>^r: LIBRARY t)K THE University of California. Received THE UNIVERSITY OF CHICAGO Founded by John D. Rockefeller. THE ORIGIN OF RHYTHMICAL VERSE m LATE LATIN. A DISSERTATION submitted to the Faculties of the Graduate Schools of Arts, Literature, and Science, in candidacy for the degree of Doctor of Philosophy (Department of Latin) by \ 3 R A n y UITIVERSITT . CHICAGO, ILL 1900. S3 not) j/>ijf//0 PREFACE. The work upon the sixbject here presented grew out of remarks made several years ago by Professor F. F. Abbott of the University of Chicago, in connection with the work of his seminar in Colloquial Latin, of which I was a member. He suggested that a metrical study of the Latin inscriptions might throw some light upon the problem of Latin popular poetry. I had not proceeded far, before I found that it would be very hard to make even a profitable classification of epigraphical verse without extending the scope of the investigation. This extension naturally brought with it a change in point of view and purpose , and has made the Christian hjanns rather than the inscriptions the centre of the discussion. The work has almost entirely been done under somewhat unfavorable conditions, in the spare moments of a busy instructor. Some undesirable effects of these conditions will, I am afraid, appear all too clearly in the treatment of the subject. Nevertheless I feel sure that any real results which I may have to offer are, to some extent at least, also due to the method of work made necessary by frequent interruptions and delays. At any rate the work would hardly have been possible without the kindness of the Latin department of the University of Chicago , in allowing me the free use of their library, and it might possibly have been abandoned in its early stages, if it had not been for the encouragement which Pro- fessor Abbott gave it from time to time The last chapter and the appendix were added at Gbttingen, where I had the privilege of obtaining a more thorough acquaintance with the views of Professor W. Meyer, to whom this branch of metrical study owes much of its progress. It seems impossible to write a metrical treatise without saying some- thing about the terminology. I have used the terms "arsis" and "thesis" as they are explained in the note on page three. In the use of the terms "rhythmical", "resolution", "elision" and others, I do not wish to be under- stood as taking a position in any controversy as to their more special 83210 IV application but have merely intended them as names for the facts and phenomena for which they usually stand. That "rhythmical" is not ne- cessarily the same as "accentual" will appear in the discussion of the accentual theory, and needs to be specially emphasized. The spelling has been made to accord with that of classical times though many of the hymns no doubt belong to a period when this spelling had changed in some particulars. The object of the review of existing theories in the first part of the dissertation is not merely to refute them, but quite as much to gain some positive results, which later (p. 52) form the starting-point of my own explanation. TEKRE HAUTE, INDIANA. J. J. S. TABLE OF CONTENTS. Introduction. rai?e STATE OF THE PROBLEM 1 THE MATERIAL 2 Tables of metrical faults and of conflicts of accent and ictus in the iambic dimeter 5 in the trochaic tetrameter 9 in the sapphic ... 10 in the asclepiadean 10 THE ACCENTUAL THEORY. General form 11 Huemer's modification of it 11 The facts are not all considered 12 Contradictions in Huemer's treatment 12 Examination of his results 12 Lewis' views. Double principle of construction 15 Initial inversions in dissyllables 16 THE "DEGENERATION" THEORY. Lucian Miiller's three tenets 17 Relation of the accentual practice to the neglect of quantity ... 17 Extent of conflict in the early rhythmical hymns 18 Was the change due to ignorance and neglect of quantity? ... 19 THE THEORY OF SEMITIC INFLUENCE. The views of W. Meyer 21 General argiiments against them 22 Supposed Semitic origin 24 of rhyme 24 of acrostichs 24 of groiips and strophes 25 of "Silbenzahlung" 25 Is the form of Augustine's Psalmus the restilt of Semitic influence? • • 26 of neglect of quantity and accentual pronunciation 29 ninstration by a similar case in English O" Meyer's more recent views. The accentual cadence at the end of the verse 30 VI Trochaic Popular Verse of the Classical Period. TWO KINDS OF VERSES 3^3 Examination of class I 34 » » n 37 CONCLUSIONS 38 The Inscriptions and Commodian. STRICTER AND FREER TREATMENT OF IAMBIC AND TROCHAIC VERSE. in classical literature 39 in the inscriptions 89 in the Christian hymns 40 CLASSIFICATION OF THE INSCRIPTIONAL MATERIAL 41 Analysis of the arses 44 The final syllable 45 The accented syllable 45 The non-accentual syllables 46 Analysis of the theses 47 The final syllable 47 The accented syllable 47 The syllable preceding the accent 47 Causes of the instability of the syllable before the accent 48 „ „ „ „ „ „ final syllable 49 CONCLUSIONS 51 The Transformation of Metrical Forms. REVIEW OF PREVIOUS CONCLUSIONS 52 L THE UNDETERMINED THESIS 52 This thesis gradually occupied by the unstable syllables 53 Table for the iambic dimeter 54 „ „ „ trochaic tetrameter 55 Accentual results of this shifting in the iambic dimeter 55 in the trochaic tetrameter 57 in the iambic trimeter 58 Comparison with metres having no undetermined theses 58 XL THE TENDENCY TO REPRODUCTION AND DIVISION 59 A characteristic of popular and unskillful versification 59 Seen clearly in the division of the asclepiadean 60 Change of rhythm following the neglect of quantity in the asclepiadean 61 in the sapphic 62 SUMMARY OF CHAPTERS I AND H . 62 III. THE LOSS OF ELISION 63 IV. THE LOSS OF RESOLUTION 64 vn The Rhythmical Imitation of Quantitative Metres. Page State of the language 66 Crippled conditiou of the metres 67 Devices employed to preserve the quantitative character of the verse ... 67 Low esteem in which rhythmical verse was held 68 Various types of rhythmical, as of metrical hymns 69 Various stages in the neglect of quantity 70 The initial impulse to rhythmical composition 71 Other instances of the free imitation of metrical forms 72 Some metrical forms better suited for this kind of imitation than others . . 73 Traces of the metrical origin in the early rhythmical compositions .... 73 Restrictions placed upon the development of rhythmical composition by the church and classical standards 74 Peculiarities of early rhythmical verse which still show the influence of classical standards 74 a. The treatment of syllables in which a short vowel is followed by a mute or a liquid 75 b. The treatment of short syllables directly before and after the accent 76 Local peculiarities 77 Difference between formal and inforaial composition 77 Difference in treatment between the iambic dimeter and trochaic tetrameter and iambic trimeter 78 GENERAL SUMMARY 79 APPENDIX I. THE TRANSFORMATION OF THE RHETORICAL CLAUSULA. The views of Havet and Meyer 83 Objections to Meyer's theory of the transformation 84 Review of the facts 84 Not all the possible rhythmical forms are used by the rhythmical writers 85 Transformation of the metrical form of the clausula similar to that found in metrical verse 86 APPENDIX IL TABLES A— F 88 triTI ^rERSITTZ Introduction. The period of change from metrical to rhythmical verse in late Latin occupies a most important place in the development of poetic forms. It is not only interesting in itself as an epoch of metrical history, but it is exceedingly instructive, on the one hand for the earlier and classical Latin , because of the light it can be made to throw upon many peculiar practices of versification, and on the other hand from the fact that it ushered in a new form of verse which is the most important factor in the development of the structure of modern poetry. Notwithstanding its great importance , however , the eifort that has been made to explain the nature of the change from metrical to rhythmical verse, has been comparatively small, and the results obtained have been far from convincing. The treatment of the subject has indeed, as yet, hardly gone beyond the stage of controversy. The discussion, for instance, of the accentual or non-accentual character of such verses as those of Commodian, Augustine and the unmetrical hymns has too often been merely a convenient pendant to the treatment of the accen- tual or quantitative character of the Saturnian metre and the question whether there was a Latin popular poetry whose principles of structure differed from those of classical poetry, and which ran along as a submerged current until the triumph of Christianity brought it to the surface again. One who sees the quantitative principle in the Saturnian metre, will find abundant material in the hymns and the hexameters of Commodian to indicate that they are at best a corrupt form of quantitative verse which, as Lucian Milller says, has come about through "ignorance or neglect of quantity." On the other hand, if one is convinced that the Sa- turnian metre is accentual, he will of course find little difficulty 1 _ 2 — in believing that the hymns and the hexameters of Commodian are accentual also. The error most commonly made in arguments proceeding from such sources , and made in this particular instance with perhaps more than the usual nonchalance, is the omission of that part of the evidence which is inconvenient to one's view of the case. A real solution of the problem before us can hardly be expected unless it is realized that it must be attacked with a mind free from preconceived theories , as a problem by itself , and that its solution involves a study of all, or practically all, of the metrical forms of the decline of Latin literature, including those of the in- scriptions. To be sure, not all metrical forms employed at any time are of equal importance, and yet the more widely a metre, however uncommon, differs from those in common use, the clearer is often the light which it throws upon the difficulties of solution. It was with these points in view that the present investigation was extended to include all those forms of verse which occur with any frequency in both the metrical and the rhythmical stage. The most important verses are naturally those of iambic and trochaic movement, since by far the greater number of the hymns were written in one or the other of these measures. This class of verse was capable, as we shall see later, of a development which made it acceptable in its unmetrical form, after other metres had lost their rhythmical identity or had fallen into hopeless confusion. Of the two characteristics which distinguish rhythmical from met- rical poetry — the disregard of quantity and the more or less complete coincidence of accent and ictus — the occurrence of the latter is practically confined to iambic and trochaic verse. Again, the different classes of material involved in the problem do not contribute equally to its solution. While the late pagan literature and the inscriptions yield abundant and valuable evi- dence, it is above all the hymns to which we must look as the scene of the change itself. Accordingly the tables which are given below and which will be referred to frequently, are for the most part confined to this field. They contain specimens of the iambic dimeter, trochaic tetrameter catalectic, sapphic and asclepiadean. But other metres, as well as other classes of material, will be referred to as occasion arises in the discussion which follows. For each of the metres included in the tables , all grades of development from the strictly metrical to the undoubtedly rhyth- mical form are presented. The classification exhibits the two — 3 — phases of the change, namely the gradual disregard of quantity and the gradual extension, in some cases, of the accentual principle. A word must be said about the composition of the tables. The iam- bic dimeter is, for obvious reasons, most extensively represented. Tables A 1, 2, 3 and 4 of this metre contain metrical specimens who.se date is more or less definitely known , arranged in groups according to tlicir age ; tables; B 1 and 2 contain metrical hymns whose date is not known or merely known to be early, arranged according to their accentual character; table C contains early metrical hymns with unmetrical theses ^) ; tables D 1 and 2 contain early and not yet fully developed rhythmical hymns, also arranged according to their accentual character; table E contains late and fully developed rhythmical liymns of the 1 th to the 1 2 th centuries. In the tables of the trochaic tetrameter, the metrical specimens from Seneca to Fortunatus are arranged chronologically, and all the other specimens of this metre, as well as those of the sapphic and asclepiadean, are arranged according to stages of development. These tables, as well as those presented elsewhere, have been compiled with considerable care and tested in various ways as to their accuracy. Still it is inevitable that they should contain mistakes, and for these the indulgence of the reader must be asked. Fortunately the arguments depend altogether on the broad, general import of the figm'es, and upon the varia- tion of one group from another , rather than upon individual specimens of verse. Hence the influence upon the final re.sult of occasional mistakes in counting would m any case be exceedingly slight. In deciding between different versions of the same hymn , the writer of the dissertation has generally been content, in the absence of a critical edition, with indicating the source from which he took the version used. Wlien a choice was possible, the older form of the hymn and the older manuscript were preferred. The "revised" forms of early hymns found in the Breviary were of course not considered. When a variant reading seriously afi'ected the question at issue, additional care was taken to adopt that reading which seemed , under the circumstances , most natural. In a few extreme cases of uncertainty, the word or syllable in question was left out of the count entirely. If it might seem that too little attention was paid to these matters, two facts may be urged in excuse, 1. the text of the hymns is still a wilderness, 2. the influence of individual cases upon the final result is, for the reasons aheady mentioned, very small. The doxologies appended to many hymns have been omitted , since they were very seldom written by the author of the hymn itself. 1) Throughout this dissertation , the term arsis is applied to the syllables under the ictus, the term thesis to those not under the ictus. 1* - 4 -^ Proper names from foreign languages whose quantity can not be deter- mined by rules of the Latin language, were omitted in the count. In the hymns of the transition period two ways of scanning a verse are sometimes possible. In such instances the preference was uniformly given to the way which resulted in the fewest metrical faults, unless it was evident from other scansions of the hymn or of the period, that the faulty scansion was probable in the case under consideration. In deciding which hymns might belong to a certain author, the prac- tice has been to admit only those which an acceptable weight of evidence seemed to attribute to him. Absolute certainty is of course in many cases unattainable , nor is it for the present purpose of the utmost importance. Still the writer has preferred to be conservative rather than liberal in this respect. The verses admitted into the tables under the names of the va- rious authors, so far as not indicated in the tables themselves, are as follows: Iambic Dimeter. Catalecta (Virgil) Baehreus P. L. M. Horace — Epodes I— X, XIV, XV. Second century fragments — Baehrens Frag. Poet. Lat. Ausonius — all in Peiper's edition. Prudentius — Cath. I, II; Peristepli. II, 1 — 188 (Dressel). Martianus Capella — II, 121; VI, 704; IX, 902 (Eyssenhardt). Ambrosius — the six hymns given to him by Kayser (Beitr. z. Erkl. d. alt. Kirchenhymneu) Deus creator. Splendor paternae, Veni re- demptor, Aeterne rerum, Aeterna Christi (from March's ed.), lam surgit hora tertia (D. I, 18). Sedulius — A solis ortus cardine (Huemer). Ennodius — Deus perenne, Votis Cypriaui, lam Christus, Nigrante tectam, Caelo ferunt Ambrosium (Hartel). Boethius — De consol. II, 7 ; III, 8 ; IV, 1 (Peiper). Fortunatus — Vexilla regis, Agnoscat omne (Leo). Bede — Hymnum canentes, Salve tropaeum, Hymnum canamus (March), Primo dies caeli (Mone I, 1). Attributed to Gregory — Wackemagel, d. deutsche Kirchenlied I, nos. 89, 90, 91, 92, 94, 96, 97, 99, 100, 102, 105. Petrus Damiani — Paschalis festi (D I, 223) Magna loannis (DI 225), Paule doctor (D I, 225), Clarisonis concentibus (M III, 521), de S. Maria (M III, 390). Anselm of Canterbury — Mone III, 1. Gallus et Vulpes — Grimm - Schmeller , Ged. d. 10. u. 11. Jahrh. p. 345. Sacerdos et Lupus — Grimm-Schmeller, Ged. d. 10. u. 11. Jahrh. p. 340. - B - Bernhard of Clairvaux — Dan. I, 227. Bonaventura — lu pa.ssioiic, Ilorae dc passionc, Dc dolore Mariae (]\Ione & Wack). Thomas Aquinas — Verbum supernura (Wack. & Dan.). Metrical hymns of the Carolingian period — all those in Poetae Lat. medii aevi I — III (Diimmler and Traube). T r c li a i c T e t r a m e t e r. Seneca — Medea 740—51; Phaedr 1201 — 12; Oed. 223 — 32 (Leo). Florus — Baehrens P. L. M. IV, 346. Terentiauus Maurus — Keil, G L. 11. 1300—1456. Pervigilium Veneris — Riese, Anth. Lat. Tiberianus — Baehrens P. L. M. Ill, 264. Prudentiiis — Cath. IX ; Peristeph. I (Dressel). Dracontius — Baehrens P. L. M. V, 128. Ennodius — Grammat. I, 4; I, 7; II, 107; It, 123 (Harlel). Fortunatus — Pange lingua (Leo). Abbreviations — W = AVackernagel, d. Deutsche Kirchenlied, vol. I. M =^ Mone, Lat. Hymnen d. Mittelalters, 3 vols. D = Daniel, Thesaurus Hymnologicus, 5 vols. March = March, Latin Hymns. Iambic Dimeter. A. Metrical — date certain. A 1. Catalecta (Virgil) .... Horace Martial Second Century P'ragmeuts . A 2. Ausonius Ambrosias Prudeiitius Martiauus Capclla .... A 3. Seduliiis Ter hora triua volvitur D I, 41 lamsextasensimsoIviturDI, 40 Hie est dies verus dei D I, 49 Ennodius Boethius Fortunatus Bade 20 203 31 38 398 188 40U 28 92 86 32 IGO 39 124 225 Metrical Fault.s. 1 Conflicts of Accen 1 cV ictus. u 1 1 2 1 7 2 2 2 2 u 1 2 u ^ u 2 59 2 5 121 74 109 14 45 6 3 18 17 12 1 7 12 84 12 11 77 35 92 4 12 90 14 13 131 50 95 9 u 13 1 4 4 1 1 2 1 9 9 49 11 1 8 7 13 12 11 14 29 12 2 1 1 1 30 12 5 9 6 1 2 96 39 7 19 A 4. Fratres alacri (Paul. Diacon) I, 41 Petre petra ecclesiae I, 136 . . Venit deus factus homo II, 248 . . Christus redemptor plebium II, 248 Laudem beati (Wal. Strabo) II, 296 Ad te polorum (Florus Lugd.) II, 537 Adest dies verus dei II, 249 . . . Ventosa cum desaeviat (Sedul. Scot.) Ill, 1 62 Amande praesul optime III, 680 . . . . B. Metrical — date uncertain. B 1. Attributed to Gregory .... Apostoloriim supparem D I, 103 Apostolorum passio D I, 101 . . Illuminans altissimus D I, 19. . Aeterne rex altissime D I, 196 . Primatis aulae caelicae M III, 503 Stephano coronae martyrum M III, 504 Bellator armis iuclitus M III, 429 Martine confessor dei M III, 430 Hymnum sacra novum die M fll, 475 Victor, Nabor, Felix pii M 111,551 Amore Christi nobilis Mill, 110 Festum beati martyris M III, 139 Festum beati martyris M III, 139 Aeternus orbis conditor MI, 31 Christi caterva clamitat M I, 45 Inluxit orbi iam dies M I, 77 . . Veni creator spiritus M I, 241 . Caeli deus sanctissime M I, 378 Plasmator hominis, deus M I, 380 Verbum supernum prodiens M I, 48 lesus refulsit omnium M I, 78 . Lucis creator optime M I, 82 Chorus novae lerusalem M I, 219 Magnae deus potentiae W I, 56 . Agnoscat omne saeculum W I, 60 In matutinis surgimus D I, 3 . . Grates tibi lesu novas D I, 47 . Nunc sancte nobis spiritus D I, 50 Rector potens, verax Deus D I, 51 Rerum deus tenax vigor D I, 52 Aeterna caeli gloria D I, 55 . . Aurora iam spargit polum D I, 56 Agnis beatae virginis D I, 94 . . B 2. Diem sacrati hominis M III, 97 . Auni recurso tempore M III, 155 Vox clara ecce intonat W I, 51 Fortem fidelem militem W I, 68 . Unam duorum gloriam W 1, 58 . Jam meta noctis transiit D I, 3 . Decus sacrati nominis D I, 8 . . Consors paterni luminis D I, 27 si 1 Metrical Faults. Conflicts of Accent & Ictus. 3 "" • o u — u — u — u u^ u — ^J — u — u u 64 1 3 2 28 13 13 16 28 1 1 7 5 3 5 20 1 11 4 2 3 IG 7 1 48 25 14 60 82 3 3 5 2^ 1 1 12 8 9 10 36 1 17 6 2 44 1 18 5 1 2 244 2 2 1 2 87 32 5 10 32 1 11 5 8 9 32 2 15 2 9 18 32 8 4 U 12 28 2 12 6 2 2 24 2 8 3 3 3 20 9 1 1 2 32 2 1 8 1 1 3 24 9 4 1 2 20 2 5 2 10 3 1 1 32 1 10 5 6 9 27 1 10 2 8 7 28 16 3 1 2 32 19 3 1 3 126 3 1 46 18 14 17 32 2 11 3 4 5 32 1 9 3 2 5 24 1 1 5 1 1 2 u. 1 1 6 3 2 2 12 1 2 4 1 2 3 16 3 1 4 1 1 2 32 18 6 4 7 16 1 7 3 1 1 20 9 5 1 1 16 1 4 4 1 2 32 2 10 4 1 4 15 1 4 1 1 o 2 1 2 32 1 15 4 7 9 e 2 1 1 1 8 3 1 2 2 e 2 1 2 2 20 1 1 5 2 2 12 4 1 2 2 32 1 12 2 14 14 20 2 1 5 2 1 24 6 1 1 16 2 5 1 4 1 1 32 13 4 1 28 12 1 1 12 2 2 1 1 20 1 5 2 1 12 4 1 Ill Trinitatis uuitas D I, 35 . . Cibis resuniptis cougriiis D. I 72 Tcruis ter horis imraerus D I, 73 lesii coroua celsior D I, 110 . . Til Christe, nostrum gaiulium M I, 229 Te, liicis auctor, persoiiant M I, 185 Beat a nobis gaudia M I, 241 . . Ad Katherinae nuptias M III, 354 Votiva cunctis orbita M III, 415 In laude regis omnium M HI, 240 Rex gloriose martyrum M III, 143 Fit porta Christi pervia W I, 47 lux beata, triuitas W I, 52 . lesu quadragenariae W I, 53 . . lam lucis orto sidere W I, 56 . lesu, salvator saeculi W I, 83 Quern terra, pontiis, aethera W I, 63 Somno refectis artubus D I, 26 . Summae Deus clemeutiae D I, 84 Te lucis ante terminum D I, 52 . Rerum creator optime D I, 53 . Nox atra rerum contegit D I, 54 Dei fide, qua vivimus D I, 71 lam cursus horae sextae D I, 71 Convexa solis orbita D I, 72 . . Deus tuorum militum D I, 109 . lesu corona virgiuum D I, 112 . C Unniotrical Theses. Deus qui certis legibus D I, 42 . Aeteruae lucis couditor D I, 39 . Fulgentis auctor aetheris D I, 43 Lucis largitor splendide D I, 1 . Certum tenentes ordinem D I, 45 Dicamus laudes domino D I, 45 Perfectum triuum nuraerum D I, 45 Tempus noctis surgentibus D I, 67 . lam Christus astra ascenderat D I, 64 Ignis creator igneus. Autipb. Bencbor. D. Rhythiuical- undeveloped. D 1. Bis teruas boras explicans D I, 23 Rex aeterue domiue D I, 85 ; IV, 20 Optatus votis omnium M I, 232 . lam Christe, sol iustitiae, M I, 91 Ad cenam agni providi M I, 217 Regi polorum debitos M III, 266 Rex Christe, Martini decus Mill, 431 Sacri triumphale tui M III, 449 . Adest dies sanctissima M III, 459 A solis ortus cardine W I, 47 . Deus, qui caeli lumen es W I, 78 Deus pater ingenite D I, 2 . . Mediae noctis tempus est D I, 42 Deus aeterni luminis D I, 68 . . 2 '^ Metrical Faults Conflicts of Accent t^ Ictus. a> a> <^ — w — \j — u u u — u — M — u u 20 5 1 1 12 8 3 2 12 2 6 1 1 3'2 2 8 1 1 12 4 1 24 7 1 24 2 1'^ 2 24 1 1 6 3 16 1 8 1 2() 1 3 1 2 8 1 \1 2 1 1 12 1 1 4 3 J^ 20 1 3 1 4 1 10 6 3 16 3 1 1 1 7 2 32 1 12 1 k; 1 6 1 IG 1 7 2 8 1 2 lii 1 6 1 16 1 2 1 1 4 2 2U 1 1 7 1 16 7 32 1 1 1 9 1 16 4 20 10 12 3 24 3 17 1 12 5 20 1 7 1 11 4 1 2 3i 2 17 1 1 10 3 1 12 I 7 1 1 5 12 1 9 1 6 12 7 1 1 8 2 IG 6 1 2 7 3 1 2 32 7 9 1 2 2 32 1 22 1 16 1 2 5 32 6 15 5 8 2 6 2 8 8 64 9 37 8 15 3 17 3 2 3 32 4 16 2 4 2 9 1 3 7 20 1 2 7 5 3 2 24 3 12 4 5 5 5 3 32 5 8 7 18 5 2 3(1 1 1 4 9 12 9 7 32 6 11 3 2 19 4 8 20 6 5 5 2 2 56 7 12 4 11 19 5 2 40 1 18 5 2 16 3 5 20 9 2 3 6 1 3 3 52 7 36 8 10 3 22 10 2 8 24 4 9 3 2 5 9 3 10 12 — 8 — Diei luce reddita D I, 68 . . . Deus qui claro lumine D I, 73 . Christe redemptor gentium D I, 78 Mysterium ecclesiae D I, 79 . . Stephano primo martyri D I, 90 Post matutinas laudes D I, 44 D 2. Christe qui lux es et dies M I, 92 Aurora lucis rutilat W I, 80 . . Nunc tibi, virgo virginum W I, 82 Magni palmam certaminis D I, 102 rex, rector regminis M III, 181 Conditor alme siderum M I, 49 . . Mysteriorum siguifer M I, 446 . . lesu nostra redemptio W I, 55 . . Hymnum dicamus domino W I, 53 . Magno salutis gaudio W I, 74 . . lam ter quaternis trahitur W I, 82 lam lucis splendor rutilat D I, 69 . Meridie orandum est D I, 72 . . . E. Bliytliiuical - Petrus Damiani . . Anselm of Canterbury Gallus et Vulpes . . Sacerdos et Lupus Bernhard of Clairvaux Bonaventura .... Thomas Aquinas . . developed. Metrical Faults. Conflicts of Accent & Ictus. 120 232 ■28& 8U l<)v 150 24 13 9 8 15 8 5 15 25 8 12 8 12 14 19 7 4 68 12) 158 53 134 86 15 42 69 lOO 33 64 65 — u _ 9 — 31 « ^ CO 05 -«J< O (>J o ^ -H OO 1^- rt -^ ^ 1 Dl c^ i:: lo CO --H ■^ -< CO — •^ 1 D ec i> CO — ' o CO CM 0-1 "—I ^ — 1 CO -H ,-H ^ SQ CO r-c rt lO ■>*> C-1 -^lOCO- t.-3 00 •* ■«< T-^ r-< CO C^ C- 03 1—1 OJ — < t> CO — CO 1 D CO ■O C:. CO lO t- C>) T — C<0 lO ■^ CO CO Ci CO r-l CJi r-H (M '^ CO -rr OJ (M m (N .— 1 1 D D (M 1 -<*«'-<-* ^« *-0<— "lOCOOCO-^OD — — —1 ^ CO — > OCI --H -^MCOCOCOCOiOOO'^— < Oi CDCOoocooo r-t ^ CO CO es uber - —.---- ^ ^ . . -- ^ V,-, . - ^^ V.V, w- «- cooJiooocococacoco-^irttOi-icoi— I t- o- era o2 i> o r- l-l CO _ -'C0ir3C005>-iC000OiO cooQco-^cococjco^o^ c3 eS — a — CO ■^ 00 i-T • o -, ■^ CO i_| ,— ( ,_! K^ CO ■«9< HH to 1;^ •- -'s! CO . -t; °Q 3 fl P Q'^ CD H u o O) C3 05 -»^ w 3 fl .„ „ _- _ ® ti (D > -i ^ 3 c3 2 u Ci-1 a 6 TO S H Cm H &H Q W fe ■ o— ^ u ±> a o P o a> :-■ u (O 00 ft 3 .- a oa CO CO -H ^-1^ ^ ,e- CO >— ' CO oi p „ ti "^ g p g C3.2 « o3 QJ c3 © " " g-^ S^a a, P c3 00 d !- •- go 3 -2 -S S ^ O ^ g ^§a-^« Oj ^ c3 3 P- pj 'O c3 tn P - K --H-T .00 w^ •S^acei^-^ S gpPpCDrf Jj S P:3P^P--=i ik, MOT .r-» -^ QJ _^ ^ -P o '-i , 3 • ;-> C3-S P.^ g^ ^ • • • • CO >— ( »o •* 1— 1 CO CJ o CO C5 10 CO 10 00 C3 g ID P -u r>- 00 g 3 >-■ 1— 1 Q g D Jp a _s IB ^'" u 'p CO ja 3 to 1— <"i— T >—' 1— ( 03 00' -*^ 3 ll a a 3 CO M P CO 00 I— 1 j" u Si -^ d P 3 "be 3 <: to ce a be ^ > c3 tS 3 t— ( 3 c3 c3 ^ jp > a p^ i-:iH t— 1 Hj 1 1 \j \j 1 u u 12 9 39 15 29 10 w u 2 1 1 1 1 4 1 3 1 6 7 27 6 12 14 20 2 10 11 1 2 1 2 1 3 8 24 21 19 13 4 1 2 23 15 1 2 1 1 3 2 5 4 4 1 2 2 5 1 8 5 9 27 30 34 16 24 23 1 1 2 1 1 2 5 C 15 18 14 16 1 2 7 11 5 2 10 3 7 16 2 4 36 20 1 16 20 8 11 15 38 36 22 41 13 18 62 37 20 1 2 2 4 8 5 10 4 5 7 3 5 12 10 2 9 5 5 4 10 9 3 8 3 4 13 6 5 1 5 5 2 4 8 9 5 7 1 15 10 1 12 44 23 9 23 47 23 28 19 4 27 98 10 15 1 Metrical F Aults. Confl icts of Accent & Ictus. 