3 1 822 02677 3515 y'^*^w*'^S*-ra^>?^?. : '^^V^ r , JBRARY UNIVERSITY OP CALIFORNIA SAN DIE<50 V. ts-j UNIVERSITY OF CALIFORNIA, SAN DIEGO 3 1822 02677 3515 s '1 THE PRACTICAL STUDY OF LANGUAGES Sweet's practical Stubs of ^Languages. By Dr. Henry Sweet of Oxford, England. xiv + 22o pp. i2tno. $1.50, net. The body of this book is given to the discussion (rein- forced by telling illustrations) of practical problems which daily confront the teacher of languages — the choice and proper use of grammar and dictionary, the selection of reading matter, the use of translation, composition, con- versation, etc. An exposition of the phonetic basis of language-study occupies the first few chapters. JBreal's Semantics. Studies in the Science of Meaning. By Prof. Michel Bkeal of the College de Fiance. lxvi+336pp. i2mo. $2.50, net. A work on the science of Significations, as distinguished from the science of Sounds (Phonetics). A knowledge of this science is of vital importance to the study of Compara- tive Philology. It shows how individual and national char- acter may be deduced from an examination of language, and discusses hundreds of interesting allied topics. The style is pleasing, and the enjoyment of the book requires no previous philological training. lOtilliatns' ©ur Dictionaries, anb ©tber jEngltsb Slanguage tropics. By Ralph Olmsted Williams. With four plates. 174 pp. i2mo. $1.25, retail. WUlltams' Some Questions of <3ooo Bnglisb. Examined in Controversies with Dr. Fitzedward Hall. 234 pp. i2ino. Si. 75, retail. Prof. Albert S. Cook of Yale, in the Journal 0/ Ger- manic Philology : " When Mr. Williams' previous book ap- peared, I said of it, ' There is a lightness of touch, a reserve of manner, a crispness of style, a seasoning of humor, an absence of hammering, which ought to make the book much more popular than I dreamed a book on such a subject could be, without suffering some species of degradation.' In the present book these qualities are equally marked." JBalg's JEur=Brgan IRoots, With their English Derivatives systematically arranged and compared with Corresponding Words in the Cog- nate Languages. Vol. I., with an index of all English derivatives. 8vo. $15.00, net, special. IRambeau & {passes Cbrestomatbie ipbonetique. French Reader, with Phonetic Transcriptions. By Prof. A. Rambeau of Johns Hopkins and Prof. Jean Passy of Paris, xxx -f- 250 pp. 8vo. $1.50, net. Some hundred pages of easy French in common orthog- raphy, and on opposite pages the same matter in phonetic script like that in Paul Passy's Maitre Phone'tique, with an introduction explaining the method of using the text. HENRY HOLT & CO. 29 W. 23d St., New York 378 Wabash Ave., Chicago THE PRACTICAL STUDY OF LANGUAGES A GUIDE FOR TEACHERS AND LEARNERS BY HENRY SWEET, M.A., Ph.D., LL.D. CORRESPONDING MEMBER OF THE MUNICH ACADEMY OF SCIENCES FORMERLY PRESIDENT OF THE PHILOLOGICAL SOCIETY With Tables and Illustrative Sluotations NEW YORK HENRY HOLT AND COMPANY 1900 PREFACE This book is intended as a guide to the practical study of languages. Its object is, first, to determine the general principles on which a rational method of learning foreign languages should be based, and then to consider the various modifications these general principles undergo in their appli- cation to different circumstances and different classes of learners. The want of such a guide has long been felt. All the works on the subject that have hitherto appeared have either been short sketches, or else have only dealt with portions of the subject, such as the teaching of classical or modern languages in schools. I have given careful attention to these questions, but have by no means confined myself to this branch of the subject. I have rather endeavoured to give a comprehensive general view of the whole field of the practical study of languages, as far as lay in my power. I have not only given special sections on the learning of dead languages and of Oriental languages, but have also added a chapter on the methods of deciphering writings in unknown languages and of dealing with unwritten forms of speech; for although such investigations have not vi PREFACE always a directly practical aim, their methods are wholly practical. This part of the book ought to be welcome to travellers and missionaries, who often feel great perplexity when confronted with the difficult problem of reducing an illiterate language to writing and analysing it grammatically. The same remarks apply with equal force to dialectologists, the results of whose labours are often worse than useless through their want of proper method. Another class of students whom I have had specially in view are self-taught learners of foreign languages, who often not only waste time, but fail to attain their aim through following bad methods and using unsuitable text-books. My examples are taken from a variety of languages, partly to avoid one-sidedness of treatment, partly to interest as many different classes of readers as possible. In discussing methods, I have drawn my illustrations from those books which I know best. The time has not yet come for an historical survey and critical estimate of the vast and increasing literature of linguistic pedagogy, either of that portion of it which deals with generalities and criticisms of methods, or that still larger portion which carries out — or pro- fesses to carry out — these general principles in practical text- books — reading-books, grammars, text-editions, ' methods,' etc. In giving warning examples of mistakes into which learners may fall, I have confined myself to those made by foreigners in speaking and writing English, for the simple reason that the mistakes made by English-speakers in the use of other languages, though in themselves equally instructive and amusing, would have no point for the majority of my readers. PREFACE vii From the point of view of the purely practical learner, my treatment may perhaps appear not only too comprehensive, but also too ideal. He will ask, What is the use of recommending a method of study which cannot be followed because of the want of the requisite helps in the way of text-books ? But this is precisely one of the objects of my book. My object is both to show how to make the best of existing conditions, and to indicate the lines of abstract research and practical work along which the path of progress lies. In the present multiplicity of methods and text-books, it is absolutely necessary for real and permanent progress that we should come to some sort of agreement on general principles. Until this is attained — until every one recognizes that there is no royal road to languages, and that no method can be a sound one which does not fulfil certain definite conditions — the public will continue to run after one new method after the other, only to return disappointed to the old routine. My attitude towards the traditional methods is, as will be seen, a mean between unyielding conservatism on the one hand and reckless radicalism on the other. There are some fundamental principles on which I insist, whether they are popular or not, such as basing all study of language on phonetics, and starting from the spoken rather than the literary language. But, on the other hand, the reader will find that while I agree with the Continental reformers in condemning the practice of exercise-writing and the use of d priori methods such as Ahn's, I refuse to join with them in their condemnation of translation and the use of grammars. As regards my qualifications for the task, I have, in the viii PREFACE first place, acquired a considerable knowledge of a variety of languages of different structure ; and in studying them I have always paid as much attention to the practical as to the purely philological questions that have suggested themselves. I may also claim the merit of having made the scientific historical study of English possible in this country by the publication of my numerous practical helps to the learning of the older stages of our language, especially Old English. At the same time, my Elementarbuch des gesprochenen Englisch has done something towards making genuine spoken English accessible to foreigners. I have, lastly, had considerable experience in lecturing and teaching in connection with various branches of the study of languages, so that this work is as much the outcome of varied practical experience as of scientific theorizing. The first draft of this work was written out as far back as 1877. but for various reasons was never published, although an abstract of it appeared in the Transactions of the Philological Society for 1882-4, under the title of The Practical Study of Language. I need hardly say that the present work is not merely an expansion of these earlier efforts, but is the result of more matured thought and wider experience, so that it is an entirely new book, except that the chapter on 'mind- training ' is taken without alteration from the first draft. Oxford, February, 1899. CONTENTS PAGE Phonetic Symbols xv CHAPTER I THE STUDY OF LANGUAGES Practical and Theoretical Study I Necessity of General Principles I Good and Bad Methods 2 CHAPTER II PHONETICS Phonetics not an Innovation 4 Fallacy of Imitation 5 Fallacy of Minute Distinctions 5 Methods of Study : Organic and Acoustic 6 Isolation of Sounds 7 Analysis of the Formation of Sounds 7 Deducing Unfamiliar from Familiar Sounds 7 Relation of Native Sounds to Sounds in General 8 CHAPTER III PHONETIC NOTATION Unphonetic Spelling ; Nomic Spelling II Fullness of Transcription 12 Relation of Nomic to Phonetic Spelling 13 Remedies : Additional Marks and Letters 14 Principles of Phonetic Notation 16 National and International Basis 1 7 x CONTENTS PAGE A Universal Alphabet Unpractical 18 Significant Sound-distinctions 18 Superfluous Sound-distinctions 19 Modifiable General Basis 19 Non-Roman Basis : Organic Alphabet 20 Analphabetic Basis 22 The Alphabetic Basis the Best 24 Universal Alphabet not suited for Connected Writing 25 Superiority of Phonetic Shorthand 26 Modified Nomic Spelling 28 CHAPTER IV FOREIGN ALPHABETS Transliteration of Foreign Alphabets 30 Orthographic Transcription 32 Nomic Pronunciation 34 Learning a Foreign Alphabet 36 CHAPTER V VARIETIES OF PRONUNCIATION Artificial Pronunciation 38 Degrees of Colloquialism 40 Standards of Pronunciation 42 Pronunciation of Rare Words 43 CHAPTER VI GENERAL STUDY OF PHONETICS Apparatus : Diagrams, Models, Phonograph 45 Experimental Phonetics 46 Phonetic Dictation 47 Advantages of Phonetics 48 CHAPTER VII BEGIN WITH THE SPOKEN LANGUAGE The Spoken the Source of the Written Language 50 Practical Considerations 51 CONTENTS xi CHAPTER VIII DIFFICULTIES OF LANGUAGE PAGE External Difficulties 54 Relations to the Native Language 54 Internal Difficulties 56 Phonetic Difficulties 61 General Difficulty of each Language 63 The Real Difficulty is in the Vocabulary 66" All Languages Equally Difficult 66 CHAPTER IX GENERAL PRINCIPLES OF METHOD Language Only Partly Rational 70 Irrational Combinations in Language : We Cannot Speak by Rule . 71 The Arithmetical Fallacy 73 Isolated Phenomena of Language : Grammar and Dictionary ... 74 The Natural Method 75 Residence Abroad 76 Speaking Foreign Languages at Home 78 Natural Aptitude 79 National Aptitude 82 One Method for All 83 The Historical Method 86 The Crude Form System 87 The Etymological Fallacy 88 Comparison with Cognate Languages 89 Comparative Philology Sometimes Useful 90 Chance Resemblances between Languages 91 Borrowed Words 91 CHAPTER X SPECIAL PRINCIPLES OF METHOD Rules ; Mechanical Isolation 93 Analysis and Synthesis 98 Paradigms 99 Learning Lists of Words 100 Detached Sentences ; Context 100 Association 103 xii CONTENTS PAGE Memory; Repetition no Interest "2 Relations between Texts, Grammar, and Vocabulary 115 Stages of Progressive Method : Irregularities 1 18 CHAPTER XI GRAMMAR Accidence and Syntax 124 Formal and Logical Syntax 125 Grammar and Dictionary 126 Accidence and Syntax Taught Together 126 Stages of Grammatical Analysis 128 Grammar Learnt Unconsciously 129 Evils of the Separation of Syntax from Accidence 131 Examples 131 Paradigms 135 Fullness of Treatment 138 CHAPTER XII THE DICTIONARY ; STUDY OF THE VOCABULARY Scope 141 Pronouncing Dictionaries 144 Fullness 145 Conciseness 146 Surveyability 147 Meanings 148 Quotations 149 References 149 Grammatical and Other Information 150 Arrangement, Word-orler 151 Logical Dictionary 153 Study of the Vocabulary of a Language 1 58 CHAPTER XIII TEXTS ; THE READING-BOOK Classification of Texts 164 Connectedness 169 Length 170 4 CONTENTS xiii PAGE Clear Context 172 Limited Vocabulary 173 The Most Necessary Elements given First 174 Familiarity of Subject 175 Simplicity of Language 177 Variety 178 Gradation of Difficulties 178 Interest 179 Literary Texts 181 Condensed Treatises 183 Subordination to Form ; Grammatical Texts 184 CHAPTER XIV RELATIONS BETWEEN DIFFERENT LANGUAGES; TRANSLATION Thinking in the Foreign Language ; Not Translating 198 Translation from the Foreign Language 199 Translation into the Foreign Language ; Exercises 202 Free Composition ; Question and Answer 206 Visualizing 209 CHAPTER XV CONVERSATION Phrase-books 212 CHAPTER XVI literature; LITERARY COMPOSITION Composition 220 CHAPTER XVII DEAD LANGUAGES External Difficulties 223 Normalizing 224 Pronunciation 225 Teaching through the Literature 227 Cross-associations with Modern Languages 227 Dead Methods in Modern Languages 230 xiv CONTENTS CHAPTER XVIII ORIENTAL LANGUAGES PA<~.E Adherence to Native Methods , . 