala for maxilla, mollis for mobilis, omentum for opimentum, otium for opitium, Pollius for Publilius (Nieb. H. R. I. n. 977),paullus for pauxillus, porcet for porro arcet (Fest. s. v. arceo, p. 15, Muller), prudens for providens, puella for puerula, qualus for quasillus, sacellum for sacraculum (comp. sakaraklum Her ekleis- sacellum Her culis, in the Cippus Abel- lanus, 1. 11), solari for sublevari, stipendium for stipipendium, sublimis for sublevimis (cf. jueTewpos), subtilis, " fine-spun," for subtexilis (comp. subtemen, tela), summus for supremus, tandem for tamendem, vdnus for vacanus, velum for vexillum, &c. This is particularly remarkable in the flexion-forms of nouns and verbs ; and, as we have seen above, the complete forms cannot be restored until we have made good the losses occasioned by this systematic abbreviation. Thus we have regularly dici-er, or even did, for dicerier ; and less commonly sumpse for sumpsisse, &c. In some cases this abbreviation will appear in a compound, though the full form is retained in the simple word. Thus, we find agriitus and cognttus by the side of notus, pejero and dejero by the side ofjiiro, and the same difference of quantity may be 282 436 CONSTITUTION AND PATHOLOGY [Cn, XIV. effected without any change in the spelling, as in niJiilum by the side of hilum. This influence of the accent is the more felt in proportion to the length of the form ; and sometimes we find two or three abbreviations in the same compound. For example, although the gen. cujus retains the original termination, this has been shortened into i in the compound: cut-cut-modi for cujus- cujus-modi (Cic. ad Att. III. 22). The Romans, however, were not satisfied with getting rapidly through their simple words and regular compounds. The same principle was applied to the parathetic formations: thus magis auctus was condensed into mactus 1 , magis volo was written malo, non volo became nolo^ and so forth ; and not only so, but we also find that in the case of quasi-compounds, made up of two or more words, ' which are not amalgamated by the loss of inflexions into one whole, some part of the termination of the first word is regularly omitted, and thus the group is subjected to the domination of a single accent. It may be sufficient to mention such words as audin = audisne, JEcere, Ecastor, Epol = [per] cedem Cereris, Castoris, s. Pollucis 2 , ho'die = hoc die, meridie = medii die, multimodis multis modis, nudiustertius = nunc dies tertius, omnimodis = omnibus modis 9 refert = rei- fert, sis = si vis, sodes = si audes, tectifractis = tectis fractis, 1 J. J. Scaliger says (Seal. Pr. p. 105): "mactum veteres Roman! vocant auctum. Herbam adultam Cato vocavit mactam, nempe quod ita aucta esset. Macta hostia cum frugibus et mola aucta erat ; sic macta ara, quod verbenis aucta et cumulata. Postea mactare hostiam pro caedere dicebant, ne scilicet caedein nominarent, quia nunquam csedebatur nisi frugibus macta esset. Nunquam autem mactabant hostiam quin dicerent ' macta esto hac mola salsa/ Sic cum Deo alicui vinum libabant macte hoc vino esto dicebant in vocandi casu, quod est TCX VL ^V grainmaticorum, nam mactus esto dicendum erat. Sic Persius : stemmate quod Tusco ramum milksime ducis, pro millesimus." This passage seems to have been taken by the compiler of the Scaligerana from Scaliger's letter to Vertun, Mus. Grit. II. p. 47. 2 It has been shown above (p. 257) that the dentals, when preceded and followed by vowels, are frequently omitted in the French forms of Latin words ; and similarly, D and T must have been dropt in the old pronunciation of some Latin words, such as pater, modo, quidem. The words Epol and Ecastor, with es for edis, &c., exhibit the same fact in the written forms of the old Latin language, and therefore complete the induction. 2.] OF THE LATIN LANGUAGE. 437 vasargenteis = vasibus argenteis, &c. Then, again, we find a number of verbal juxta-positions, for we cannot term them com- pounds, belonging to the same class : such are pate-facio = pa- tere-facio, sci-licet = scire licet, vide-licet = videre-licet, &c. It has been shown above, that many verbs in -do, -eo, -lo, -so, may be explained in the same manner ; and that a similar analysis may be applied to the secondary tenses of every verb. It is not necessary to pursue this part of the subject any farther; for we can scarcely read a page of Latin without finding some proofs of the general rule \ $ 3. Ancient Testimonies to the difference between the spoken and the written Language. II. But although there is much abbreviation in the written forms of the Latin language, the orthography of the Romans expressed much more than their articulation. This is more con- spicuous in proportion as we take a more polished and advanced period of the language. Before proceeding to demonstrate this from the metres of the comedians 2 , it will be convenient to adduce some passages, in which the difference between the written and the spoken language of ancient Rome is expressly recognised. When Cicero's Crassus (de Oratore, III. 11, 41) is speaking of the true mode of pronouncing Latin, he says : " I 1 The reader might bo referred for further instances to a paper on the " Ausfall oder Verwandlung der Consonanten durch Zusammenzie- hung oder Assimilation in dcr Lateinischen Sprache," in the Rheinisch. Museum for 1839 (pp. 42 81) ; but, although most of the words there enumerated are cases of contraction, the author, Professor Schwenck, has not been happy in his restorations. In the same volume of the Rhein. Mm. p. 297, there is a criticism on Prof. Schwenck by Dr. Duntzer. 2 The first attempt, so far as I know, to apply this very natural and obvious test of the old colloquial pronunciation of Latin, was made by Mr. Hallam in his View of the State of Europe during the Middle Ages, Vol. III. p. 316, where he says : "a decisive proof in my opinion of the deviation which took place, through the rapidity of ordinary elocution, from the strict laws of enunciation, may be found in the metre of Terence. His verses, which are absolutely refractory to the common laws of pro- sody, may bo readily scanned by the application of this principle." But perhaps every observing reader of the Latin dramatists, especially since the time of Bentley, may have arrived at some similar conclusion. 438 CONSTITUTION AND PATHOLOGY [Cn. XIV. do not like the separate letters to be either pronounced with pedantic accuracy, or slurred over too carelessly." This shows that, though an uneducated countryman might represent by his articulation too little of the written word, it would be a fault, on the other hand, if the scholar recollected too much of his spelling. Again, Suetonius, who had seen the chirograph of Augustus (Vit. Octav. c. 87), writes thus about his method of spelling (c. 88) : " He did not strictly attend to orthography, that is, the method and laws of writing as taught by the grammarians ; on the contrary, he seems rather to adopt the opinion of those who think that we should write just as we talk. For as to his often changing or omitting not letters only, but whole syllables, this is a common inaccuracy ; nor would I remark the fact, did it not appear strange to me that he should have superseded a con- sular legate as being illiterate, because he saw in his handwriting ixi for ipsi" From this it is clear, that in the time of Augustus people did not pronounce as they wrote. Quintilian, too, ex- pressly tells us (Inst. Orat. XL 3, 33), that, " although it is necessary, on the one hand, to articulate every word, yet it is wearisome and disgusting to take account of every letter, and as it were to reckon them up : for not only is the crasis of vowels very common, but even some of the consonants are disguised when a vowel follows;'* and then he quotes the examples of both ecthlipsis and synalcepha in Virgil's multum ille et terns. Much to the same effect are Cicero's remarks about the conglu- tinatio verborum or avoidance of the hiatus by a kind of crasis or synizesis (Orator, c. XXIII. 78), and he says expressly that the Latin language repudiates a concurrence of vowels (Orator, c. XLIV. 150 : " quod quidem Latina lingua sic ob- servat, nemo ut tarn rusticus sit, qui vocales nolit conjungere"). From these and other passages which might be quoted, we conclude that the written language of Rome could not be taken as a standard of even the most exact and careful pronunciation of educated men living in the city itself, whose mode of pro- nouncing was strikingly different from that of the provincials (Cicero, de Oratore, III. 11, 43, cf. Brutus, c. LXXIY. 259) 1 . Accordingly, the colloquialisms of the country people must have 1 On the difference between the lingua urbana and the lingua rustica, see Adelung, Mithridat. II. p. 464, and the works quoted by him (p. 467). 3.] OF THE LATIN LANGUAGE. 