1 w u 1 u u u u 16 6 14 10 2 9 29 8 23 29 7 14 6 6 6 11 7 2 5 2 4 6 D 5 4 2 4 2 11 8 8 9 8 6 1 4 9 16 2 1 1 1 4 1 1 11 IS 1 2 1 5 7 1 3 2 18 24 1 1 3 5 22 8 1 28 2 5 14 3 1 67 102 2 12 11 4 6 15 1 7 3 14 IB 4 13 21 6 2 4 7 8 2 2 7 12 7 7 15 1 4 3 5 9 4 5 12 15 7 8 21 6 1 2 9 14 1 10 16 15 7 1 6 19 2 9 5 3 9 3 2 17 14 8 8 20 1 5 9 5 2 7 3 15 10 7 3 1 6 14 11 9 2 4 6 9 18 15 3 5 1 2 2 10 2 1 7 9 1 12 15 3 3 7 7 13 6 3 17 21 11 8 24 5 — 11 — The Accentual Theory. The broad and general form of the "accentual theory" is that along with the sermo plebeius went an accentual popular poet- ry, traces of whose influence are seen in the prominence of the accent in the verse of the comic poets, and pure specimens of which are found in the trochaic soldiers' songs, lampoons &c. pre- served by Suetonius and a few other writers. This popular poet- ry, it is said, lived on among the common people, though rarely represented in writing, all through the classical period ; and when Christianity, which first gained its hold among the lower classes, became supreme, the popular method of constructing verse also gained the supremacy through the hymns, which according to this theory exhibit practically the popular poetic form turned to new uses. To such a view several unanswerable objections can easily be made. If the Christian hymns grew up as a popular accentual form of poetry while Christianity still abode with the lowly, why is it that Ambrosius and most of the other early hymnists wrote only quantitative hymns? Further, as is seen from the tables just presented, there is a large body of hymns (Iambic Dimeter C and D, Trochaic Tetrameter B and C), which are neither wholly quan- titative nor wholly rhythmical , but evidently represent a transi- tional stage and plainly suggest a gradual development from the quantitative to the rhythmical form. Huemer's Theory, — Considerations like these have led to the modified accentual theory of Huemer (Untersuchungen liber d. altesten lat.-christl. Rhythmen). While still holding to the belief in the existence of a popular poetry which was constructed according to accent and not according to quantity, along with the opinion that the Saturnian metre is accentual, he admits that the Christian hymns were, to start with, quantitative, but maintains that, as a result of their use in the church, they gradually made concessions to the popular manner , in consequence of which the various stages of transition just mentioned were brought about — "dass die lateinische Hymnendichtung von Anfang an zwar nicht accentuierend, sondern quantitierend gewesen sei , dass sie aber die verschiedensten Concessionen an die Rhythmik in Folge ihrer Bestimmung als Liederdichtung fiir den Grebrauch in der Kirche gemacht habe und allmahlich dieser accentuierenden Richtung ver- - 12 - fallen sei" (Untersucliungen , p, 5.). This modification makes the theory appear in general to conform to the outward facts , and we shall accordingly be obliged to examine it more in detail in order to test its adequacy. The reader of the "Untersuchungen" just mentioned is struck by two noteworthy facts. In the first place, Huemer ignores the hymns written in the sapphic and asclepiadean metres, and dis- misses the hexameter of Commodian without giving any reason for doing so (p. 9.). Now the hexameter and the sapphic do not at any time within the period of transition show a tendency toward greater harmony of accent and ictus. A glance at the list of sapphic hymns in the tables just referred to will convince us that the proportion of conflicts between accent and ictus remains much the same in all classes of these hymns, both metrical and rhythmical. As for the hexameters of Commodian and other rhythmical specimens of this measure, it is clear from the investigation of W. Meyer (Anfang u. Ursprung d. lat. u. griech. rhyth. Dichtung, p. 300 ff.), that they are no more accentual than quantitative hexameters are. The asclepiadean attained harmony only in its second half in the later metrical and the rhythmical specimens. It is not hard to see that if all these metres had been considered by Huemer, they would have been, to speak moderately, very inconvenient to his theory. They show the opposite of what he believes , and so he confines himself to iambic and trochaic verses. The other thing that strikes the reader in the "Untersuchungen," is the change from the confident spirit with which the author her- alds his theory in the first few pages to the rather colorless and non-committal way in which the results of the investigation are summed up (pp. 59 — 61). On pages 5 and 18 it is stated that the hymns were origi- nally metrical, but gradually made concessions to the popular pronunciation until they fell completely under the domination of the word-accent. On pages 20 and 59 the change is said to be due to the increasing strength of the verse - ictus , which caused even metrical poets to make concessions to it to the extent of avoiding conflict between it and word -accent at certain places in the verse, and of admitting longs in place of shorts in the thesis. In the rhythmical poetry then, the verse-ictus became so powerful, according to Huemer, that short syllables were put in place of former longs in the arsis,, and thus, as it were, lengthened (p. 20). Huemer's investigation had shown him, no doubt, that the theory of accentual influence was in itself insufficient, since, even -la- in the metres examined by him, the number of conflicts between accent and ictus did not materially decrease anywhere except in the trochaic tetrameter, the iambic verses continuing to have a considerable proportion of conflicts even in their rhythmical form (cf. these metres in the tables). As a result he was obliged to shift his base of operations from the influence of popular accentual pronunciation to the influence of the verse-ictus. But this is only an awkward attempt to avoid the issue. For how can there be a prominence of the verse - ictus apart from the prominence of the syllables standing in the arsis'? In metrical poetry this promin- ence depends upon the prominence of a series of longs ^), recurring at certain intervals, over the intervening syllables, which are very largely short ; in accentual poetry it depends upon the prominence of the accented and emphatic syllables over the unaccented and unemphatic syllables. To say then that verse-ictus brought about rhythmic poetry is to assign as a cause what is merely an atten- dant phenomenon. If on the other hand we take Huemer's statement made in the beginning of the "Untersuchungen" that rhythmical poetry came about through concessions to popular accentual pronun- ciation, a mere glance at the tables previously presented, or at his own (Unters. p. 32), will convince one that the facts are against the theory. For with the exception of the trochaic tetrameter and, to a less degree, the iambic trimeter, every rhythmical verse - form shows abundant and regular conflicts between accent and ictus , and it is surely contrary to common sense to suppose that the accent caused rhythmical poetry, when rhythmical poetry itself is not accentual to any marked degree. But the unsoundness of the theory is shown still more clearly by the fact that among the most perfectly accentual specimens are some of the metrical hymns (cf. Iambic Dimeter B 2 and C, Trochaic Tetrameter A and B in the tables). In fact in both the trochaic tetrameter, which reached the highest accentual character, and in the iambic dimeter, which took some steps in this direction, no progress seems to have been made by the rhythmical hymns over the more advanced metrical ones ; rather , if anything , the 1) This is of course uot to be interpreted as excliuliug the possibility of a resolution of the long syllable in the arsis. Anapaests as well as spondees may take the place of trochees and still leave the verse trochaic in rhythm through the preponderance of trochaic feet. If we were to introduce a sufficient number of anapaestic feet, the rhythm would no longer remain trochaic, since the verse-ictus would naturally fall upon that part of the successive feet, where the largest number of long syllables is found. — 14 — fully developed rhythmical hymns show a slight retrogression from the stage of accentual development reached by the metrical hymns. Such a state of things is wholly at variance with the theory of accentual influence. For one can see that the accentual character might not be fully developed while the verse remained metrical, on account of the restrictions imposed by the metre; But one must at any rate expect that , when these restrictions were removed, the accent would gain control of the verse and change its character completely. Surely, if the accent was able, even when hampered by the rules of quantity, to accomplish the progress toward an accentual character as we see it in the met- rical hymns , it would do incomparably more when left to itself, as it is in the rhythmical hymns. But it did not do anything under these favorable conditions : hence we must conclude that what happened under unfavorable ones was not its work. But even the statement to which Huemer at last confines himself, that in the end there were two classes of hymns, the one merely insisting on harmony between accent and ictus in certain parts of the verse , the other striving for complete harmony (Unters. p. 59) , is misleading. Our tables (as well as his) show that to the latter class of hymns belong only those in trochaic tetrameter, while the iambic dimeter and the second half of the asclepiadean belong to the former, and the sapphic, the first half of the asclepiadean and, as Meyer has shown, the hexameter, do not make any appreciable effort at avoiding conflict in either their metrical or their rhythmical forms. The percentage of words showing a conflict between accent and ictus found by Huemer (p. 32) in the early rhythmical hymns which he has examined is as follows : Iambic Trochaic 10.04 27.8 14.9 31.12 5.5 55.3 5.6 29.7 14.9 37.4 9.7 36.2 9.6 9.04 4.7 11.3 6.1 16.6 21.4 Average 25.99 + Average 7.88 — — IB — Of course there are wide variations in each metre, due no doubt to a variety of causes. Some of the writers of rhythmical hymns in iambic dimeter were in constant danger of slipping into the trochaic measure, as is seen by the fact that they admit trochaic lines occasionally^). This admission was accomplished Ijy lopping off the first syllable of the iambic verse. When that happened, or even if the first syllable of the iambic dimeter was merely considered an anacrusis, the verse was practically trochaic, and the same characteristics might be expected as are found in the second half of the trochaic tetrameter. But in spite of the variety which thus exists within the sep- arate metres , the general statement is no doubt true , that the degree of harmony attained between word-accent and verse - ictus depends upon the kind of metre used, and is not the result of a common influence like the accent, which would afi^ect all kinds of verse alike. Lewis' Theory. — A few additions have recently been made to the accentual theory by Charlton M. Lewis (Foreign Sources of Modern English Versification). He holds with Huemer that in the time of Augustine "what actually appealed to the ear in the reading of these hymns was not the quantity but the metrical ictus" (p. 26) , without of course showing how a fixed metrical ictus would be maintained in hymns which were not yet accentual, at a time when as he says (p. 24) natural prosody was dead. Sedulius and Fortunatus , he thinks , paid attention to accent as well as to quantity in the composition of their hymns, a view which we shall meet again in connection with another theory. How easy it is to fall into this way of thinking, can be seen from the fact that it has even been thought that Horace paid atten- tion to both accent and quantity in the composition of the sapphic line. Now it is true that accent and ictus coincide nearly always in the end of the iambic and trochaic verses of Fortunatus and Sedulius, as well as of their contemporaries and successors. It is also true that the sapphics of Horace nearly always , and those of Seneca with perhaps one exception (Ag. 865), have their 1) Mone III, 449, 1 6. Turba laudem concinit. Dan. I, 70, 1 5. Hora prima psallimus. Dan. I, 72, 1 2. Christus deprecandus est. Dan. I, 102, 1 27. Ensis et craticulae. Dau. I, 44, 1 7. Sed oremus sedule. There is a case of a suppressed thesis in Dan. I, no. 66, 1 1. lam cur- sus borae sextae. — 16 — accents on certain syllables. Bat these facts need not have come about because these metrical poets recognized the accent as a metrical principle. If this had really been the case, it could only have res- ulted from an increased prominence of the accent, and then it is assing strange that it should not have occurred in other metres. Over nine-tenths of the poetry of Sedulius and Fortunatus was written in the hexameter or the elegiac distich, and these verses are not a whit more accentual than they had been at any time in their history. More than that, our common sense tells us that such a practice as wri- ting poetry according to two different principles at the same time, could never become a universal and permanent custom, to say nothing of the peculiar assumption that , while the accent played a very subordinate part in early and classical Latin and now plays a subordinate part in the Romance languages, there should have been a period between these two, when it was strong enough to usurp the supremacy of the verse. The coincidence between accent and ictus which is found in the later quantitative iambic and trochaic metres, was due to a far different and very natural cause, as we shall see (p. 52 ff.). A more original contribution is made by Lewis in his theory of "initial inversions." His view is that the normal form of rhyth- mical verse was accentual throughout, and that the conflicts which occur so frequently, in the first half of the rhythmical iambic di- meter for instance, were merely "interruptions to the smooth flow of the ideal rhythm," which "occur often enough to prevent mo- notony" (p. 39). It requires great courage to state such a theory after finding that Adam of St. Victor has 71 single and 89 double inversions, i. e. altogether 249 inverted feet, in a total of 310 iam- bic eight - syllable lines, while in 318 trochaic lines of eight or seven syllables by the same author there are but 28 inverted feet. Can it be that the iambic line was in its ideal state nine times as monotonous as the trochaic line? For this difference holds not only in a single writer but , as can be seen from the tables , for the later iambic and trochaic verses generally. And what shall we say of the other metres which continued in their rhythmical state to have the same number of conflicts (Table of iambic di- meters D 1 , sapphic B , and asclepiadean B and C) which they had had in their quantitative state? Lewis also makes the point that, as early as Sedulius and Fortunatus, the inversions always fall on dissyllables (p. 27), and that "in the care with which the accent of polysyllables is pre- served, there is a peculiar significance" (p. 29). It is perhaps — 17 — true , as Lewis suggests , that the departure from the accentual rhythm is less striking wlicn a word of two syllables has the ictus on the last, than when a word like "errorum" has the ictus on the first and last syllables. But the fact in the case is that in those quantitative iambic dimeters and second parts of trochaic tetrameters which show an agreement of accent and ictus at the end, there is no possible chance for a conflict of accent and ictus in other words than those of two syllables, since a word of three or more syllables never has the accent on a short syllable unless this syllable is followed by another short, and for words of this kind there is no place in the metrical form of the verses mentioned. A word like "errorum", moreover, is excluded from these quanti- tative verses by the fact that in order to have harmony of accent and ictus, the end of the line must be formed either by a poly- syllable, or by a dissj-llable or a pol^^syllable plus a monosyllable. When either of these endings occurs, there is left only a^ ^ u _^ H or -z- u ^ ^ in the two cases respectively. The divided form of the first part of the trochaic tetrameter (— ^ _ ^ || _ u — ^) which had become almost the only form used by the time of Fortunatus, does not admit conflicts at all, except when a monosyllable stands before the caesura, and then only in words of two syllables. The "Degeneration" Theory. Opposed to the theory just treated is the older view — usu- ally a general impression rather than a theory — that rhythmical poetry is the result of a negligent and corrupt pronunciation that came with the decline of culture and the conquests of the barba- rians. This view has been elaborated by Lucian Miiller (d. satur- nischeVers, p. 28 ff. and de re metr. 2nd ed., p. 554 fF.), who holds that the change from metrical to rhythmical poetry was due to igno- rance or neglect of quantity (d. sat. V. p. 28; de re metr. p. 554); that the earlier rhythmical poetry showed all the conflicts between verse-ictus and word-accent which are found in metrical poetry generally (d. sat. V. p. 29); and that only later on, after the ne- glect of quantity had become general, did the practice originate of writing poetry according to the accent, a practice which was brought about by the need of having something to give character and firmness (einen Halt) to the verse (d. sat. V. p. 30 and 31). Of these three tenets the last is shown to be false by the relation which we have found existing between the neglect of quantity and the coincidence between accent and ictus. As we 2 - 18 - have seen, tlie accentual development not only, to a large extent, preceded the admission of unmetrical syllables, but actually reached its height in the metrical hymns (cf. pp. 13—14). The second contention, namely that the earlier rhythmical poetry showed all the conflicts between verse-ictus and word-accent which are found in metrical poetry generally , is probably true to the extent that between the rhythmical and metrical hymns of the same period there was perhaps, if we make allowance for in- dividual peculiarities, no great difference so far as the extent of the coincidence between accent and ictus is concerned. This appears when we compare with one another a number of early hymns, metrical and rhythmical, which are referred to in the writings of about the middle of the sixth century. All but the last one of these hymns are mentioned in the regula of Aurelianus, bishop of Aries f 555 A.D. ; the last hymn is quoted by Cassiodorus, in the Exp. in Psalt. CI (Migne LXX, 707) which was written about 550 A. D. Of these hymns the first three belong to Am- brosius and are therefore earlier and, generally speaking, less faulty and less accentual than the others. number of conflicts be- tween accent and ictus. lam surgit hora tcrtia 32 1 37 Splendor paternae gloriae 32 1 27 Deus creator omnium 32 1 29 Hie est dies verus dei 32 45 Ter hora trina volvitur 32 22 lam sexta sensim volvitur 40 4 13 Fulgentis auctor aetheris 20 9 18 Deus qui certis legibus 20 10 15 Aeterne lucis conditor 24 21 17 E,ex aeterne domine 64 72 25 Bis ternas horas explicans 32 36 24 The statement made by Miiller is not true , however , in the sense in which he himself understands it , namely that the early rhythmical hymns are no more accentual than quantitative poetry had always been — „so wurde audi das beiden klassischen Sprachen gemeinsame Princip, den poetischen Iktus moglichst in Widerstreit zum grammatischen Accent zu bringen, am Ende der metrischen Eeihe aber ein Monosyllabum mit vorhergehendem mehrsilbigen Wort zu meiden, unverandert in die rhythmische Poesie iiberge- — 19 — nommen." (D. Sat. Vers, p. 29). For, as is clearly evident from the tables previously presented (compare la. Dim. A 2 and A 3 with A 1; als(') Trocli. Tetram. A), at the time of the earlier rhyth- mical h^-mns all iambic and trochaic verses had in large measure become accentual. The lirst of the three statements , that the origin of rhyth- mical poetry was due to ignorance or neglect of quantity, calls for a few words of comment. The ignorance and neglect of quan- tity which Miiller mentions, he attributes to the decline of culture and to the growing influence of the provinces where Latin was spoken. That the pronunciation of Latin in the provinces differed in many respects from that at Rome , that distinctions of quantity, if we are to draw an inference from the modern Romance lan- guages, were perhaps less marked, is no doubt true. To be sure, the Latin in which the hymns were written is still Latin, it is not Romance. This fact does not, however, exclude the possibility of a lack of sensitiveness in the pronunciation of quantity. And, this tendency being at work , it would be made much easier for the rhythmical practice to gain the upper hand as it did. Yes, in the course ot time and among people who did not have a specific knowledge of the character of ancient verse, a factor like this might be able unaided to break down the classical standards. But did it as a matter of fact do this? We have already seen that it could not have given the verse its accentual quality for the simple reason that this was established apart from, and to a large extent before the admission of unmetrical syllables, in which alone the working of such a tendency as that under dis- cussion would appear. Was it then this dialectal lack of sen- sitiveness to quantity which biought about the neglect of quantity as we find it, for instance, in the hymns and inscriptions ? It is a fact, as can be seen from the tables previously pre- sented, that the extent of the neglect of quantity differs widely in different parts of the verse. For instance, it is usually greater in the third arsis of the iambic dimeter and the third and seventh arses of the trochaic tetrameter , and it is hardly ever found in the last thesis of either kind of verse. Besides this the individual hymns show very marked individual peculiarities. From an exam- ination of the tables we find that all these peculiarities are particularly prominent in the rhythmical hymns which are but partly developed , that is, which still belong to the transitional stage, the very time when soihe force or influence was at causing a neglect of quantity. //^ 2 * f -Tl-TIVERSITY — 20 — Now if the lack of sensitiveness to quantity were the only, or even the leading cause of this neglect, the differences spoken of ought not to exist, at least not to such an extent as they (jlo. For the want of a sense of quantity would surely not manifest itself regularly and freely in certain parts of the verse and very rarely, if at all, in others. Why then does this unevenness exist in the verse, why do we so rarely have a long syllable in the last thesis ? This we might perhaps, for want of a better explan- ation , consider to be an inherent and inalienable feature of the verse itself, if it were not for the fact that we do have a small number of hymns which show a very close approach to such a condition of affairs as we should expect if the lack of sensitive- ness to distinctions of quantity were the sole cause of rhyth- mical poetry. Three may be mentioned as especially remarkable^): Mone I, 447 Archangelum mirum magnum, Dan. I, 68 Deus aeterne luminis, Mone I, 96 Summi largitor praemii. The first of these has ten cases of a long syllable in the last thesis in a total of 48 lines , the second has five in 24 lines and the third has ten in 20 lines. There are some other hymns which, as the tables show, have scattering cases of this kind, and they occur mainly, though not exclusively, in the less developed rhyth- mical hymns. This is but natural. For in the first place metrical forms were then nearer being in a state of flux than before or after , and had less power of resisting any sort of tendency or influence. And, besides that, it is only in this earlier period that the dialects which later developed into the various Romance languages , were still sufiiciently like High Latin to influence its pronunciation. Later on, by the tenth and eleventh centuries, the vernacular had become so different from Latin that it was felt to be a different tongue and in fact was used by a different class of people. As soon as this divergence had reached a certain point , the two tongues could of course no longer influence each other in pronunciation. The fact which interests us at present is that the number of those hymns which seem to have been written by persons with very slight sensitiveness to quantity, is so small as to justify us in calling them the exceptions which prove the rule. Our conclu- sion is therefore this. The "neglect of quantity" as Lucian Miiller 1) Compare also the Hymnus abecedarius contra antitrinitarios (Boucherie, Melanges latins et bas-latins) which has fifteen cases of a long penultimate syl- lable in 280 iambic dimeters. — 21 — understands and uses that term, would have been capable in itself of brcalving down the structure of Latin poetry, but the fact that its intiuence is unmistakably and positively seen only in a very few liymns in the most susceptible period of linguistic and metrical development, and in most of the h3'mns is not seen at all, shows this "neglect" to have been a comparatively unimportant factor. There is the further fact to be accounted for , that during the same time in whicli rhj-thmical hymns were composed, metrical hymns continued to be written ; indeed some metres , such as the hexameter, which were very frequently used, are almost without exception found in the metrical form. And more than that, some- times the same men who wrote metrical hexameters and penta- meters, also wrute rhythmical iambic dimeters or trochaic tetra- meters. This was the practice of Aldhelm , Boniface , Paulus Diaconus , Petrus Diaconus and probably of Bede ^). With the writers of the Carolingian period, there was the further distinction that the iambic dimeter was usually employed in its metrical form, while the iambic trimeter and trochaic tetrameter were more often used in their rhythmical forms. All this points to the same conclusion which we reached be- fore, that there was a diiference in the treatment and development of the ditferent metres , and that no general influence can have caused the phenomena which we find in rhythmical poetry. The Theory of Semitic Influence. Before we pass to an examination of the literary remains which furnish evidence as to the nature of the change from