233 Texts 234 CHAPTER XIX GENERAL CONSIDERATIONS Time and Effort . . 236 Results ; Stages and Degrees of Knowledge 238 Epitomes and Note-books 242 The Subject-matter of the Texts 243 Teaching Children 244 Methods for Adults ; Self-instruction 247 CHAPTER XX ORIGINAL INVESTIGATION Decipherment 256 Help afforded by Comparative Philology 259 Decipherment a Practical Problem 262 All Text-reading Implies Originality 263 Text-editing : Original Research 263 Investigations of Unwritten Speech 264 Collecting Materials 267 Principles of Collecting 271 CHAPTER XXI mind-training ; classical and modern languages .... 273 Appendix 279 PHONETIC SYMBOLS Phonetic writing enclosed in ( ). Length marked by doubling, strong stress by ( • )> medium by ( : ), and weak by ( - ) before the syllable. ; then. a as in 'cut; ' also short of (aa). aa ,, 'father.' a ,, French ' pate.' a ,, French 'sans.' a ,, 'bird.' ae ,, 'man.' c = front stop. 5 as in German ' ich.' d — emphatic Arabic d. *\ • r e as in French ' ete ; ' also = (e). e „ ' men.' e ,, French ' vin.' a ,, 'sofa.' 6 ,, 'men, air.' 3 „ ' German 'sagen.' h = Arabic throat-sound ha i as in French ' fini ; ' also = (*')• i ,, 'fin.' i „ Welsh 'dyn.' j „ 'you.' J = front stop voice, k = deep Arabic k. \h = Welsh //. n as in Italian 'ogni.' o ,, French ' eau ;' also = (p). o ,, German 'stock;' also 'not.' 6 ,, French 'son.' ce ,, French 'peur.' ce ,, French 'un.' o ,, 'not ' ; ' saw.' ,, French 'peu.' s = Arabic emphatic s. f as in 'she.' vfh x y 3 ' thin.' French ' sou ; ' 'good.' 'what.' German ' loch. French 'une.' ' rouge.' also = (»). THE PRACTICAL STUDY OF LANGUAGES CHAPTER I THE STUDY OF LANGUAGES Practical and Theoretical Study It is hardly necessary to enlarge on the distinction between the practical and the theoretical study of languages — between learning to understand, read, speak, write a language on the one hand, and studying its history and etymology on the other hand. But it is important to realize at the same time that the practical study of languages is not in any way less scientific than the theoretical. The scientific basis of the practical study of languages is what may be called ' living philology,' which starts from the accurate observation of spoken languages by means of phonetics and psychology, and makes this the basis of all study of language, whether practical or theoretical. The opposite of living is 'antiquarian' philology, which regards the present merely as a key to the past, subordinating living to dead languages and sounds to their written symbols. Necessity of General Principles The first thing, therefore, is to determine the general prin- ciples on which the practical study of languages should be based. It is evident that if these principles are to be really general, they must be based on a survey of the whole field of I B 2 THE PRACTICAL STUDY OF LANGUAGES languages : that is, while giving due prominence to French and German, as being the two modern languages most generally studied in this country, we must not neglect the remoter languages, confining ourselves, of course, to an examination of a sufficient number of typical ones. 1 Having settled our general principles, the next thing is to consider what modifications, what special combinations of them may be required under special circumstances. It is evident that a method which suits an inflectional language may require modification when applied to a language of a different character; that learning to read a dead language is a different process from learning to speak a living one ; that self-instruction and teaching children in school require different text-books, and so on. As the tendency at present is to exaggerate rather than under- rate these differences, I shall confine myself as much as possible to general principles, leaving special modifications and applica- tions to be made by others. It would, indeed, be presumptuous in me to say much about such subjects as the school-teaching of languages, in which I have no practical experience — at least as teacher. I am not much concerned with such questions as, Why do we learn languages ? Is learning languages a good or a bad training for the mind ? Is Greek a better training for the mind than German or mathematics ? I start from the axiom that as languages have to be learnt, even if it turns out that the process injures the mind, our first business is to find out the most efficient and economical way of learning them. Good and Bad Methods The plan of this book involves, to some extent at least, a criticism of existing methods. In this connection it is significant to observe that though there is great conservatism in scholastic circles — as shown in the retention of antiquated text-books, in the prejudice against phonetics, and so on — there are, on the other hand, many signs of dissatisfaction with these methods. 1 Besides English, French, and German, I have drawn my illustrations chiefly from those remoter languages of which I have some practical knowledge, that is, Sanskrit, Welsh, Old Irish, Finnish, Arabic, and Chinese. THE STUDY OF LANGUAGES 3 This dissatisfaction is strikingly shown by the way in which new ' methods ' are run after — especially the more sensational ones, and such as have the good fortune to be taken up by the editor of some popular periodical. But none of these methods retain their popularity long — the interest in them soon dies out. There is a constant succession of them ; Ollendorff, Ahn, Prendergast, Gouin — to mention only a few — have all had their day. They have all failed to keep a permanent hold on the public mind because they have all failed to perform what they promised : after promising impossibilities they have all turned out to be on the whole no better than the older methods. But the return to the older methods is only a half-hearted one : even Ollendorff still has his adherents. In fact, things are altogether unsettled, both as regards methods and text-books. This is a good sign : it gives a promise of the survival of the fittest. Anything is better than artificial uniformity enforced from without. The methods I have just mentioned are failures because they are based on an insufficient knowledge of the science of language, and because they are one-sided. A method such as Gouin's, which ignores phonetics, is not a method : at the most, it gives hints for a real method. Gouin's ' series-method ' may in itself be a sound principle, but it is too limited in its appli- cations to form even the basis of a fully developed method. A good method must, before all, be comprehensive and eclectic. It must be based on a thorough knowledge of the science of language — phonetics, sound-notation, the grammatical structure of a variety of representative languages, and linguistic problems generally. In utilizing this knowledge it must be constantly guided by the psychological laws on which memory and the association of ideas depend. CHAPTER II PHONETICS The main axiom of living philology is that all study of language must be based on phonetics. Phonetics is the science of speech-sounds, or, from a practical point of view, the art of pronunciation. Phonetics is to the science of language generally what mathematics is to astronomy and the physical sciences. Without it, we can neither observe nor record the simplest phenomena of language. It is equally necessary in the theoretical and in the practical study of languages. Phonetics not an Innovation The necessity of phonetics has, indeed, always been tacitly recognized — even by its opponents. Even such a simple statement as that ' English nouns take -es instead of -s in the plural after a hiss-consonant ' involves elementary facts of phonetics ; the terms ' vowel ' and ' consonant,' ' hard ' and 1 soft,' all imply phonetic analysis. What the reformers claim is not that phonetics should be introduced — for it is there already — but that its study should be made efficient by being put on a scientific basis. In fact, phonetics is almost as old as civilization itself. The Alexandrian grammarians were not only phoneticians — they were spelling-reformers ! Few of those who mechanically learn the rules of Greek accentuation by way of gilding the refined gold of their scholarship have any idea that these to them unmeaning marks were invented by the Alexandrian grammarians solely for the purpose of making the pronunciation of Greek easier to foreigners. The Romans, too, were pho- neticians : they learnt Greek on a phonetic basis, as far as their lights allowed them. The Sanskrit grammarians were still better phoneticians. It is the unphonetic, not the phonetic methods that are an innovation. The efficient teaching of phonetics is impeded by two popular fallacies. PHONETICS 5 Fallacy of Imitation The first of these is that pronunciation can be learnt by mere imitation. This is as if fencing could be learnt by looking on at other people fencing. The movements of the tongue in speaking are even quicker and more complicated than those of the foil in fencing, and are, besides, mostly concealed from sight. The complicated articulations which make up the sound of such a French word as ennui cannot be reproduced correctly by mere imitation except in the case of an exceptionally gifted learner. Even in the case of children learning the sounds of their own language, the process is a slow and tedious one, and the nearer the approach to maturity, the greater the difficulty of acquiring new sounds. Indeed, the untrained adult seems to be often absolutely incapable of imitating an unfamiliar sound or even an unfamiliar combination of familiar sounds. To the uneducated even unfamiliar syllables are a difficulty, as we see in 1 familiarizations ' such as sparrow-grass for asparagus. 1 Even those who devote their lives to the study of languages generally fail to acquire a good pronunciation by imitation — perhaps after living ten or twenty years in the country and learning to write the language with perfect ease and accuracy. Fallacy of Minute Distinctions The second fallacy is that minute distinctions of sound can be disregarded — or, in other words, that a bad pronunciation does not matter. The answer to this is that significant dis- tinctions cannot be disregarded with impunity. By significant sound-distinctions we mean those on which distinctions of meaning depend, such as between close and open e in French pecker, pecker. We see from this example that significant sound-distinctions may be very minute — or at least may appear so to an unaccustomed ear. To a native ear they always seem considerable. Thus to English people the distinction between the vowels of men and man, /lead and had, seems a very marked one, while to most foreigners it seems but a slight one : many Germans are apt to confound head, had, hat under the one pronunciation het. 1 I knew a child who used to make giraffe, facsimile, chiffonier into edgiruff, face smile, and shove anear respectively. 