439 been still farther removed from the written language of the day, and are less to be inferred from it. The true way of considering the Latin language, if we wish to realise to ourselves its spoken form, is to regard it as strug- gling with the fetters of the Greek metrical system. $ 4. The Poetry of the Augustan age does not represent the genuine Latin Pronunciation ; The poetry of the Augustan age shows us, that the Greek rules of metre are observed with greater strictness by the Romans, who adopted them, than by the Greeks themselves. With the Roman poets the trochaic dipodia, that important rhythm in lyric poetry, always appears under the form of trochee + spondee ; whereas in the Greek system there was nothing to prevent the dipodia from being pure. Take, for instance, the Sapphic verse: Horace's second foot is always a spondee, Sappho's as often a trochee. The same minute accuracy, or rather sameness, is observable in their anacrusis. In Horace's Alcaics the anacrusis at the beginning of the first three lines is rarely a short syllable ; but in his Greek models he would as often find a short syllable as a long one 1 . All this 1 The remarks in the text refer to a mode of scanning the Sapphic and Alcaic stanzas, which is not in accordance with the common doctrine, but which is, I think, demonstrably correct. The Sapphic and Alcaic stanzas differ only in a varied arrangement of the same elements ; and the first three lines of the Alcaic stanza begin with an anacrusis, which the Sapphic rhythm excludes. If we call the dactyl A, the trochee B, and the anacrusis #, the law of the verse appears in the following simple formulae : (1) Sapphic stanza : 2 B + A + 2 B (ter) 2 A. (2) Alcaic stanza : x + 2 B + 2 A (bis) # + 4B. 2A + 2B. Thus, for example, the Sapphic contains three lines like Jam sa\tis ter\\ris nivis \\ dtque \ dirce ||, and one like terndt \ firbem \\ ; where, it will be observed, the second member of the trochaic as well as of the dactylic dipodia is always a spondee. The Alcaic has two lines like Vt\des ut \ dlta || stet nive \ cdndidum \\ , one like Sil\va6 Ia\loran\\t6s ge\luque ||, and one like Flfimina \ constite\\rint actito. With regard to the Sapphic verse, in particular, it will not perhaps be easy to correct errors which are 440 CONSTITUTION AND PATHOLOGY [On. XIV. leads to the inference, that the poetry of the Augustan age was recited with a pedantic accuracy at variance with the genius of the language ; and as the German opera-singers at the present day soften down their gutturals in order to accommodate their language to the flowing rhythm of Italian music, so the Romans, in the days of Horace and Virgil, were proud of their foreign fetters, and were glad to display the ascendancy which van- quished Greece had gained over the minds of her rude con- querors. J 5. which is rather to be derived from an Examination of the Comic Metres. This refined and mincing pronunciation was, of course, less compatible with the colloquialisms of comedy than with the elegant stiffness of copied heroic or lyric poetry. Consequently, though the comedians borrowed their metres from the Greeks, they were content to pronounce the words as they were uttered by the common people ; and as the busy talkers of the forum were wont to clip and contract their words, so the syllables usually omitted in speaking were not taken into account on the comic stage. When, therefore, we can recognise the law of the verse in a Latin comedy, but find that the syllables, as they stand written in many of the lines, are more numerous than is necessary for the feet of the verse, we may safely conclude that the superfluous syllables were omitted in the pronunciation of the actor ; and if by him, a fortiori, that they were habitually slurred over by the majority of his audience. This opinion will be confirmed, if we discover, on farther enquiry, that the syllables so dispensed with are not found in the corresponding forms exhibited by the modern idioms which derive their origin from the language of ancient Rome. sanctioned no less by the practice of schools than by the well-known jingle of the Anti- Jacobin ; but it is not to be borne that this ignorance should exalt itself to dogmatism. In the third number of the Classical Museum (pp. 338, sqq.) there is an article in which we are told that the Sapphic verse, " recited with the true metrical quantity and the natural spoken accent," will read thus : Jdwm sattees \ taerees \\ nivis autque \ deerce, &c. ; and that the following is a Sapphic of the same kind : che il gran sepolcro liber o di Christo ! And this is delivered, not as a modest sugges- tion, but as a decree of oracular wisdom. 5.] OF THE LATIN LANGUAGE. 441 The following instances, few out of many, may be sufficient to establish this 1 . Let us first take some of the short impera- tives, which are, by the nature of the case, especially liable to hurried pronunciation. As our look ! has degenerated into lo !, and the Latin vide has become the Italian ve', and the French voi or v* (in voi-ci, v'la) ; so in Terent. Adelph. II. 2, 31, it is clear that we must pronounce this line : Labdscit : fin* hoc hdbyo : ve 1 si sat placet. Here, also, we have Italian abbio. Similarly, as Cicero tells us 2 that cave ne eas was pronounced cauneas, we see that the follow- ing line (Phormio, V. 1, 37) must be pronounced : Sed per deos atqu' hommes, medm's' home, cdu resciscat quisquam. This line also furnishes the French abbreviation hommes. A question might arise whether deos might not be a monosyllable = dyos, as in Plaut. Trin. 520, and homines a dissyllable = homines; but the commonest rules of emphasis plead for the arrangement which I have proposed. It is impossible that deos should be a mere thesis, and that an accent should fall on atque. Then, again, as the French say tai, it is clear that tace is a single long syllable in the following line (Adelph. II. 4. 16) : At ut omne reddat omne reddet tai-mod', dc suire hdc sequor. Which line also furnishes us with the imperative suire for sequere, if we may in this case also follow the French analogy. In general there seems to have been a tendency towards softening down the guttural into its ultimate form, the vowel i. This has obvi- ously taken place in faire and ceil, derived from facere and oculus ; and not only is the imperative tace a monosyllable, but also its indicative facet, as in the following line (Adelph. IV. 5,5): Tait: c&r non lud* hunc dPquantisper melyus est. 1 The reader, who desires a more copious induction, may refer to the well-known essays of Bentley and Hermann ; to some compilations, de- rived from these and other sources, in the Journal of Education (Vol. II. pp. 344, sqq.)> and in the Penny Cyclopaedia, s. v. Terentian Metres; and to Ritschl's valuable Prolegomena to Plautus. 2 De Divin. II. 40, 84 : " Quum M. Crassus exercitum Brundisii imponeret, quidam in portu, caricas Cauno advectas vendens, Cauneas clamitabat. Dicamus, si placet, monitum ab eo Crassum, caveret ne iret" 442 CONSTITUTION AND PATHOLOGY [On. XIV. Where for aVquantisper compare Italian alcuno, and the French aucun, from aliquis units. It can scarcely be doubted that Adelphi, III. 2, 20, was pronounced as follows : 'AcTlescent* ips j 3npV ceilos: posthac praecip'tem dar&m; and that in III. 2, 37, lacrymas is a dissyllable after the ana- logy of larme, and of serment from sacramentum. Similarly, in Heaut. V. 5, 16, quoted below, as the ictus falls on facile, we may conclude that it was pronounced as a single long syllable. Festus tells us that there was a form facul, and facile appears as a mere anacrusis in the Scipio epitaph (c. 5) ; above, Ch. VI. 20. Perhaps the most singular instance of this omission of the guttural is furnished by the French faible from flexibilis ; for in this there is a double collapse. The imperatives obi, redi, are monosyllables with the omis- sion of the unnecessary b and d (Adelph. II. 1, 13, and 36), and jube throws off its b (Adelph. V. 6, 1), as it does in the perfect, &c. The phrase bono animo es is shortened for the same reason as the other imperatives. In Plautus (Rudens, III. 3, 17) it forms a cretic : 'O salutis meae spes tac* dc bon-ame 6s. We observe the same sort of abbreviation in a number of nouns of common occurrence ; such, for instance, as express the nearest degrees of family relationship. The compound parricida indicates a contraction of pater analogous to the French pere, and the word was probably so pronounced in such lines as (Adelph. I. 1, 51) : Hoc pater ac domlnus interest : hoc qui nequit ; i. e. Hoc pre ac donnus interest : hoc qui nequit. and (Adelph. I. 2, 46) : Natura tu illi pater es, consiliis ego ; i. e. Natura tu gli pere es, consigns ego. where the ictus falls upon it. In the latter line, as tu is em- phatic, an elision would be inadmissible ; we must therefore pro- nounce illi either as the Italian gli or as the French lui, and this gives us another modern analogy. In the former line domi- nus is probably a dissyllable following the analogy of domina, which becomes donna in Italian, and dame in French. Similarly, $5.] OF THE LATIN LANGUAGE. 443 Jtomines is a monosyllable in the passage quoted above from the Phormio ; animus becomes ame ; femina, femme, &c. That puer was often a monosyllable appears from the forms por, pora, which occur in inscriptions, from the compounds Luci- por, Marcipor, &c., and from the Spartan Trolp for Trats. In Heaut. V. 5, 16, we should read or pronounce as follows : Gnate myo pol tt do pollam [or pwellam] ttpidam quam tu fail ames. The mood of ames shows that the emphatic illam would be as out of place here, as it is appropriate in the following line. And do, which we should have expected in the first instance (cf. Andr. I. 5, 60 ; II. 2, 15), has been turned into dabo, partly from a confusion between the readings dopuellam and daboillam, and partly by an anticipation of dabo in v. 19. With regard to the monosyllabic ti for tibi, the Romans frequently omitted b in the middle of a word : this is most common in the dat. and abl. pi. of the first declension, and is also observable in the French derivatives ; such as oil and y from ubi and ibi. For the change of puer into por, we may also compare the transformation of fuere and fuerent into fore, forent. Perhaps two of the most striking instances of this dipt pro- nunciation are afforded by the scansion of the particles quidem and modo, in both of which the d is omitted. With regard to the former even Bentley remarked that it must be frequently a monosyllable in Terence (ad Andr. I. 3, 20). The following reasons have been adduced to prove that it was so in general. (1) The analogy of item, shortened from itidem, will support the pronunciation of qvCem for quidem. (2) As it is an enclitic, and is regularly attached