met- rical to rhythmical verse , it will be necessary to mention the attempt whicli has been made to explain this change independently of, and without reference to the previous metrical forms of the language. This is the theory of W. Meyer (Anfang u. Ursprung d. lat. u. griech. rhythm. Dichtung). Meyer has after extended and thorough investigations satisfied himself that neither the "accentual theory" nor the "corrupt pronunciation theory" can explain the origin of rhythmical poetry, and he thinks he has found this origin in the Semitic, and particularly the Syrian and Hebrew church poetry. „Als ich erkannt hatte," he says (p. 113=377), „wie ahnlich die altesten Rythmen der Lateiner und Griechen einander in vielen inneren und ausseren Stlicken seien, wie aber dennoch weder 1) The passage iu which he inentious amoug his works a "liber hymnorum diverso metro sive rhythmo," is usually interpreted in this way. — 22 ~ gleichzeitiger einheimischer Ursprung, noch Uebergang der ryth- mischen Dichtung von den Lateinern zu den Griechen oder umge- kehrt angenommcn werden konne, war ich lange in peinlicher Un- ruhe ; endlich bekam ich Licht , als ich dieselben Formen in den Dichtungen der semitischen Christen aus friihester Zeit wieder- fand und mir vergegenwartigte , wie lebhaft in den ersten Zeiten der geistige Austausch der Christen der verschiedenen Nationen war, und immer fester wurde die Ucberzeugung, dass weder die lateinische noch die griechische Rythmik ein einheimisches Gewachs sei, sondern dass das Grundprincip der rythmischen Dichtung nebst manchen auffallenden Aeusserlichkeiten mit dem Christenthum von den Semiten zu den Lateinern einerseits und zu den Griechen andererseits iibergegangen sei. Durch jenes semitische Vorbild wurden diese Volker angeregt, die Quantitat der Silben nicht mehr zu beachten, woraus die Aussprache nach dem Wortaccent sich von selbst ergab , dagegen auf die Silbenzabl zu achten , die Zeilen in Gruppen oder Strophen zu schliessen, die Gruppen oder Strophen durch Akrosticha oder Reim zu binden. Diese Elemente finden wir fast alle schon bei den Syrern." This is in reality an article of faith rather than a theory. For our ignorance of the practices which obtained in the singing of the early church, and especially our ignorance of the nature of the music, make it very hard to prove or to refute it. Still, we are at present concerned only with the Latin side of the pro- position, and from this side, at least, some very weighty arguments of a general nature can be urged against it. To begin with, Latin was a fully-developed language with an extensive literature, and its poetry had for a number of centuries been consistently quantitative. There can be but little doubt that substantially the same distinction between long and short syllables that we find in poetry, appeared also in the pronunciation of the language in its non-poetical uses. Further, in the Semitic countries of the Eoman empire, it does not appear that the vernacular was displaced by Latin, or that Romance dialects were formed. Con- sequently the Latin spoken in these provinces must have been the Latin of the officials and soldiers sent from Rome, and must have remained comparatively free from Semitic corruption of its pro- nunciation. There was therefore, from this side, no stepping-stone between Syrian and Latin which would have facilitated the passage of influence from one to the other. In spite of this, however, it must be admitted that at first sight Meyer's theory has a certain plausibility. For while the — 23 - early Christians were hostile to the pagan literature and especi- ally to the pagan poetry, they did stand, through tlioir religion, in close relation to the Syrians. The practice of composing liymns for the use of the churches no doubt originated in the East , and it would go without saying that this with other religious practices would naturally be adopted in the West just as the principles of the faith had been adopted. Yes, we are expressly told that the singing of hymns and psalms introduced by Ambrosius , was done after the Eastern manner, and we have two contemporary wit- nesses who, independently of each other, testify to the fact that it was from this introduction by Ambrosius that church - singing spread over the West^). But instead of being favorable to Meyer's theory, this fact is really against it. For this very Ambrosius who introduced the Eastern manner of church-singing , himself wrote hj^mns for the use of his church, and these , as well as most of the other early hymns, follow the Latin rules of quantity as strictly as we could expect from writers belonging to the age in which they were composed. It is not till a century or more later that rhythmical hymns can be shown to have existed. But these rhythmical hymns have the same measure and rhythm as the earlier , quantitative ones and it would be a bold assertion to make, that when Latin hymns in iambic dimeter already existed in some abundance, those who wrote rhythmical hymns of the same pattern took as their models the Syrian hymns and not the Latin hymns which were already in use in their own churches. If then a Syrian influence is not apparent in the hj-mns of the man who introduced Syrian church-singing into the Latin churches, we surely can not assume Syrian influence in the hymns of the next few centuries, which were modeled upon his so closely and faithfully as to be called by his name, "hymni Ambrosiani." "We may draw two conclusions from this evidence, first that in the introduction of church-singing from the East, no method of versification was adopted which was diff'erent from that which was current in Latin , and second that no melodies or musical 1) Aiigiistiu. confess. IX. 7. "Tuuc hymiii et psalmi ut canerentur secuudura morem orieutalium iiartium, ne populus maeroris taedio contabesceret, iustitutum est; et ex illo in hodieruum retentum, multis jam et paene omnibus gregibus tuis et per cetera orbis imitautibus." Vita S. Ambrosii auctore Paulino eius uotario, cbap. 13. "Hoc in tempore primum antipbonae, hymni ac vigiliae in ccclesia Mediolanensi celebrari coeperunt, cuius celebritatis devotio usque in hodieruum diem uou solum in eadem ecclesia, verum per omues peue occideutis proviucias mauet." — 24 — schemes were introduced of a kind sufficiently pronounced to shape or change the character of the verse. But however much Meyer may disagree with the other theo- ries of the origin of rhythmical verse, he has in common with them a weakness for passing over lightly whatever is inconven- ient in the evidence. For although Latin hymn - writing , as we know it, began with Ambrosius and it is mainly in hymns of the same pattern as his that we see rhythmical verse developed, Meyer refers to this whole part of the subject only in a very general and perfunctory way in a few short chapters at the end of his treatise. His main attention, on the Latin side, is directed to the irregular hexameters of Commodian and Augustine's Psalmus contra partem Donati, even though these are surely not the direct ancestors of the rhythmical verse of the Middle Ages. Still, as it is from an examination of these and a few later specimens of a similar character that he arrives at his conclusion previously mentioned, it is but fair that we should examine briefly the va- rious points at which he assumes an influence of Syrian or Hebrew poetry. These points are: 1. Rhyme; 2. Acrostichs ; 3. Groups of lines and strophes; 4. Equality of the lines in respect to the number of sjdlables (Silbenzahlung) ; 5. Disregard of quantity, and as a result, accentual pronunciation (1. c. p. 113=377). In regard to rhyme he is himself in great doubt; and his own statement will be a sufiicient refutation of his claim. After re- ferring to the fact (p. 379) that the Arabians had a highly devel- oped rhyme about 500 A. D., he says: „Allein die Araber waren damals noch so isolirt, dass nicht daran zu denken ist, von ihnen batten Commodian oder Augustin ihren Tiradenreim gelernt. Bei den Hebraern und bei den Syrern finden sich lleime, doch nur in solcher Ausdehnung, dass man darnach wohl die Reimprosa und den Reim in den Hymnen der Griechen sich erklaren kann. Da- gegen kann der auffallige Tiradenreim bei Commodian und Au- gustin, sowie die ziemlich ausgebreitete Reimprosa der Lateiner aus den geringen Ansatzen der Hebraer und Syrer, welche wir kennen, nicht erklart werden" ^). Acrostichs are found in a few instances in Latin before the time of Commodian, as Meyer admits. The abecedaria are in all probability a product of the Semitic languages. They exist in 1) More recently Meyer seems to have thought of the rhyme in Latin prose as preliminary to the poetic rhyme in point of development, cf. Gott. gel. Anz. 1893, p. 8. For the view that the poetic rhyme sprang from the rhyme of prose, see Nordeu, die Autike Kuustprosa, Appendix. — 26 - Hebrew and Syrian in considerable abundance, and as we gather from a passage in Augustine (Enarratiu in Psalm. 118, sermo 32,8), they were somewhat common in his day in Punic — and in Latin. The passage is interesting also in showing that Augustine could not have had the practice from Hebrew at first hand. It reads : ''Quod autem de alphabeto hcbraeo, ubi octoni versus singulis sub- iacent litteris , atque ita psalmus totus contexitur, nihil dixi , non sit mirum, quoniam nihil quod ad istum proprie pertineret inveni : non enim solus habet has litteras. Illud sane sciant qui hoc in graeca et latina scriptura, quoniam non illic servatum est, invenire non ptissunt, omnes octonos in hebraicis codicibus ab ea quae illis proponitur littera incipere ; sicut nobis ab eis qui illas novcrunt litteras, indicatum est. Quod multo diligentius factum est , quam nostri vel latine vel punice, quos abecedarius vacant psalmos , fa- cere consueverunt. Non enim omnes versus donee claudatur pcri- odus, sed solos primos ab eadem littera incipiunt, quam praeponunt." But even if we grant that the abecedaria in Latin owe their origin to Semitic poetry , they are purely external ornaments easily adopted as an occasional practice into any language , and therefore of very little consequence as showing an outside influence. In the matter of groups and strophes Meyer's contention is that the system was new in hexameter verse (p. 371). But as it had existed in other kinds of Latin verse before that time, and as he does not show that the Semitic languages possess a dactylic hexameter , there is certainly no justification for assuming this outside influence. The fact, however, that in the Carmen Apulo- geticum of Commodian there is almost invariably a full sentence- pause at the end of every second line, is striking. Yet , strange as this grouping of the hexameter is , it is easy to find other Latin examples of it both before and after Commodian. First of all , there are the Disticha Catonis which may not be so old as the poems of Commodian, though it is hard to believe that the collection originated in a single age or with a single man, either in form or in substance. We find the hexameter distich also in a number of the longer tomb-inscriptions, as for instance Biicheler, Carm. Ep. 422, 429, 434, 454 and 512, Several of these are as early as the second century. The last one (CIL 8, 7156) is, more- over, an acrostich, and in its versification can well be placed by the side of the poems of Commodian (cf. Teuff'el under Com.). The other two points at which Meyer assumes the influence of Semitic poetry, namely the equality of lines in respect to the number of syllables (Silbenzahlung), and the disregard of quantity — 26 — witli resulting pronunciation according to accent, are really the fundamental ones, for if his assumption is correct, there would in these cases be a complete transformation of Latin verse both in form and in internal character. The number of syllables in Commodian Meyer finds varying from 13 to 17, and neither he nor any one else thinks that this number is due to any other cause than that Commodian imitated the classical hexameter. With the Psalmus of Augustine , it must be admitted , the case is somewhat diiferent. It is not modeled after any of the commonly employed quantitative metres , and the quantity is al- most completely disregarded, so far as the structure of the verse is concerned. The number of syllables in the half- lines of the Psalmus varies from 7 to 11 according to Meyer, and it is in this respect in open conflict with all the other rhythmical verse in the language from that time on, with the exception of that scanty part, in which the hexameter is imitated. Can it then be assumed that the form of Augustine's verse was a sporadic case which arose under the influence of Syrian church-poetry? If we look about for a time in Augustine's life , when he could have come into contact with and have been influenced by the Syrian manner of church-singing, our attention is naturally fixed upon the latter part of his stay at Milan, and especially the time of his conversion to Christianity. He several times in his works refers to the profound impression which the singing in the church made upon him. One of the passages is Confessiones IX, 6, where he refers to the occasion of his baptism. At the beginning of the next chapter (7) he says that church - singing had been in- troduced at Milan a year before. The date of Augustine's baptism is April 25, 387. Soon after this he departed for Africa which he did not reach till the fall of 388 , on account of the delay caused by the death of his mother. Till 391 he lived in retirement in Africa. Then he became presbyter of the church at Hippo and two years later , in 393 , six years after his baptism , wrote his Psalmus against the Donatists. Augustine was well acquainted with the metrical theory as it was handed down by grammarians. He even wrote six books De Musica in which, of course, he treats metre fully, and in the conventional manner. We know of no influence from a Syrian source which could have in any way disturbed his metrical prin- ciples. For the hymns sung in the church at Milan , which pro- duced such a profound impression upon him , were written with — 27 — strict regard for quantity^). He was ignorant of Hebrew (ci, p. 25) and certainly equally ignorant of Syrian, since that was a lan- guage which was of very much smaller consequence to him. We know however from the passage just referred to that there were psalmi abecedarii in both Punic and Latin. In all proba- bility Augustine was acquainted with Punic, that being very like- ly the language of many of his flock. But of what nature these Punic psalms were , we know only from what he tells us about their acrostichs. As for the Latin psalms, they may quite as well have been quantitative , as rhythmical , indeed the pre- sumption is naturally in favor of the former supposition. But even if these other Latin psalms should have been of the same kind as that of Augustine, they were in all probability mere imi- tations of the latter , for the way in which he himself refers to his own psalm, and the explanation which he finds it necessary to make (Retractationes I, 20 quoted in the note below), lead us to the conclusion that he is departing from the beaten path. Yet even if this were not so, it would hardly be possible for any one in Augustine's place and with his previous training, to neglect the ordinary rules of metre without doing it consciously and pur- posely. We should expect him to explain why he did so un- heard-of a thing as to write poetry contrary to all the rules of versification. And fortunately he does make this very explanation, but he merely excuses his failure to employ a regular metre, by expressing the fear that he would in that way be compelled to use words which would not be understood by the common poeple ^) The P s a 1 m u s is of course, first and last, a polemic, and not a work of art. Above all he wished to reach the masses. And to accomplish this purpose, he did what popular preachers have always done, he dropped the conventional forms. This de- parture allowed him on the one hand to admit such language as was most easily understood, and on the other hand attracted attention 1) That the hymns simg at Milan in 380 — 7, when Angustiue was there, were by Ambrosins himself, is apparent from a passage in his st-rmon against Auxeutius preached during this year: "Hyniuorum quoque mcorum canuinibus deceplura popnlum ferunt. Plane hoc non abnuo. Grande carmen istud est quo nihil potentius." 2) Retractiones I, 20 -'Volens etiam causam Donatistarum ad ipsius humillimi vulgi et omnino imperitorum atque idiotarum notitiam pervenirc et eorum quantum fieri posset per nos inhaerere memoriae, psalmum qui eis cautaretur, per latinas litteras feci. Ideo autem non aliquo carminis genere id fieri volui, ne me necessitas metrica ad aliqua verba quae vulgo minus sunt usitata, com- pelleret." - 28 - by its very novelty, and thus produced more rapid and decisive results, exactly the thing which he would strive for in his conflict with the heretics. This explanation is not only in accord with the facts of the case, but with Augustine's own words. It will be objected that this explanation does not explain the peculiarities of the Psalraus, namely the fact that its lines and half-lines are fairly even in length and that they have accent- ual trochaic endings. As for the former it is useless to go to a foreign language in search of a prototype, when we know that to be sung to the simplest of melodies , as this no doubt was , a poem must have frequent repetition of similar parts. Lewis (and before him Ebert, Allgem. Gesch. d. Lit. d. Mittelalters 1,251) is probably not far from the right track when he proposes to draw a conclusion from the form of the Psalmus as to the na- ture of the music (Foreign Sources of Mod. Engl. Versif. p. 50). It remains however to set right an impression which has so far existed as to the nature of the trochaic cadence at the end of each half-line. This has simply been accepted as being accentual without much further inquiry, Meyer says for instance (p. 288): „Von Reobachtung der Quantitat ist auch im Schlusse keine Rede, hochstens dass die zweisilbigen Schlussworter mit langer vorletzter Silbe bedeutend zahlreicher sind als die mit kurzer." His expla- nation of the cases where the final word apparently has the accent on the antepenult is probably correct. Nearly all of these are words like iudicio, hodie, quotidie, gladium, in which if the "i" be- fore the last syllable be read as a consonant, the accent will not only be on the penult , but the latter will be long by position. But not enough attention has been called to the fact that after this is done, we have only 31 short penults left in a total of 534 endings. The smallness of this number invites a closer examina- tion which shows that the cadence both at the end and in the middle of the line, is rudely quantitative, instead of being merely accentual. Of the 31 cases of a short penult, the short vowel is in 5 cases followed by another vowel, in 10 cases by a sonant, in 13 cases by a liquid and in only 3 cases by a mute. It is clear that if one is not trying to be strict according to the rules in the matter of quantity, a short vowel followed by another vowel, a sonant or a liquid can, especially if it is under the accent, with a little effort be made to appear longer than the rules would have it. A mute, on the other hand, will not permit this extension. If we may say so then, of the 31 cases of a short penult 28 may, if we take no account of the rules of prosody, be treated as — 29 — common. We shall see later that a practice of the same kind was followed in some of the earliest rhythmical hyiims. If it seems strange that Augustine should make such a hold departure from a norm which he himself recognized, it should be recalled that he was an African and no doubt quite accustomed to that hybrid style of composition midway between prose and poetry which we meet so frequently among the inscriptions from that province. If the strong opposition to the establishment of an orthodox system of doctrine had continued longer, perhaps these beginnings of a new kind of verse which we see in Augustine, might have become important for the later development of Christian poetry in Latin. As it was, the growing importance in the West of a central ecclesiastical power and a single doctrine , and the absorption by the church of the upper classes of the pagan popu- lation seem to have been decisive for the future of poetic forms, resulting, as they did, in the adoption of the pagan metres. If a Semitic influence is not to be assumed in the outward structure of the verse in rhythmical poetry , much less can we assume such an influence upon its internal character. Meyer's assumption that, as Semitic poetry pays no attention to the quan- tity of the syllables, and as Latin rhythmical poetry usually does the same, the one was the cause of the other, is a mere assertion without any proof to back it. He is at pains to tell us several times (p. 303, 370, 377) that the Latin verses in question, being constructed without regard to quantity, were read according to the word -accent. He did not stop to consider, apparently, that the change of which he speaks would involve a complete revolution in the pronunciation of Latin. For that is certainly what it means when we say that whereas Latin poetry had previously taken account only of quantity, it began through Semitic influence to neglect the quantity and to be read only according to the accent (Meyer p. 377). Further than that, Meyer's assumption that Semitic poetry changed the character of Latin pronunciation through its influence on Latin poetry, is a flat contradiction of the well-known usage of poets in matters of pronunciation. Poetry does not anticipate changes in pronunciation, but often lags behind when the changes have been accepted elsewhere, and occasionally does not give up an old pronunciation for generations after it has disappeared from ordinary speech ^). 1) Compare the poetic pronunciation of "wind" and of the preterite and participial ending"-ed." — 3U — An illustration from a modern language shows in a convincing manner how untenable Meyer's general proposition is. When the new English poetry was built up after the Norman Conquest, it was done mainly in verse - forms borrowed from French , which, from an English point of view, has no accent to speak of. The habits of life and thought as well as the language of the culti- vated classes in England bad all been French for several centuries before Chaucer. The prospects for an influence of the French manner of treating the accent in poetry were surely as favorable as can be imagined, many times more favorable than the prospects for an influence of Syrian or Hebrew pronunciation upon Latin had been; for in the Latin-speaking part of the Roman empire the Syrian language was at most spoken by a handful of foreign residents, and even Hebrew was spoken by a comparatively small number of poeple who were, moreover, stubbornly opposed to Christianity. And yet when English poetry emerged from its French influences, the various forms of verse which it had adopted were constructed according to the stress-accent no less than the old Germanic measures had been. We must hold that if any Semitic influence was exerted upon Latin poetry, it was slight and purely external and confined to such phases as could easily appear in a translation; in other words that it did not aff'ect the real structure of Latin verse. More recently Meyer has added to his list of proposals another for the explanation of the accentual ending of the line in rhythmic poetry. In a review of Havet's La Prose Metrique de Symmaque (Gott. gel. Anz. 1893) , he states at some length his view that accentual cadences at the ends of clauses and sentences in prose grew out of the previously existing quantitative cadences. According to his view those endings in which the accent and quantity had coincided were continued as rhythmical endings, while the others died out '). He ends by saying (p. 26) : ;,Das ist die Geschichte der rythmischen lateinischen Prosa. In der Zeit, in welcher der quantitierende Satzschluss in den accentuierenden sich verwandelte , also in der zweiten Halfte des vierten Jahr- hunderts, hat nach meinen Aufstellungen die sog. rythmische Dichtung der Lateiner ihren Anfang genommen. Zwei Haupt- bestandtheile derselben habe ich gefunden : in Caesur und Zeilen- schluss wird ein bestimmter Tonfall der Accente beobachtet , vor diesem geregelten Schlusse wurden weder Quantitat noch Accent 1) For a discussion of the cadences in prose, see Appendix I. — 31 - beobaclitet, sondern nur Silben gezalilt. Die obigen Darlegungen zeigen , wie die Lateiner jcner Zeit dazu kamen , die Schliisse in den neuen Dicbtungsformen nacb dem Tonfall dcr Accente zu regeln. Wie sie auf den merkwiirdigen Einfall gekommen sind, vor diesen Schliissen nur Silben zu ziiblen, das babe icb friibcr scbon angedeutet und wcrde ich bald beweisen. Damit wird die Fragc nacb dem Urspi'ung der rytbmiscben lateiniscben Dicbtweise erlcdigt sein." The cbange was in Meyer's opinion due to tbe increased pro- minence of tbe accent. „Quantitat und Accent batten den Kampf, den sie in alien Sprachen ewig fiibren, innerbalb der lateiniscben Spracbe im Verlauf des 3. und 4. Jabrbunderts mit dem Ergeb- nisse ausgefocbten, dass beim Sprecben der Accent viel deutlicber bervortrat als die Quantitat der einzelnen Silben" (p. 19). This is an accentual tbeory pure and simple, and is subject, witb sligbt modifications, to tbe same objections as its fellows. In tbe first place, if tbe accent was prominent enougb to determine tbe character of the ending, it is strange that it should in some of tbe metres have allowed tbe same number, if not a larger number of conflicts to stand in the rest of tbe line than bad stood there before tbe time of this assumed change. If tbe accent really acquired tbe super- iority necessary to accomplish what Meyer supposes it accom- plished in tbe end of tbe verse, it would have made itself felt in the beginning and tbe middle. Further, we have on the one band numerous rhythmical hymns which do not have the accentual close any more than the metrical hymns of tbe same period, and on the other hand there are nu- merous metrical hymns which have a perfect accentual close, even though they are constructed on the principle of quantity. That it is possible to write poetry in which consistent attention is paid to both accent and quantity, is perhaps proved by tbe clever experiments of Tennyson, but whether such a method of versification could ever become an extensive practice, and that in the Dark Ages , is rather more than questionable. But even if we admit the possibility of such a practice becoming general, the small regard in which rhythmical versification was commonly held by metrical poets (pp. 68—9), would hardly allow anyone capable of writing metrical hymns, to choose voluntarily to adopt an element which he felt to be peculiar to rhythmic verse. The tables of conflict between accent and ictus , moreover , lead us to conclude that tbe movement toward the avoidance of conflict at tbe end of the line was native to quantitative verse in Latin, since - 32 — it bad set in long before rhythmical hymns were written. The tendency is indeed as decided before the time of Ambrosius as after it. Conflicts of Accent and Ictus in the Quantitative Iambic Dimeter. number of verses. Nun iber. Perce itage. first 2 foet. last 2 feet first 2 feet. last 2 feet. Classical period (A 1) 292 77 248 24- 76 + Time of Ambrosius (A2) 1014 365 480 43 + 57- Sedulius to Bede (A 3) 740 338 155 69- 31 + Carolingian period (A4) 344 215 75 74 + 26- Perhaps the most objectionable feature of the theory is the mechanical patching up of a verse out of two such widely sepa- rated and uncertain elements as the syllabic principle of the Sem- itic languages for the first part of the verse, and Latin accent- ual pronunciation for the end. Such a method of explanation is hardly admissible as a last resort, but is not to be thought of in the presence of the fact that side by side with the various stages of development in the rhythmical measures there are contemporary stages of development in the metrical form which show the same peculiarities in the relation between accent and ictus in all parts of the verse. Trochaic Popular Verses of the Classical Period. The earliest specimens of popular verse which have been con- sidered in close connection with the later rhythmical poetry, are a small number of trochaic tetrameters and septenarii, mostly found in Suetonius' Lives of the Caesars. They have been adduced as examples of popular poetry par excellence, and much reli- ance has been put upon them by the champions of the accentual theory. These lines do, as a rule, show no conflict between accent and ictus. At the same time, however, the great bulk of them, in fact all but two later and irregular specimens in Vopiscus' Vita Aureliani, show no greater violation of quantity than that of allowing the choice of long or short in all theses except the last. In the arses they have everywhere either a long or two shorts. This shows that they are not rhythmical but follow quan- titative rules. If we leave out of consideration for the present the irregular lines in Vopiscus , these verses may be roughly divided into two classes, w^hich differ from each other in several respects. Class I. The one class of verses avoids hiatus, but also, if we allow an easy synezesis in a few instances, avoids resolution of long syllables, and has almost perfect accord of accent and ictus. The verses are 1. Suet. lul. 49. Caesar Gallias subegit, Nicomedes Caesarem ; Ecce Caesar nunc triumphat qui subegit Galliam, Nicomedes non triumphat qui subegit Caesarem. 