6 THE PRACTICAL STUDY OF LANGUAGES Nor can we tell a priori what sound-distinctions are significant in a language : a distinction that is significant in one language may exist as a distinction in another, but without being signifi- cant, or one of the sounds may be wanting altogether. Thus in ordinary Southern English we have no close e at all ; while in the North of England they have the close sound in such words as name without its being distinctive, for it is simply a con- comitant of the long or diphthongic sound of e. Experience shows that even the slightest distinctions of sound cannot be disregarded without the danger of unin- telligibility. The friends of the late GuSbrand Vi'gfiisson, the well-known Icelander, still remember how he used to complain that the country people round Oxford could hardly be made to understand him when he asked for eggs : ' I said ex — I ought to have said airx.' Here the remedy was almost worse than the disease; and yet what suggested eks to an English ear differed only from the correct pronunciation in having whisper instead of voice in the first as well as the second consonant ! Methods of Study : Organic and Acoustic The first business of phonetics is to describe the actions of the organs of speech by which sounds are produced, as when we describe the relative positions of tongue and palate by which (s) is produced. This is the organic side of phonetics. The acoustic investigation of speech-sounds, on the other hand, describes and classifies them according to their likeness to the ear, and explains how the acoustic effect of each sound is the necessary result of its organic formation, as when we call (s) a hiss-sound or sibilant, and explain why it has a higher pitch — a shriller hiss — than the allied hiss-consonant (/) in she. It is evident that both the organic and the acoustic sense must be cultivated : we must learn both to recognize each sound by ear and to recognize the organic positions by which it is produced, this recognition being effected by means of the accompanying muscular sensations. We all carry out these processes every day of our lives in speaking our own language. All, therefore, that we have to do in the case of familiar sounds is to develope this unconscious organic and acoustic sense into a conscious and analytic sense. PHONETICS 7 Isolation of Sounds The first step is to learn to isolate the sounds and to keep them unchanged in all combinations and under all the varying conditions of quantity and stress (accent). Thus the learner may lengthen and isolate the vowels in pity, and observe the distinction between them and between the vowels of /// and peat. This method of isolation is a great help in learning foreign sounds. A teacher of French who has learnt to cut up such a word as ennui into (aa, nyy, ii) will, without any knowledge of phonetics, be able to give his pupils a much better idea of the pronunciation of the word than by repeating it any number of times undivided. Analysis of the Formation of Sounds The next step is to learn to analyze the formation of the familiar sounds. This analysis must be practical as well as theoretical. It is no use being able to explain theoretically and to hear the distinction between a breath consonant such as (f) and the corresponding voice consonant (v), unless we are able to feel the difference. Let the beginner learn to isolate and lengthen the (f) in life and the corresponding (v) in liver till he can feel that while (f) is articulated in one place only, (v) is articulated in two places — not only between lip and teeth, but also in the throat. If he presses his first two fingers on the ' Adam's apple,' he will feel the vibration which produces the effect of voice in (v), which vibration is absent from (f). If he closes both ears, he will hear the voice-vibration very distinctly. Deducing Unfamiliar from Familiar Sounds The great test of the practical command of such a distinction as breath and voice is the power it gives of deducing un- familiar from familiar sounds. Repeat (vvff) several times in succession, and try to carry out a similar change with the voice- consonant (1), and the result will be the Welsh (1/;) in llan. Again, to get the German or Scotch (x) in loch it is only necessary to exaggerate and isolate the ' off-glide ' of the (k) of the English lock. Often, indeed, mere isolation is enough to deduce an apparently unfamiliar sound. Thus the peculiar obscure a 8 THE PRACTICAL STUDY OF LANGUAGES and peculiar (s)-sound in Portuguese, as in amamos, are simply the first element of the diphthong in English how and the second element of the English (t/) in chin, which is distinct from the (/) in Jish, being really a sound intermediate between (/) and (s). It is interesting to observe that hearing such an unfamiliar sound as (1//) is a hindrance rather than a help to the beginner, who, hearing a sound which is partly a hiss and partly an (1), tries to do justice to the acoustic effect by sounding separately the familiar English hiss (b) in think and an ordinary voice (1), so that he makes (L$an) into (blsen). This is an additional argument against the imitation fallacy. But, as already remarked, the acoustic sense must be thoroughly trained, for in many cases the acoustic does help the organic analysis. 'Listen before you imitate' is one of the axioms of practical phonetics. Relation of Native Sounds to Sounds in General Before beginning the study of foreign sounds, it is important to get a clear idea of the relations of our own sound-system to that of sounds in general, and especially to learn to realize what is anomalous and peculiar in our own sound-system. Thus, when the English learner has once learnt to regard his (ei) and (ou) in such words as name and so as abnormal varieties of monophthongic close (ee, oo), he will find that much of the difficulty of pronouncing such languages as French and German will disappear ; he will no longer have the mortification of betraying his nationality the moment he utters the German word so. Indeed, speakers of the broad London dialect in which (ei) and (ou) are exaggerated in the direction of (ai) and (au) often become unintelligible in speaking foreign languages. Two young Englishmen abroad once entered into conversation with a French cure, and one of them had occasion to use the word beaucoup ; the Frenchman was heard repeating to himself (bauky) and asking himself what it meant. Each language has its own ' organic basis,' and the organic bases of French and English are as distinct as they can well be. Hence the importance of a clear conception of the character of each basis, and their relations to one another. CHAPTER III PHONETIC NOTATION Next to analysis, the most important problem of practical phonetics is that of sound-notation, or spelling by sound. The first and most obvious advantage of a phonetic notation is that the learner who has once mastered the elementary sounds of the language, together with the elementary symbols of the notation he employs, is able to read off any phonetically written text with certainty, without having to burden his memory with rules of pronunciation. To such a student the distinction, for instance, between close and open e and o in Italian offers no difficulties : he learns from the beginning to pronounce each word with the correct vowel. Another advantage of a phonetic notation is that as the learner sees the words written in a representation of their actual spoken form, he is able to recognize them when he hears them with comparative ease — or, at any rate, he is better prepared to recognize them. Most English people, when they first go to France, are unable to understand a word of the language when spoken, however well they may be able to read it. This is simply because the unphonetic French spelling they are used to represents not the spoken French of to-day, but the French that was spoken in the sixteenth century — being a very bad representation even of that. But if a foreigner has learnt to decipher such written forms as (aksebo) or ( a k s e bo !), (ktkstksa, kjtski), he would certainly be better prepared to understand them when spoken than if he had first to translate them in his mind into (aa ka ss ei bou) or something of that kind. Phonetic notation helps the ear in many ways. The spoken word is fleeting, the written word is permanent. However often the learner has the elements of such a word as ennui repeated to him, it is still a help to have the impressions of his car confirmed by association with the written symbols of such a io THE PRACTICAL STUDY OF LANGUAGES transliteration as (aanqi). If the phonetic notation, instead of confirming, corrects an impression of the ear, its utility becomes still more manifest. Teaching by ear alone throws away these advantages. It is certain that even the quickest linguist is helped by phonetic notation. Even if it were not absolutely required for the purpose of saving him from mishearings and mispronunciations, it would still serve to strengthen his hold of the spoken word. The consideration that the written word is permanent is enough to refute the objection sometimes made to phonetic spelling, namely, that it makes the language more difficult to understand. It is clear that if the learner cannot solve such a riddle as (aksebo) at his leisure, he will certainly not be able to solve it when he has only the fleeting impression on his ear to rely on. Unphonetic Spelling ; Nomic Spelling The question of phonetic notation is complicated by the fact that the traditional or ' nomic ' orthography of most languages is only partially phonetic. But even French and English are not wholly unphonetic. Even in English we find hundreds of such spellings as send, if, not, which even the most radical spelling-reformer need not alter, together with many others which would require only a slight change to make them wholly phonetic. Indeed, it is easy to see that a wholly unphonetic system of spelling — one in which every word was written with an absolutely arbitrary combination of letters — would be too much even for the most retentive memory. But even a little unphoneticness may cause a good deal of confusion and perplexity, as we see in the case of the two pronunciations of Italian e, o, z and of Welsh y — a language whose spelling is often said to be entirely phonetic. The want of stress-marks in English, and still more in Russian, is one of the greatest obstacles to learning to speak these languages, and sometimes gives rise to ludicrous misunderstandings. Thus a German staying in an English house, when summoned to dinner, told the servant that he was ' occupied ' and could not come yet ; but he put the accent in the wrong place, the result of which was that the assembled company was startled by the information, ' Please, sir, Dr. A. says he's a Cupid ! ' As Dr. A. was short and stout, amazement soon yielded to amusement. PHONETIC NOTATION n So difficult is the Russian stress, that an Englishman in Russia, when asked by another Englishman who was learning Russian to give him some simple rules for the accent, told him to try and find out what syllable the accent ought to fall on, and then to put it on some other syllable. Although German stress is on the whole regular, yet such a distinction as that between Hibersetzen, ' leap over,' and iiber'setzeti, ' translate,' is puzzling enough to the beginner. Fullness of Transcription Besides unphonetic writing which is positively misleading, there is another way of being negatively unphonetic by simply suppressing — not perverting — the phonetic information required. Thus, when a foreigner has to read aloud about ' the reform- bill of 1830,' it is no help to him to have it phonetically transcribed into (fts rifombil 9v 1830), if the numerals are not transcribed in full at the same time. It is still worse when an Englishman has to read straight off in French such a number as 1789. Vietor and Dorr are quite right in giving such texts as the following in their Englisches Lesebuch — except, of course, that it ought to be in phonetic spelling : — ' In the course of last month Jack saved elevenpence. Out of this he bought a few steel pens, for which he paid threepence, and a pot x of ink, which cost him twopence. The rest of his money was then just one small silver coin ; what is its name ? ' But they spoil it all by going on to give such 'texts' as the following : — Addition Table. ] and 2 and 3 and 4 and I are 2 1 are 3 I are 4 1 are 5 2 » 3 2 „ 4 2 „ S 2 „ 6 3 >, 4 3 » 5 3 ,. 6 3 ,» 7 4 » 5 4 „ 6 4 » 7 4 „ 8 5 „ 6 5 ,. 7 5 „ 8 5 -» 9 6 „ 7 6 „ 8 6 „ 9 6 „ 10 7 „ 8 7 „ 9 7 » 10 7 „ 11 8 » 9 8 „ 10 8 „ n 8 „ 12 9 „ 10 9 ,, 11 9 » 12 9 ., 13 10 » 11 10 ,, 12 10 „ 13 10 „ 14 11 „ 12 11 „ 13 11 ,, 14 11 11 IS 12 » 13 12 „ 14 12 „ 15 12 „ 16 Ought to be ' bottle.' 12 THE PRACTICAL STUDY OF LANGUAGES So also such a formula as a 2 + 2ab + b 2 ought to be also written in full (ei -skwe9 :plas tuw :ei bij :plas bij -skwea), although, of course, it would be out of place in an elementary book. On the same principle such contractions as lb., cwt., oz., fiy., ought to be written in full. But here a caution is necessary. It would be quite wrong to expand P.M. into (poust mi'ridjem), for we always pronounce this contraction literally — (pij em). M.A. may be read either as (:maast3r sv aats) or (em ei), as also M.P. and many others, the literal pronunciations being the most common. Relation of Nomic to Phonetic Spelling The first and most obvious objection brought against the use of a phonetic notation in teaching a foreign language is the danger of confusion between the phonetic and the nomic spelling of the language. A-priori theorists have argued that the result of beginning with a phonetic spelling will inevitably be ' to spoil the learner's spelling for life.' But all who have ever given the phonetic method a fair trial maintain that this objection has no practical weight. They assure us that their experience shows that when a language has once been thor- oughly mastered in a phonetic notation, the learning of the ordinary traditional spelling offers no difficulty : those who have begun phonetically end by spelling orthographically just as well as those who began at the same time with the ordinary spelling, and learn no other spelling, and thus were able to give much more time to it. The explanation of the quicker progress of the phonetic learners is, of course, that they are able to grasp the general idea of sound-representation easier and quicker by beginning with an easier — that is, a phonetic — spelling. It cannot, of course, be denied that the study of such a language as French would be easier if the divergence between its nomic and its phonetic spelling did not exist. But the difficulty of which this divergence is the expression is not the fault of phonetics : ignoring phonetics does not get rid of the divergence between the spoken and the written sounds of the language. All we can do is to minimize the difficulty; and the" first step towards this is the adoption of a phonetic transcription. PHONETIC