to certain words, in the same way as ircp, 76, &c. in Greek, it seems reasonable to suppose that it would be peculiarly liable to curtailment. Now, if we retain the full form of quidem with some of these words, we alter their quantity, and so sacrifice the principal word in order to preserve a mere appendage. Thus, ego-quidem, or eg-quidem, is marked equidem in books on Latin prosody, and siquidem, quandoqui- dem, are marked siquidem, quandoqutdem, although the true quantity of the separate words is si, quando ; and though in other compounds quandoque, quandocunque this quantity is invariably retained. It follows, therefore, that quandoquidem must have been pronounced quandoqu'em ; siquidem, slqu'em ; 444 CONSTITUTION AND PATHOLOGY [Cm XIV. and equidem, eqifem ; just as me quidem must be scanned me qu'em in Pers. I. 10 : Littera : per me quidem slnt omnia protinus alba. In the same way it is manifest that modo must often have been a monosyllable : see e. g. Ter. Andr. II. 1, 2, and II. 4, 6. In the languages derived from the Latin the compound quomodo is represented by como Sp., come It., and comme Fr. ; in which the d is omitted, and in the last, as in the old French cum (be- low, 9), the syllable is dropt altogether. The knowledge of this abbreviated pronunciation enables us sometimes to correct a faulty reading. But although KitschV was well aware that modo was monosyllabic, and though one of his best MSS. in Plautus, Trin. II. 4, 179 = 580, gives the reading si for st, he has allowed actumst to stand when actum sit would improve both the metre and the syntax : L. Set, Stasime, abi hue ad meam sororem ad Calliclem: Die, hoc negoti quomodo actum sit. St. Ibitur. The scanning is obviously : quom'do actum sit. It is to be re- marked, however, that the d of modo, quomodo is never omitted in writing, and there is, therefore, no justification for the absurd proposal that immo or imo, which is obviously the adverb of imus, should be regarded as a mutilation of in modo l . 6. The French Language is the best modern representative of the spoken Latin. III. We may now pass, by a natural transition, to our third source of information respecting the constitution of the 1 Classical Museum, III. pp. 291 297. The author of this sug- gestion must have learned in his younger days that an ablative of manner repudiates any preposition ; and the violation of this rule in the case of modus, above all other Latin words, can only spring from a sort of delusion, fostered by a habit of self-reliance, which has survived the possession and reasonable consciousness of knowledge. The same writer, I am told, has published an expensive edition of Cicero's orations against Verres, in the introduction to which he states that the defendant, a Roman patrician, a Cornelius in fact, had no family name. This is an addi- tional proof, if proof were needed, of the laxity of our Latin scholarship. 6.] OF THE LATIN LANGUAGE. 445 Latin language that which exhibits it pathologically, or in its state of disorganisation or decay. It will not be expected that I should here show at length how the Romance languages were formed from the Latin. It will be sufficient to point out some of the reasons for believing that the French language is a better living representative of the pronun- ciation of the ancient Italians than the language which is now spoken in the peninsula itself; and, in conclusion, to state briefly what was the process of the disintegration, and in what degree the modern differed from the ancient form. As the Romans successively conquered the different nations which formed the population of Italy, they gradually included within the limits of a single empire a number of different tribes, who spoke idioms, or dialects, differing but little from the lan- guage of the Romans themselves. It is not, therefore, surprising that a gradual amalgamation should have taken place, and that every Italian should have spoken, with only slight variations of accent, one and the same Latin language. The language of Rome itself the language of government, of literature, and of law . would, of course, be independent of these minor differences. Every educated man and every public functionary would refer to this unvarying standard, and would speak or write, in some cases with pedantic accuracy, the language of the senate-house and the forum 1 . Accordingly, the inhabitants of the provinces, i. e. the foreign subjects of the Empire, would hear nothing but pure Roman Latin ; and, if they learned the language of their rulers at all, they would at least learn it in the best form. Their position in this respect differed materially from that of colonists, even in ancient times. The colonists of our day, and especially the English emigrants, present a material contrast to the case of the Roman provincials. For, while the colonists who sailed from Corinth or Athens were of all classes oi TVXOVTCS . our modern colonists are generally those who are either not able to live at home, or, at all events, who practise trades incon- sistent with a high amount of educational polish. We find, there- fore, that colonial English represents only the vulgar colloquial 1 Scaliger partly saw this ; ho says (Prima Scaligerana, p. 99) : "Lingiuc nostrse Gallicso potior pars ex publicis instruments quse Latino scribebantur conflata est." 446 CONSTITUTION AND PATHOLOGY [Cii. XIV. language of the mother-country ; whereas the Roman provincials spoke a language derived imperfectly, it might be, but still derived from the polished and elegant diction of proconsuls, jurisconsults, negotiatores, and publicani. The Gauls, in particular, were remarkable for their tendency to assimilate themselves, in their language and usages, to the Romans. In an inconceivably short space of time the province Gallia was completely Romanised 1 . Their own language was out of the pale of civilisation : in fact, they had no mother-tongue to struggle for. A language is only dear to us, when we know its capabilities, and when it is hallowed by a thousand connexions with our civilisation, our literature, and our comforts. So long as it merely lisps the inarticulate utterances of half-educated men, it has no hold upon the hearts of those who speak it, and it is readily neglected or thrown aside in favour of the more cultivated idiom, which, while it finds names for luxuries of civilisation before unknown, also opens a communication with those who appear as the heralds of moral and intellectual rege- neration. The Greeks and the Jews had good reasons for lov- ing the language of their ancestors, and could never be induced to forget or relinquish the flowing rhythms of their poets or the noble energy of their prose writers. The case was not so with the provincials of Gaul. Without any anterior predilections, and with a mobility of character which still distinguishes their modern representatives, they speedily adopted the manners and the words of the Romans ; and it is probable that in the time of the Empire there was no more difference between the grammatical Latin of Lyons and Rome, than there is now between the grammatical French of St Petersburg and Paris. 1 How completely this was the case even in Cicero's time may be inferred from what he says in his Or at. pro Fonteio, 1, 1 : " Referta Gallia negotiatorum est, plena civium Romanorum. Nemo Gallorum sine cive Romano quidquam negotii gerit; nuinmus in Gallia nullus sine civium Romanorum tabulis commovetur, &c." For the literary culture of Gaul some hundred years later, the reader may consult the commentators on Juvenal, I. 44; VII. 147, 8; XV. 111. Gibbon, who perceived that the language of Virgil and Cicero completely superseded the Celtic idiom in Gaul (Vol. I. p. 64, Milman), extends the same remark to other provinces to which it is not equally applicable. 7.] OF THE LATIN LANGUAGE. 447 7. The modern Italian not equally so : and why. From what I have just said, it should appear that the Latin spoken in Gaul was upon the whole better and purer than the Latin spoken in the municipal districts* of Italy during the time of the Empire. Let us, however, suppose that they were only equally good. Then, if it can be shown that the disturbing causes were greater and more efficacious in Italy than in Gaul, we shall still have a greater surplus of good Latinity in the latter. Before the Italian language revived as a vehicle of literary communication, the peninsula had been subjected to a series of invasions, which had modified and corrupted in no slight degree the speech of the country people. This was effected not only by the influence of the conquerors, but also by the infusion of a con- siderable amount of foreign population. In Lombardy and other parts, where the invaders formed a permanent settlement, the change was most sensibly and durably felt ; whereas Tuscany, which had been screened by its position from any permanent or extensive occupation by the northern tribes, was not exposed to this corruption of its familiar language, and its greater wealth, its commerce, and its independence, preserved among its inha- bitants a residuum of the old Latin literature and civilisation. When, therefore, vernacular composition revived in Italy, it was emphatically Tuscan. It is true that the new literary language spread itself over the whole of Italy, and that there were varieties of accent in the different districts 1 . Still, how- ever, a purity of Tuscan phraseology is essential to literary cor- rectness : and whatever a man's native accent may be, he must accommodate it to this court-language. It follows, therefore, that the pronunciation of modern Italian must be syllabic. In other words, it must be more akin to the studied accuracy with which the Romans of the Augustan age pronounced their Grae- i On these differences of Italian articulation Matthseus ^Egyptius writes as follows (ad S. C. de Baccli. p. 145) : " Quosdam audias ore adstricto, et inter dentes, dimidiata yerba tanquam invitos, et cum quadam parsimonia efferre, ut Ligures : quosdam ore patulo et laxo, claraque et sonora voce, animi sensus effundere, ut Neapolitan! faciunt : medios inter hos Senenses, queis Musa declit ore rotunda loqui. Adderem Florentines nisi ex imo gutture pronuntiantes originem adhuc ostenderent Phoenician!