2. Suet. lul. 51. Urbani, servate uxores, moechum calvom adducimus. Aurum in Grallia effutuisti, hie sumpsisti mutuom. 3. Suet. lul. 80. Gallos Caesar in triumphum ducit, idem in curiam; Galli bracas deposuerunt, latum clavom sumpserunt. 3 — 34 — 4. Marius Plotius (Keil. GL. VI, 461). Postquam Crassus carbo factus, Carbo crassus factus est. 5. Velleius Paterculus II, 67. De germanis, non de Gallis, duo triumphant consules. 6. Suet. Caligula 6. Salva Roma, salva patria, salvus est Grermanicus. 7. Suet. Galba 6. Disce, miles, militari, Galba est, non (raetulicus. Class II. Different from the verses just mentioned and showing cases of conflict between accent and ictus , as well as numerous resolutions of long syllables, are these : A. Porphyr. ad Horat. Epp. I, 1,62. Rex erit qui recte faciet, qui non faciet non erit. B. Dio Cassius 43, 23 — Teufiel, Gesch. d. rom. Lit. 11,2. Plecteris si recte facies, si non facies rex eris. C. Porphyr. ad Horat. Epp. II, 3, 417. Habeat scabiem quisquis ad me venerit novissimus, D. Schol. ad luvenal. V, 3. Aliud scriptum habet Sarmentus, aliud populus voluerat. Digna dignis ; sic Sarmentus habeat crassas compedes. Rustici, ne nihil agatis, aliquis Sarmentum alliget. E. Suet. lul. 80. Brutus quia reges eiecit, consul primus factus est; Hie quia consules eiecit, rex postremo factus est. To these may be added, though no doubt belonging to an earlier age : F. GeU. Praef. 19. Nil cum fidibus graculost, nihil cum amaracino sui. The same form is further found in the scrawls on the walls of Pompeii. G. Bucheler, Carm. Epigr. 230. CIL IV, 1830 add. p. 217. Futuitur cunnus [piljossus multo melius [qu]am glaber ; E[ad]em continet vaporem et eadem v[ell]it mentulam. H. Biich. C. E. 231. CIL IV, 1939. — fueere quondam Vibii opulentissumi , Non ideo tenuerunt in manu sceptrum pro mutunio Itidem, quod tu factitas cottidie in manu penem tenes. L Biich. C. E. 232. CIL IV, 1234. Pupaque bela is, tibi me misit tuus es(t) : vale. Cf also under Biich. C. E. 233. The question which of these two classes represents the usual popular manner , is not so hard to answer as it may seem — 35 — at first sight. Of the verses of the first class we are told explic- itly in the case of nos. 1, 2 and 5, that they were sung by the soldiers following the triumphal car of their general. Of no. 7 it is said that it spread instantly through the camp (statimque per castra iactatum est) when the new commander gave the first example of his strictness the day after his arrival. The sentiment is much the same as in the case of the songs sung in the triumphal procession and it is, without doubt, like them, a true specimen of the soldier song, though we can not be sure whether it also was sung on the march or not. Of the occasion of the use of no. 4 we know nothing. It is worth noticing, however, that in its sen- timent and the form which the thought takes , it has much in common with one of the soldier songs already spoken of (no. 1). In the case of no. 6 we are told that the pocple of the city sang it in a chorus heard on all sides , as they hurried to the Capitol by night at the report of good news about the health of German- icus. Of the lines under no. 3 we are only told *'et ilia vulgo canebantur." From the passage that immediately precedes, we should judge that the poeple sang these verses to show their dissatisfaction with Caesar. There are two possible views of the origin of these verses. They were either sung by the spectators at the Gallic triumph as an answer to the song of Caesar's sol- diers , or they were composed by other soldiers of Caesar in the triumphs immediately succeeding that over the Gauls ^). In either case, expressing exactly the strong feeling of dissatisfaction on 1) By the Gauls whom Caesar led "in curiam," probably the Transpadane Gauls are meant whom he admitted to citizenship. In that case "Gallos in tri- umphum ducit" might refer to his legions which were raised or recruited in Hither Gaul. Or it might refer to captives from Farther Gaul that graced his triumph. In the latter case no distinction would be made between the two kinds of Gauls, but that was natural enough with the aristocratic Roman vagabond, to whom both alike were barbarians when his own exclusive rights were at stake. The song would then express the feeling aroused among the inhabitants of the city at seeing the prominent part foreigners were beginning to play in the government. There are two circumstances which point to the Gallic triumph as the date: 1. The present tense "ducit in triumphum"; 2. The mere fact of the reference to the triumph. The bestowal of the citizenship and the right to membership in the senate had occurred several years before the triumph, to be sure (note also the perfect tense in the second line), but that was a fact that was ever before the eyes of the populace; that particular triumph, on the other hand, was a matter of mere temporary interest, and its details would very quickly fade from the memory, all the more so as it was directly followed in the same month by three other triumphs, and not so long a time after by a fifth (Suet. lul. 37). If 3* — 36 — the part of the pocple, they would easily become current and remain alive for a considerable time. So far then as we have any knowledge of the occasions on which the verses of the first class mentioned above had their origin , they all agree in this : they were in each case sung in chorus by a large number of poeple in close touch with one another, and whose feelings had in some way been aroused and directed to a common object. In addition to this, of the six examples about which we have any circumstantial knowledge, four were sung by soldiers or citizens in motion toward a common goal, and a fifth (no. 3) was in all probability at least directly called forth by a procession of this kind. Four of the six were certainly sung by soldiers and a fifth may have been. And the one which was perhaps not sung on the march (no. 7), has the same charac- teristics as those which were , and therefore helps to show that there was very likely a peculiar manner of composition em- ployed in soldier songs. We notice the same peculiarity when we compare our own soldier songs with other songs — everything of the kind that originates with soldiers necessarily has a strong marching rhythm, which is nothing more than a relatively greater stress on the syllables that bear the ictus. Is it then strange, that all soldier songs in Latin should have harmony of accent and ictus? The metrical form of the two songs which were not, or may not have been sung by soldiers (no. 3 and 6), is likewise easy to explain. If the first conjecture in the case of no. 3 is correct, these verses are simply a sort of answer or echo to the marching songs of the soldiers, or at any rate were influenced by the occa- sion and caught the rhythm of the march. On the other hand, that the song of a crowd of people moving along in the night with a common feeling and purpose, as is the case of the German- icus song (no. 6), should fall into the marching rhythm of its own accord, would not be strange, if we leave entirely out of account the fact that those most concerned about Grermanicus and most likely to start a demonstration like that described by Sue- we adopt this view, we should owe this song to the "licentia" of the crowd that watched the procession. A second interpretation is, however, possible, that is, that the song originated with the Roman soldiers in some one of the triumphs just spoken of, which close- ly followed the one over the Gauls. Being Roman citizens of long standing themselves, these soldiers would of course have the same prejudice against the admission of outsiders to the citizenship that the Roman populace had. — 37 — i<5piius (Caligula 6) and Tacitus (Annales II, 82), would be precise- ly the old veterans who liad served under him and wore now spending their last days in idleness at the capital. The other point in which the first class of popular verses differs from the second, is that the former have practically no resolution of long syllables. This is only another result of their origin and use. It is inconvenient and confusing for a large number of people to sing or shout anything in unison which does not have a regular alternation of stressed and unstressed syllables. Any one will understand this who has compared the clear-cut, satisfactory rendering of a simple "college yell" of even rhythm with the confusion produced by one which undertakes to convey an idea in every-day English and thus makes necessary an irreg- ular succession of accented and unaccented syllables. If we now examine the verses of the second class, which have conflict of accent and ictus and a considerable number of resolutions, we find the conditions of the first class largely or entirely lacking. E was found written upon the statue of Brutus. This and the Pompeian inscriptions were certainly never shouted or sung in public by a large body of people, if indeed any of them were ever spoken at all. Children's songs like A and C and the restored parody B, naturally' lack the force of feeling and earnestness, the movement toward a common end, and above all the regularity of movement necessary to make their rhythm severely regular or strictly accentual. The naturalness of this difference will be evident to any one that compares the Mother Goose rhj^mcs with the soldier songs in our own language. F is in all likelihood a century or more older than the Caesar songs. It has the manner of the old dramatists — Baehrens thinks it is "fortasse ex Naevii ludis." Gellius calls it "vetus adagium." The only one whose origin has any similarity to that of the first class , is D , Schol. luv. V, 3. It is directed against a single person and the tone of criticism in it is much the same as in the soldier songs. On the other hand, the circumstances were entirely difi'erent. In the case under discussion the speakers were no doubt seated or standing — "per ludos quibus primum (in) XIIII ordinibus sedit, haec a populo in eum dicta sunt" — and not, as in the case of the other songs, moving with a common purpose to a common end. From the circumstances it is quite certain that the verses did not originate with soldiers. They smack rather of the sentiments of the petted and idle rabble of the city. Then, it is not probable that any large number of - 38 - people who were well acquainted, or even knew each other by sight, would be within easy speaking or hearing distance of each other at the games. And finally, these verses were spoken (dicta), not sung, as the others were. We are not, therefore, surprised that they show cases both of conflict and of resolution. Conclusions. — The conclusions we come to, then, are these. The popular poetry of classical times was quantitative, that is, the arses were long. The harmony between accent and ictus was not a characteristic of popular poetry, but of a single branch of it, namely of the soldiers' songs and of others arising under like conditions, and was the result of peculiar circumstances. The same circumstances seem to have operated to prevent the resolu- tion of the long syllables. The usual form of the popular trochaic tetrameter appears to have been something like the old comic form of this verse. Corroborative evidence of the latter conclusion is furnished in abundance by the later iambic and trochaic tomb- inscriptions , the shorter and more popular forms of which are invariably composed in this loose form (cf. Biich. C. E. nos. 234 — 247 and 117—211)0. 1) The two late examples of popular songs (Vopiscus, Vita Aureliani, 6), wbich are usually considered in this connection, need liberal padding before they can pose as trochaic tetrameters or septenarii even of a degenerate kind. If we accept the necessary additions usually made, they show externally about the same characteristics as the earlier soldier songs except that one of them has two short arses and the other one has several resolutions. It is to be borne in mind, how- ever, that they are not really genuine soldier songs, though they have something of the old military form. They are really "ballistia et saltatiunculae" and re- present a class of their own, which springs from circumstances similar to those of the soldier songs in so far only as the song is accompanied by bodily move- ment. Aside from the mere fact of the movement, however, they have in common with the soldier songs neither the regularity nor the movement toward a common end of a multitude actuated by strong feeling. The short arses can be duplicated from the inscriptions of the period , and will not cause us to hesitate when we consider that the verses are somewhat later than Commodiau. The Inscriptions and Conimotlian. As is well known, iambic and trochaic metres arc represented in Latin by freer and stricter imitations of Greek models , the former in the republican drama and Phaedrus, the latter in all the other poetry from the time of Catullus on. The iambic tri- meter (senarius) as well as the catalectic trochaic tetrameter (septenarius) exists abundantly in both the strict and the free form. Of the iambic dimeter , however , there is barely a trace here and there in the drama. This metre was first employed to any considerable extent, so far as we know, by Laevius, who pro- bably belongs to the generation preceding Cicero. The score of iambic dimeters found among his fragments (Baehrens, FPR), show no affinity in their structure to the verse of the dramatists. They regularly avoid a long syllable in the second thesis of the dipody, and very rarely admit a resolution, that is, they follow only the strict form. And the same statement is true of all the subsequent uses of this metre. The treatment of these three kinds of verse in the inscrip- tions corresponds closely to that in literature. For it goes without saying that the imitation of a metre means its imitation as a whole , and not the formation of a new metre , a dimeter or a trimeter for instance , out of imitated elements , or dipodies, either free or strict. So we find that the inscriptional dimeters, like their relatives and prototypes in literature, show very rare cases of resolution and equally rare cases of the employment of a long syllable in the second thesis of the dipody. In 37 dimeters belonging to a relatively early period (Bitch. C. E. 217 — 221), there are only five cases in which the second thesis is long. Of the two later exam- ples (Biich. C. E. 222 and 223) , the former shows four long syl- lables in this location in its eight verses, and the latter four short syllables under the ictus in five verses. But even these late specimens do not show any frequency of resolution. Number 223, in fact, which is purely rhythmical, has none at all. - 40 - Of the 180 or more inscriptions in iambic trimeter, the vast majority show no regard whatever for short theses, except in the penultimate syllable of the verse, and nearly all have more or less resolution. Generally speaking, there are three classes of iambic trimeters (senarii): 1. Those following the early dramatic style, with an "anceps" in every thesis except the last, and with full sway of resolution; 2. Those which show regard for strict rule in both arses and theses, and very seldom allow a resolution; 3. A development, or rather, corruption of the second class, with infrequent resolution, but waning regard for quantity, especially in thesis. In general it is the larger and more pretentious in- scriptions like Biicheler 19, 20, 104 and 111, that show very few or no violations of the strict rules of quantity and at the same time avoid resolutions. And it is a noteworthy fact that nearly all of the popular tomb-inscriptions of a line or two each (Biicheler 117 — 211), belong to the first of the three classes mentioned. When taken together these two observations indicate that the popular conception of an iambic verse was the senarius as it was handed down by comedy, while the use of the newer and stricter form remained more or less confined to the cultivated classes. Of the inscriptions in trochaic tetrameter (septenarii) , those from Pompeii have already been mentioned in another connection (pp. 34 & 37). They have the loose comic form. The same is also true, as in the case of the iambic trimeter, of the shorter trochaic tomb-inscriptions. Correct examples according to the strict form, with few or no resolutions, are furnished by three inscriptions of a religious or semi-religious character (Biicheler 227, 228, 229). The thing that interests us for the present, however, is not so much the fact that certain classes of the people used certain kinds of metre, for the early comic form can not possibly have been the model of the rhythmical church poetry, nor can it even have had any appreciable influence upon it. The strong tendency of church poetry to follow the models already existing in its field, is well known. Poets in other fields may originate or introduce new poetical forms with much greater freedom, and yet we know how rarely even this has come about, especially in Latin litera- ture. What shall we say then of religious poets who are strictly confined in form as well as in thought? The actual practice of the early Latin hymnists in this par- ticular is exceedingly clear and convincing. Ambrosius wrote the first metrical hymns which we certainly know to have existed, and which were extensively used in the church. He employed __ 41 - the iambic dimeter of strict form. It is snrely not a mere coin- cidence that the vast majority of the hymns for three or four centuries were written in this metre. On the other hand the iambic trimeter was very little used in the hymns, and then only in its strict form, though the inscriptions show it to have been quite frequently employed by the people. The trochaic tetrameter, as far as the inscriptions show , was much less of a favorite. Still it was quite extensively used in the hymns. Besides this, the form of all these metres as we find them in the early hymns, shows that they have nothing in common with their popular form in the inscriptions. In the first place , the iambic and trochaic verses of these hymns are strict in their treatment of the theses, leaving but one in each dipody undeter- mined. In the second place, the most distinctive feature of the popular form of the iambic and trochaic verses, which they firmly retained as we have seen, was their wide liberty of resolution. In the strict form, the iambic dimeter together with the other metres examined, admits of resolution, though very sparingly. But as soon as this measure was employed in hymns designed for the use of the church, we find even this restricted liberty aban- doned. Ausonius, for instance, has 54 resolutions in 398 dimeters and Prudentius has 23 in 400, but Ambrosius has but two in a total of 188 lines, both occurring in the same hymn, "veni re- demptor gentium," and later hymns as a rule avoided them as much as possible (cf. Appendix II). The same is true of the trochaic tetrameter, with the difi'erence that the change is less marked, because the strict form of this metre never admitted re- solution even to the same extent as the iambic metres. All this is just the opposite of what we should expect, if we were looking for the influence of the current popular form of iambic and trochaic verse upon the hymns of the church. Though we can not , therefore , get much direct information from the inscriptions , of the course and cause of the changes which these hymns underwent , we may still feel confident that the sporadic and individual departures from quantity which the inscriptions show , will throw much light on the nature of those changes, containing as they do, in a great variety of compositions, the germs of the later development. In making the tables that are to follow, the plan has been to include every thing that could be scanned^), omitting only 1) It will be necessary to say a word about the composition of the tables. The task is made difficult in the case of many inscriptions on account of the changes - 42 — those Inscriptions , or lines , or parts of lines where the author either had no definite conception of the metre employed or made no effort to maintain its structure. Such cases it would.be man- ifestly unfair to include , since they would show nothing as to the handling of the metre. With these exceptions^), the tables of unmetrical scansion from the inscriptions include all the iambic trimeters (senarii), iambic dimeters and trochaic tetrameters (sep- tenarii) found in Biicheler, Carmina Epigraphica; further, all the that were very often introduced between the time when the verses were originally composed, and the time when they were cut into the stone. These irregularities are, of course, especially frequent in the inscriptions on tomb-stones, and are of several kinds. First, they may be due to the workman who carved the stone. In such cases they are usually mistakes in spelling, transposition of words or their trans- fer to another line, the use of one word tor another of similar meaning, &c. Or the mistakes may be due to the person who had the carving done. This class of cases is very prolific in troublesome difficulties. Confusion is usually caused by adapting some lines that originally belonged to another tomb or were taken from some other source , and changing them not only to suit the name of the person on whose tomb they were to stand, but frequently also to fit the different circumstances. Being done by unskillful people who had little command of lan- guage and still less control over the metre, the result is what one might expect. There are all grades of changes and it is a fortunate case when a simple addi- tion was made in the middle or at the end of the verse. Frequently a person seems to have had put upon the stone something he recalled inaccurately from memory. Or a person with some knowledge, appa- rently, of metrical rules, will in his alterations avoid faults in quantity, but will change the length of the line, sometimes making the he.xameter, for instance, five, but more often seven, and occasionally even eight feet long. While most of the difficulties enco untered arise from one or the other of the sources mentioned, instances are not wanting where persons of the kind just mentioned will attempt an original composition, to classify which often taxes the imagination severely. 1) The following is a list of the passages which have been excluded : Inscriptions in iambic trimeter (senarius) — Biicheler C. E. G8, 72*, 74, Tl^-^-^, 80*, 85S 86S 92»2, 95, lOl'', 102i, 107S 108^ llO^-*, 120S 125^ 134^, 141', 158»-8, 172*, 180^ 186, 190^ 191, 206. Inscriptions in hexameter — I century A. D. — Bucheler C. E. 375", SGTS II century A.D. — Biicheler C. E. 4743-*, 476' ^ 479. III century A.D. — Biicheler C. E. 496=, 497'-3, 508, 51P•»•^ 512^ 513'-", 516*«, 5273*, 5414.10^ 5491^ 5531^ 5551^ 5551, 557, 558, 560A». 