NOTATION 13 The next question is, Which should be learnt first ? This amounts practically to the question, Which associations ought to be strongest ? Clearly those with the sounds : in speaking the associations between sounds and ideas must be instan- taneous, while in reading or writing we have time to stop and think. This is the order we follow in learning our own language : we speak before we spell. If children learnt by eye first, they would never speak properly — they would speak like foreigners who have begun with the literary language. The same kind of reasoning which forbids us to begin with the nomic spelling, forbids us also to learn the two simul- taneously. The only way of avoiding cross-associations is to begin with one of them and use it exclusively, and then — either for a time or permanently — use the other as exclusively. As we have seen, there is every reason why we should begin with the phonetic spelling, which, when it has served its purpose, may be put aside entirely. The relation between phonetic and nomic spelling is analogous to that between the tonic sol-fa notation and the ordinary staff-notation in music. The advocates of the former notation argue that the first thing is to learn the thing itself in the easiest way possible. They then go on to state as a fact, the result of experience, that when the thing music is once learnt, it does not matter so much what notation is used. The result of beginning with the tonic sol-fa notation is that thousands who would be quite unable to learn music from the ordinary notation, master it perfectly on the new system, and are then able with a little practice to read music at sight from the staff notation, so that even if their sole object is to learn the latter, they save themselves much toil and trouble by beginning with the tonic sol-fa notation. Remedies : Additional Marks and Letters The difficulties caused by unphonetic writing may be met in a variety of ways. Such a difficulty as that of the place of stress is only a negative one, and can easily be remedied by the addition of accents or other marks without any alteration of the nomic spelling. Nor does this kind of difficulty involve the same 14 THE PRACTICAL STUDY OF LANGUAGES amount of cross-association as the confusion between close and open e in Italian. Still worse are cross-associations involved in such a group of spellings as the English plough, enough, trough = (plau, inaf, trof), or those two which made the witty French philosopher express a wish that the plague (pleig) might take half of the English people, the ague (eigju) the other half. The defects of such comparatively phonetic orthographies as the Italian can be easily remedied by the application of diacritics as in ora (close), oro (open), or by the use of italics, which may also be used to indicate ' silent letters.' But any system which involves retention of the nomic spelling practically breaks down in the case of such languages as English and French. Here we must sooner or later come to the con- clusion that instead of trying to teach pronunciation not through but in spite of the nomic spelling, it is better to start with an entirely new phonetic spelling. The defects of the ordinary Roman alphabet may be supple- mented in a variety of ways : — i. By adding new letters — either entirely new, or taken from other alphabets : /, 3, 17; b, •$, 6, 8. 2. By adding diacritics : a, 6, h. 3. By utilizing superfluous letters : c, q, x. 4. By turned letters : s, o, j. 5. By italics and capitals : a, s, r. 6. By digraphs : th, dh, ny, Xh. Of these expedients the first is the most popular. As a general rule, the more ignorant and inexperienced the reformer, the more reckless he is in adding new types, although nothing is more difficult than to invent a new letter. The main objec- tion to new types is, of course, the trouble and expense of procuring them. The same objections apply also, though in a less degree, to diacritics, which, as Ellis says, 'act as new letters.' The best known of the diacritic alphabets is Lepsius's Standard Alphabet, in which seventeen diacritics are used above and fourteen below the letters, the number of lower-case letters employed being more than 280, of which 200 have to be cut specially for each fount. The four other expedients have the advantage of not requiring new types to be cut. PHONETIC NOTATION 15 Principles of Phonetic Notation The first requisite of a good alphabet is that it should be capable of being written and read with ease and written with moderate quickness. Simplicity. — For ease of reading, it is desirable that the letters should be as simple as is consistent with distinctness. From this point of view, the Roman letters are superior to the black-letter or Gothic forms still used in Germany, as we see especially in the capitals. Dots and other diacritics, which must be made small, tend to indistinctness. Compactness. — Ease of reading depends also greatly on compactness. Hence syllabic systems of writing like Sanskrit, in which such a syllable as skra is expressed by a single cha- racter, are in many respects easier and pleasanter to read than the corresponding Roman transcription. It is often a matter of surprise that the Chinese characters try the eyes so little, in spite of the great complexity and minute distinctions they often involve. The reason is that every word is represented by a compact square character, all the characters being of uniform size, the strain on the eyesight being further reduced by the arrangement of the characters in perpendicular columns. The superiority of the syllabic principle is strikingly shown by the fact that both the Protestant and the Catholic missionaries in Canada use syllabic alphabets in teaching the Crees and other native tribes to read, on account of the length to which the words run when written in Roman letters. These alphabets consist of simple characters expressing consonants, such as V» turned different ways — < > — to indicate what vowel follows. Joining. — Ease and quickness of writing require that the letters should be easily joined together, as may be seen by comparing a passage written in Greek letters with one in Roman letters. The most accurate way of estimating the comparative merits of letters as regards ease and quickness of writing is to count the number of strokes of which they are composed on some uniform plan. Thus i without a dot consists of one stroke, scrpit ^ of two, s of four. But this method of calculation leaves out of account the ' aerial movements ' of the pen from the line of writing to the diacritic and back again. We see now that writing the single letter s takes as much time as writing the five letters seeee ! 1 6 THE PRACTICAL STUDY OF LANGUAGES Printed Forms. — In printing, the complexity of the letters does not influence speed or ease : the main thing is to have as few types as possible. This is an additional reason for abolish- ing the use of capitals in phonetic writing — except for special distinctions. In printing it is easiest to have the letters detached. This is highly objectionable in writing, but is generally an advantage in reading. As regards the relations between the written and printed forms of the letters, it is evidently desirable to avoid un- necessary deviation without, on the other hand, attempting to make print into a — necessarily imperfect — imitation of handwriting. The disadvantages of such an attempt are well shown in Arabic, with its superfluous distinction of initial, medial, and final forms of one and the same letter, the maximum of discomfort being reached when the short vowels are indicated by diacritic strokes printed on separate lines, so that the reader is sometimes in doubt whether the diacritic is to be read above the consonant of the line he is reading or below the consonant of the line above. Some phonetic transcriptions — such as that of the Swedish Dialect Society and of Trautmann in his Sprachlaute — con- sist entirely of italics, so as to diminish the difference between the written and printed characters as much as possible, and also to make the phonetic writing stand out distinctly in a page of Roman type. But as italics are required for a variety of other purposes, and as it is a waste of existing material not to utilize the distinction of Roman and italic, it seems better to make the more .legible Roman the basis, and use italics for various supplementary purposes; it is always easy to mark off pho- netic writing by enclosing it in ( ). The transcription of the Danish Dialect Society Dania is so far an advance on the other italic systems that it utilizes Roman letters for special dis- tinctions of sound. Having thus determined the general principles on which the choice of symbols is founded, we come to the still more diffi- cult question, how to use these symbols — what sounds or what phonetic functions to assign to them. PHONETIC NOTATION 17 National and International Basis The most obvious way of making an unphonetic orthography phonetic is to select some one out of the various traditional representations of each sound, and use that one symbol exclu- sively, omitting at the same time all silent letters, and adding marks of stress (accent) if necessary, as in the following speci- men of Ellis's ' English Glossic : ' — ' Ingglish Glosik iz veri eezi too reed. A cheild foar yeerz oald kan bee taut too reed Glosik buoks.' A system which, like Glossic, writes short and long vowels with totally different symbols (i, ee) is only half-phonetic : it is phonetic on an unphonetic basis. Again, this unphonetic English basis breaks down altogether in some cases. It fails, for instance, to supply unambiguous symbols for the vowels in child and book, full and the consonant in the, which Ellis writes (dh). The following specimens of French and German spellings formed in a similar way on the basis of the respective nomic orthographies of these languages are taken from Soames's Introduction to the Study of Phonetics : — ' Deu pti gars> , i, pronouncing them all (s), and recognizing them only by the form of their symbols. But even in an extreme case like this there ought not to be any great difficulty in establishing visual associations between the Arabic letters and their transliterations s, s, z, z, 6, 8 (or ]?, $). This, however, only elicits fresh objections. The opponents of transliteration say, ' This would be plausible enough if we had only one fixed transliteration to learn ; but unfortunately 32 THE PRACTICAL STUDY OF LANGUAGES almost every text-book has a special transliteration of its own : one cannot even get a grammar and a dictionary with the same transliteration. It is therefore impossible to carry out your advice of keeping to one transliteration till one has mastered the Arabic alphabet.' The multiplicity of transliterations is certainly to be deplored, but it is no more an argument against the principle of trans- literation than the multiplicity of phonetic notations is against the phonetic method. The same influences which are steadily bringing us nearer to our ideal of a general basis of phonetic notation will doubtless bring about uniformity in the tran- scription of remoter languages as soon as the results of our experience with European languages become known to Orien- talists and others, who are still hampered by bad traditions and the unscientific methods of their native authorities to a degree which is incredible to those familiar with the phonetic method as applied to European languages. The great safeguard against confusions that arise from con- flicting transcriptions is the principle already insisted upon — that of beginning with the language itself, which of course means beginning with a mastery of its sounds. The beginner in Arabic who has once learnt to distinguish saif, ' summer,' from saif, ' sword,' by the combined associations of the peculiar sound and the special muscular sensations which accompany the utterance of the ' emphatic ' s, will be independent of trans- literations, for the ideas of ' summer ' and ' sword ' will at once suggest to his mind combinations of sounds as well as com- binations of letters, the former associations being the stronger and more direct : he will be in quite a different position from the student whose only definite associations are with the written and i — ?