/' 448 CONSTITUTION AND PATHOLOGY [On. XIV. cised poetry, than to the natural articulation of the ancient Italians. It has been truly said, that the Italian language can- not be pronounced both well and quickly. This is only another expression of the fact, that a literary language, which is not natural, can only be articulated syllabically. The qualification of lingua Toscana in bocca Romano, is another illustration of the same fact ; for here we have a recognition of the truth, that the modern Italian is a written language to be pronounced according to its syllables, and that of the accents, in which it can be pro- nounced, the best and sweetest is that of a well-educated inha- bitant of the pontifical metropolis. ^ 8. Dialects of the French Language. Very different was the case of the Gauls. After living for several hundred years under the dominion and influence of the Romans, and having lost their Celtic language and in a great measure their Celtic character, they were invaded and partially conquered by a confederation of German warriors, who called themselves Franks, a name indicating their bold and martial character 1 . The domination of these rude conquerors did not 1 It has usually been supposed that the word Frank denotes " free- man," so that "French" and "Latin" would, when referred to their ety- mology, appear as synonymous terms. This is not, however, the original meaning of the word Frank : though, in a secondary sense, the word has borne this signification. In the Teutonic languages, to which it belongs, the word fra-n-k, or frak, is equivalent to ferox, and signifies " bold," "warlike," "intrepid" (see Thierry, Lettres sur VHistoire de France, Lettr. VI. p. 436, Bruxelles ed.). The name, therefore, according to its original signification, refers to the martial qualities, just as the name of the Rasena (which may also be compared with the Hebrew y*n) expresses the rapid movements of warlike hordes (cf. Joel II. 4). Some nations have derived their name from their physical characteristics. Thus, as wo have seen (p. 29), the Pelopes and Pelasgians of Greece got this appellation from the sun-burnt complexion of the colonists from Lydia. And there can be little doubt that the ivory shoulder of the mythical Pelops was suggested by the white necks of those Asiatics, who wore high dresses (Thucyd. I. 6), and consequently did not expose the whole of their person to the sun. That men and women differed in complexion in Greece, and that a sedentary in-door's occupation might produce a dif- ference of colour, is clear from the remark in Aristophanes (Ecclesiaz. 385) that the parliament of women looked like an assembly of cobblers : S.J OF THE LATIN LANGUAGE. destroy the Roman texture of the language which was spoken by the inhabitants of Gaul. At first both the conquerors and the conquered retained their own idioms ; and the lingua Francisca, or Francica, of the German invaders flourished by the side of the lingua Gallica, or Gallicana, of the conquered provincials. In time, however, as there was much more literary culture among the latter, and as the priests and scholars of the age were all furnished by the district in which the Franks had settled, the standard of diction would be sought in the language of the more educated class, and the Roman language, more or less corrupted, would gradually become the medium of communication between the conquerors and the conquered. As might have been expected, this gradual adoption of the Roman language by the Teutonic invaders gave rise to a number of dialects. Of these the most refined and polished was that which was spoken by the inhabitants of the south-eastern dis- trict of France. Many causes conspired to give this idiom an earlier development. The south-eastern provincials were more completely Romanised in the first instance l ; they were less sub- jected to foreign invasion than the other inhabitants of France ; the Burgundians and Visigoths, who settled among them, were more adapted to social life than their German brethren, and more readily assimilated their language and customs to those of their subjects ; and when at length Provence became a part of the Frankish dominions, the conquerors were no longer unruly German barbarians, but the civilised and Romanised subjects of a ov yap aXX* V7rep(pvs as \evK07r\T)drjs rjv Iftf'iv 6p,i\ia. I remember that on one occasion, when a highland regiment landed in kilts from the West Indies, where they had worn trousers, it was remarked that their faces and legs did not match. Ethnical names, in addition to their primi- tive meaning, are often used as expressive of certain qualities, whether the use is complimentary or not. Assassin, Gascon, Vandal, and Goth, are attributive words in our own language ; the word Slave has been derived from the low estate of the Sclavonians ; and even in ancient times, Kap, K/jqs, IIa(Xaya)i>, Mvcror, "SvftaptTrjs, SKvQrjs, &c., were terms significant of qualities. The German confederacy of the Franks seems to have corresponded to that of the Isccevones; those of the Saxons and Thu- ringians to the Ingcevones and Herminones respectively. (See above, p. 6S). i It is right, perhaps, to say, that Marseilles in particular was rather Grsecised than Romanised : see Cic. pro Flacco, 26, 36. 29 450 CONSTITUTION AND PATHOLOGY [On. XIV. regular monarchy, The happy climate of Provence, and the wealth and commerce of the people, contributed to foster and en- courage those arts which can only flourish in a genial soil ; and we are not to wonder if the provincials outstript the northern Gauls in intellectual tastes as well as in physical comforts. The connexion between Provence and Catalonia tended to increase the civilisation of the latter. But, in reference to the present object, to discover a Romance language which shall most accurately represent the spoken language of the Romans, we may safely dismiss the Spaniards ; whose language, already corrupted by the invasions of the Suevians and Visigoths, has been still farther disorganised by the pervading and durable influence of the highly civilised Arabians. The people of Provence were keenly sensible of the difference between their own language and that of their Franco-Gallic rulers. The names, by which they distinguished their own country and that of the French, referred to the differences of the idioms spoken in them. It is singular that this difference should have been expressed in terms of the affirmative particle, which they had respectively adopted. Drawing a line through Dau- phine, Lyonnais, Auvergne, Limousin, Perigord, and Saintonge, the country to the south of this was called Langue d'oc, the dis- trict to the north of the line was termed Langue d'oyl. JSTow, although the differences between the Langue d'oc and the Langue d'oyl consisted mainly in the greater or less development of the Latin element in each, it is to be remembered that these affirmative particles are both due to their Teutonic affinities 1 . And here is the inconsistency ; the words oc and oyl are equally Frankish or German, and yet the people of the Langue d'oc dis- tinguished their language from that of the Langue d'oyl by calling it Roman, lemozi, provensalesc ; and they termed them- selves Provinciales, i. e. Romance, Provincice inquilini, as distin- guished from the Francigence of the north. l According to Grimm (D. Gramm. III. p. 768), oyl is ja il, and oc is ja ich ; the only difference between them being, that the affirmative is combined with the first person in the one case, and with the third person in the other. To me it appears that oyl is simply the affirmative wel or wohl (for this power of the initial o see above, p. 49), and that oc is the German auch - etiam (Phil. Mus. II. p. 345). 9.] OF THE LATIN LANGUAGE. 451 $ 9. But all these Dialects were closely related to the Latin. But whatever were the distinctions between the languages of the northern and southern inhabitants of the province of Gaul, it is clear that the language of the whole country was to the middle of the ninth century A. D. a very near approximation to the Latin. We have the original of an oath which was sworn at Strasburg in 842 A. D., by Lode wig, king of Germany. This interesting document, which is expressly stated to have been in the Romania lingua, is in the following words L : " Pro Deu amor et pro Christian poblo et nostro commun salvament, disi di en avant, in quant Deus savir et podir me dunat, si salvarat io cist meonfradre Karle, et in adjuda et in cadhuna cosa, si cum om per dreit sonfradre salvar dist, in o quid il mi altresi fazet : et ab Ludher nul plaid numquam prindrai, qui, meon vol, cist meon fradre Karle in damno sit" It appears from the context of the history, that the oath was couched in this language in order that it might be understood by the French subjects of Karl le Chauve. It was, therefore, the common language of the country ; and as it is free from Germanisms, and exhibits only those corruptions of the Latin for which it is easy 1 Nithardi Hist. ap. Scr. Rer. Frantic. VII. p. 26, quoted by Thierry, Lettres sur VHistoire de France (lettr. XL). Substituting the Latin words which come nearest in etymology to the words of this frag- ment, we have : Pro Dei amore et pro Christiano populo et nostro com- muni salvamento, de isto die in ab-ante, in quantum Deus sapere et posse mihi donabit, sic salvare habeo ego ecc' istum