560C^ 565«, bdV-^*, 5686. IV century A.D. — Biicheler C. E. 575», 579', 593*-3, 594, 596'*«', 5979, 598», 604, 607«, 608», 613'-»-««9, 621, 622", 623^ 624, 626, 627, 629'-8, 641», 642S 651, 653', 658', 661», 673'". — 43 — hexameter inscriptions of the first four centuries of our era accord- ing to Bilcheler's arrangement, as follows : First century 351—421, 349 tf. (Pompeian), 806. Second century 249—50, 270—2, 422—86. Third century 251-60, 273-7, 343—4, 487—569, 656, 731. Fourth century 261—8, 278—87, 301—10, 325-30, 570-655, 657—79. In the case of Commodian, the first 275 lines of the Carmen Apologeticum. The unmetrical syllables have been divided into five classes according to the place in the word where they occur: 1. Final; 2. Directly before the accent; 3. Under the accent; 4. Directly after the accent (except final syllables); 5. The remainder, being syllables more than one place before the accent. I. The first table includes all metrical faults whether (A) a short for a long in arsis , or (B) a long (or two shorts) for one short in thesis , or (C) a short for a long in thesis (this in the hexameter only). I II TTT TV Total Com- troch. Total A. Short for long cent. cent. cent. cent. hex. mod- la. trim. la. dim. te- ia. & in arsis. hex. hex. hex. hex. inscr. ian. trara. tro. Final 9 44 73 48 174 136 21 4 25 Before Accent . . 4 8 12 24 18 8 H Under „ . . 2 6 19 27 54 20 17 1 1 19 After „ . . 1' 2 1 5 9 5 3 1 4 Remainder . . . 1 2 5 7 6 1 1 Long for short in thesis. C. Short for long in thesis. Final 11 56 154 82 303 243 306 7 22 335 Before Accent . . 6 19 20 45 93 104 2 6 112 Under „ . . 2 16 31 22 71 114 49 1 7 57 After „ . . 1 Remainder . . . 2 8 4 141 30 7 7 Final 5 9 17 31 27 Before Accent . . 5 10 21 36 14 Under „ . . 2 3 10 10 25 11 After „ . . 4 4 Remainder . . . 2 21 1 — 44 — II. In the second table the shorts for longs in arsis and the longs for shorts in thesis are in each case placed side by side for comparison. The shorts for longs in thesis in the hexameter are thus omitted, as they do not admit of comparison with anything in the iambic and trochaic metres. short arses. long theses. s. a. l.t. 3. a l.t. s. a. l.t. 1 1 s. a. l.t. I. cent . hex. II. cent, hex. III. cent, hex. IV. cent. , Total hex. he^;Fonr centuries. Final ...... 9 11 44 56 73 154 48 82 74 303 Before Accent . . . 4 6 8 19 12 20 124 45 Under „ ... 2 2 6 16 19 31 27 1 22 54 71 After „ ... 1 2 1 5! 9 Remainder .... 2 2 8 5' 4 7 14 Comi nod. la. trim. la. dim. Troch. tetr. Total ia. & troch. Final 136 243 21306 4 7 22 25 335 Before Accent . . . 18 93 8 104 2 6 8 112 Under „ ... 20 114 17 49 1 1 1 7 19 57 After „ ... 5 3 1 1 4 1 Remainder .... 6 30 1 7 1 7 Totals Short arses Long theses Inscript. hexa^meters 268 433 Commodian 185 480 la. & troch. inscr. 57 512 III. In the third table the total metrical faults , in both arsis and thesis, of the second table, are given according to posi- tion in the word and kind of metre, as well as the percentage of each class to the whole in each kind of metre. Short arses long theses. Ia. & troch. Hex. inscriptions. number Final 360 Before accent Under „ After „ Remainder . 120 76 5 8 7o 63 + 21 + 13 + 1- 1 + number 477 69 125 9 23 7« 68 + 10- 18- 1 + 3 + 569 100 701 100 Commodian. number "/o 379 57- 111 17- 134 20 + 5 1- 36 5 + 665 100" It will be noticed that while the two kinds of inscriptions and Commodian do not differ very materially from each other — 45 - with regard to the proportion of total faults in a particular word- position (Table III), the distribution between arses and theses in the case of the iambic and trochaic verses is different from that in the hexameter. In the iambic and trochaic inscriptions only 57 of 569 faults are in the arsis , while in the hexameters of the inscriptions, 268 of 701 , and in Commodian 185 of 665 are found there. This is clearly due to the fact that so many of the iambic and trochaic inscriptions followed the comic style , in which all the theses except the last are undetermined. Further, the iambic trimeter has a proportionally much smaller number of short finals in arsis when compared with the total number of unmctrical arses, namely 21 out of 50, than has the hexameter, 174 out of 268 in the inscriptions and 136 out of 185 in Commodian. This state of things is no doubt due to the position of the caesura, which makes the rhythm of the greater part of the iambic trimeter trochaic. In fact the part of the trimeter after the caesura is in all respects equivalent to the se- cond half of the trochaic tetrameter catalectic. This method of construction of course tends to bring a larger number of final syllables into the thesis. The trochaic tetrameter inscriptions, where the rhythm is completely trochaic, have all their unmetrical final syllables in thesis. In the iambic dimeter inscriptions, on the other hand, where there is no trochaic caesura, four of the five short arses are final syllables. In the hexameter , however, the prevalent masculine caesura brings the final syllable under the ictus, which in most cases practically fixes the rhythm also of the foot preceding and following. This accounts partially no doubt for the very large proportion of short final syllables under the ictus in the hexameters. Compared with the number of short final syllables in the arsis, the number of short accented syllables found there is very small, except in the case of the trimeter (and tetrameter) where the position of the caesura, as just pointed out, works the other way. The figures are : Inscript. hex : 54 accented 174 final 268 total. Commodian: 20 „ 136 „ 185 „ Trimeter: 17 „ 21 „ 50 „ But if we take the proportion of short accented syllables in arsis to the total number of faults in accented syllables, the trimeter stands midway between Commodian and the hexameter inscriptions : — 46 — Hex. insci'ipt. 54 short acctd. arses to 125 total acctd, faults. Commodian 20 „ „ „ „ 134 „ „ „ la. trim, inscr. 17 „ „ „ ;, 66 „ „ „ These figures themselves almost exclude the hypothesis of accentual influence in the inscriptions or Commodian. However, there is still another method by which a possible influence of the accent may be tested, that is , to compare the number of short accented syllables in arsis with the number of short syllables immediately before or after the accent, found in that position. For if there is an accentual influence , these latter syllables are least likely to be recognized by the ictus. The proportion is : Inscr. hex. 54 accented to 33 non - accentual ; Commodian 20 „ „ 23 „ „ Trim, inscr. 17 „ „ 11 „ „ These figures show the number of accented shorts to be in general somewhat larger, but they do not tell the whole story, for the syllables after the accent in words of two syllables , are the finals , and of course these can not be separated from finals in general , and are included under that heading. Words of two syllables, furthermore, have no syllable before the accent. They are therefore entirely debarred from contributing to the showing against the words whose short accented syllable falls under the ictus. If we allow for this, we have the fact that the proportion of short accented syllables under the ictus is no larger, and if anything even smaller, than that of the syllables with least stress, the shorts immediately preceding or following them. There was, however, a slight gain from one century to the other in favor of the short accented syllable under the ictus. The figures for the hexameter inscriptions by centuries are: I cent. — 2; II— 6; 111—19; IV— 27. The gain is one in pro- portion as well as in absolute number : I cent. 2 acctd. out of a total of 12 short syllables under the ictus. -Lln"n » ;j5o„ „ nnn HI „ 19 „ „ . 103 „ „ . r, . IV „27„ „ „97„ „ „„„ But this gain was practically all made under the first and fifth ictuses , the latter of which always , and the former nearly always stood over accented syllables, in perfectly metrical verse as well. — 47 — rr, , 1 , , T 1 ii • , I cent. II cent. Ill cent. IVcent. Total short syls. under the ictus ^^ rp inQ 07 JShort accen ted 1. ictus 1 1 4 7) r 2. n 1 1 2 3 » n 3. » 1 1 » J) 4. n 1 4 8 » n 5. » 2 5 13 « w 6. n 1 3 2 These figures are again not to be taken absolutely , but in proportion to the total short syllables under the ictus in their respective centuries. The extent of the accentual influence seems then to have been confined to this , that it became slightly easier , as time went on, to admit a short accented syllable into the arsis at such points in the verse as already in their pure metrical form had a coin- cidence of verse - ictus and word - accent. The traditional accumu- lated stress at these points seems to have somewhat obscured the quantitative character of the syllables which came to stand there. If we turn to the metrical faults in thesis , we find the final syllables everywhere by far the most fully represented : Inscrip, hex. 303 of a total of 433 Commodian 243 „ „ 480 la. & troch. inscr. 335 „ „ 512 Generally speaking and taking into account the peculiarities of the various metres, the unmetrical long syllable with the accent is found in the thesis in about the same proportional frequency as the unmetrical short syllable with the accent is found in the arsis : Inscr. hex. 71 out of 433 Commodian 114 „ „ 480 la. & tro. inscr. 57 „ „ 512 The accented syllables seem therefore to take without much effort or refusal the share of blunders that fall to their lot, whether in arsis or in thesis. The syllable before the accent is rather better represented in the faults of the thesis than of the arsis, and this is especially true of the iambic and trochaic verses : Inscr. hex. — 45 out of total of 433 unmetr. long theses. Commodian — 93 „ „ 480 „ „ „ la. & tro. inscr. — 112 „ „ 512 n n n In the hexameters this class of faults is not quite so nume- — 48 — rous as the unmetrical accented syllables in thesis, but in the iambic and trochaic verses about twice as numerous. In the theses of the spondaic feet of the hexameter, the faulty syllables preceding the accent are more numerous than the faulty accented syllables : Inscr. hex. — 36 to 25, Commodian — 14 „ 11. The comparison between syllables preceding the accent and accented syllables is, however not altogether fair, as all words of two syllables can contribute only to the showing of the ac- cented syllables. But the fact that the syllable before the accent holds its own in spite of this , shows a considerable preference for it as a faulty thesis. This preference increased both absolutely and relatively from century to century, as appears in the hexa- meter inscriptions : Longs for shorts in thesis — I cent. n „ III „ IV „ Shorts for longs in thesis — I » n , III „ IV „ ., ., The increase occurs mainly before the fifth and sixth ictus, where the accented syllable before which it stands is regularly under the ictus. In these two positions are found : Of the longs for short II cent. 2 out of 6 III „ IV „ Of the shorts for longs II „ III , IV „ In Commodian the proportion is 80 out of 93, 14 for the two kinds of theses respectively. The fact that the quantity of the syllable before the accent, whether long or short, was more easily neglected when the follow- ing, accented syllable was under the ictus, accounts for the great number of faults of this class in the iambic and trochaic verses, where of the two arses before and after a short thesis, one or the other and often both almost necessarily have the accent. This instability of the syllable preceding the accent needs little explanation. It is shown by abundant examples in early out of 13 6 „ . 80 19 , „ 204 20 „ « 128 . . 2 5 „ . 13 10 „ „ 35 21 „ „ 48. 9 W J5 19 15 „ „ 20 1 „ . 5 8 „ „ 10 16 „ „ 21. at of 93, and 6 out of '^\ B « A ;? y* or THR — 49 — V ^ Latin and is, in fact, a universal tendency of pronunciatTon, wnen- ever the fullowing syllaLle is raised to greater prominence , as is dune here when in addition to being accented , it stands under the ictus and coincides with "the pulse-beat of the verse." "Through its enhanced prominence this syllable anticipates the attention which would otherwise be given to the preceding syllable, and the pronunciation of the latter is consequently slighted. The final syllable, which has already been touched upon sev- eral times , is by far the most important part of the word in the investigation of the period before us. In it occur over 63% of all faults in the iambic and trochaic verses, over 68% in the inscriptional hexameters and nearly hl^jo in Commodian. Not only is this large proportion of faults found in the final syllable, but the final is also the first syllable in point of time which fur- nished a considerable number of faults, its great prominence as a faulty syllable beginning as early as the second century. The faulty final syllables in the inscriptional hexameters are as follows : I. cent. — short arses 9, long theses 11. II. „ — „ „ 44, „ „ 56. HI. „ - . „ 73, „ „ 154. IV. „ - „ „ 48, „ _ , _ 82. Of the total number of metrical faults (including the spondaic feet) the proportion of faults found in final syllables is : I. cent. — 20 out of 25. 11. „ - 105 „ , 149. III. „ — 236 „ „ 360. IV. „ - 147 „ „ 253. Commodian — 406 „ „ 719. In proportion to the total number of faults , as these figures show, the number of faults found in final syllables does not in- crease , nor does it decrease in any great degree , but maintains nearly the same level during the four centuries. To be sure, the final syllable has , in the nature of the case, certain advantages over the others as far as number is concerned. Eveiy word has a final syllable though it may not be represented in the other classes spoken of, such as the syllable under or before the accent. We should therefore expect rather more faults in the final than in any other syllable. Monosyllables properly belong here and have been included, since they present the same general difficulties of adjustment as final syllables. But even if we take these things into account, the result is still very striking. Really, however, the important thing for the present 4 - 50 — purpose is that the final syllable is actually to be credited with something like two thirds of all the faults. Whether it got them by numerical superiority or otherwise, is fur the present, a matter of indifference. Still it may not be out of place to mention sev- eral causes that may help to account for the faulty scansion of a final syllable, one grammatical, another metrical, a third lying in the nature of the syllable in respect to pronunciation. In the first place the modification of words due to declension and conjugation, is in the majority of cases confined to a change of the final syllable. The difficulty in using a highly inflected language consists largel}^ in the fact that this modification often appears slight, and out of all proportion to its importance. The outward expression of this condition of things is seen in the fact that the languages mentioned tend to become analytic as soon as an epoch in their history arrives which places them in the hands of people lacking the ability to appreciate and preserve the fine and accurate distinctions which the mechanism of these languages has come to express. This experience the Latin language was be- ginning to have after the Silver Age, and was already having in fact, so far as many individual composers of inscriptions were concerned. What was more natural than that the distinction between ''aqua" and 'aqua," for instance, should fade and that the ablative idea should be expressed by a prepositional phrase. By the same process the distinction between forms like "reddis" and "reddit," "laudat" and "laudant" would cease to be vital, even if it was not lost altogether , and the personal pronouns would be needed for difi'erentiation. In fact the use made of these analytic tools in late Latin shows that the synthetic mechanism had already given out. In the second place there may be a metrical cause. The final syllable is hardest to provide for in scansion, not only because its metrical character changes in the inflections , but also because in the majority of cases the writer has to take into account the beginning of the next word. In the body of a word a vowel al- ways has the same companion consonants and its metrical status is fixed, but at the end this status may depend on the initial letter of a word which may stand in another clause , or even in a distinct sentence. That this constant necessity for looking ahead must have been the cause ol much anxiety to the occasional writer of verses who slowly groped his way from word to word, can not be doubted. The evidence of it remains to us in such scan- sions as — 51 — desine iam flerc poenam non sentio mortis poena fuit vita requies mihi morte parata est. Buchelcr 507. Germaniae meruit speculator et cunicularius, legionis initium vitis vitac imt finis. Bitch. 522. iste quater denis et quattuor mensibus annis. Bitch. 528. Then, in the third place, a final syllable except when it ends in a mute, is easily drawn out long or cut short by the exigencies of the sentence. Any one who will watch the fate of some English word, as for instance "who" or "and," in the course of a conversation , will realize the possibilities for the corruption of quantitative verse that lie in mere elocution. Again, this fault is of course much less likely to occur in the body of a word, where the vowels have their ever present limits, and it is, natu- rally, most easily made in the case of a final vowel. Summary. — The main results obtained from the study of the inscriptions of the first four centuries and the hexameters of Commodian, are briefly the following : I. From one half to three fourths of the faults occur in the final syllable. The great prominence of the syllable in this respect dates from the second century and is maintained through the following centuries. II. The accent does not seem to have been an influence of an}' moment in the introduction of faults. Only in cases where the structure of metrical verse made accent and ictus coincide, is there a gradual increase of short accented syllables under the ictus, and this toward the end of the period under consideration. III. The instability of the syllable preceding the accent in- creased toward the end of the period, especially in the theses preceding those arses in which the structure of the verse makes accent and ictus coincide. The Transformation of Metrical Forms. From the discussion of the various theories of the origin of rhythmical verse in Latin, several facts about the change have already become plain to us. In the first place, coincidence be- tween accent and ictus is a characteristic of metrical as well as rhythmical verse, and came about independently of the neglect of quantity. The disregard of quantity cannot have been simply due to a general lack of sensitiveness to distinctions of quantity, for certain parts of the verse were much more subject to this neglect than others, while still others were altogether free from it. Further, coincidence between accent and ictus is found to have arisen in some classes of verse, while it did not arise in others, and in still a third class of metres , came about only in part of the verse. This phenomenon can not be explained by an}^ of the theories which have been advanced, because the influences to which they respectively attribute the coincidence of accent and ictus, would operate upon all kinds of metre alike, and this could surely not result in the wide diversity just mentioned. It will be ad- mitted that the conditions which made possible this variety of results, must have been of such a character that they could affect one class of verse, while they could not affect a second, and but partially affect a third. What were the conditions which could make for such a change? I. The Undetermined Thesis. As a result of our study of the inscriptions we found that in iambic and trochaic verses two parts of the word were especi- ally liable to cause metrical faults — the final syllable , which was responsible for about two thirds of all the faults in the in- — 53 — scriptions , and , later and to a less extent , and mostly in the theses, the syllable preceding the accent. These phenomena were so wide-spread, that we must assume that the tendency they represent was gradually gaining control of the pronunciation of all classes of the people. By the fourth century, as one can easily convince himself, there are very few met- rical inscriptions that are altogether free from these faults. Even the liexameter of literature began to 3'ield here and there to the instability of the syllables just mentioned [cf. Carmen contra Paganos, (Baehrens P L M, III, p. 287 fF), of ca. 394 or 895 A.D. — 11. G. 11, 26, 31, 35, 41, 44, 45, 46, 50, 52, 73, 110, 111, 117; also, of probably a little later date, Aegritudo Perdiccae (Baehr. PLM, V, p. 112). 11. 7, 15, 31, 58, 93, 107, 125, 131, 201, 241 ; Orestis tragoedia (ibid. p. 218) 11. 82, 83, &c]. To be sure, with most reputable poets this tendency had not yet, at the time of the earliest hymns , gained control to the extent of forcing itself into their writing in contravention of the established met- rical rules. But that it was present and made itself felt even with them, is clearly seen from certain changes that took place in the structure of the iambic and trochaic verses. According to the strict form of these verses every second thesis , or one in each dipody , might be either long or short. That this was one of the privileges of these measures and con- tributed greatly to the poet's freedom of expression , is evident. These undetermined theses furnished so maTiy breathing -spells to the verse in its arduous metrical task. But to the conscientious poet of the fourth and following centuries they also became a deliv- erer in the time of need. For when there was doubt about the quantity of a syllable, it might in most cases easily be made to fall in the undetermined thesis, and would in fact naturally grav- itate in that direction. And thus it was that the unstable final syllables and the unstable syllables preceding the accent gradually established a practically exclusive claim to the undetermined theses. The extent to which this occurred in the history of the iambic dimeter, will appear from the following table (for a detailed table of the individual hymns see Appendix 11, table E): 54 - Iambic Dimeter: Syllables in the undetermined thesis. Number. before | ^^^] \ Qti^g^g accent. i Perceutage. before accent. final. I others. Metrical: A 1) Classical period A 2) Time of Ambrosias A3) Sedulius to Bede A 4) Carolingianperiod B) Of uncertain date 20 166 116 7- 55- 103 675 244 10 + 66 + 101 579 60 14- 79 + .48 264 32 14- 77- 295 1330 134 17- 76- 38 + 24- 8 + 9 + C) With unmetrical theses (early) 33 165 6 16 + 81- 195 396 710 694 71 6 20- 35 + 73- 63 + 3- Rhythmical: D) Early E) Late 7 + 1- Several things are clear from this table. In the earlier met- rical period , to the time of Sedulius , it was especially the final syllable which made for the undetermined thesis. This is in accord with what we found true in the inscriptions, where this syllable was by far the most unstable. There was of course a limit, in a short verse like the dimeter, beyond which this shifting of the final syllable could not go without causing a neglect of quantity. Tlie highest point is reached in those hymns which were making their last frantic efforts to remain quantitative, hav- ing already given up the theses. It is further apparent that the tendency taken as a whole had practically reached its possible limit in the metrical form by the time of Sedulius. Beyond the stage reached at that time , the later metrical and early rhyth- mical hymns did not go. The end had indeed almost been reached, when the proportion of lines whose undetermined thesis was not filled by an unstable syllable had decreased from 38% to 8^0. In the trochaic tetrameter the decrease was still more marked, as the following table will show: — 55 — lines. First Half. Sooond III ilf. •^efo'-fl final, accent. | others. before accent. final. others. Metrical : A 1) Before Pruden- tius A 2) Prudcntius to Fui-timatus A 3) IJnoertain date 286 308 216 57 68 19 132 136 157 131 104 40 28 69 31 193 200 182 99 39 3 B) With unmetri- cal theses 169 10 138 21 29 136 3 Rhythmical: C) Undeveloped D) Developed 124 504 2 26 119 326 3 50 28 191 96 310 1 In the second lialf of this metre Ennodins and Fortunatus have already reached the point vv^here only unstable syllables are found in the undetermined thesis. This continued to be the uni- versal practice of writers both metrical and rhythmical. In the first half the final sjdlable in the course of time all but usurped an undisputed claim to the undetermined thesis at the expense of all other syllables. The result of this shifting in the iambic and trochaic verses was that new caesuras arose, little by little, before and especially after the undetermined theses. When the final syllable stands in the thesis, tlie iambic dimeter would have the form m_w — i^||_u'^, and the trocliaic tetrameter would be cut into its four dipodies, _ u — ^ II — u — M II — w _ iri II _ w ^. If the syllable preceding the accent in words of either three or four syllables stands in the undetermined thesis, the lines would have the form i=^_u_||ii_u.ii and — u— II- — u— -II _u— 11^ — w^ respectively. Words of five or more syllables at the end of the verse , may be left out of account for the present because of their rare occurrence. It is necessary only to notice that they would prevent conflict of accent and ictus at the end of either verse. Let us now examine the possible types in detail. If the di- meter takes the form i^_u— ^||— u^, harmonj'- of accent and ictus will of necessity exist just before the caesura. The begin- ning of the verse will be - | - u or :^^u or - — j , and only in the last of the three cases could there be conflict of accent and ictus. After the caesura we may have ^^^ or iu j h ov —\ ^ ^, - 56 - with conflict only in the last case of the three. If the dimeter takes the form y- _ u — || V- A o ^ , the first half may be either ^ \ ±u — or ^Au|— with harmony , or 5 _ | u _ with a double conflict of accent and ictus. In the second half, in this case, there can be no conflict , since only — — u|^ or ^Auii would be possible. There could be conflict at the end of the verse then in only one possible case out of five and at the beginning in only two possible cases out of six. Conflict depends in the first foot on a word of the type — — , in the second on u — preceded by -^ — or ii I — |, and in the last foot on ^^ preceded by a monosyllable long by nature or position. Of these there can be no doubt which would occur most frequently. The only place where a spondaic word could occur would be in the first foot. An iambic word could occur unconditionally in the first and conditionally in the second or last, where its admission would depend on what preceded. A pyrrhic word could occur on one condition in the first and last. It would have to be followed by a position-making word in the first, and preceded by a monosyllable long by nature or position in the last. In the second it could occur on two conditions , viz. when a following word made position and before it stood the dissylable ^ — or two monosyllables ^ ' — \. In the first foot we have then one conditioned and two un- conditioned chances for a conflict, and in the second and the last, two conditioned chances each. Of these conditions, furthermore, the hardest to fulfill is that required of the last foot, on account of the comparatively greater scarcity of suitable monosyllables, when compared with the frequency of ^— or -^ | — |, or following words that make position (in case of the pyrrhic) , which are re- quired in the first and second feet. Further than that, in the iam- bic dimeter of the by far most widely represented type ^ — u _ ^ [| — u ^, in which alone , of the types considered , a conflict could occur at the end, there is no good place elsewhere — since a mono- syllable does not as a rule stand before the caesura — for words of the type —^^. These would of course naturally go to the end, and thus still further narrow the possibility^ of conflict at that point. At the same time there was a natural limit to the full accentual development of the verse-end in the iambic dimeter. It would be practically impossible always to introduce into so short a verse enough words of three or more syllables with the ending — ^~, to have one for the end of each line , since these words would on the average occupy nearly or quite half of all the avail- 57 - able space. In the trochaic tetrameter and the iambic trimeter this difficulty did not exist on account of the greater length of these lines. Still the tendency in the dimeter would necessarily be in the direction of avoiding conflicts at the end of the verse, and of having them, if at all, in the beginning, especially in the first foot. A glance over the table of conflicts in the iambic dimeter (p. 5ff.) shows that this is what actually took place. In the case of the trochaic tretrameter of the tj^e _ u — ^- 1| _ u — >^ II — u — ^ II — u -^, no conflict whatever is possible in the first three dipodies. They will either be —uJa^, _ | w 3l m, _>jii-^, — ' ^|j!