*** Orthographic Transcription If the national alphabet itself is phonetic, the transcription will be phonetic also : it will be a key to the pronunciation, and at the same time it will be a key to the original spelling of each word, so that any one who is acquainted both with the method of transcription and the national alphabet will be able to trans- literate the transcription back into the original writing. If the national alphabet is unphonetic, but only moderately FOREIGN ALPHABETS 33 so, the most obvious course is to follow the same method as in reproducing the manuscript spellings of dead languages ; that is, to add the necessary diacritics, or make whatever modifica- tions may be found convenient for the purpose of indicating pronunciation, so that all that is necessary to transliterate back into the national writing is to ignore these supplementary dis- tinctions. If the national writing makes unphonetic distinctions by having two or more letters or combinations of letters to express the same sound or sound-group, then the diacritics will have an orthographic, not a phonetic value, and will therefore be ignored except as giving the key to the original writing. We thus have a distinction between a purely phonetic and an orthographic transcription, the characteristic of the latter being that it can always be transliterated back into the national writing whether the latter is phonetic or not. It need scarcely be said that every orthographic transcription ought to be phonetic at the same time, or at any rate not markedly un- phonetic, although in many cases it is most practical to sacrifice rigorous phonetic consistency whenever an unphonetic detail of transcription does not cause real difficulty. Thus in tran- scribing German it is better to keep the distinction between sz and ss in fusz, musz, miissen, than to run the risk of subsequent confusion by writing fuss, muss ; for such a spelling as fuss is only a compromise between fusz and the fully phonetic (fuus), and not even a beginner would think of trying to pronounce sz exactly as it is spelt. The method of orthographic transcription has been success- fully applied to Persian by H. Barbs, a full account of whose transcription by K. Feyerabend will be found in Phonetische Studien, iii. 162. Persian in itself is generally considered one of the easiest and simplest of languages, but in its written form it is distinctly a difficult language because of the irregularity, complexity, and ambiguity of its alphabet and orthography. Without the help of a skilled and patient teacher it is hardly possible to learn it in its nomic form, because, as Feyerabend remarks, 'one can only read out of it what one has already learnt and knows.' Persian has the disadvantage of being written with an alphabet in every way alien to its genius — the Arabic. Hence such a defect as the omission of the short vowels — which in Arabic occasions much less difficulty than might be supposed because of the regularity and symmetry of the Arabic vowel-system — becomes very serious in a language u 34 THE PRACTICAL STUDY OF LANGUAGES like Persian, where there are no rules for determining a priori the vowel-structure of a word, as is to a considerable extent the case in Arabic. Persian is, besides, full of Arabic words, which are written in the Arabic orthography, while the pro- nunciation is only imperfectly preserved. The slavish applica- tion of Arabic rules of orthography to Persian words is a further source of unphonetic spellings. Barbs' transcription seems fully to solve the double problem of giving a phonetic tran- scription which can at the same time be transliterated back letter for letter into the national writing. The student begins with a Reader in which all the texts are transcribed on these principles. When he has gone through it, he begins again, and at the same time he is gradually introduced to the Persian alphabet and the rules of Persian orthography. Then a parallel Reader in the Persian writing is put into his hands, and the work of deciphering begins. Feyerabend assures us, as the result of personal experience, that this causes no difficulty in the second third of the first year's course ; for, as he says, 'we soon learnt to recognize our old acquaintances in their new dress.' Nomic Pronunciation The principle that in learning a language through written texts we should strengthen our associations with the characters by associating each character with its proper sound, and should avoid giving the same sound to letters which are pronounced differently (p. 30), cannot always be carried out literally. Sometimes the learner has not access to a native teacher or to reliable information about the pronunciation. These diffi- culties are of course greatly increased if he is learning a dead language. Under such circumstances the learner need not hesitate to make up a pronunciation of his own on the principle of accom- panying every written distinction with a corresponding difference of sound, so as to strengthen as much as possible his visual associations. Many foreigners have begun English in this way, pronouncing, for instance, knowledges three syllables (knovledge), not because they thought this was the real pronunciation, but simply as a means of fixing the spelling in their minds. FOREIGN ALPHABETS 35 G. von der Gabelentz — who united many of the qualifications of the theoretical and the practical linguist— goes a step further, and advises the beginner in Arabic who cannot pronounce p 'routsaundz) in MF 1895. 4. 2 See J. Passy's paper (la dikte fonetik) in MF 1894, pp. 34, 50. 48 THE PRACTICAL STUDY OF LANGUAGES the dormant faculty of phonetic observation. At first the dictation should be in the pupil's native language, and he should be expected only to write down the significant distinctions of sounds in some easy Romic notation without any attempt to mark stress or intonation. It is surprising to see what mistakes are made, partly through confusion with the nomic spelling, partly through complete absence of the faculty of observing even the broadest distinctions when unaided by visual associa- tions. When the pupils can write with fluency and correctness on this basis, they should be trained to add stress-marks, and then simple tone-marks. Then the same stages should be repeated in the foreign language. Advanced pupils in general phonetics may be cautiously exercised in writing down nonsense-words consisting at first of a certain limited number of sounds. Thus the teacher may tell them that all the vowels will be narrow, that there will be no mixed vowels, no front consonants, and so on. For this advanced dictation the organic alphabet should be used. Advantages of Phonetics The first and most evident advantage of phonetics is the independence it gives us. In the first place, it makes us inde- pendent of residence abroad. Even if the learner intends to go to the country where the language is spoken, it is a great advantage to him to start with a thorough practical knowledge of the sounds in which he is to practise himself. Secondly, phonetics makes us independent of native teachers. It is certain that a phonetically trained Englishman who has a clear knowledge of the relations between French and English sounds can teach French sounds to English people better than an unphonetic Frenchman who is unable to communicate his pronunciation to his pupils, and perhaps speaks a dialectal or vulgar form of French. Again, phonetics enables an intelligent adult to get a sound elementary knowledge of the sounds of a foreign language without any help from outside— that is, if he has an adequate phonetic analysis and transcription to work with. But the gain of a phonetic grasp of a language extends far beyond such special considerations. A secure grasp of the sounds of a language is a great strengthening of the mastery of GENERAL STUDY OF PHONETICS 49 its forms and meanings. A minute discrimination of similar sounds in closely allied languages is the surest safeguard against otherwise inevitable confusions, as when we keep up the slight distinction between the Norwegian and the Swedish (ii) in hus, 'house,' the Swedish sound being more advanced and nearer (y). Hence also the literary and aesthetic use of phonetics. Phonetics alone can breathe life into the dead mass of letters which constitutes a written language; it alone can bring the rustic dialogues of our novels before every intelligent reader as living realities, and make us realize the living power and beauty of the ancient classical languages in prose and verse. Phonetics is not merely an indirect strengthener of gramma- tical associations, it is an essential part of grammar itself. It enables us to state grammatical and philological laws with a brevity and definiteness which would be otherwise unattainable, as when we condense the information that under certain circumstances in a given language d becomes /, g becomes k, and b becomes /, into the simple statement that ' voice stops become breath.' In Eliot's Finnish Grammar (p. n)we find the following statement : ' The final e of a dissyllabic stem disappears in nouns before terminations commencing with t, and in verbs before terminations beginning with k or ;?, pro- vided that e is preceded by any simple consonant but k, p, v, m, or by a double consonant of which the last letter is t or s (except ht). Thus from the stem une, ' sleep,' vuore, ' mountain,' vete, 'water' (nominative vest), come the forms ttnta, vuor/a, vetta . . .' If in this statement we substitute for the negative and purely abstract conception of ' any simple consonant but k, /, v, m,' the positive enumeration of the consonants left after this subtraction, namely r, /, s, t, n, we are able to simplify it still further by saying that in nouns e is dropped before / when the e is preceded by a forward consonant, the evident reason being that these consonants are formed in the same place as /. A knowledge of sentence stress and intonation is not only an essential part of elocution and correct pronunciation, but is also an integral part of the syntax of many languages. 1 In short, there is no branch of the study of language which can afford to dispense with phonetics. 1 See my New English Grammar t Part II. CHAPTER VII BEGIN WITH THE SPOKEN LANGUAGE The second main axiom of living philology is that all study of language, whether theoretical or practical, ought to be based on the spoken language. The distinction between the literary and the colloquial form of the same language has considerably complicated the problem of learning languages. This distinction is not solely the result of the use of writing and printing, for even such unlettered savages as the Andaman islanders have an archaic poetical dialect which differs considerably from their ordinary spoken language ; but writing — and, still more, printing — have naturally increased the divergence. In many Oriental languages the divergence is so great that the colloquial is no longer a mere variation of the literary form, but the two practically constitute distinct, mutually unintelligible languages. The Spoken the Source of the Written Language In European languages, where the difference is much less, most grammarians tacitly assume that the spoken is a mere corruption of the literary language. But the exact contrary is the case : it is the spoken which is the real source of the literary language. We may pick out the most far-fetched literary words and forms we can think of, but we shall always find that they are derived from the colloquial speech of an earlier period. Even such forms as thou hast, he hath, were ordinary colloquialisms a few centuries ago, though they now survive only as fossil, dead colloquialisms side by side with the living colloquialisms you have, he has. Every literary language is, in fact, a mixture of colloquialisms of different periods. Every literary language must indeed in its first beginnings be purely colloquial. It is certainly difficult to realize that such a language as the classical Italian of Dante and Petrarch was originally nothing but a rough attempt to write down what were then considered the slovenly colloquialisms of Late 50 BEGIN WITH THE SPOKEN LANGUAGE 51 Latin ; but nevertheless such is the origin not only of Italian, but of all the other Romance languages as well. The tradi- tion of the origin of Italian is still kept up in the word for ' translate,' namely volgarizzare, literally ' make popular.' Accordingly, it is now an axiom not only of Romance philology, but of philology generally, that the real life of language is better seen in dialects and colloquial forms of speech than in highly developed literary languages, such as Greek, Latin, and Sanskrit. Practical Considerations Important as this principle is from a scientific point of view, it is still more so from a practical one, and for the following reasons : — If we compare the written and spoken language of a given period, we shall find that the literary language is full of super- fluous words and phrases, which the spoken language nearly always gets rid of. Thus in the English spoken language the idea ' sky ' is expressed by this word only, while in the literary language it may also be expressed by heaven, heavens, firma- ment, welkin. So also the form hath was still used in literary prose in the last century in such phrases as the author hath . . . , and it is still used in poetry and in the liturgical language of the Bible and Prayer-book, while in the spoken language the only form used is has. Again, nothing is more difficult than to give definite grammatical rules for the use of the subjunctive mood in literary English ; in the spoken language the subjunc- tive is not used at all except in a few perfectly definite con- structions, such as if it were. So also in spoken French the two most difficult tenses of the verb, the preterite indicative and subjunctive, have been supplanted by the perfect. So completely is the preterite obsolete that Passy, in his translation of the Gospel of Luke into modern French, discards it entirely, as in the beginning of the parable of the vineyard : oen am a plaate yn vifi, i 1 a lwe a de vinro, e il e parti pur loota (20. 