meumfratrem Carolum, et in adjutu et in quaque una causa, sic quomodo homo per directum suum fratrem salvare debitus est, in eo quod ille mihi alterum sic faciet ; et ab Lothario nullum placitum numquam prendere habeo, quod, mea voluntate, ecc' isti meo fratri Carolo in damno sit. It is not necessary to enter upon any lengthened discussion of the corrupt Latinity of these words. That salvar-ai, &c., are salvare-habeo, &c., is well known. It appears from the oldest forms of the words that the French eel, cest (cist), Italian quello, questo, are the compounds ecc' ille and ecc' iste respectively. For, as in Proven9al we have aisso, in old French aezo, into which co enters, so we have icel and icest, anterior to eel and cest. Similarly id is ecc' ibi. Of altresi, which is common in Italian, Varchi says : "Altresi e Provenzale, non Ispagniuolo, e gli antichi nostri scrivevano altresie, e non altresi" Comp. altrettale, altrettanto. The French aussi represents altresi with the usual change of I into u. 292 452 CONSTITUTION AND PATHOLOGY [On. XIV. to account, it furnishes us with a distinct confirmation of the opinion, that we ought to seek in the language of France for the best modern representative of the language of ancient Italy. Among the political or official terms, which the Franks adopted from the Latinised inhabitants of Gaul, and which show the extent of the influence to which I am referring, not the least interesting are the titles maire and bailli, which designated the primary and secondary offices in a municipality or district. The former name is a corruption of the Latin adjective major, and it was originally used as an epithet to the term prcepositus, which has left its traces in the French prevSt and our provost. Hence, it happens that mayor in England and provost in Scotland are synonymous designations for the chief of a municipal body. On the other hand, the word bailli, It. balio or bailo, is derived from the Latin bajulus, sometimes corrupted into baillivus, and denoted the secondary officer or deputy. According to its ety- mology bajulus for bar-iolus (cf. pejor for per-ior) denoted a bearer of burdens, and so the word is connected with (pep-, (pop-, bhri, fer, bar-dus, bar-o, /3aa-rd^a), &c. (Doderl. Syn. u. Et. I. 151). In his official duties, therefore, the bajulus or baillivus was a charge d'affaires, one who bore the weight of office on behalf of others. And not to speak of the profound and solemn meaning of the phrase in Isaiah IX. 5 : IDpttrby iTIJtf 2pn ^Hfll, "and the government shall be upon his shoulder," we may re- member that the Arabic \\$ Vezir, which signifies " vicarius prin- cipis," is derived from the verb " , vazara, which means "sus- tinuit onus grave." The relation between the Scottish baillies and their provost is precisely that which subsisted between the baillivi and their major, or prcepositus^ or prcepositus major, namely, the latter was the chief, and the former his vicars or deputies. Thus we find the major or prcepositus in a cathedral, by the side of the bajuli or baillivi conventuales or confratrice; we have major domus in the royal palace, by the side of the bajuli depalatio; and in general, wherever there was a term of authority, the bailli represented the vice-comes, vice-gerent, deputy, or regent. The military use of the term major be- longs to the same application of the Latin word. While the lieutenant-genera], or lieutenant-colonel is the deputy of the full 9.] OF THE LATIN LANGUAGE, 453 general or colonel, the wer/or-general or major is so called in reference to the rank immediately below him ; thus the major- general is the prcepositus or maire of the colonels in his division, and the major is the prcepositus of the captains in the regiment, just as the serjeant-wio/or is the chief of the Serjeants, and the drum-major of the drummers. In its lowest application the term baillie or " bailiff" still signifies a deputy, and the mere "tipstaff" or " catchpole" is called by this name because he is the sheriff's officer, or the deputy pro re nata of that prcepositus of the county or district. The difference between the modern Italian, considered as the offspring of the new Tuscan literature, and the old French, regarded as a scion of the Roman language which was spoken in the province of Gaul, consists in the fact to which I 'have already adverted namely, that the former would reproduce the mincing and pedantic pronunciation of the literary Romans, while the latter would retain the genuine colloquial utterance of the free colonists of the empire. It is worthy of observation that the French language itself enables us to illustrate this difference. If we examine the French language as it is, we shall often find double forms of derivatives from the Latin. Now in every one of these cases it is remarkable that the older word that which belongs to the oldest and most genuine vocabulary differs most from the written form or syllabic pronunciation of the Latin original. Thus chanoine, chetif, chez, chose, hotel, naif, Noel, pitie, pousser, from canonicus, captivus, casa, causa, hospes, nativus, natalis, pietas, expulsare, are older forms than cano- nique, captif, case, cause, hopital, native, natal, piete, expulser. (See A. W. Schlegel, Observations sur la Langue et la Litte- rature Prov. p. 44.) The fact is, that the latter were derived from the written, the former from the spoken language. 10. Leading Distinctions between the Roman and Romance Idioms. The manner in which the transition from the Latin language to the French may be supposed to have taken place is well known, and very easily described. In this place we must be contented with a few brief remarks ; for it would be an idle attempt to discuss as a secondary matter the details of a subject which admits of such ample illustration, and which has already 454 CONSTITUTION AND PATHOLOGY [On. XIV. been treated at great length, though with various degrees of success, by Raynouard, Schlegel, Diez, Ampere, Fuchs, and Lewis. The tendency of the spoken Latin language to clip and mutilate itself began at an early period to militate against the regularity of the grammatical forms. With regard to the verbs, it has been shown above that the organic inflexions had been in a great measure superseded by secondary or compound tenses before the commencement of the classical age; and that the person-endings are obliterated, or deformed by inconsistencies, in the oldest specimens of the written language. In regard to the verbs, then, the change from the Roman to the Romance is merely a further development of that which was already in operation. The Roman case-system was in itself more complete than the conjugation of the verb ; and therefore we may expect to find greater changes in the French noun as compared with the Latin. In general it may be remarked, that when the tendency to abbreviation has commenced its action on the flexio- nal forms of a language, certain devices are at once adopted for the purpose of preventing any syntactical obscurity. Indeed, the logical or syntactical development of a language is gene- rally benefited by the change ; and where the etymological organisation becomes imperfect, the literary capabilities of the particular idiom are extended and confirmed. There is good reason for believing, that in the spoken language of the ancient Italians the difference between the sub- jective and objective cases of the noun was at an early period neglected or overlooked (see Lepsius, ad Inscript. p. 120). At any rate, it is clear that this was the first step towards the breaking up of the Roman case-system. The accusative case was substituted for the nominative, and all the subordinate relations were expressed by prefixing prepositions to this new crude form of the noun. We observe a tendency of the same kind in vulgar English ; and perhaps this passage from the sub- ject to the object may be explained on general principles, without any reference to the want of grammatical education on the part of those in whom it is most observable. Connected with this employment of prepositions, to give definiteness to the crude forms of nouns, is the use of the old Roman demonstratives ille and ipse to mark a definite object, as contrasted with units and 10.] OF THE LATIN LANGUAGE. 455 aliquis-unus, which denote indifference. This is, of course, identical with the use of the definitive article in the Greek and other languages ; and the Komance languages owe much of their acknowledged perspicuity to this adaptation. It is true that the artifice is not applied with the logical subtilty by which the employment of the Greek article is distinguished; but any deficiency in this respect is amply compensated by the strictly logical order of the sentences in which the words are arranged. It is not necessary in this place to say much on the subject of the Komance verb. Where the tenses have preserved the forms of the Latin verb, we observe a systematic abbreviation. Labials are absorbed, according to the practice so remarkable in Latin ; final syllables are dropt, and the accent is thrown for- ward. We sometimes find that what appears to be an arbitrary corruption is really only an attempt to represent in writing some genuine articulation of the old Latin ; thus we have seen above (p. 244) that a palatal may take the place of a labial in French, when the latter is followed by i, as in sapiam = sapjam, Fr. sache (cf. ravir and arracher from rapio and arripio). We see the process of this change in the Proven9al. Thus, we have in the celebrated prison-song of Richard Cceur-de-Lion : " Or sapchon ben miei horn e miei baron Engles, Norman, Peytavin, e Gascon, Qu* ieu non ai ja si paubre companhon, Que per aver lo laisses en prison." Where sap-ch-on-sap-i-ant=sachent : " Know all my lieges and my barons true From