-^ or Au— li^, but in no case can an accented syllable stand in the thesis. In the final dipody we may have A u ^, -2- u 1 ^ or _ | 5 v, a conflict occurring in the last case. As in the iambic dimeter, the conflict in the final word depends on having a monosyllable long by nature or position for the pe- nultimate word. But in this verse there are six other places where the long monosyllable can be used to advantage, so that the share of long monosyllables which would fall to the lot of the last dipody, would in any case be small. And besides tliat, just as in the dimeter of the first type, there is in all the rest of the trochaic tetrameter of this , the prevalent type , no good place for a word of the type a ^ h, which just fits the end. We should then in the pure examples of this type hardly expect con- flicts in any part of the verse. In the trochaic tetrameter of the type — u— ||ii-3.u_-^|| _ u _ II ^ i u '^-, the second and fourth parts do not allow of con- flict , while the first and third may have it or not according as they are made up of single words of the type - u — and two words i u I — , or of two words — | ^ — Here again there is only one chance out of three for a conflict , in only two feet of the verse. The actual occurrence of conflict in the trochaic tetrameter, in the different stages of its development, is as follows : Metrical: verses. 1. foot 2. 3. 4. 5. 6. 7. A 1) Before Prudentius A 2) Prudentius to Fortu- natus A3) Of uncertain date 286 308 216 80 49 22 125 1 1 100 33 1 1 33 100 116 16 37 4J 7 14 B) Unmetrical theses 169 11 20 7 4 8 Rhythmical: C) Undeveloped D) Developed 124 547 6 19 3 14 3 12 1 58 1 58 That this development of the iambic and trochaic verses in an accentual direction did not occur in the rhythmic verses alone, but in the metrical ones as well, and that it really reached its height in the metrical hymns , has been pointed out in another connection (pp. 13 — 14), If our conclusions have been correct, this must necessarily be so, for it was the strict metrical form re- quiring all the arses and half of the theses to be of a certain fixed quantity and allowing the liberty of long or short in only one syllable in each dipody, that gradually drew the unstable syllables to the latter place, and thus of necessity brought on coincidence of accent and ictus at certain points in verses of these classes. An additional confirmation of the correctness of this position is derived from a comparison of several metres which were used more rarely in a rhythmical form. The iambic trimeter had its undetermined theses , one of which had always stood before the caesura, making the second half of the verse equivalent to the second half of the trochaic tetrameter catalectic. In this part the conflicts would then of course practically be confined to one foot , the fourth of the trimeter (corresponding to the sixth of the tetrameter). In the first half of the trimeter the first foot would be subject to prac- tically the same conflicts as the first foot of the dimeter , while in the second foot of the trimeter no conflict could occur, owing to the trochaic ending before the caesura. The following table of I'hythmical trimeters shows the location and number of conflicts. MI, 387 Ad eaeli DI, 237 Felix per MI, 145 Salve crux Mill, 406 caeli „ 239 s. Blasi „ 145 Sanctorum „ 61 Andrea p. March: Aurea luce no. of verses. 72 40 20 24 12 20 52 28 1. foot. 2. 3. 4. 5. (i. 36 10 15 3 9 3 19 5 4 3 14 4 22 6 11 The sapphic had no undetermined theses in the body of the verse , and consequently no accentual development. The table (p. 10) shows that this metre had the same number of conflicts in about the same location in both the metrical and the rhyth- mical forms. — 59 — The same is true of the rhythmical hexameter , in which the coincidence of accent and ictus which had for a long time existed in the last two feet of the metrical form of the verse, is main- tained, but the conflicts that occur in the remainder of the verse are just as conspicuous, if not more conspicuous, in the rhythmical than in the metrical form (cf. Meyer, I.e. p. 36— 39 = 300—303). Not only then has a vital connection been shown to exist between the decrease of the conflict of accent and ictus, and the gravitation of the unstable syllables of the word toward the undetermined theses, but the tables of the actual occurrence of conflict between accent and ictus show the extent of tbis conflict to have decreased pari passu with the spread of the other tendency just mentioned, and what is still more convincing, to have decreased in exactly the same way in which it would be expected to decrease as a result of that tendency. II, The Tendency to Reproduction and Division. Though the gravitation of unstable syllables toward the un- determined theses was no doubt the main ^) factor in the trans- formation of the metrical form of the verse, there were other forces at work which exerted more or less influence, according to circum- stances. The movement toward uniformity in the relation between accent and ictus at the end of iambic and trochaic verses, was in all probability at least accelerated by the natural prominence of the verse-ending , which causes it to adhere in the mind and thus easily leads to a more or less exact reproduction of it. When a certain verse-ending, constructed with a particular relation between accent and ictus, had once gained a very decided numerical suprem- acy , its complete control would naturally be more or less a question of time and convenience. A second result also to which, as we have seen, the existence of undetermined theses leads, name- ly the tendency to cut a verse up into shorter parts and to reproduce these broken verses with some regularit}', is at the same time a common characteristic of most of the popular and 1) The vigorous operation of this tendency can be seen from the unusually large number of such verse-endings as "dum pudet" (— | ^ ^), in Ambrosius and other early metrical hymns. That this ending had nothing pleasant about it for Roman ears is shown by the fact that its occurrence is extremely rare among the writers of earlier periods. What recommended it to Ambrosius and his fol- lowers was the fact that it allowed a final syllable to stand in the preceding un- determined thesis. — 60 — unskillful forms of verse generally. This tendency owes Its origin, on the one hand, to the fact that it is easier to reproduce short than to reproduce long verse-elements, and on the other hand, to the primitive pleasure in the rapid jingle that it intro- duces. Both of the influences mentioned, toward reproduction of the verse-ending and toward division of the line into smaller parts, were no doubt at work in all kinds of verse that underwent a transformation in their late metrical stage. To be sure, the effect of these influences could not , in the nature of the case , be as definite and decisive as that caused by the movement of the un- stable syllables toward the undetermined theses, since the former influences are largely matters of taste and preference , while the latter furnished a means of escape from a difficulty and was to that extent a necessity ^). And still, when circumstances were especially favorable, the results of the former may be seen quite as unmistakably as those of the latter. A notewortly example of the former influences is furnished by the lesser asclepiadean verse, which had no undetermined theses, and therefore furnishes an un- clouded view of the working of the principles under discussion. The later metrical forms of this metre present a development which is altogether unique. Hymns like Dan. I, 217 and 203, which clearly stand near the verge of the metrical practice, show an extraordinary prevalence of a secondar^^ caesura in each half of the verse, which results in a division of the line into four parts of three syllables each ^), The principal cause of this seems to have been, in the first half, the similarity in condition and surroundings of the third and sixth syllables. Both these syllables have the ictus, the former being made still more prominent, on the side toward the end of the verse, by the two short syllables 1) The numerical preponderance gained in the course of time by that form of trochaic tetrameter in which the verse is cut into its dipodies, was no doubt to some extent due to the identity of three of these parts. 2) As an illustration, the first two stanzas of the former hymn may be given. Festum nunc celebre magnaque gaudia Compellunt animos carmina promere, Cum Christus solium scandit ad arduum, Caelorum pius arbiter. Conscendit iubilans laetus ad aethera, Sanctorum populus praedicat inclytum, Concinit pariter angelicus chorus Victoris boni gloriam. — 61 — in thesis which follow it , the latter , on the same side , by the main caesura. In the second half of the verse , the first and fourth syllables also have the ictus and are similarly situated on the side toward the beginning of the verse, the first being pre- deced by the main caesura, the fourth by the two short syllables in thesis. That prominent syllables so placed should, with un- skillful versifiers, in the course of time come more and more to stand , the former pair at the end of words , the latter pair at the beginning, is onl}' natural and exhibits an extension to similar- ity of surroundings and quantitative prominence of the same principle which underlies rhyme and alliteration. The situation thus created could not last long, nor indeed could this minute division become a universal practice, owing to the impossibility of finding suitable words. The movement seems, however, to have gone far enough to enable it to fall in with the other movement which we found in the very common iambic and trochaic verses, and which resulted in making all verse-endings consist of polysyllabic words with the accent on the antepenult (cf. p. 55 2*.). The natural result of the whole process was of course, in the asclepiadean , a decrease of conflicts at the end of the second half, and an increase of conflicts at the end of the first half of the verse. The later metrical lines, in fact, have no conflicts in the former, and invariably have conflicts in the latter position (cf. the table of conflicts in the asclepiadean, p. 10). An examination shows that while in quantity the verse ad- hered to the old rules , it had at the same time attained quite a hard and fast form when read according to accent. With the possible accents the verse is now -^a. — 5 u _ || i' 5 ^j a^u.^. Of course an asclepiadean constructed in that way, at once lost its original rhythmical character as soon as the quantity of the syl- lables was disregarded and the accent was left as the sole rhyth- mical element of the verse. But in spite of the difl'erence in rhythm between the classical asclepiadean and the verse just mentioned, the descent of the latter from the former is evident from several facts. In the first place, most rhythmical hymns of this class are written in strophes of four lines, of which each of the first three has the same number of syllables as the lesser asclepiadean and the fourth line has the same number as the gly- conic. Now this very combination of verses had been the favorite asclepiadean stanza in the metrical hymns and is quite common also in the classical and post -classical pagan writers. But still — 62 - more convincing, perhaps, of the close connection between the rhythmical and metrical verses under discussion is the fact that at least one hymn (D I, p. 18G Cyxilla) which, while abounding in metrical faults, still clearly exhibits the rhythm of the old ascle- piadean metre, has a doxology appended to it, which is thoroughly unmetrical and has the trochaic - dactyilic rhythm which is found in all the later rhythmical forms ^). The change in rhythm which attended the change from the metrical to the rhythmical form of the asclepiadean has a counter- part in the history of the sapphic, though the identity is , from the regular and unvarying form of the sapphic strophe, much more clearly evident ''^). In the case of the sapphic , moreover, the new accentual rhythm was alread}^ inherent in the metrical form since the time of Horace and Seneca (cf. p. 15), and a trans- formation like that of the metrical asclepiadean was not first necessary to make the rhythmical form possible. If one reads the sapphic metre accentually, it is found to have a distinct rhythm : lam satis terris nivis atque dirae Grandinis misit pater et rubente Dextera sacras iaculatus arcis Terruit rirbem. This is a well-known fact and has even led to the extraordi- nary theory that Horace wrote his sapphics with attention both to accent and to quantity^). The real reason for this concomitant "accentual rliythm," is that in the sapphic line as it had existed in Latin since Horace's da}^, with the fourth syllable invariably long , with a fixed caesura after the fifth syllable and with the avoidance, common to Latin verse, of a single monosyllable before the caesura and the verse-end , the accent must per force fall on the syllables on which it is found. 1) A comparison of the last stanza proper with the appended stanza will make this clear ; Te nostra iubilet gloria perpetim, Qui solus dominus, triniis et unus es: lugi imperio saecula coutinens, Et cuncta dominans, omnibus imperans. Praesta, ingenite, per unigenitum, Regnans qui perpetim cum sancto spiritu, Olympum contines iugi iudicio, Sustentans aridam perenne imperio. 2) cf. Mone I, 20 note. 3) Eickhoff, d. horazische Doppelbau d. sapphischen Strophe. — 63 — To sum up briefly the changes in the late metrical hymns thus far mentioned, we have found that the iambic, trochaic and asclepiadcan verses reached a more or less complete harmon}^ of accent and ictus at the end of the line , while conflicts between accent and ictus remained in certain other positions of the line, and in some of these metres , like the iambic dimeter and ascle- piadcan, actually became more numerous in these latter positions. In the sapphic and hexameter practically no change took place in the relation between accent and ictus. In this condition the various metres were found when the practice of disregarding quantit}^ became general and accent was left as the only element which could give them a rhythmical character. Then the rhythm at the end of the iambic dimeter, of the two halves of the trochaic tetrameter, of the iambic trimeter and of the second half of the asclepiadcan naturally remained as it had been before , since in these positions the accent had in the metrical forms always coincided or had come to coincide with the ictus. Except in the end of these verses and parts of verses, there was , to be sure , no fixed place for the accent. But the part in which there was harmony of accent and ictus was relatively so large and prominent compared with that in which there was not, that the latter was naturally lost in the former, as far as the rhythm was concerned. In the rhj^thmical hexameter, on the other hand, only the last two feet would have a fixed rhythm, which was the same as that of the metrical hexameter. All the rest of this verse, therefore, failed to acquire a definite accentual rhythm, as it was too long to fall in with the rhythm of the last two feet. As a result , the hexameter found less favor with rhythmical writers than any of the other metres under discussion. In the sapphic and the first half of the asclepiadcan, the accent had or came to have a fixed place in the verse, with a location, however, which difi'ered from that of the metrical ictus. Hence the rhythm in these metres was difi'erent in their rhythmical from what it had been in their metrical form. III. The Loss of Elision. The instability of the final syllable in the course of time quite naturally brought about another change. To the already existing hardship caused by the necessity of looking ahead to the next word , whose initial letter might determine the length of a final syllable, was now added the fading away of the final con- — 64 — sonant ^). Under such circumstances the elision of the final vowel or vowel + m before an initial vowel or A, became a burden. For on the one hand, with writers ill-trained in metrical theory and practice, there was need of no little providential calculation to allow under certain conditions for the substitution in the metre of the first syllable of a following word for the final syllable preceding it, and to make sure that they obtained the quantity desired, and on the other hand, the fading of the final consonants made a distinction between the cases where elision was allowed and where it was not allowed, increasingly difficult. In this condition of affairs the path of least resistance clearly lay in the direction of avoiding the circumstances that would lead to elision. And we find in the late quantitative hymns but very rare occurrence of this phenomenon. Only here and there a hymn has something like the old-time proportion. Nearly all of them, however, have no elision whatever, or at most a single case of it. (For a detailed statement see table B in Appendix II). IV. The Loss of Resolution. It has already been shown that the metrical, as well as the rhythmical form of the hymns was based upon the stricter , not the looser form of Latin iambic and trochaic verse. Not only was this strict form followed , but one of the few liberties allowed it , that of occasional resolution of a long syllable into two shorts, was, as we have seen (p. 41), at once given up when the metres began to be employed in hymns intended for the use of the church. Thus while Ausonius shows considerable freedom in this respect, and Prudentius also makes frequent use of reso- lution, Ambrosius has but two cases in his 188 lines. And not only was it abandoned by Ambrosius, but it was very rarely re- sorted to in the hymns thereafter (for a detailed statement see table A in Appendix II). The surrender of this liberty by the hymns from the very start, was in all probability due largely to the same cause which we saw operating in the case of the soldier songs. All irregu- 1) Examples of this are seen in the earliest rhymes: Fortunatus: -Ms, -is-it, ens-et, -i-it, -a-at. Hymnum clicat: -em-et, -ur-us, -I-it, -it-is, -um-ur, -at-a, -um-u, -am-a, -us-um. Incliti festum: -em-et-e, -a-am, -i-it, -o-um, -ur-um, -es-e, -is-it. Sedulius: -e-em, -i-it, -a-at, -is-i, -um-o, -us-um, -e-es, -ant-am, -is-it, &c. Jam sexta sensim: -ur-us, -i-it, -es-e, -us-o, est-em, -o-um, -e-er, -ax-am. — 65 — larity of rhythm clearly tends to confusion when the hymn or song is to be sung in chorus by a number of pocple. Another influence which would tend to keep uniform the hymns written for church use and which is also partly to be held reponsible for the lack of variety in the metres employed, is that the hymns were in all likelihood written to fit a fixed rhythmical scheme or melody, to which they were to be sung. We know how potent an influence for uniformity this was with all lyric metres of antiquit}^, and especially in Latin. To the average body of people it is immensely more difficult to learn a new tune than to sing new songs to an old tune. Then to say nothing of the harm to the music itself, the embarrasment which one singer would naturally feel if he found himself parting com- pany with the rest, or found himself at the end of the line sud- denly a syllable ahead or behind as a result of an irregularity of the metre , and the general confusion apt to arise from the singing of a h3^mn of uneven rhythm by a chorus, are circum- stances which would themselves account for the almost complete absence of resolution in the Latin hymns, even if we did not have the examples of the earlier lyric measures of the language be- fore us. The Khythmical Imitation of Quantitative Metres. To understand fully the position of the early writers of rhyth- mical hymns, it will be necessary to keep in mind the conditions which at that time surrounded the writing of quantitative poetry; for that the rhythmical in some way grew out of the metrical practice, we may now assume without fear of serious contradiction. The ancient culture was still in a sense the ideal, and education still followed ancient lines , though the spheres of its influence were becoming more and more narrowly confined. Schoolbooks were shortened and simplified — always the symptom of an attempt to attain in a shorter time or under less favorable con- ditions the mastery which previous ages possessed. The study of men and their works had indeed long been superceded by the manual which contained in abstract all that it was necessary to know. A disproportionate amount of attention was paid to the mechanical structure of a literary product. This was the result partly of a want of thoughts which called for expression , and partly of the condition of the language. That faultlessly correct observance of quantity, for instance, which had once been the second nature of the cultured man , had no doubt become rare even in the instructors and the models of good taste , to say no- thing of the poeple in general. If acquired at all, it must have been the result of long and patient effort. Some parts of the language were indeed falling into irreparable decay, as for in- stance the endings of the words. No living pronunciation was in this latter point any longer a safeguard against metrical faults. "Whoever wished to write quantitative verse correctly, was obliged, in this particular at least, to learn it purely from the rules. In other words, Latin had for metrical purposes become in part a dead language. - 67 — We must not suppose, however, that the zeal of the verse- writers abated in the least in the face of these difficulties. Some there were, to be sure, who proved unequal to the task wherever special vigilance was demanded. But in general those who pro- fessed to write metrical poetry were fully conscious of all the obligations imposed by the rules of the art. We are somewhat surprised to see how closely the late metrical hymns, for instance, observe the rules of quantity and hiatus (for a detailed statement see p. 5ff. and Appendix II, F). But the difficulties encountered by these metrical hymnists did not lie in the language alone. Every detail in the whole process of transformation which has been traced in the preceding chapters, the shifting of the unstable syllables, the breaking up of the verse into equal parts, as well as the loss of resolution and eli- sion, served to cramp the verse itself and to hinder the poet's free- dom of expression. The possible combinations of words had become very few indeed, and the cases were surely not rare within a single short hymn , when the poet would be obliged to go out of his way to find a word or a form of expression that would at the same time suit the thought and the crippled metre. And all this at a time when it became from year to year a greater problem to suit the words to the intelligence of the poeple. We can best realize the hardship of metrical versification, if we observe some of the methods employed by the writers. The most common cause of faults was, as we saw, the fading of the final consonant and of the distinction between long and short in the final vowel. An unusual effort was necessary , therefore , at once to avoid hiatus and to make sure of a long sj^llable in the arsis. The device employed in the attempt to do this, was to have as few as possible of the initial syllables in thesis begin with a vowel. In the case of Horace, Ausonius and Prudentius, the propor- tion of initial vowels to the total number of words beginning in thesis in the iambic dimeter (not counting the first syllable of the verse), is from about one out of four to one out of five and a half. But as early as Ambrosius we find it about one out of 13, in Sedulius it is one out of 27, and in the hymns attributed to Gregory, one out of 12. Individual poets with a facile pen, like Fortunatus, and individual hymns like "iam sexta sensim solvitur" and "deus qui certis legibus" keep close to the old proportion, but on the average the proportion is much lower in the metrical hymns than in the earlier and classical writers. According to 5* — 68 — the date and stage of development , the proportion is as follows (for a detailed statement see Appendix II, D) : Classical Time of Ambrosius 5G out of 275 or 1 out of 4.9 + 176 )) » 922 )j 1 « „ 5-2 + 83 n n 611 « 1 „ „ 7.4- 35 n H 320 V 1 « „ 9-1 + Sedulius to Bade Carolingian period Metrical hymns of ) -, p-o -.070 1 q n 1 /. . , — 152 „ „ 1373 „ 1 „ „ 9.0 + uncertam date ) Metrical hymns with ) -.0 1 a7 1 q q . . 