9). According to Passy (Elementarbuch, 156), it occurs only in comic imitations of the South French dialect. Even in German the complicated rules for the inflection of proper names — Luise, gen. Luise'ns, Cato, Cato's, plur. Cato'ne, Leibjdtz, plur. Leib- nitz 'e — are swept away bodily in the spoken language, which, as a general rule, does not inflect such words at all. 52 THE PRACTICAL STUDY OF LANGUAGES Again, in literature the context is often vague, as in the Homeric meropes anthropoid where meropes may mean any quality that can be predicated of men generally. So also in the Sanskrit Vedas we have whole hymns, which, when epito- mized, leave not much more than ' the bright shiner (that is, the sun) shines brightly.' In simple colloquial prose, on the other hand, the meaning of a word is generally quite clear from the context. The spoken language, too, is far stricter in its use of epithets : it hardly ever introduces an adjective or other qualifier except to convey some definite information. Thus in ordinary speech we do not talk of ' the bright sun ' or ' the silver moon,' simply because the epithets convey no information — tell us nothing that is not already implied in the words sun and moon themselves. Even such a phrase as ' the sun shines brightly' has an uncolloquial ring about it, although it is not exactly anti-colloquial. We could say ' the moon is bright to-night,' because this really conveys information. The spoken language also prefers a simple paratactic arrangement of sen- tences. The complicated periods of literary prose would, indeed, often be unintelligible in speech. We see, then, that the advantage as regards clearness and definiteness is on the side of the spoken language : by starting from the spoken language we have less to learn, and we learn it accurately. Everything therefore points to the conclusion that in learned foreign languages we should follow the natural order in which we learn our own language : that is, that we should begin with learning the spoken language thoroughly, and then go on to the literary language. The psychological arguments for beginning with the spoken language are precisely analogous to those for beginning with a phonetic transcription (p. 12) : if we learn the literary and the spoken language simultaneously, cross-associations are inevi- table ; and the only possible way of avoiding or minimizing these cross-associations is to learn the two forms of the language separately. The question, which of the two we ought to begin with, is easily answered. It is evident that our strongest and most direct associations ought to be with the spoken language, for in speaking we must nave all our associations between ideas and words in perfect working- order : we have no time to pick and choose our words BEGIN WITH THE SPOKEN LANGUAGE 53 and constructions, as when we are writing. So also when others are speaking to us, we must understand each sentence at once, or the whole statement becomes unintelligible, while in reading, as in writing, we can pause and consider as often as we like. If, then, we first get a thorough knowledge of the spoken form of the foreign language, and then proceed to learn its literary form, we shall be in exactly the same position as regards relative strength of associations as the natives themselves : we shall think in the spoken language, because our associations are directly with it,, while at the same time we are able to understand the literary language, and, with a little effort at first, to write it ; but we are no more able to speak the pure literary language than a native is. As it is, we too often reverse the process, and so do foreigners who learn English. They first of all imprint firmly on their memories the obsolete phraseology of the Vicar of Wakefield, or, at the best, of Washington Irving's Sketch-book, then add a few choice Shakespearisms, and finally season this heterogeneous mixture with such modern colloquialisms as they can gather from the pages of Punch and Dickens. The result is always unsatisfactory, and often leads to unintelligibility. Thus I remember a case in which a German, on being asked how a certain lady was, replied that she was (rsept). As he tapped his forehead at the same time, the Englishman thought he meant to say that she had had a rap or knock on the head ; but after a long discussion and many vain attempts to get at his meaning, it turned out that he was thinking of Shakespeare's phrase in Macbeth, ' how our partner's rapt ' ( = transported, in an ecstasy), and meant to convey the idea that she was out of her mind. Another foreigner, a Spaniard, was observed to speak English with perfect grammatical correctness, but with a curious old-fashioned stateliness of diction, which was at first assumed to be the natural accompaniment of the blue blood of Spain ; it turned out, however, that the sole source of his colloquial English had been the dialogues in Dr. Samuel Johnson's Rasselas. I remember myself that when I first began to talk German, I was complimented on the poetical diction I used. It is said that when Sir Walter Scott talked French to the ambassadors of Charles IX., they were amused and often puzzled to hear a Scotch adaptation of the language of Froissart and Joinville. CHAPTER VIII DIFFICULTIES OF LANGUAGE Learning a language means overcoming difficulties, and each language has its own peculiar difficulties. External Difficulties Some of the difficulties may be purely external — due not to anything in the language itself, but to the circumstances under which it is learnt. Perhaps there is a want of text-books and other helps; the beginner is perhaps met with the cheerful warning, 'You will have to make your own dictionary, you know.' Or there may be text-books, grammars, dictionaries in plenty, but not in the learner's native language ; thus no one can learn Finnish without knowing Swedish, and to many languages Russian is the only key. The difficulties caused by the written form of the language, such as the complexity of its alphabet — which, again, may be the result of the writing being partly hieroglyphic — the ambiguity or unphonetic character of its orthography, are all purely external : Arabic is still Arabic when transcribed into Roman letters, nor is Japanese any the more Japanese for being written in a mixture of disguised hieroglyphs and syllabic alphabetic writing, both borrowed from China. No existing system of writing is anything but an external disguise borrowed from some other language : Arabic is disguised Syriac writing, and the Russian alphabet is Byzantine Greek. Relations to the Native Lano-uao-e There is another class of difficulties which may be regarded as partly external, partly internal — those which depend on the 54 DIFFICULTIES OF LANGUAGE 55 relations of the foreign language to the learner's native language, especially as regards similarity in vocabulary and structure. We are naturally inclined to assume that the nearer the foreign language is to our own, the easier it is. A Spaniard soon learns to understand Portuguese, and a Portuguese soon learns Spanish enough to understand it, a Dane soon learns to understand Swedish, and an Englishman soon learns to understand broad Scotch, because in all these pairs the two languages are practically only dialects of one another — in other words, because knowing Spanish or Danish or English implies knowing two-thirds of Portuguese, Danish, or Scotch respec- tively. Hence also we are often told that ' Italian is very easy if you know Latin and French.' Hence also Old English (Anglo-Saxon) is easier to a German than to an Englishman, so that, as I have remarked in the preface to my Anglo-Saxon Reader, ' he (the German) is able to acquire a practical know- ledge of it from a crabbedly theoretical exposition of it that would baffle an English learner.' But this very likeness is often a source of confusion. It is a help to the beginner who merely wants to understand the allied language, and is contented with a rough knowledge ; but it is a hindrance to any thorough knowledge, because of the constant cross-associations that are sure to present themselves. Thus in German werden is present and infinitive, wo?'den is past participle ; but in Dutch worden is equivalent to the German werden, while the Dutch werd is the preterite, being equivalent to German ward. And yet the general resemblance between German and Dutch is much less than that between such a group of languages as Danish, Norwegian, and Swedish. The resemblance between these three is, indeed, so strong that it is practically impossible to keep them apart : a foreigner who has learnt to speak Danish fluently, and then goes on to learn Swedish, will soon lose the power of speaking the former language, and will not regain it till he has forgotten his Swedish. A further study of Norwegian, which is intermediate between Danish and Swedish, will cause still greater confusion. Differences in the vocabulary are an even greater snare than differences of grammatical structure, because they cannot be brought under definite rules. Thus it is very difficult for an English speaker to realize that when a Frenchman ' demands permission,' he does not mean to imply the slightest impera- tiveness. It is dangerous to guess at the meanings of words in 56 THE PRACTICAL STUDY OF LANGUAGES closely allied languages, or in languages between which there is any borrowing of words ; thus in German gottesdienst means ' divine service,' but in Dutch godsdienst has the wider meaning ' religion.' So also in Swedish rolig means ' pleasant, amusing,' while in Danish and Norwegian it has only the older meaning 1 quiet, tranquil,' in accordance with its derivation from ro, ' rest,' cognate with the German ruhe. Hence a Dane would be puzzled if a Swede told him that he had found the Carnival or the Lord Mayor's Show ' rolig.' In learning a remote, unconnected language the difficulties are reversed. The beginning is much more difficult, and, of course, it takes a much longer time to understand the language. But when the initial difficulties have been once overcome, it is easier to get a minutely accurate knowledge of the language, because the learner is less disturbed by cross-associations. Internal Difficulties We will now consider those difficulties which are, in the strict sense of the word, internal — inherent in each language apart from external circumstances and from its varying relations to other languages. The difficulties of language in general may be classed under the four heads of (i) logic or reasonableness, (2) definiteness, (3) fullness of expression, and (4) simplicity. (1) As regards logic, most untrained minds regard every- thing in a foreign language that differs from their own as essen- tially irrational. But apart from such prejudices, there are some grammatical constructions, some methods of expression in special languages, which all foreigners — as well as unpreju- diced natives of a philosophic mind — would agree in consider- ing irrational. Such a construction is that by which in classical Arabic the numerals from three to ten are put in the feminine before masculine nouns, and in the masculine before feminine nouns, as in \ala\atu ba?iina, ' three sons,' arba'u bandtin, ' four daughters.' The contradiction here is purely formal. We have an example of an equally striking logical contradic- tion in the French plus de soupe ! ' no more soup ! ' an expres- sion which every Englishman would naturally and instinctively use instead of the correct encore de la soupe ! Such construc- tions are absolute paradoxes. As an example of an ordinary DIFFICULTIES OF LANGUAGE 57 irrational construction we may quote the English use of up in pack up, lock up, wrap up, which is opposed both to common sense and to the usage of most other languages, in which the literal translation of ' pack up ' would mean the exact oppo- site^ — ' unpack.' Of antigrammatical constructions — those constructions which cannot be parsed in accordance with the general grammatical rules of the language in question — some are logical and rational in themselves, such as the construction of a singular collective noun with a plural verb or a word implying plurality {the com- mittee are of opinion that . . . many cattle), while others are irrational, such as that almost incredible German con- struction in ich hade kommen miissen, ' I have had to come,' where the infinitive miissen is used as if it were a preterite participle. Over-abstraction sometimes leads to difficulties which defy direct logical analysis, such as the curious use of the verb ' to be ' in the passive, which is common in the Celtic languages, as in the Old Irish cein both oc aurgnom doib, ' while they were being served (waited upon),' literally ' while it-was-being-been with-serving to-them,' as if we were to say in Latin dum eratur ministrando eis. The use of the preterite in English and other Arian languages to imply rejected condition in such sentences as if I