England, Normandy, Guienne, Poitou, I would not leave the poorest of my train In dreary dungeon for the love of gain." The g which represents the Latin -tis in the second person plural of all present tenses of French verbs, except in the cases of etes and faites for estis and facitis, is not equivalent to ts t as some have supposed, but stands for the dental sibilant, which followed the t in the older Romance languages ; thus we have avetz=habetis before we find avez, and even etz for estis before etes. That z is merely an s, so written after -t, is clear from its similar appearance as a plural affix to nouns and participles, as in gentz-gentes-gens, toutz-tous t escriptz=ecris, &c. Generally, the number of compound or auxiliary tenses is very much 456 CONSTITUTION AND PATHOLOGY [On. XIV. increased in the Romance as compared with the Latin verb. In addition to the verbs sum and fui, we find that habeo and sto are regularly pressed into the service. Verbs in their first for- mation construct their perfect and future tenses with the aid of habeo ; for the past participle with habeo makes up the former (as j'ai aime = ego habeo amatum), and the regular future consists of a combination of the same verb with the infinitive (as faimer-ai=ego habeo amare). This analysis of the Romance future was probably known to Sainte Palaye, who cites the main proof of it, namely, the fact that the infinitive was sometimes separated from its auxiliary by the interposition of another word (see Bopp, Annals of Oriental Literature, p. 45). But the formal enunciation of this view was first made by Raynouard (see Grammaire Romane, p. 221 ; Lewis On the Romance Languages, p. 194) ; and there cannot be the least doubt of its truth. This is shown not only in the tmesis, to which I have referred, but also by the varying forms of the future in the dif- ferent Romance languages, which correspond to the varieties in the form of the present of habere. Thus, on the one hand, we find : " et quant cobrat Tauran, tornar Van e so poder per fe e senes engan" =. " et quand recouvre Tauront, tourner Yont en son pouvoir par foi et sans tromperie." "E pos mon cor non aus dir a rescos, Pregar vos ai, s'en aus, en ma chansos." = "et puisque mon desir je n'ose dire a c&cheite, prier vous az, si en ose, en ma chanson." On the other hand, we see that the present of the verb, corresponding to habeo in each of the Komance languages, is duly represented by the corresponding affix of the future. Thus we have : ITAL. SPAN. Pnov. FRENCH. ho he ai ai amer-d, amar-e, amar-ai, aimer-ai ; and similarly of the other persons. In Italian the future also ex- hibits the longer forms in aggio or abbo, as in dir-aggio, " I shall sa,y,"far-abbo, "I shall make" (d.far-ebbe, &c.). It is obvious that the same explanation must apply to the secondary tenses. For if the future aur-ai, aur-as, aur-a, aur-em, aur-etz, aur-an t is compounded of the infinitive aver and the present ai, as, a, avem, avetz, an, it is clear that the conditional aur-ia, aur-ias, 10.] OF THE LATIN LANGUAGE. 457 aur-ia, aur-iam, aur-iats, aur-ian must be made up of the same infinitive and the subjunctive present a-ia, a-ias, a-ia, a-iam, a-iatz 3 a-ian. And thus amar-ia will not represent amarem, as some writers have supposed, but will exhibit the same agency of the auxiliary verb as the future amar-ai. The indeclinable words in the Romance languages are parti- cularly interesting, as examples of the manner in which frequent use contributed to the abbreviation of phraseology in these idioms. In some shorter words the alterations are very slight, as in a for ab l , done for tune, avant for ab-ante, av-ec from ab-esc for ab-usque (cf. the Provencal duesc for de-usque, Ray- nouard, Gramm. Rom. p. 318), ailleurs for aliorsum, dorena- vant for dehora in ab-ante, mais for magis, jamais for jam magis, ensemble for insimul, de-main for de mane, moms for minus, quand for quando, car for quare, derriere for de retro, assez for ad satis, si for sic, whence ain-si for in-sic, souvent for subinde, dont for de unde, maintenant for manu tenens, or for hora, desormais for de ista hora magis, trop for the Low Latin troppus, " a large number" (as " si en troppo de ju- mentis." Lex Al. ap. Raynouard, Gr. Rom. p. 317), pret, Ital. presto for proesto, &c. Other particles are much more cor- rupted : per becomes pour ; post is changed into puts ; prope into pres, whence au-pres, a-pres, &c. ; secundum passes through segont into selon ; meme, from medesimo, brings us back to the corrupt form met-ipsissimus ; sine is shortened into sans or lengthened into senza ; while aut appears as ou, ubi and ibi are turned into ou and y ; paucies, which was pauc in the Romance languages, becomes peu in French, &c. I should be inclined to place tot, Rom. tost, Ital. tosto, in the former class, but Mr. Lewis says {Rom. Lang. p. 248) that "no probable explanation of its origin has hitherto been given." I am not aware what interpretations have been proposed, but it seems to me obvious that tosto is merely the adverb of the corresponding adjective, derived from the Latin tostus, and signifying " swift," " sudden," " all in a heat," " hot with haste," so that it is syno- nymous with ardens. i In the Provengal language, as in modern French and Italian, a or db bore the sense of apud=ab-ad, and signified location in all its forms : av-ant = ab-ante, devant = de-ab-ante, av-ec = ab-usque, &c. 458 CONSTITUTION AND PATHOLOGY [On. XIV. J 11. Importance and value of the Latin Language. In the preceding pages I have indeavoured to write the history of the Latin language, and to characterise its peculiari- ties, from the earliest period of its existence down to the present time, when it is represented by a number of daughters, all re- sembling their mother jnore or less, and all possessing in some degree her beauties and defects. Of these, it can hardly be doubted that the French has the best claim to the primogeniture and inheritance. The Latin and French languages stand related to one another, not only in the connexion of affinity, but still more so in the important position which they have occupied as poli- tical and literary organs of communication. They have both striven to become the common language of civilised and educated men ; and they have had singular recommendations for the office which they partially assumed. For power of condensation, for lucid perspicuity, and for the practical exposition of common matters, there are few idioms which can compete with the Latin or the French. In many particulars they fall far behind the Greek and the German ; in many more they are surpassed by the English ; and it seems now to be determined that neither Caesar nor Napoleon was destined to reverse the decree of Providence, that man, though the one reasoning and speaking creature, should, in different parts of the world, express his thoughts in different languages. If there is one idiom which seems both worthy and likely to include within it the articulate utterances of all the world, it is our own, for we, too, " are sprung of earth's first blood," and the sun never sets upon our Saxondom. But the dignky of our English speech, and its wide diffusion by means of our commercial enterprise and missionary zeal, do not suggest any argument or motive, which should in- duce us to neglect or discourage the study of the old Koman literature. Though the Latin tongue will never again become the spoken language of Europe, there is no reason why it should not resume its place as the organ of literary communication, why, with its powers of conciseness and abbreviation, and with its appropriation of all the conventional terms of science and art, it should not still flow from the pens of those who have truths and facts to communicate, and who are not careful to invest or disguise them in the embellishments of some modern and fashion. 11.] OF THE LATIN LANGUAGE. 459 able style. This at least is certain, that the Latin language has struck its roots so deeply and so permanently in our own lan- guage, that we cannot extirpate it, if we would ; for we must know Latin, if we would thoroughly understand our own mother- tongue ; even those who are least learned, and most disposed to undervalue classical attainments, are very liable to further what others would call the corruption of our language, by the intro- duction of new terms erroneously formed after a Latin model 1 ; and whatever changes may take place in the professional edu- cation of Englishmen though the Universities may cease to bestow the highest degrees in their faculties upon those who have passed through the Latin exercises of their schools though the meeting of Convocation may never again be inaugurated with a Latin sermon at St. Paul's though a study of Justinian and Gaius may be pronounced of no use to the lawyer though even Roman history may lose its general interest though phy- sicians may decline to prescribe and apothecaries to dispense ac- cording to the phraseology of a Latin materia medica though the House of Commons may no longer bestow the sanction of parliamentary applause on well applied quotations from the clas- sical authors still, a competent acquaintance with the language and literature of ancient Rome will be indispensable to every one, who lays claim to a complete cultivation of his reason and taste, and who wishes either to understand and enjoy the writings of our best authors, or to enrich the English language with new examples of its capacity for terse arguments, happy expressions, and harmonious periods. 