1 ,, — 18 „ „ 167 „ 1 „ „ 9.3 - unmetrical theses ) Early rhythmical — 160 „ „ 744 „ 1 „ „ 4.6V2 We see that the proportion of words beginning in thesis with an initial vowel was about the same in the metrical dimeters of classical times, which were practically unaffected by metrical and linguistic difficulties, and in the rhythmical dimeters, where these difficulties no longer existed ; but that in the later metrical hymns, where an effort was necessary to maintain the quantity, the ave- rage proportion was only about half as high, and in many individ- ual cases much lower than that. That under conditions like these the metrical form should have been abandoned in the hymns , is not strange. The wonder is rather that it did not die out altogether. But it continued to hold its own, and for four or five centuries after Ambrosius, practically the only hymns of which we know the author, are me- trical in form. Further than that, every revival of culture meant a revival of metrical versification. These facts can be well ex- plained only on the assumption that it was considered a nobler occupation to write metrical than to write rhythmical poetry. The rhythmical hymns are, as a fact, classed with the "cantica vulgarium poetarum," ^) and rhythmical poems are given to the world with an apology for their form, which is spoken of as prose rather than poetry^). With the special help of God, one poet hopes to be able to write metrically^), while another feels sure that even if his verses can not be called metrical, they may at any rate be classed as rhythmical ^). Still another ^) is surer of his results : 1) Bede in Keil, G. L. VII, 258. 2) The end of the Carmen de Synodo Ticinensi , (Waitz, SS. rerum Langob. et Italic, p. 181) A. D. ca. 698. Compare the heading of a rhythmical poem Poet. Lat. med. aev. I, 79 "Alphabetum de bonis sacerdotibus prosa compositum." 3) Smaragdus, ca. 800 A. D. Poet. Lat. med. aev. I, 619 (2. Stanza). 4) Milo, 9!^ century. Poet. Lat. med. aev. Ill, 674, 1. 1037. 5) Paulus Albarus, mid. of 9. century. Poet. Lat. med. aev. Ill, 129, 1. 21. — 69 — Ergo vos cigni lautique decore pavones, Cum suavi meatim philoinela clucite carmen Et pedibus metricis rithmi contemnite monstra. The ability to write metrical poetry was a badge of culture, as it were, and placed the poet in direct line of succession to the metrical nobility of the past. Skill in writing verses had in former times been one of the accomplishments of every man of good breeding, and it must never be forgotten that all eyes were still turned backward. Up to the ninth or tenth century rhyth- mical hymns were in all probability less numerous than metrical hymns , and were certainlj^ as a general thing held in smaller esteem. But with all that, rhythmical versification was better suited to the circumstances and destined in the end to survive its competitor. The question just how rhythmical h}'mn - writing began, is hard to answer on account of the almost utter absence of reli- able dates for the early specimens. One thing seems however to be certain. The early rhythmical hymns did not arise out of a single metrical type, but reflect the various forms of the metrical hymns during the whole of their transformation, which , as we saw, resulted in harmony of accent and ictus at the end of the verse. Some of the rhythmical hj^mns of the iambic dimeter pattern, for instance, have conflict of accent and ictus at the end of the line to a considerable extent. Of the rhythmical hymns which grew out of the trochaic tetrameter this is not true, to be sure, but neither is it true of any of the later metrical specimens in this metre. For certain reasons which we mentioned on p. 57, the trochaic tetrameter developed along this line more rapidly than the iambic dimeter, so that the former verse had completely finished its accentual development before rhythmical hymn-writing began. The whole transformation of the metrical forms, iambic as well as trochaic, seems to have practically reached its conclusion within less than a century from the time of Ambrosius. Marti- anus Capella no longer has conflict of accent and ictus at the end of his iambic dimeters , and Sedulius has but nine cases of it in 92 lines. A little later than that, about the year 470, we have a rhythmical poem by Auspicius, composed of 158 iambic dimeters (six irregular lines being omitted from the count), in which but a single line ends in a conflict. To be sure we have metrical hymns in iam- bic dimeter later than this, which have conflict of accent and ictus at the end of the verse, but the cases of conflict are in most of them — 70 — very few, and to be accounted for either by the difficulty of avoiding conflict altogether in the metrical form of this metre, or by a possible imitation of the older manner; for we must re- member that lij-mns of all patterns were constantly used by the church. Likewise there were no doubt rhythmical hymns with this conflict, composed later than the fifth century, though in the case of these the definite dates are so few, that we are unable to show it. The list of hymns previously mentioned (p. 18), whose date is not later than the middle of the sixth century, will serve for comparison. Of the two rhythmical specimens among these hymns, one has three conflicts of accent and ictus at the end of the line, the other has eight. The metrical hymns (omitting those which belong to Ambrosius) have conflict at the end of one , two, three, five and fourteen lines respectively. This wide variety of the early hj'mns, both rhythmical and metrical, in respect to the degree of coincidence of accent and ictus found in them , shows that the two kinds of verse did not develop separately. But that the younger form, which hardly had recognition and was but just finding its waj^ , should have influenced the old established form with the teaching of the schools and the practice of centuries behind it, is out of the question. If we examine the hymns which seem to represent the be- ginnings of rhythmical poetry, we find two quite distinct classes. In the one class the unmetrical syllables apparently entered against the will and intention of the writer, who meant to adhere to the quantitative form , but was unequal to the problem pre- sented by the condition of the metre and of the language. This is the same class of metrical faults which we find here and there in the more strictly metrical hymns and to a greater or less extent in the other poetry of the time. We may say in a general way that all those unmetrical hymns are to be classed under this head, in which a comparatively small number of faults is found practically confined to one or two points in the verse. This includes such hymns as Mone I, 91, III, 431 and 459, Wackernagel no. 101, and all those hymns which have metrical arses and unmetrical theses (cf. the tables p. 7, Iambic Dimeter C and parts of D). But while faults like these, no doubt, prepared the way for a more complete departure from metrical form, it would hardly be possible for rhythmical poetry , such as we have it in the early rhythmical hymns , to develop , as long as the purpose of the writers was to write quantitative verse. For either would — 71 — adverse criticism and reference to models and rules again and again correct tlie faults which occurred, as was the case with the hexameter, or the task would finally be given up in despair. A voluntary departure from quantity is not to be thought of in the case of a writer whose efforts arc directed to writing metrical verse. There remains the possibility that the metrical form was given up for such reasons as those which prompted Augustine to write his Psalmus contra Partem Donati. If we stop to think however, we see that such undertakings as the humble and prosaic task of enlightening the ignorant or preacliing against the pagans or heretics , were very different things from the lyric poetry in which the soul reached its heights of exaltation. Horace was sure- ly neither the last nor the only Latin writer who had a "musa pedestris." If we consider further that rhythmical poetry could not get a standing , or get any sort of recognition through its form , but that onl^^ the inherent power and beauty of a rhyth- mical hymn could give it recognition and acceptance in spite of its lowly dress, we are all the more convinced that the writing of rhythmical hymns can not have come from any conscious or unconscious departure from a metrical standard so long as this was considered the correct and desirable form by the writer in question. This initial impulse is rather to be found, as has already been implied, in the production of a rhythmical masterpiece. Inspired by the music and the content of the songs which were sung at the services , well acquainted with their outward form and general manner , though , it may be , entirely ignorant of the rules of the metre , some poet of native genius in all probability produced the first great rh^'thmical hymn, which by its excellence gained recognition and invited to further imitation along the same line. This may well have been the history of such a hymn as "Rex aeterne domine," with its slight irregularities in the length of the line and in the penultimate syllable of the verse, which however in spite of these pecularities evidently stood in high regard, to judge from the frequent references to it, and is called "hymnus ille praeclarus" by Bede in distinction from other rhyth- mical poems (Keil, G. L., VII, 258). When this point was once reached, and the free imitation of metrical h^'mns had gained a certain recognition as a method of hymn -writing distinct from that employed in the metrical hymns , then the way was open also for those versifiers whose knowledge of ancient prosody was - 72 - not sufficient to help them over the existing state of the language, and also for those who , possessing this knowledge in a sufficient degree, shrank from the hindrances which the cramped form of metrical verse placed in the way of free expression. The free imitation of metrical poetry was not by any means an unknown practice in the late centuries of the empire. The effort was of course to reproduce in a general way the impression which the original made upon the ear. In doing this the imitator naturally reproduced those parts most accurate^ which impressed the ear most, i. e. the parts preceding the caesura and the end of the line. Commodian's imitation of the hexameter was, no doubt, of this kind ^). So also was that curious reproduction of the spirit of Plautine verse in the Querolus, whose author tells us "prodire autem in agendum non auderemus cum clodo pede, nisi magnos praeclarosque in hac parte sequeremur duces." Whether the "magni praeclarique duces" are Plautus and Terence or men of the author's own day, it is clear that he considered his manner of composition to be not without precedent. At any rate, to treat the Querolus simply as a piece of highly rhythmic prose, as Norden does (Antike Kunstprosa p. 630), is certainly impossible, since the Plautine verses and verse-endings found in the Querolus in every sentence, were of a kind not used in prose. Free imi- tations of the hexameter and the elegiac distich are found abun- dantly among the tomb -inscriptions of nearly all periods. They vary in degree of accuracy from prose with a few reminiscences of the beginnings and endings of verses (Biicheler 699, 708, 733, 1581 = CIL. V, 6742; VI, 18086; Rossi II, 294; Fabretti 456), to poems constructed as regularly as the verses of Commodian (CIL. VIII, 7156 = Biicheler 512). Free imitations of iambic verses among the inscriptions are more rare. It will suffice to mention the two most extensive specimens of this class (CIL. VIII, 646 and 647). In the sixth century we meet an instance the same kind of imitation in the Vita SS. Abbatum Agaunensium, chapter 10 (Arndt, Kleine Denkmaler aus d. Merovingerzeit, p. 20). Another is the epitaph of St. Felix f 724 (Waitz, rer. longob. et ital. p. 375, cf. also Meyer, Anfang u. Ursprung). A higher form of imitation than that last mentioned , or perhaps rather a development along rhythmical lines , is seen in the riddles printed at the end of Meyer's treatise. Meyer refers 1) cf. Meyer's exposition of the characteristics of Commodian's verse (Anfang u. Ursprung d. lat. u. gr. rbyth. Dichtung). - 78 - to the fact (p. 279) that the first half of these latter verses is uniformly composed of six, the second half of eight syllables. This restriction of the hexameter to a single type was no doubt the result of the very sound feeling that variable metres like the hexameter and the iambic and trochaic verses of the drama were not suited to being reproduced rhythmical^. Many of the inscrip- tions , even of those otherwise metrically correct , show how easy it was in the case of these metres to wander from the legitimate number of feet. It was far different in the case of lines with a fixed number of syllables and a hard and fast internal composition, such as the verses of the hymns had become. These latter verses had, besides , another advantage in the brevity of the recurring ])arts, which in the trochaic tetrameter had come to consist, for the most part of four, respectively three syllables. Another ad- vantage which the common metres of the hymns possessed was the fixed relation which had grown up between the accent and the arsis. In addition to all these formative and restrictive in- fluences, there was the fact that the hymns were intended for a definite place in a fixed ceremony , and for a definite manner of execution, for the writing of hymns placed the same restrictions upon the poet that the painting of an altar-piece places upon a painter. From all these circumstances it came about that only those metres which were used in the hymns, developed rh^^tbmical forms that possessed sufficient vitality to survive and supplant the met- rical forms out of which they grew. That this off - hand imitation of metrical hymns must leave its traces in the rhythmical hymns produced, and that there must be differences between these early rhythmical hymns and those of later centuries , is natural, and we should be surprised if it were not so. The prevalence of the conflict between accent and ictus at the end of the line in the early rhythmical hymns after the pattern of the iambic dimeter, has already been referred to. The continuance of this conflict resulted naturally for the metrical hymns from the conditions of their construction. But for the rhythmical hymns no such restriction existed, since they were no longer bound by the rules which involved it (cf. pp. 56—7). Its presence in the rhythmical hymns can only be explained on the ground of imitation. The departures from the quantitative form in the early rhythmical hymns are, on the whole, considerably less extensive than in the later ones, and one often finds in the former whole lines and sometimes several lines in succession which have no — 74 — metrical faults at all. In an age when such models and standards as existed were still classical and the vast majority of the hymns in use were still metrical, too wide a departure from these com- monly recognized standards would at once attract unfavorable attention, and no doubt bring upon it the verdict of lawlessness and bad taste. We find the line between what was allowed and what was not allowed in rhythmic verse, distinctly drawn, for instance, by Bede (Keil, Gr. L., VII, 258): "Metrum est ratio cum modulatione, rhythmus modulatio sine ratione. plerumque tamen casu quodam invenies etiam rationem in rhj^thmo, non artifici mo- deratione servata, sed sono et ipsa modulatione ducente, quem vulgares poetae necesse est rustice, docti faciant docte. Quomodo et ad instar iambici metri pulcherrime factus est hymnus ille praeclarus, Rex aeterne domine et alii Ambrosiani non pauci. item ad formam metri trochaici canunt hymnum de die iudicii per alphabetum, Apparebit repentina." That the church was fully alive to the difference between good and bad rh^'thms , appears from canon 24 (23) of the council of Tours (A. D. 567) which makes plain to us the grounds on which hymns were excluded from the service. "Et licet libros Ambrosianos habeamus in canone, tamen quoniam reliquorum sunt aliqui, qui digni sunt forma cantare , volumus libenter amplectere praeterea, quorum auctorum nomina fuerint in limine praenotata ; quoniam quae fide constiterint, dicendi ratione non obstant.'' Pro- vided the hymns are orthodox in sentiment, no attention is to be paid to their manner of expression, only to the question whether they are "digni forma cantare." What the standard of excellence in form was in those days is apparent in the light of the passages quoted on pp. 68 — 9. At the same time there were, as we know, rhythmical hymns among the "hymni Ambrosiani" (cf. the passage from Bede just quoted — "et alii Ambrosiani non pauci."). "Digni forma cantare" can therefore only refer to those rhythmical hymns which were worthy to be placed by the side of the metrical ones in their form , i. e. those whose departure from quantitative correctness was not so apparent. As good taste had formerly placed a check uj)on the intro- duction of metrical feet into prose , so now it seems to have set a limit beyond which unmetrical syllables should not be admitted into rhythmical poetry. Of course this was a matter for the ear rather than for the metrical intelligence to decide, since, if our - 75 - theory of "free imitation" is correct — and the expressions" sono et ipsa modulatione ducente," "ad instar iambici metri," "ad for- mam metri trochaicr' in the passage just quoted from Bede , give it additional confirmation *) — rbythmical poetry was based upon the natural judgment of the ear instead of being based upon rules. Nevertheless standards must be conformed to , and in this case the standards were the impressions produced upon the ear by the metrical hymns. It is interesting to notice how these early rhythmical poets avoided the charge of "rusticity" in cases where the temptation to admit unmetrical sjdiables was unusually great. The case which at once occurs to us is the third arsis of the iambic dimeter. We have seen (pp. 56 — 7) that the difficulty of this syllable for the metrical form of the verse in question lay in the great demand it made upon the words which had both a short penult and a long antepenult. The strain to which the metrical form was subjected, was therefore greatest and the tendency to the unrestrained ad- mission of unmetrical syllables was strongest at this point. While a free admision of these syllables was a necessity, it was no less a necessity that the verse should remain acceptable to the ear, so as not to fall back into the contempt from which it had but hardly risen. What happened was that the language went back at a bound to its condition before the introduction of Greek metres had fixed the quantity of doubtful syllables and had divided them into two opposing camps ot iong and short. Where a promiscuous admission of short syllables would have appeared lawless and would have provoked criticism , a select class of them seem to have entered unchallenged. These were the syllables in which the short vowel was followed by a sonant or a liquid, i. e. by consonants whose sound is capable of extension. So, for in- stance , by holding fast the first syllable of "dominus," we can make it resemble a long syllable much more nearly, than is pos- sible with the first syllable of "capitis," where the short vowel is followed by a mute. 1) cf. Walafrid Strabo, de rebus ecclesiasticis, ch. 25. "In officiis quoque quae beatus Benedictus abbas, omni sanctitate praecipuus , ordinavit, hymni di- cuntur per boras canonicas: quos Ambrosianos ipse nominans , vel illos vult in- telligi quos confecit Ambrosius, vel alios ad imitationem Ambrosiaiiorum compo- sites." Cf. earlier in the same chapter "hymni metric! ac rhythmici in Ambrosi anis officiis." — 76 — In some of the early rhythmical hymns of the iambic dimeter pattern, short vowels followed by any one of the latter consonants are hardly admitted to the third arsis at all, and the place, when not occupied by a long syllable, is nearly always filled by a short vowel followed by a sonant or a liquid. This is notably the case in tlie hymn "rex aeterne domine" which, as will be recalled (cf. p. 18) is the only rh^^thmical h^^mn in a total of ten hymns laid down by bishop Aurelian in his regula about the middle of the sixth century, and which was also selected by Bede as a specimen of a rhythmical iambic hymn "pulcherrime factus." This hymn has 14 short syllables in the third arsis , of which in 13 cases the short vowel is followed by a sonant or liquid and in only one case by a mute. In the only other rhythmical iambic hymn which we know with certainty to be older than the middle of the sixth century, "bis ternas horas explicans," there are seven short syl- lables in the third arsis, in all of which the short vowel is followed by a sonant or a liquid. We have already seen (p. 28) that the same rule holds in the Psalmus of Augustine, where of a total of 534 trochaic verse-endings, 31 have the penultimate syllable short. Of these 31 there are but three cases in which the short vowel is followed by a mute (vetet, cruce, datumst). Another factor which helped the rhythmical hymns to repro- duce approximately the effect of their quantitative models , was the accent. We found in the later inscriptions in hexameter, that unmetrical syllables seem to have attracted less attention in those arses where the accent and ictus regularly coincided. If unmetrical arses of this kind came nearer to producing the same impression which metrical arses produced, it would also be true that short syllables immediately before or after the accented syl- lable would, when employed as arses, produce the impression of a more decided and radical departure from the quantitative model. This would be true of all short syllables in the positions mentioned, except when these syllables stood at the end of the word where, as we saw, an extensive departure from the metrical form had already existed for a long time. The final syllable in itself was, moreover, in this present period so thoroughly uncertain, that the feeling for its ancient quantity must have been all but lost in ordinary pronunciation. But aside from the final syllable, other short syllables immediately preceding or following the accent, were as a general thing very rarely admitted to the arsis in the rhythmical hymns. There are cases, to be sure, where these syl- lables are admitted somewhat freely, but they are comparatively — 77 - few. It is interesting to find a close connection between the latter practice and that of admitting long syllables to the last thesis of the verse. The hymn "summi largitor praemii" which we found (p. 20) to have 10 long syllables in the last thesis, has the scan- sions supplicamus , piaculis , rogare , capiamus , gloriatur. The trochaic hj-mn "Archangelum mirum magnum" which was men- tioned in the same connection, has 13 unaccentual short syllables in the arsis. The practice is found in other cases as can be seen from the detailed statement (Appendix II, F), but the hymns in which such scansions are of more than occasional occurrence, are rare. The restricting influence of classical models and traditions is very apparent when we observe the peculiarities of versification in Britain and Ireland, where classical standards had not become so firmly rooted. In the five poems found among the letters of Boniface (Jaff'e, Mouumenta Moguntiua), which probably belong in part at least to Aldhelm, there are 122 unaccentual scansions with a short sj'Uable in the arsis in 586 lines of eight syllables. In 76 additional lines of the same length found here and there at the ends of letters in the same collection, there are 15 such scan- sions. In the Antiphonarium of Bangor there are three rhythmical hymns of the pattern mentioned , all of them abounding in scan- sions of the kind under consideration. In the hymn on St. Comgill, there are 58 in 200 lines of eight syllables, in a series of hymns "ad horas," there are seven such scansions in 40 lines, and in the "Memoria abbatum nostrorum," seven in 42 lines. The hymn on St. Patrick, though trochaic, also has a considerable number. Perhaps the most extraordinary example is a poem of unknown origiu, an abecedarius against the anti-trinitarians , (published by Boucherie , Melanges latins et bas-latins) which has 280 lines of eight syllables and 81 unaccentual scansions with a short syllable in the arsis. It is an interesting fact that none of the poems mentioned is, in all probability, later than the eighth century, and some are perhaps as early as the sixth. If there were local differences in the handling of the metres in the early rhythmical period , differences can also be traced be- tween the method of rhythmical versification employed in the hymns and that employed in other branches of writing, especially those in which the form is usually subordinate to the content. The letter of Auspicius to Arbogastes, though as early as 470 A. D., shows no preference in the character of the short syllables admitted to the third arsis of its rhythmical iambic dimeters, such as we found in the hymns of the period. Four of the five poems referred to — 78 — above, which go under the name of Aldhelm, are occasional pieces of narrative and personal compliment, such as might be sent in a letter to a friend. This, we saw, was actually the use which was made of the shorter pieces found in the same