knew, implying ' I do not know,' is not wholly irrational, but certainly shows a certain intellectual clumsiness, as compared with the sensible Arabic use of two words for ' if,' one of which (lau) always implies rejection of the condition, so that there is no occasion to throw the distinction on the verb. Some difficulties are what may be called ' negatively illogical.' Thus to a foreigner the distinctions of gender in German and Old English by which hands are feminine and fingers mascu- line, while feet are masculine and toes feminine, appears to ' have no sense in it.' I remember a young Welshman correct- ing me, when I called the pair of bellows y megin instead of y fegin, by saying, ' We call a pair of bellows a she, sir ; ' he was then evidently struck by the absurdity of it, for he added after a pause of reflection, ' I don't know why we do so.' The diffi- culties connected with grammatical gender are purely mechanical difficulties, which cannot be overcome or evaded by any exertion of the reasoning faculties. Another — and perhaps the greatest — source of difficulty is 58 THE PRACTICAL STUDY OF LANGUAGES that the same fact may be regarded from a variety of different points of view, all of which are perhaps equally logical and reasonable. Thus in such a sentence as ' she held her hands before her face,' we should expect those languages which use the accusative case to express motion and the dative to imply rest to put face in the dative, as the hands are supposed to be at rest ; but in German the accusative would be used in such a construction, showing that the speakers who first framed this construction were thinking of the movement which brought the hands before the face rather than of the resulting position of rest. This difference of point of view is one of the chief sources of difficulty in idioms. Thus in French the idea of ' back numbers ' of a periodical is expressed by (kohksjo dy ■zurnal), where the element of ' backness ' is entirely ignored, the whole idea being approached from a totally different point of view. Sometimes the difference of point of view is the result of different circumstances or way of life, as when a German translates ' he followed me all over the house ' by ' he ran after me through all the rooms,' because Germans generally live in flats, and seldom occupy a whole house. (2) As regards definiteness, one language may make more minute distinctions than another. Hence to an ordinary Englishman who contents himself with roughly designating objects in space as 'this' or 'that,' or as being 'here' or ' there,' the threefold distinction involved in the Scotch this, that, yon, or here, there, yonder, the Latin hie, iste, ille, or the Welsh yma, yna, aew, occasions great difficulties — especially some of the special idiomatic uses of the Welsh aew — although he cannot help admitting that the threefold division is in some respects logically superior to his one twofold one. Want of definiteness, on the other hand, may cause just as much difficulty. How often in speaking a foreign language do we hesitate, vainly trying to find a word or phrase which corresponds definitely and exactly to the idea in our mind, till at last we have to fall back on a periphrase ! Those who have lived long abroad sometimes hesitate even in speaking their own language, because they feel tempted to use some foreign word, such as the German gemiithlich or the French flcLner. Nouns, such as the German philisier and the French fl&neur, are, indeed, so easily incorporated into the native speech that they soon become actual denizens, unless some translation or DIFFICULTIES OF LANGUAGE 59 adaptation takes their place, as when philister is adopted in the form of Philistine. This want of definiteness may sometimes amount to positive ambiguity, as in the English use of will and shall to express wish and compulsion on the one hand and futurity on the other, an ambiguity which is completely avoided in German by the use of werden to express pure futurity only. This makes an Englishman hesitate sometimes to use wollen or sollen in German where he ought to do so ; he does not feel the slight shade of wish or compulsion implied by the substitution of these auxiliaries for werden, and is therefore afraid of introducing an anglicism. A frequent source of indefiniteness and ambiguity is reliance on the context. In all languages a word may have a great variety of meanings distinguishable solely by the context, as when in English we apply the adjective sharp to knives, dis- tinctions, answers, and tempers. But the function of gram- matical forms is also largely dependent on the context, as we see in the English inflectional -s in sheep's, trees, he knows. In Chinese this reliance on the context is carried to extreme lengths : thus sam yuet, literally ' three month,' may mean either ' three months ' or ' the third month,' and laii lau, literally ' old old,' means ' to treat old people as they ought to be treated (that is, with respect),' the first lau being converted into a transitive verb ' to old.' (3) Fullness of expression may go to the extremes of redun- dance on the one hand, as in the reason why, my future address will be ... , and ellipse on the other, as in at his brother's (house). It is not these clearly marked cases, but the less defined ones, which cause real difficulty. Thus many of the Greek particles seem redundant and superfluous when compared with those of most other languages. So also do many of those used in classical Chinese, especially the finals, which practically in many cases seem to a foreigner to be little more than marks of punctuation, serving to show that the sentence is completed. But Old Chinese in most cases is almost incredibly concise and elliptical. Thus it has no word for the pronoun of the third person in the nominative — that is, it has no word for he, she, it, they, the absence of a pronoun being supposed to imply the third person; but not content with this, they omit the pronoun freely in the other persons as well, whenever the 60 THE PRACTICAL STUDY OF LANGUAGES context seems to allow it, so that, for instance, yuet may mean not only ' he says, she said, they will say, one may say,' and so on, but also ' I say, we have said,' etc. (4) Simplicity of expression implies in the first place regu- larity. As every one knows, irregular inflections are one of the most formidable difficulties in the study of inflectional languages. Simplicity also leads to generalization and abstraction, which, when unfamiliar, may require an effort to grasp, as in the many idiomatic uses of the Chinese indefinite pronoun ck, which has the function of making the preceding word or word-group into a noun of general meaning, so that, for instance, govern ce means ' the abstract conception of government,' able mend fault ce means ' one who is able to reform his faults,' grass firewood ce means ' cutters of grass and gatherers of firewood.' The opposite extreme of want of abstraction which leads to over specialization is a more frequent source of difficulty. It is most clearly seen in those savage languages, which often have no word even for so concrete an idea as that of ' washing,' but only separate words for ' wash the hands,' ' wash the feet,' ' wash dishes,' and so on. In the language of Tierra del Fuego no verb implying place can stand alone — the point of the compass must be indicated : they cannot say ' he stood ' by itself, but only ' he stood in the north, in the south . . . ,' these local deter- minations being used also in a variety of metaphorical uses, ' in the north,' for instance, implying ' away from the fire.' But want of abstraction is by no means confined to savage languages. Even in English we have no word to express the ' running ' of a horse : we must define the pace as trotting, galloping, etc. German has no general word for ' handle.' In Swedish there is no general word for ' aunt ' or ' uncle,' these ideas being expressed by contractions such as father-sister, mother-sister {faster, moster), so that it is always necessary to state expressly whether the maternal or paternal aunt or uncle is meant, just as in the older languages. One of the greatest sources of difficulty is that caused by superfluous distinctions — that is to say, distinctions which are invariably and unmistakably shown by the context, such as the Swedish and Norwegian distinction between ja and jo in the sense of ' yes,' the former being used after a positive, the latter after a negative question. Equally superfluous is the German distinction between herauf and hinaufi ' up towards the speaker,' DIFFICULTIES OF LANGUAGE 61 ' up away from the speaker.' In such cases the fact that the distinction is always implied unambiguously by the context makes the foreigner inclined to ignore it; unless, indeed, he carries it too far, saying, for instance, gerade hinaus instead of gerade aus in the sense of ' straight on.' The use of the sub- junctive mood in indirect narration is almost equally superfluous ; it is instructive to observe that modern French, which is other- wise strict enough in its use of the subjunctive, has in this case substituted the indicative, a change which also took place very early in the transition from Old to Middle English. Some minute distinctions may be justified logically on the ground that they do sometimes express shades of meaning which are more or less independent of the context, and may yet be, on the whole, practically superfluous. This is the case with the difficult Welsh distinction of four verbs ' to be ' — sydd, mae, yw, oes — whose use depends on subtle distinctions of definiteness and indefiniteness, emphasizing the predicate and so on. As these verbs are incessantly employed in the numerous substitutes for ' yes ' and ' no,' it is impossible in Welsh to express simple affirmation or negation without a thorough knowledge of the syntax. Another equally fruitful source of difficulty is unnecessary complexity. This is frequent in numerical expressions, such as threescore and ten for seventy, French quatre-vingt-onze, 'four- twenty-eleven ' = ' ninety-one,' Danish halvtredsindsfyve, ' half three times twenty,' that is ' threescore minus half a score ' = 50, with which compare German halb zwei, ' half two ' = ' one and a half.' Very curious also are Finnish numerals, such as kaksikymmentd, 'two tens' = 20, yksikolmatta, 'one of (the) third (set of tens) ' = 21. Even the English numerals are complex as compared with the Chinese ones, such as \ip n, ' ten two' =12, r\ti jip yit, 'five ten one ' = 51. The difficulty of the English vocabulary is the result of the complexity of its root-system, as shown in such groups as sun, sol-ar, helio-cent?ical, and sour, acid, oxy-gen. In German or Greek two roots would suffice for these six words. Phonetic Difficulties As regards phonetic difficulties — difficulties of pronunciation — there are three main considerations. The first is, that the difficulty of a sound depends more than anything on whether it 62 THE PRACTICAL STUDY OF LANGUAGES is familiar or unfamiliar, which is not an intrinsic, but a relative or, we may almost say, an external difficulty. To the unphonetic learner all unfamiliar sounds are difficult, or even impossible — at least, he thinks so. This applies also to unfamiliar combina- tions of familiar sounds. Thus even initial (ts) may be difficult to English speakers, as well as such combinations as (ft J) in Russian, because, although (ts) is a familiar combination, it is unfamiliar when initial. Hence a language may have a very simple and normal sound- system, and yet be difficult to pronounce, as we see in the case of Finnish, where it is necessary to make a strict distinction between long and short vowels, double (or long) and single consonants in unstressed as well as stressed syllables, the stressed syllable — which is always the first in the word — having a very strong stress, the others a very weak one, besides being uttered with great rapidity, so that the only way to keep up the necessary distinctions of quantity is by making the short sounds excessively short ; hence such a word as opettamattomuudessansa, 1 in his want of instruction (in his ignorance),' requires much practice. As the number of distinctive sounds of natural occurrence is rather limited, there is always an a priori probability of meet- ing at least some familiar sound in every new language. Hence there is, on the whole, a tendency to a balance of difficulties in foreign languages. Thus the English speaker meets his soft and hard th and his w hardly anywhere till he comes to Arabic, where, however, the first two are lost in most of the modern dialects. The Dane, again, finds his ' stodtone ' again in the Arabic hamza, and something, at least, of the sound of his ;■ in the Arabic 'en, for I certainly hear the same kind of throat-con- traction in both sounds (I mean the Copenhagen r), although the Danish phoneticians do not agree with me in this. The second consideration is, that no sound that actually exists in a language for any length of time can be intrinsically difficult ; for sounds are so easily and so imperceptibly modified in their transmission