1 It would be easy to cite a long list of words in -ation, which are not formed from Latin roots, and are certainly not due to the Latin scholar- ship of those who first used them. The verb " to base" for "to cause to rest on a basis or foundation" is a modern corruption so common that I cannot hope to have avoided it in my own writings, though I am quite aware that according to all analogy " to base" or " abase" must mean " to depress" or lay low, not " to build up." INDICES. ETHNICAL NAMES, AND NAMES OF PLACES OK COUNTRIES. Abella, 115, 127 Herminones, 42, 68 Rasena, 69, 448 JEnus, 7 Hirpini, 60 Raetians, 67 JEqui, 5 Humber, 63 Roma, 60 Agathyrsi, 41 Iguvium, 79 Romanus, 414 Aidio\{/, 30 Ingaevones, 68, 71 Sabinus, 7 Alba, 5 "I OH/, Javan, 39, 41 Sacs, 41 Ambrones, 62 Iran, 40 Sauromatae, 41 Angli, 71 Languedoc, 450 Saxon, 41 Anio, 7 Larissa, 13 Sclavonian, 61 Apulus, 5 Latinus, 6, 61 Scolotae, 44, 58 Argos, 13 Lavinium, 6 Servians, 66 *Apioi, 41 Leleges, 63 Sintians, 39 Asia, 41 Ligyes, 63 ZKvOat, 40 Atella, 127 Lithuanian, 61 Thames, 46 Auruncus, Auaou>, 4 Maidoi, 39 Thracians, 39, 42 Caere, 166 Massageta3, 42 Thuringians, 42 Cascus, 5 Mysi, 39 Thyrea, Thyraeon, 13 Chawilah, 57 Northumberland, 63 Thyssagetae, 41 Courland, 60 Opicus, Oscus, 4 Tiryns, 13 Cumberland, 63 Pahlavi, 41 Toltecs, 14 Daci, Danes, 43 HdpQos, 39 Tvpprjvos, 12 Dorian, 42 IleXao-yos, 29, 448 Tuscus, 16 Eden, 56 IleXoi//-, 29, 448 Umbri, 62 Etruscus, 69 Piceni, 8 Veneti, 66 Falerii, 243 Pomeranians, 70 Vindelici, 67 Frank, 448 Prussians, 70 Volscus, 5 Getae, Goths, 43 Quirites, 60 Wineds, 66 II. SCYTHIAN WORDS. aba, 54 brix-aba, 54 Grau-casis, 53 Apia, 49 Dan-ubius, 46 halinda, 54 ara, 54 Dnieper, 47 Hypa-caris, 46 araxa, 54 Dniester, 46 Hypan-is, 46 Araxes, 48 dun, 46 Is-ter, 45 Arimaspi, 52 enarees, 54 Maeotis, 51 Artemis, 51, 54 Eri-danus, 48 masadas, 51 Artimpasa, 51 Exam-paeus, 54 Octa-masadas, 49 Borysthenes, 47 Ger-rus, 46 Oito-surus, 49 462 INDICES. Oior-pata, 53 Porata, 46 Tami-masadas, 51 Panticapes, 47 Rha, 48 Tana-is, 47 Papseus, 48 Rho-danus, 48 Temarunda, 51 pata, 63 Sparga-pises, Spargapi- Tyres, 46 phru, 54 thes, 51 xa, 54 phry-xa, 54 Tahiti, 48 - III. UMBRIAN WORDS. The Alphabetical List in pp. 99101, and the following. abrof, abrons, 91, 280 hont, hondra, huntra, hon- pune, pus, pusnaies, 85, aferum, 85 domu, 85, 315 punus, 102 ahaltru, 102 kapire, 82 purtinsus, 85 anzeriates, 88 karetu, 85 pusei, puze, 85, 97 ape, 85 Krapuvius, 91 pustru, 85 arsie, 85 kupifiatu, 85 seritu, 88 arsmo, 97 kurnase, 95 sevakni, 94 arepes, 96 kutef, 93 sevum, 8, 93 arveitu, 82 mers, 96 skrehto, skreihtor, 86 arves, 96 nep, 83 stahito, 82 arvia, 92 nerf, 97 steplatu, 95 buf, 91 nume, 84 subator, 97 dersecus, 97 okris, 84, 92 subokau, suboko, 96 dupursus, 99 orer, 98 sue-pis, 96 enetu, 88 ose, 98 tases, 94 enumek, 94 paker, 96 tertie, 99 erar, erer, 96 parfa, 95 tera, 82 erek, eront, 85, 315 peica, peiqu, 95 tesenakes, 90 eras, 85 pepe, 102 tesva, 95 eso, 85 pernaies, 88 titis, 102 este, 88 persei, 97 tota, 84, 93 etre, 99 peskier, 85 tover, 85 feitu, 92 persklum, 88 treplanes, 90 ferine, 92 pesetom, 85 tuplak, tupler, tuves, 99 festira, 82 peturpursus, 99 vas, vasetom, 85 fos, 96 pihatu, 82 vatuva, 91 fri, 96 pir, 98 vehiies, 90 frite, 96 poe, 84 veres, 89 frosetom, 85 poplus or puplus, 84 virseto, 85 furenr, 82 portatu, 96 vitlup, 83 futu, fututo, 85 pre, 89 ulo,96 habe, &c. 85 prumum, 99 uru, 96 heris, 92 prusesetu, 81 ustentu, 92 heritu, 92, 98 pufe, 85 INDICES. IV. OSCAN WORDS. The Alphabetical List in pp. 105116, and the following. 463 seteis, 123 him, 127 piei, 120 aisken, 126 iok, ionk, 120 pis, 111, 121 akenas, 132 kadeis, 121 pistia, 132 aktud, 124 karneis, 119 pod, 121 akum, 127 kastro, 122 poizad, 126 allo, 126 kebnust, 126 pokapit, 121 amirikatud, 126 kensam, kensaum, 125 pomtis, 124 amma, 131 kensazet, 125 pon, 125 amnud, 120 kenstom, 126 post-esak, 121, 317 am pert, 123 kenstur, 125 praefukus, 127 angit, anget, 119 kerus, 131 prsesentid, 126 anter, 119 keus, 125 preivatud, 124 araget, aragetud, 119 kom, 124 prof-tuset, 129 AusU, 112 komenei, 120 pru, 121 Bansaa, 125 komono, 120 pruhipid, 121 Bantins, 125 kontrud, 123 prumedikatud, 123 Degetasius, 119 kvaisstur, 83 pruter, 125 deivaid, deivast, 119 ligis, ligud, 127 puf, 126 dikust, 124 likitud, 122 regator, 132 dolum, 120 loufir, 122 Q[uaestor], 119 egmo, egmazum, 120 maimas, 119 sakaraklum, 435 eituam, eituas, 121 mais, 119 senateis, 120 eizazunk, 127 mallum, malud, 121 set, 127 embratur, 88 manimaserum, 127 siom, 120 estud, 121, 126 meddisud (pru-), 126 sipus, 124 esuf, 126 medikatud, 126 skriftas, 127 esak, 121 mesene, 90 slagis, 129 etanto, 123 minstreis, 123 suae, 120 Evklus, 131 molta, 119 tadait, 122 famelo, 126 moltaum, 123 tanginud, 119 fefakust, 122 neip, nep, 127 teforom, 48, 132 feihos, 129 nesimois, 127 toutiko, 126 fiisna, 129 op, 124 tribarakat, 129 flusare, 90 pa, 126 tuset, 125, 129, 184 fortis, 123 pam, 125 valaemom, 122 fuid, fust, 127 Patana, 132 verehasius, 132 Futris, 131 Perna, 132 vinkter, 126 Herekleis, 435 pertemust, 120 umbrateis, 120 herest, 123 perum, 120 urust, 124 hipid, 121 petiropert, 124 zikolom, 124 464 INDICES. V. ETRUSCAN WORDS. The Alphabetical List in pp. 151164, 182, 183, and the following. achr, 175 hintha, hinthiu, 186 ril, 305 Afuna, 188 ipa, 186 sains, 186 ama, 179, 187 ir, 168 Saturnus, 146 Ancaria, 149 Janus, 143 Secstinal, 142 Apulu, Aplu, 148 Juno, 145 Sethlans, 145 Aril, 163 Jupetrul, 170 sie, 168 Aritimis, 37, 50, 54 Kalairu, 169 Soranus, 148 Aruns, 103 kemulmleskul, 187 Sothina, 148 Ausil, 113 kethuma, 167 stem, 178 Caecina and Caecilia, 188 Kupra, 107, 145 sver, 176 Caphatial, 178 lar, 150 Summanus, 144 caratse, caresri, carutezan, lat, 178 suthi, 175 184 lauchme, 102 Tanaquil, 71 ceca, 170 lautn, lautnescle, 179 tanna, 177 cechaze, 170 lisiai, 167 tephral, 48 cehen, cen, 175 Mantus, 148 telur, 186 Ceres, 149 maram, 167 tenilaeth,'17S cerurum, 186 Mars, 146 Thalna, 145 chfinchfe, 102 mathu, 167 Thana, 178 clen, 171 Matuta, 147 thaura, 176 ever, 176 Menerfa, 146 Thekinthul, 179 eca, 175 Merqurius, 150 thmtflaneth, 179 ein, 186 mi, 167 thipurenai, 153, 168 Elchsntre, 142 murzva, 186 thues, 185 epana, 168 nac, 175 thura, 176 Epure, 170 nastav, 168 Tina, 143 erai, 168, 175 nesla, 175 tulati, 187 ersce, 175 nethu, 168 Turce, 170 etera, 171 Nethuns, 148 tree, 174 ethe, 168 Nfatia, 178 Turms, 150 eu, 179 ni, 167 Velthina, 188 Feronia, 147 Nortia, 149 Velthu, 168 flenim, 179 phleres, 173, 376 Vedius, 145 fuius, 169 phruntac, 161 versus, 150 hareutuse, 184 Phupluns, 192 Vertumnus, 146, 386 heczri, 186 Porsena, 16 Usil, 113 helefu, 168 Rasne, 188 Utuze, 142 ayaX/ua, 152 ay/oios, 268 atojy, 256 50 aipeto, 92 VI. GREEK WORDS. altra, 50, 162 alwvj 409 i'Xts, 256 vj 160 a t 199 aW, 331 j>e/oa, 2/ucpa f 268 d-jrcopa, 4 "Apre/us, 54, 129 ;'yuepos, 268 ow, 98, 338 ao"i/0tjXos, 164 rjiretpos, 269 oupavds, 425 "ATpOTTOS, 150 Oa'Xao-ffa, 252 t/l/TOS, 312 aurds, 315 Oappelv, 255 6ci\ts, indecl. 291 7rapaoTJ;i/i, tru/xTrapa- /SXlTTW, 241 6^i/ap, 304 o-TiJi/at, 298 BoWopos, 47 Bi/'/o, 243 irapt'xw, 379 /8oDs, 155 0$s, 185 7rap0c'j/os, 242 /30W7TIS, 37 0^T6S, 125 ' rt i n^? /3t/pyos, 47 Qvpa, 255 Trarjj'p, 48 yeXe'oj/res, 159 . 0c'p?j^, 256 7rei0a>, 384 ye(pvpa, 418 6wy^a, 268 rre'Xayo?, -TreXayios, 305 yrjpucov, 149 Wios, 288 Tre/Xapyds, 30 yXu/cuppia, 251 'I/xepa, 268 irt'Xios, 7reXi5i/ds, 29 dai'ip, 255 /caipds, 392 Trepf JJAU, 392 5a:puoi/, 255 /caXo's, 255 -TTlTrpdcTKU), TTpia/JLOl, 388 <5aXis, 107 K611/OS, 314 -TTiTTTW, 384 <5aXds, 255 Kepavvvfjii t 392 iriW, 102 dcnravr), 168 /cXyco, 62 'wXj/ptjs, 256 xJaXjs, 255 iroV/uos, 251 Trouj/xa, 406 6e, 162 ^i-TrXa'ortos, 417 Xe'cos, 288 polpdo-s, 250 SiXo/j-rivia, 158 XiVpa, 6 o-aij/w, 258 6pw, 76 Xoi^opos, 160 creX/j/i;, 146, 157 dwpov, 256 Xo^tas, 25 vtytov, 164 cap, 386 /ua'i/rts, 149, 245 SouMva, 147 t-yprjyopOai, 383 p-ofprus, 304 os, 48 tt/apyrjs, 304 /xe-roTTfj, 35 Tpe'xw, 76, 387 ej/e'y/ceii/, 149 /i>}, 338 Tupavi/os, 13 tTTO/xai, 263 fj.ijvu(a t 147 Tupo-ts, 13 e7riTf5etos, 291 pvpioi, 264 uto's, 170 epis, 267 f0p.o's, 168 U7TVOS, 253 fpxonai, 76, 387 i/eco, 148 Xiao-tos, 417 tare, 321 feVos, 206 0oX/co's, 242 6uxcpj5, 304 'OapiW, 244 0i'/w, 169, 347 ecus, 288 oao-i?, 49 x wai, 168 CCOVTOU* 268 dco us, 257 Xo/ f 5 30 466 INDICES. \dpfjiti, 253 ^QafjLaXos, 76) 167 X<*v>o X w/ P a > 268 XeiXos, 168 XtXtot, 379 oJ/cectvos, 425 Xe'o>, 373, 418 X"V, 379 (oveonai, 388 MV> 155 XXuyjoSj 158 CO(p\U> } 383 VII. LATIN WORDS. a, ab, abs, 330 antiquus, 268 caput, 54 ac, atque, 429 antrum, 412 caput, capud, 299 accerso, 352 anus, 163 career, 435 accipiter, 155, 300 apud, 332 cardo, 269 accuso, &c. 265 arboresco, arboretum, 396 carmen, 299, 406 acer, 404 arbustum, 301 carnifex, 419 acerra, 219 arcera, 204 castrum, 412 acervus, 195, 404 arcesso, 352 castus, 53, 122 actus, 124, 270 asper, 429 catamitus, 256 acua, 250 aspernor, 392 caterva, 195, 404 ad, 331 assiduus, 204 catervatim, 289 adeps, 92 atrox, 404 cauneas, 441 adhaesum, 153 auctumnus, 386 celer, 25, 302 adoro, 216, 375 augur, 263 celsus, collis, &c. 171 adulo, 258 aula, 435 ceremonia, 406 advena, 403 auriga, 426 cerno, 392 seditimus, 263 auris, 256, 426 cerus manus, 198 aeger, 263 autumo, 263 ceu, 179 aeneus, ahenus, 82 aurum, 113 ceva, 155 sereus, aeneus, 256 avena, 157 chorauloedos, 198 sestimo, 262 aveo, 75 cieo, cio, 381 seternus, 146 bajulus, 452 cimeterium, 265 ager, 296 bellum, 240 cippus, 243 ala, 423 berber, 196 circa, circum, circiter, 335 alimentum, alimonia, 406 berbex, 54 citus, 381 aliquis, 322 bibo, 102, 383 civis, 125, 303, 404 alms, 313 bis, 240 clam, calim, 291 almus, alo, 388 bitumen, 241 cliens, 64 alumnus, 97, 406 blandus, 424 clipeus, 271 amanuensis, 332 bonus, 240 coelebs, 257 ambitus, 212 bos, 303 coemo, 390 ambo, 327 bruma, 434 coena, 106 amicus, 256 caducous, 256 cohors, 82 amo, amor, 54, 389, 410 caedo, 392 colonia, 249 amoenus, 160 caeruleus, 259 cominus, eminus, 291 ampirvo, 198, 9 Caius, 286 comissari, 81 amsegetes, 213 calvitur, calumnia, 204 compaseuus, 212 ancilla, 149 canis, 302 concapes, 211 anfractus, 213 cano, 387 confuto, 376 anguis, 157 cantilena, 214 congruo, 247 annus, 163, 425 canus, candidus, 53, 106 considero, 376 ante, 332 capesso, 352 containinare, 435 antid, 306 capio, 390 contemplor, 376 INDICES. 