collection. The abecedarius against the anti-trinitarians reminds one of the Psalmus contra Partem Donati and Commodian's Carmen Apologeticum. The long hymni in honor of St. Patrick and St. Comgill are com- binations of biography and eulogy rather than hymns. To make the list complete , we have a cosmography, a Prankish product of about the seventh century, which is written in unutterably bar- barous trochaic tetrameters (printed by Pertz, Abhandl. d. Berl. Akad. Jahr 1845). In looking over this list, the thing that attracts attention is the large proportion of verses of the iambic dimeter pattern. It is really what we ought to expect , for a short verse which was so late in reaching its full accentual development, and which always showed great irregularity in the relation between accent and ictus in its first half, was naturally open to the peculiarities of scansion just mentioned to a greater extent than the longer and more ac- centual measures which grew out of the trochaic tetrameter and the iambic trimeter. This fact, together with the restrictions which the extensive employment of the iambic dimeter in the hymns placed upon its free rhythmical development, had the result, in the Carolingian period, of quite a general return to the met- rical form of this metre. While the other two metres mentioned were quite generally used in a rhythmical form by good writers during this period , the iambic dimeter appears in the rhythmical form but rarely. Some writers even use the metrical form of the latter measure side by side with the unmetrical form of one of the others. It may have been this condition of things which helped to rob the iambic dimeter of its almost absolute supremacy and to make the trochaic verse the favorite measure in the great rhyth- mical period of the 12 th and 13 th centuries. General Summary. A brief summary of the various stages in the change from quantitative to rhythmical poetry, which have been traced in de- tail in the preceding pages, will perhaps contribute to a more definite conception of the development as a whole. The change mentioned was found to have resulted from causes within the Latin language, with very slight traces of influence from without. The rhythmical iambic and trochaic verse of the hymns grew out of the stricter form of these metres and is, so far as form goes, in no sense a survival of a popular method of versification. The coincidence of accent and ictus, which is usually consid- ered to be the chief distinguishing feature of rhythmical poetry, is merely a phenomenon appearing in certain classes of verse, such as the iambic dimeter and trimeter, the trochaic tetrameter and the second half of the asclepiadean, where circumstances were peculiarly favorable to its development. This characteristic is not found developed in the rhythmical forms of some other verses, such as the dactylic hexameter, the sapphic and the first half of the asclepiadean. To be sure, the rhythmical forms of the latter metres did not posses the same vitality as the rhythmical forms of the iambic and trochaic verses , though a considerable number of examples is found. This fact was due to several causes. First, the metres referred to were in their metrical form more complex and difiicult than the iambic and trochaic verses, and therefore were not, from the start, employed with the same frequency for religious purposes. Then, it is no doubt true that after the quan- titative structure of the various metres had broken down , those verses which had in the late metrical stage developed a more or less complete coincidence between accent and ictus , had by virtue of this fact a great advantage over those in which a struc- tural principle ceased to exist as soon as the rules of quantity — 80 — were uo longer observed, and over those in whieli the change from metrical to rhythmical form involved a change in rhythmical char- acter. A study of the origin of rhythmical poetry must then necessarily be , to a large extent , a study of the development of iambic and trochaic verses, and especially of the process by which they gained their great advantage over the other verses , namely the development in their metrical form of a certain degree of coincidence between verse-ictus and word-accent. The element in the iambic and trochaic metres which was favorable to the development of their accentual quality , was the presence , in their metrical form , of a syllable in each dipody which might be either long or short. As the metrical inscriptions of the first four centuries after Christ show, two syllables of the word were especially unstable as to their quantity, the final syl- lable and the syllable preceding the accent. In the effort to ad- here to the strict metrical construction of the metres referred to, these unstable syllables gradually found their way more and more into the positions of the verse where distinction between long and short was not required. This tendency is noticeable to some extent in the metrical hymns of Ambrosius and even earlier. After that time it spread until , in the later metrical and most of the rhythmical hymns, practically all the undetermined theses were filled by unstable syllables of this kind. Such a construction of the iambic and trochaic metres nec- essarily brought with it coincidence of accent and ictus in certain parts of the verse, notably at the end, and practically restricted the prevalence of conflict between accent and ictus to a limited and fixed number of feet. So in the iambic dimeter conflict was practically confined to the first two feet, in the trochaic tetra- meter to the first, second and sixth, and in the iambic trimeter to the first and fourth. The last two feet in each of these metres became, in the course of time , entirely free from conflict between accent and ictus. When this whole process was almost completed , the rhyth- mical form of versification became a recognized method of hymn- writing, and the accentual quality which had been gradually de- veloped in the metrical forms of the iambic and trochaic verses, fell as a heritage to their rhythmical forms. This passage from the metrical to the rhythmical form was gradual. The immediate circumstance which gave the advantage to the rhythmical form and insured its final success in the contest with its rival, the met- rical form, was the cramped and mechanical method of construe- — 81 — tion into which the metrical verses had fallen. The chief cause of this narrowing and cramping was , in the iambic and trochaic verses, the effort to maintain their strict quantitative character, which resulted, as we have seen, in placing the metrically unstable syllables in the undetermined theses. In the case of the asclcpia- dean a similar result was produced, in the later metrical period, by the peculiar symmetrical character of the verse, which, in the hands of unskillful writers, caused it to tend toward a division into four parts of three syllables each. The same tendency to division and the repetition of like parts no doubt had its share in helping to establish in the middle of the first half of the trochaic tetrameter the caesura which clung to that part of the verse throughout its rhythmical career. But meanwhile other influences also were at work. The res- olution of long syllables , a privilege which iambic and trochaic verses had always had, was almost entirely abandoned, as soon as these verses were employed in hymns for the use of the church. Further, the growing instability of the final vowel and the gradual fading away of the final consonant made impossible the distinction between syllables which could and those which could not be elided, and led everywhere to devices for avoiding the necessity of eli- sion. Though the verses were thus cramped and lost their liber- ties, their obligations in the matter of quantity and hiatus still continued to be observed. To make sure, for instance, of a long syllable in the arsis, when the syllable came at the end of the word and was therefore often of uncertain quantity , the device was resorted to of having the following word begin with a con- sonant as frequently as possible. From this state of things a welcome means of escape was offered to many a toiling writer by the new rhythmical practice of writing hymns, which though still regarded with suspicion, was tolerated within certain limits. Since they were, in addition to this restriction, at first free imitations of the quantitative hymns then in use in the churches, many of these early rhythmical hymns bear distinct traces of their origin. They not only admit fewer unmetrical syllables than the later specimens of their class , but in a place like the third arsis of the iambic dimeter, where the strain upon the metrical form and the tendency to admit unmet- rical syllables were naturally greatest, they compromised upon those short syllables which are followed by a sonant or a liquid and are therefore, under stress, capable of being drawn out some- what beyond the length of other short syllables. This preference 6 — 82 — for syllables of medium length, together with the other relics of quantitative verse, gradually disappeared when rhythmical hymns became more numerous and when, partly also through the waning of classical influences in general , a distinction of form was no longer made. Then, when rhythmical verse had reached the end of its de- velopment and had hardened into a fixed form , as we see it in the tenth and eleventh centuries, the time was at hand for a new age of inspiration which was to use the various lines and parts of lines in new combinations and for new purposes. Appendix I. The Transformation of the Rhetorical Clausula. As is now well known, most of the Latin prose of the Empire has a pronounced rhythm, i. e. those parts of the clauses and sentences immediately preceding the rhetorical pauses show a regular succession of long and short syllables. From about the fourth century onward, moreover, the ictus in these cadences coincided more and more with the accent of the words. And later when distinctions of quantity disappeared, the accentual rhythm remained. Outwardly we therefore have the same facts that we found in the ends of the contemporary iambic and trochaic verses. It remains to inquire into the nature of the change and, if possible, to find its cause. That there was a close evolutional relation between the accentual clausula of the Middle Ages and the quantitative clau- sula of the Latin prose of earlier centuries, was noticed by Havet (La Prose metrique de Symmaque, 1892). A detailed theory of the nature and cause of the change from the one form of the clausula to the other was then published by W. Meyer in his review of Havet's book in the Gottingische Gelehrte Anzeigen for 1893, p. 1—26. Contrary to all his previous views on the origin of rhythmical verse, Meyer concludes that the change in prose was due to the increased power of the accent, which in the 3rd and 4th centuries, according to his view, became so prominent „dass beim Sprechen der Accent viel deutlicher hervortrat als die Quantitat der einzelnen Silben" (p. 19). Those forms of the metrical clausula in which the accent and ictus coincided, he thinks were thus preserved in the rhythmical period, while the others died out. The fact that this coincidence between accent and ictus arose while quantity was still rigidly adhered to he explains as follows : ;;Die neuen Accentformen werden noch lange Zeit mehr oder minder achtsam nach der alten Art mit langen und kurzen Silben ausstaffirt" (p. 20). 6* — 84 - This theory has the weaknesses which we found in accentual theories elsewhere , for it assumes the very thing that most needs proof, i. e. an accent sufficiently strong not only to overcome the method of pronunciation which must have been current in Latin, but also — a still harder thing — to overthrow in theory and practice the metrical system which had been built upon this pro- nunciation. If Meyer reversed his theory and said that the old quantitative forms were fitted up with accented and unaccented syllables , it would impress us as presenting a more plausible historical development. But even then, like Huemer's theory of the transformation of verse-forms, it could explain the phenomena only in certain cases, — which of course amounts to no explana- tion at all. We found that the accentual theories had to he rejected in the case of verse, because in some verse-forms no accentual de- velopment took place from quantitative to rhythmical poetry, and because in others, where it did take place as in the iambic and trochaic metres , it sometimes affected the whole verse and some- times only the end of it. If we attempt to explain the transfor- mation of the rhetorical clausula by the theory of accentual pro- minence, we meet difficulties of a similar kind. Three forms of the clausula had become all but universal, 1. _vj ^; 2. — u u — ; 3. — u — ii (or in its fuller form — u u — ^). With the exception of the third, these forms were freely modified and varied by the resolution of the long syllables. As in poetry however, this occurred more and more rarely as time went on, and by the fifth century resolution seems to have been almost entirely confined to the penultimate syllable of clau- sula no. 1. A further liberty which the first and second forms had from the beginning, was that the syllable between the first two longs might itself be long. The most common form of no. 3 in the earlier time (Cicero, Seneca, Pliny) was _ u _ ii. About the second and third centuries A. D. a cretic developed before this ending, but it did not attain a fixed character, as either of its two long syllables, or even both of them, might at any time be replaced by short syllables (cf. Meyer's article just referred to). This lengthened form of no. 3 (^^ — — ^ — ^) gradually displaced the older , simple form, though the latter does not entirely disappear. Accordingly, the forms of the quantitative clausula with which we have to deal at the beginning of the rhythmical period are ^ 85 -- 1, _£.~ — ^SUd.^ , 2. ^(j _-t.vjii, Of the quantitative clausulae as they existed at the beginning of the fifth century, there are, as Meyer explains, six rhythmical representatives : Three of no. 1 : a. r(j r>j \ r\j r^ r^ ; b. rij ^ r^ \ r^ r\J ] C. rOr\^|(-vrvrv r^. Two of no, 2: a. rOrC'Irv^rGur^; b. rOurv|rC»u rv. One of no. 3 : rij \j r\j \ r^ r>j r>j r\j. An examination of the writers whom Meyer classes as rhyth- mical — Sedulius, Ennodius, Fortunatus — shows that of these six forms only three, la, 2a and 3, are found in abundance, while the other three (with the exception of 1 c, which is rather common in Ennodius) occur only in a few scattering instances, as do also several forms not mentioned by Meyer, for instance rsj r>J I rO r\j , r\j r^j | rO u rv/ , rO rv I i^ rO rv u no &C. i QC lact then is that only one of the possible rhythmical varieties of each of the three original quantitative types is found commonly employed in the time of the transition from the quantitative to the rhythmical character. A count from each of the three authors gives these results: Sedulius Ennodius Fortunatus 1 a. rO rv I r^ rO r^ 81 37 0» 1 1 ■» I / ry/'-t times "dicoiis"before\ o ,. ., ,,,„-,.> Q lb. rvur^lrvjr^ 7( a quotation ) ^ <-^'°^^ ^"' ^ ^ 1 C. rCi rvi I rvj r\j r^u r\j 4 24 o 2a,. r^rsj\r\jr(j^rjror^r>j 68 18 80 rO rv I rO rv 1 1 r^ u r>j \ r\j rsj r>j 1 3 rOr>^|rC/urv 2 rOurv;|'"vrOurv 1 1 r^r\j\r\jrsjJ^^jrsj 8 r\jr\j\r\jrun>j^^r>j 1 ^urv|o>^rvrv^r\^l 2 It is strange that the words of his former polemic against the accentual theory (Abh. d. bayr. Akad. XVII, p. 9) did not occur to Meyer in this connection: „Dass aber wirklich Virgil, Ovid und ihre Nachfolger im Hexameterschluss nicht Ueberein- stimmung der Wort- und Versaccente erstrebten, geht daraus hervor, dass sie auch SchlUsse wie ,res repararej Tyndaridarum j — 86 — armamentis', obwohl hier der Wortaccent treffllch gewahrt wurde, dennoch nicht minder gemieden haben als jene , aquae mons'. Nur rhetorische Grlinde waren es also, um derentwillen erst diese feinen Dichter die Regel ausbildeten, der Hexameter solle weder mit einem einsilbigen noch mit einem vier- oder mebrsilbigen Worte schliessen'^ The conditions in the two cases — the hexameter-ending and the rhetorical clausula — are identical, with the exception that it can not have been "rhetorical grounds" which caused the dis- appearance of the endings — u — | — .y. and — u — | —u^ in prose, since they are, besides being accentually unobjectionable, at least as well balanced as the endings which were preserved. If we look at the problem from the quantitative side, we find that the disappearance of the two varieties of clausula just mentioned was a gradual one. In writers of the fourth century like Ausonius , Symmachus and Sulpicius Severus , whose lapses from correct quantity are surely rare enough to entitle their endings to he called quantitative, the two varieties mentioned are proportionally very much less numerous than in the earlier writers. Of the writers of the fourth century, Ammianus Marcellinus alone stands near the older practice. and and 21 31 21 12 17 31 8 18 9 10 3 1 5 1) Only the endings before the heavy pauses were couuted. The endings — <-< I — 1 — — and — ^1 — 1 — ^ ~ which may belong to either column, were omitted. The figures for Cyprian were taken from the article of Meyer. The numbers of the second column must of course be taken in their relation to those of the first column, from ten to twenty pages from each author being included in the count. Cicero ^) 67 Seneca 65 Pliny 35 Minucius Felix 59 Cyprian 90 Firmicus Maternus 66 Ausonius 75 Ammianus Marcellinus 44 Symmachus 90 Sulpicius Severus 80 Sedulius 93 Ennodius 97 Fortunatus 88 - 87 — The phenomena which we have here are exactly parallel with those in iambic and trochaic verse. There also the accentual de- velopment had made great progress by the year 400, though the rules of quantity were, all in all, strictly observed; there as here, by the time of Sedulius the transformation had practically reached its end. The cause of the transformation was the same in prose as in poetry (cf. p. 52 ff.). If we allow the unstable final syllable of the word to stand in those places where it was not necessary to distinguish between long and short, and where it would naturally be placed in the attempt to maintain the correct quantity, then we shall have, in each of the three types of the quantitative clausula, those particular varieties which we find continued in the early rhythmical period, the accent in each case coinciding with the ictus: a !_/ I _a_ \j_ a \j_ I _a_ ^ \j^ a , Somewhat later , when an eiFort to preserve the quantity ceased to he made, a restriction to these three forms was of course no longer necessary, and we find a period of several cen- turies beginning about the time of Gregory the Great, when they were not so strictly and exclusively used. Finally they were restored to their former supremacy in the rhythmical form by the rules of the dictatores in the 11th and 12 th centuries. At what time the coincidence between accent and ictus began to he consciously felt as the structural principle in verse and prose, is a question into which we need not here enter. Only of this we may be quite certain, that it was not before but after the beginning of the "rhythmical" practice. I I List of Tables: Appendix II. A. Cases of Resolution (iambic dimeter). B. „ „ Elision „ „ C. „ „ Hiatus „ „ D. Total words beginning in theses (except the first) aad words „ rj n with a vowel (iambic dimeter). E. Syllables in the interior undetermined thesis arranged according to their location in the word, (iambic di- meter). F. The Metrical faults in arsis according to their location in the word (iamb. dim.). Iambic Dimeter. A. Ifletrical -- date certain. A 1. Catalecta (Virgil) . . . . Horace Martial Sec. Cent. Frag A 2. Ausonius , . Ambrosius Pxudentius Martianus Capella . . . . A 3. Sedulius Ter hora trina volvitur . . lam sexta sensim solvitur . Hie est dies verus dei . . Ennodius ....... Boethius Fortunatus Bede D. Begin- E. ' Syllables F. Word position 0^ CO d ning in in undeter- of faulty syllables > 1 CO the sis. mined thesis. in arsis. O 03 c g c (V 35 "* > o ce a "3 to o3 C9 g 3 < CQ d o is 'a J3 a Si to ti a 3 C3 •T3 a o % 20 __ 2 6 21 6 3, 5 12 203 4 33 190 33 13 120 80 31 U 1 26 7 4 15 12 38 12 11 38 10 26 12 398 54 71 357 77 34 253 111 1 1 188 2 7 1 189 15 24 130 34 2 1 400 23 64 363 84 44 265 91 28 13 1 27 92 2 81 3 14 73 5 2 32 4 28 3 4 23 5 36 2 1 23 6 7 29 1 32 1 5 44 3 4 14 14 160 8 116 15 14 134 12 1 39 1 1 25 3 1 27 11 124 1 8 99 18 23 96 5 3 1 225 14 195 32 34 183 8 1 89 A 4. Fratrcs .ilacri (Taul. Diai'on) Petro pi'tra ('cclcsiae . . Von it (lens fact us lioino . . Cliristiis re(leini)tor plebinir. Laiulein beati (Wal. Strabo) All te pulorum (Floras Liigd.) Ailest (lies voriis dei . Voiitosa cum desaeviat . . Aniande praesul optime . . B. Metrical — date uncertain. B 1. Attributed to Gregory Apostolorum supi)arem . Apostolorum passio . . lUuminans altissimus . . Aeterne rex altissime . . Priraatis aulae caelicae . Stephano coronae martjrnm Bellator armis inclitus Marline confessor dei . , Hymnum sacra novum die Victor, Nabor, P'elix pii . Amore Cliristi nobilis . . Festum beati martyris Festum beati martyris Aeteraus orbis conditor . Cliristi caterva clamilat . luluxit orbi iam dies . Veni creator optime . . Caeli deus sauctissime Plasmator hominis, deus . Verbum supernum prodiens lesus refulsit omnium . Lucis creator optime . . Chorus novae lerusalem . Magnae deus potentiae . Agnoscat omne saeculum In matutinis surgimus . . Grates tibi lesu novas Nunc saucte uobis spiritus Rector poteus, verax deus Rerum deus teuax vigor . Aeterua caeli gloria . . Aurora iam spargit polum Aguis beatae virginis . . B 2. Diem sacrati hominis . . Anni recurso tempore . . Vox clara ecce intonat . D. Begin- ning in tiiesis. E. Syllables in undeter- mined thesis. F. Word position of faulty syllables ill arsis. 2M pyj 82 28 24 20 32 24 2(1 'd-I o2 i2r. ■A2 o2 24 IG 12 1( 82 IG 20 IG 82 15 32 8 178 35 34 38 28 18 14 25 22 18 29 2G 22 28 115 28 27 11 15 IH 12 85 14 17 13 24 y 34 5 8 8 14 9 86 16 13 10 193 20 20 13 21 16 14 22 16 10 23 18 24 25 77 24 26 22 12 7 14 2(1 13 13 12 28 12 21 7 5 5 14 9 17 16 19 15 14 - 90 Fortem fidelera militem Uiiam duorum gloriam lam meta uoctis transiit Decus sacrati nominis Cousors paterni luniinis In Trinitatis unitas Cibis rcsumptis congruis Ternis ter horis numerus lesu corona celsior . . Tu Christ e, nostrum gaudium Te, lucis auctor, personant Beata nobis gaudia . . Ad Katberinae nuptias . Votiva cunctis orbita . . In laude regis omnium Rex gloriose martyrum . Fit porta Christi pervia . lux beata, trinitas . . lesu quadragenariae . . lam lucis orto sidere . . lesu, salvator saeculi . . Quem terra, pontus, aetbera Somno refectis artnbus Summae Deus clementiae Te lucis ante terminum Rerum creator optima Nox atra rerum contegit Dei fide, qua vivimus . lam cursus borae sextae Convexa solis orbita . Deus tuorum militum . lesu corona virginum . C Uniiietrical Theses. Deus qui certis legibus Aeternae lucis conditor . Fulgentis auctor aetheris Lucis largitor splendide . Certum tenentes ordinem Dicamus laudes domino . Perfectura trinum numerum Tempus noctis surgentibus lam Christus astra ascenderat Ignis creator igneus. D. Rliythinical- nndeVeloped. D 1. Bis teruas boras explicans . Rex aeterne domine . . . D. Begin- E, Syllables F. Word position CO d ning in in undeter- of faulty syllables o u a o o < a o 03 to thesis. mined thesis. in arsis. o o > is '3 c O "5 '■a -a a a a 5 o V a o u 3 a o OS I. Oj 'C3 c -d c o 32 1 1 20 2 4 28 28 6 '24 7 6 22 12 1 7 12 2U 1 16 2 4 16 V2 6 4 8 2(J 2 13 3 17 12 1 14 2 1 11 12 12 3 9 2 82 4 16 6 26 12 8 1 12 2-4 1 1 16 3 6 18 24 22 2 7 17 2 24 1 14 1 4 20 1 ir. 9 1 1 15 1 20 1 11 1 3 17 3 1 12 1 4 6 6 1 12 1 10 3 2 10 1 8 2 8 20 2 12 2 10 10 1 1 l(i 14 1 6 10 IG 12 3 3 13 2 2 32 1 2 17 3 32 1 16 13 2 6 10 1 Ki 3 10 3 13 8 8 3 5 1 10 10 2 14 1 IG 6 2 14 12 10 1 2 10 1 1 20 1 9 1 19 2 16 9 1 15 32 4 19 1 5 27 1 1 16 5 1 16 20 17 3 1 19 24 19 2 1 23 3 20 1 21 1 2 17 1 2 32 20 1 4 2b 1 1 1 12 1 8 1 4 8 1 2 12 3 10 2 3 9 1 1 12 11 2 2 10 2 16 1 15 1 6 9 1 3 32 1 2 18 3 8 22 2 32 3 1 28 2 2 28 2 1 1 3^ ■» 2 1 31' ^ & 9 15 8 4 13 64 3 1 14 44 . 8 23 39 2 9 22 1 - 91 Optatus votis omnium . . lam Cbriste, sol iustitiae Ad ceuam agiii provitli . Ilegi polorum tlebitos . . Rex Christe, Martini decus Sacri triiimpbale tui . . Adest dies sauctissima A sol is ortns cardine . . Deus, qui caeli lumen es . Deus pater ingenite . . Mediae noctis tempus est Deus aeterni luminis . . Diei luce reddita . . . Deus qui claro lumine Christe redemptor gentium Mysterium ecclesiae . . Stephano primo martyri . Post matutinas laudes. . D 2. Christe qui lux es et dies Aurora lucis rutilat . . Nunc tibi, virgo virginum Magni palmam certaminis rex, rector regminis . Couditor alme siderum . Mysteriorum signifer . . lesu nostra redemptio . . Hymuum dicamus domino Magno salutis gaudio . . lam ter quaternis trabitur lam lucis splendor rutilat Meridie orandum est . . £. Rhythmical- developed. Petrus Damiani . . . Anselm of Canterbury . Gallus et Vulpes . . Sacerdos et Lupus . . Beruhard of Clairvaux Bonaventura . . . . Thomas Aquinas . . . M D. Begin- E. Syllables (F.Word position a> d nin % in in unde ter- of faulty syllables u > o H O 3 tiiesis. mined tiiesis. in arsis. O 4J o5 5 5 o "rt u O) 1 rt c CO a a> o o cs a < 32 'J) o !2 1 a 5 a a V L, ^ 0) -53 O O OJ OJ >'^ a ^ t-i .2 § oj ^ 32 2 1 22 2 5 24 3 7 3 20 1 19 2 5 14 1 1 8 24 6 17 4 6 16 2 4 8 o2 1 10 28 8 8 23 1 5 6 5 8(1 2 44 8 8 24 4 4 9 1 32 4 27 4 5 22 5 3 5 3 4 1 20 2 10 4 15 1 5 5(i 6 2 38 10 7 48 1 5 14 2 1 40 2 3 33 4 2 31 4 2 6 20 2 9 2 4 14 2 2 ] 4 52 2 5 41 7 7 41 4 9 2 9 3 2 24 1 5 34 9 6 6 12 3 2 4 36 32 2 4 25 7 2 1 5 le 1 11 4 1 14 1 1 1 (| 24 2 26 6 4 12 8 6 1 I) 32 1 2 13 33 15 9 22 1 8 9 3 I 32 21 8 5 25 2 5 1 10 1 2 16 11 5 3 11 2 1 7 2 24 1 2 15 3 3 21 2 17 4 1 44 1 3 27 7 10 34 7 23 1 I 24 3 8 1 2 22 2 1 11 1 32 2 18 7 14 \^ 6 17 3 20 5 19 7 5 15 3 1 6 2 20 1 16 5 7 13 3 7 2 1 3-2 4 15 5 4 28 3 9 20 1 3 14 4 3 17 8 10 32 2 4 23 6 8 24 7 16 2 1 48 1 37 5 4 44 4 11 16 1 9 1 9 14 7 3 12 1 3 1 1 11 7 12 2 9 3 7 5 2 5 120 34 86 23 1 67 1 3 232 62 170 28 9 111 8 288 125 160 3 44 1 137 6 10 80 20 58 2 7 2 53 2 192 81 110 1 29 2 100 3 6 150 58 92 15 1 57 1 7 24 6 18 6 14 1 Kf ig^t^r 'n UC BERKELEY LIBRARIES