467 contio, 257, 435 dunque, 198 fraus, 430 contra, 335 e, ex, 331 frausus, 217 coquus, 250 eapse, &c. 316 frustra, 160 cor, 299 Ecastor, c. 436 f ui, 348 coram, 291 eho, ego, eja, &c. 310 fundus, 270, 418 corvus, 155 elementum, 140 fungor, 390 cosol, 257 emo, 388 Gaius, 246 creo, cresco, 396 enim, 126 gena, 245 cms, 161 enos, 195 generosus, 300 cubo, 374 eo, 382 genus, 245 cuicuimodi, 436 equidem, 443 gigno, 388 cujus, &c. 321 equus, 404 glisco, 171 culmen, 171 ercticisco, 210 globus, 244 cum, 335 erga, 289, 335 Gnams, 246 camera, 115 eruditus, 141 gnarures, 279 cunae, 435 escit, 201, 346, 396 grarium, 245 cuneus, 195 esum, 345 gruma, 61, 270 cunque, 325 et, 331, 429 grus, 157 cupio, 390 exiguus, 434 habena, 157 cur, 249 exilis, 434 habeo, 378 curia, curiatius, 24 exim, 287 hsereo, 247, 378 custos, 122, 298 existimo, 262 helup, 168 de, 333 explodo, 265 here's, 378 debeo, 76, 380 explore, 374 Herminius, 25, 141 debilis, 76 facesso, 352 heri#6 decumanus, 269 facie, 196 herus, 25 deliro, 271 facul, 229, 442 hibernus, 244 demo, 333 fagus, 242 hie, 310, sqq. demum, 322 famelicus, 302 hiems, 300 denique, 322 familias, 279 hinc, 288 denuntio, 235 Fatua, 156 hir, 92 deploro, 376 fatuus, 92 hodie, 435 dequim, 231 favor, 410 homicida, 431 desidero, 376 febris, 403 homo, 25 dextra, 96 fendo, 397 honestus, 301 di, 286 fera, 241 honor, 410 dice, 196 ferio, 397 Horatius, 25 diffensus, 207 fero, 310, 397 hortus, 212 difficultas, 267 filius, 169, 347 hospes, 206 Digentia, 255 finalis, 422 hostis, 206 digitus, 405 findo, 158, 256 humus, 76, 167, 247 diligo, 387 fio, 347 idem, 316 dissicentes, 97 flamma, 402 ideo, idoneus, 291 divido, 158 flecto, 386 idus, 158 do, 372 foecundus, 348 igitur, 204, 289, 342 dolus, 430 foedus, 423 llithyia, 265 Domitius, 257 foemina, 348 ille, 310, sqq. domo, 374 foetus, 348 im, 234 donee, 322 folium, 239 imitor, 158 donum, 256 forceps, forf ex, forpex, 297 imo, 444 duco, 229 forem, 349 imperator, 111 Duillius, 240 fovea, foveo, 156 imus, 435 dumtaxat, 231 frangere, 239 in, 331 302 468 INDICES. inclitus, 271 luscus, 25 nix, 297 inde, 287 luuci, 235 non, 338 induce, 222 lympha, 255 nonus, 327 indulgeo, 76, 423 macte, 286, 436 nostri, nostrum, 309 infensus, infestus, 398 mala, 149, 435 nudiustertius, 327, 436 infit, 399 mando, 149 num, 339 iniquus, 262 manifestus, 304 nuncupo, 210 inquam, 112, 249, 342, 352 manus, 147 nuntius, 257 inquilinus, 249 Marcipor, &c. 443 ob, 334 inquire, 262 mare, 75, 304 obedio, 265, 334 instar, 291 Mars, Mayors, &c. 146, obesus, 334 intelligo, 387 150, 244 obliquus, 159 inter, 332 massa, 267 obliviscor, 396 interatim, 208 materia, -es, 3'02 obrussa, 81 interea, 317 mea, &c. 317 obstetrix, 297 interpres, interpreter, 419 meditor, 420 obstinere, 92, 199, 334 invitus, 93 mel, 299 obsto, 298 ipse, 316 melior, 244 occentare, 214 ipsippe, 316 mens and animus, 146 occultus, 250 iracundus, 364 mentum, 149 octavus, 327 is, 315 merces, 298 odi, 399 iste, 310, sqq. mergus, 97 officina, 276 item , 443 meridie, 236 officium, 298 iterum, 255 mile, miles, 264 oleaster, 158 judaidiare, 267 Minerva, 146, 404 olfacit, 255 jugerum, 124, 269 minister, 123 olim, 314 jurgium, 213 minus for non, 338 ollus, 314 juvenis,302 mis, 308 omentum, 435 juxta, 335 modo, 443 omnimodis, 436 lacesso, 352 mollis, 435 onus, 410 lacryma, 255 moneo, 378 onustus, 410 lanius, 158 monstrare, 147 oportet, 76 lapicidinae, 435 multimodis, 436 oppidum, oppido, 93, 334 lapiderum, 281 muscipula, 431 ops, 4 lappa, 243 musso, 81 optimus, 329, 334 largus, 151 namque, 322 os, 299 larva, 151 nanxitor, 209 oscines, 95 lentus, 393 narro, 61 osem, 198 levir, 255 naufragus, 431 otium, 435 liber, loeber, 122, 206 navalis, 423 pagunt, 205 librarius, 231 ne, nee, and non -quidem, palam, 291 ligare, 255 339 paries, 264 limes, 269 nee, 98, 210, 338 parochus, 379 lingua, 255, 423 necesse, 306 pars, 291 lino, 393 negligo, 98, 387 partim, 275 lira, 159 nego, 98, 422 patefacio, 352, 437 lis, 224, 259 negotium, 98 paullus, 435 longinquus, 256 negritu[do], 98, 260 pectuscum, 21 lorica, 256 negumo, 199 pedester,4l6 Indus, 141, 160 nempe, 322 pejero, 332 luervem, 195, 404 nequinont, 251 pejor, 328, 452 lupus, 251 Nero, 54 pelagus, 305 luridus, 158 nihilum, 436 per, 332 INDICES. 469 perennis, 163 purgo, 426 rupitia, 215 pergo, 264 quaero, 352 rursus, 150 pestis, 404 qualus, 435 rus, 247 peto, 384 quando, 323 sacellum, 435 pigeo, 122 quandoquidem, 443 sacramentum, 231 pipulo, 113, 214 quantus, 301 saeculum, 146 pilumnus, 199 quapropter, 317 sagitta, 75 plaustrum, 413 quare, 249 sanates, 204 pleores, 173, 195 qua propter, 121 sarpta, 211 plorare, 173, 200, 374 queo, 382 scando, 397 plumbum, 245 ques, 232, 320 scilicet, 273, 362 poema, 266 qui and quis, 232, 320 scribere, 240 polenta, 271 quia, 320 scriptor, scriptura, 360, 411 pollex, 297 quidam, 322 se, 308 pomoerium, 335 quidem, 322 secundus, secutus, 364 pondus, 418 quippe, 323 securis, 75, 107 pone, post, 335 quispiam, 322 sed, sine, 208 pono, 395 quisquam, 320 sedulo, 231, 430 pontifex, 419 quisque, 322 semel, simplex, &c. 327 populus, 192 quisquis, 322 Semones, 196 porcet, 435 quivis, 324 sempiternus, 146 porrigo, 85, 178 quomodo, 444 senex, 253, 297 portus, 207 quoque, 325 sequester, 416 posco, 76 quoquus, 250 sero, 393 posthac, &c. 121, 317 quorsus, 150 sesquipes, sesquipedalis, postliminum, 335 radere genas, 218 425 pote, potis, 306 rastrum, 412 sestertius, 212 prse, praeter, pro, propter, ratio, 305 severus, 8 332 re=aW, 337 sibilo, 164 praedium, 122, 298 re=rei, 310 sibus, 124 praesens, 298 reapse, 316 sicilicus, 124, 270 praetor, praetura, 360 redantruo, 199 silva, 27 praetor, &c. 411 refert, 310 simul, 291 pravus, 54 regio, 76 sino, 393 precor, 76 rego, 387 siremps, 236 prehendo, 82, 399 religio, 407 sis = si vis, 436 pretium, 420 remus, 402 socer, 296 primus, 327 reor, 305 sodes, 436 proceres, 25 repetundarum, 363 solari, 435 procul, 291 repudio, 226 sollemnis, 114 proelium, 264 res, 246, 303 sellers, 114 proficiscor, 261 rettuli, 257 solus, 314 proletarius, 204 reus, 206, 305 sonticus, 206 promulgare, 244 ricinium, 218 speres, 279 propinquus, 256 rixa, 267 sperno, 392 propitius, 257 robustus, 300 sponte, 314 propter, 291 rorarius, 265 Spurius, 26 propterea, 316 rostrum, 412 squama, 250 prosper, 429 rota, 255 statim, 336 prudens, 435 ruber, 6 sterno, 256 puella, 415,435 rudimenta, 141 stipendium, 435 puer, 443 rudo, 384 stipulus, 95 puniceus, 243 ruma, 61 sto, 373 470 INDICES. strenuus, 114 tis, 308 velum, 435 suad, 120 Titus, 26 veneo, venumdo, 352 sublimis, 435 toga, 164, 423 venilia, 29 subtilis, 435 topper, 243 ventus, 152 sueres, 281 torquular, 249 venum, 360 sum and fio, 349 torres, 168 Venus, 244, 300 summus, 435 tot, 93 venustas, 301 suovetaurilia, 42S totus, 93 ver, 386 supellex, 297 trans, 335 veratrum, 412 ta=da, 196 trebla, 90 verecundus, 364 tabeo, 48 tripudium, 198, 226 verto, vertumnus, 386 taciturnus, 417 trucido, 404 veru, 242 taedet, tardus, 122 tuber, 244 vestri, vestrum, 309 tandem, 435 tugurium, 212 veto, 374 taxo, 231 ullus, 322, 338 vicus, vicinus, 212 tectifractis, 436 ultro, 314 videlicet, 352, 437 tellus, 93 unus, 264, 327 viduus, 158 temno, 269, 392 urvo, 124 viginti, 240 tempero, 392, 420 usque, 321 villa, 212 tempestas, 205, 300 uterque, 327 vindex, 111 templum , 89, 144, 269, 392, utpote, 324 vindico, 352 405 vacca, 4 vir, 53 tempus, 392 van us, 435 vis, vires, 318 tendo, 178 Varro, 54 virgo, virago, 54 teneo, 379 vasargenteis, 437 vitricus, 242 tenus, 291 vates, 244, 302 vitta, 49 tepidus, 48 vaticinari, 244 vivus, 97, 250 terra, 93 vehemens, 32 vox, 403 Tiberis, 171 Vejovis, 145 zona, 295, 403 VIII. LATIN TERMINATIONS. a, 403 dus, 404 ies, 403 ago, 300 ea, 403 igo, 300 alls, 414 edo, 300 ile, 414 anus, 414 ensis, 415 ilis, 414 ao (verb), 373, 422 eo (verb), 377, 423 ilius, 416 ar,300 er, 296, 300 in, 300 aris, 414 ero (verb), 420 inus, 414 arium, 413 es, 302, 403 io (verb), 382, 422 ax, 403 ester, 416 io, 406 b, 298 etum, 414 ior, 410 ber, bra, brum, 413 eus, 403 is, 301, 403 bills, bundus, 76, 426 ia, 403 isso (verb), 394 bulum, 413 ico (verb), 421 ito (verb), 419 c, 298, 403 icus, 403 it-s, 24, 264, 420 cer, cris, crum, 413 ido, 300 ius, 403 culus, 413, 414= idus, 404 1, 300 cus, 403 idius, 416 lentus, 393 d, 298 iensis, 415 Ius, leus, 405, 414 INDICES. 471 men, mentum, 299, 405 rum, 405, 412 tro (verb), 420 mnus, mna, mnum, 406 s, 405 trum, 411 mon, 405 sco (verb), 395 tudo, 300, 411 monia, 400 strum, 412 turio (verb), 421 mus, ma, 402 t, 298, 405 turus, tura, 411 n, 406 ta, 405 tus, ta, turn, 405 ndus, 301 tat-, tut-, 301, 410 tus, tus, 403 no (verb), 385, 391 ter, tor, 296, 300, 411, 416 v, 303 nt, 417 timus, tinus, 417 ulo (verb), 421 nus, na, num, 405 tio, 407 um, 412 on, 407 tis, 298, 404, 415 urio (verb), 421 or, 300 tivus, 404 us, -i, 296 osus, 409 terium, torium, 411 us, -uris, 300 r, 300, 405 tn, 417 vus, 404 rt, 298 trie, 297, 411 x, 297, 403 IX. FRENCH WORDS. a, 457 Besan^on, 240 choux, 251 abeille, 252 biais, 259 chose, 453 age, 258 bougre, 259 cochere (porte), 90 aigre, 252 Bourges, 257 combler, 241 ailleurs, 457 Briancon, 252 comme, 444 aime, 257 cage, 241 corbeille, 252 aimerai, 454 Cahors, 257 courtois, 257 aimois, 241 canonique, 453 dais, 252 ainsi, 457 car, 457 dame, 442 ame, 442 case, 453 demain, 457 Anjou, 257 cause, 453 derriere, 457 Aout, 252 Cavaillon, 240 desormais, 457 apotre, 258 eel, cet, 451 devant, 332 arracher, 455 cendre, 257 diacre, 259 Arras, 257 chacun, 451 Dijon, 241 assez, 257, 457 Chalons, 257 dit, 252 aucun, 259, 326, 442 chambre, 241 done, 457 aumone, 259 changer, 241 dont, 457 aune, 259 chanoine, 453 dornavant, 457 aussi, 451 chaste, 251 droit, 287 autel, 259 chataigne, 251 ensemble, 457 autre, 259 chaud, 251 ensevelir, 244 Autun, 252 1 cheoir, 251