from generation to generation that their retention, unchanged for only a few generations, is enough to prove that they cannot be difficult in themselves. Thus, if the two Arabic throat-sounds, the ha and the '■en, were as difficult in themselves as most foreigners imagine them to be, they would not have been preserved, as they have been, unchanged in Arabic for at least ten thousand years. Nor do Arab-speaking DIFFICULTIES OF LANGUAGE 63 children find them so difficult to learn as some of the other consonants, such as the deep k. Lastly, practical training in general phonetics gets rid of many difficulties at once, and tends to make a complete mastery of the pronunciation of a foreign language simply a matter of practice and perseverance, ample time for which is afforded by the difficulties of mastering the grammar and vocabulary of the language. As our knowledge of phonetics and our methods of teaching it are gradually perfected, the easier it will be to clear away the remaining difficulties, especially if the practical study of phonetics is begun young enough — that is to say, in the nursery. General Difficulty of each Language In estimating the general difficulty of one language as com- pared with others, it is necessary once more to insist on the elimination of all external and irrelevant considerations, such as those caused by a defective or complicated system of writing, by want of grammars and dictionaries, by want of suitable texts. Latin is difficult partly because most of its literature is rhetorical and artificial — hardly ever naive and simple. Brown- ing and Hegel are difficult and obscure writers, but that has nothing to do with the question whether English and German are in themselves difficult. Old Slavonic, on the contrary, is comparatively easy partly because most of its literature consists of translations of ecclesiastical writings. Gothic is easy because the whole language — texts, grammar, glossary, and all — can be comprised in one volume, and this in addition to the texts being mostly Biblical translations. Most people, if asked what constitutes the real difficulty of such a language as Greek or Sanskrit, would answer without hesitation, ' the complexity of its inflections.' Most school- boys have wondered how the Greeks ever could have learned to conjugate the verbs in -mi. These people assume that all inflectional languages are necessarily difficult, and that the only real progress in language as regards ease of learning is getting rid of inflections. They are inclined to assume that a language such as Sanskrit or Russian, with its eight cases, must be more difficult than one which has only four, such as German, and that Finnish, with its fifteen cases, must be nearly twice as 64 THE PRACTICAL STUDY OF LANGUAGES difficult as Sanskrit — at least, from the point of view of noun- inflection. But when we look a little closer into the question, we see that there are generally compensations for an increased number of inflections. We find that, as a general rule, the greater the number of cases, the more regular they are, and, what is equally important, the more distinctive in form, and therefore the easier to remember. Thus in Finnish all ablatives end in -Ita — which under certain definite and simple phonetic conditions is regularly modified to -ltd — all ( translatives ' end in -ksi without any distinctions of gender, the endings being the same in the plural as in the singular ; the only difficulty in Finnish are the changes undergone by the stem, which, though often consider- able, are not so difficult as in more advanced inflectional languages. In Sanskrit there is much more irregularity than in Finnish, but many of the endings — such as -bhyas — are so full- sounding and heavy that they are as easy to remember as if they were independent words. German, on the other hand, has only four cases, which are expressed by a very limited number of endings : -e, -en, -em, -es, -er. But this formal sim- plicity is in itself a source of difficulty, for most of these endings have such a multiplicity of grammatical functions that they lose all individuality and become mere abstractions, which are abso- lutely meaningless apart from their context. It is a question whether the modern German inflections are not as difficult as the Finnish. The German dialects seem to think the noun- inflections difficult, for most of them get rid of them more or less completely. Again, the Finnish inflections enable the language to dis- pense with prepositions to a great extent. Thus ' without money ' is expressed by putting money in the ' caritative ' case, or, in other words, making without money into mone\-without, so that having fifteen cases, which sounds so formidable at first, means, from this point of view, having only fifteen prepositions in common use. The result often is that a grammatical cate- gory which in English can be expressed only by a variety of prepositions of complicated meanings and functions is in Finnish expressed by a single case which is often as distinct and tangible as an independent word. We thus arrive at the conclusion not only that a larger number of inflections does not necessarily increase the difficulty of a language, but also that inflections may in some respects be DIFFICULTIES OF LANGUAGE 65 easier to learn than the prepositions, particles, and auxiliaries which take their place in ' analytical ' languages such as English and French. No inflections can possibly be more difficult than the English distinction between will and shall in the future, or the French uses of the prepositions a and de. Then, again, inflections are not the only formal irregularities in language. The student of spoken English has not only to learn the syntactical use of will and shall, but has also to learn to recognize these words in their various formal disguises in' such combinations as (ail, ai wount, ai Jaant), and so on. So also French, after substituting de for the various inflections of the Latin genitive, goes on to develope fresh irregularities, such as du, des. The epithet ' analytic,' too, is often applied too sweepingly. If we compare Italian with Latin, we see that the loss of the cases is to a great extent compensated, as regards irregularity and complexity of form, by the difficulty of the verbs, and by the various forms of the pronouns and the other new develop- ments. It is clear, therefore, either that the intrinsic ease of Italian as compared with Latin has been exaggerated, or that it is the result of other changes than mere loss of inflection. If, indeed, we put ourselves in imagination in the place of an intelligent Asiatic who knows nothing of any European language, we shall have reason to doubt whether Italian is, after all, easier than Latin. The comparative ease of Italian to Europeans is mainly the result of purely external conditions, the most important of which is that most of those who learn it, really know it partially beforehand through knowing French and Latin — languages which no European can help learning to some extent through the French and Latin words imported into his own vocabulary. If inflections and grammatical irregularities were the main cause of difficulty, then Chinese ought to be the easiest language in the world, for it has no inflections, no grammatical genders, no irregularities of form, and its particles and auxiliaries are few in number : Chinese grammar is all phonology and syntax — there is no accidence whatever. And yet the construction of classical Chinese is as difficult as that of Latin, quite apart from any external difficulties. 66 THE PRACTICAL STUDY OF LANGUAGES The Real Difficulty is in the Vocabulary The fact that the languages commonly learnt by Europeans belong mostly to the same Aryan stock, and have besides a large vocabulary in common of borrowed Latin, French, and Greek words, is apt to blind them to a recognition of the fact Ithat the real intrinsic difficulty of learning a foreign language lies in that of having to master its vocabulary. Mastering the vocabulary of most European languages means simply learning to recognize a number of old friends under slight disguises, and making a certain effort to learn a residue of irrecognizable words, which, however, offer less difficulty than they otherwise would through being imbedded in a con- text of familiar words. The higher vocabulary of science, art, and abstract thought hardly requires to be learnt at all ; for it consists either of Latin and Greek terms common to most European languages, or of translations of them. It is very different with a remote disconnected language such as Arabic or Chinese. The abstract vocabulary of Arabic shows Greek influence, although this affords very little practical help ; but the terminology of Chinese philosophy and science is independent of Western influence, so that every extension of the vocabulary requires a special effort of memory and reason- ing. The task of mastering such languages is literally an endless one. Enough Arabic grammar for reading purposes is soon acquired, the construction being always perfectly simple — at least in ordinary prose, but the student may read one class of texts for years, and then, when he proceeds to another branch of the literature, he may find that he can hardly understand a word, this being almost entirely the result of the unfamiliarity of the new vocabulary required. In short, we can master enough of the grammar of any language for reading purposes within a definite period — gene- rally less than six months — but we cannot do the same with the vocabulary unless it is already partially familiar to us in the way that the vocabulary of Italian is to all English speakers. All Languages Equally Difficult All these considerations, if summed up impartially, lead us finally to the conclusion that, as regards ease of learning, all languages are intrinsically on a level — they are all equally easy DIFFICULTIES OF LANGUAGE 67 or equally difficult ; that is, of course, if we rigorously eliminate all external considerations, and disregard the special relations between individual languages. But as it is practically impossible for any one who has not an equally perfect knowledge of all languages to test this by experience, it must remain an abstraction, like the dogma of the absolute regularity of sound-changes. We may also say of the dogma of the intrinsically equal difficulty of languages, as of that of the absolute regularity of sound-changes, that even if it is not true, it has a certain value as a corrective to one- sidedness and inaccurate reasoning. The external considerations have been already discussed. One reservation only remains to be stated. When we talk of the difficulty of a language, we must strictly define the limits of the language ; we must be careful in speaking of a language to make sure that we are not really speaking of a group of languages, or — what is the same thing from our present point of view — a group of dialects. Thus an ordinary Greek grammar would give us a very exaggerated estimate of the difficulty of the verbs in -mi if we reckoned up all the divergent forms without regard to difference of dialect. It must also be remembered that the Homeric dialect is a confused mixture of forms of different periods and dialects with artificial mon- strosities invented by grammarians : it never could have been an actual language. So also Italian is not so difficult as its conventional grammars are. The conclusion to which we have just arrived is strengthened by some a priori considerations. The history of grammatical irregularities is very instructive from this point of view. The tendency of unrestrained phonetic change is to cause increasing complexity and irregularity in language. The origin of inflections is to be sought mainly in phonetic changes which caused originally independent post-positions to become incor- porated into the preceding word, as we see in the Icelandic reflexive inflection -sk, which is simply a shortening of the reflexive pronoun sik, as in buask, ' prepare oneself,' whence our verb busk. The phonetic changes which brought inflections into being tend to complicate more and more both the inflections themselves and the inflected words. We see the result in such English forms as (vvuman), plural (wimin), where there are only traces left of the original Old English forms wlfmami, plural 68 THE PRACTICAL STUDY OF LANGUAGES wlfmenn, the second element of (wimin) being also completely isolated from modern English (maen) or its plural (men). So also in Old Irish ben, ' woman,' plural mna, where the change of b into m is purely phonetic. So, again, in Welsh the word potatoes was borrowed in the form of tatws, which was regarded as a collective plural, from which on the analogy of native words a new singular was formed by vowel-change and the addition of -en, the whole word being afterwards shortened to tysen, the singular being thus completely isolated from the plural, as if they were unconnected words. In languages as they exist, such difficulties are allowed to accumulate up to a certain point. When they threaten to undermine the whole structure of the language — as they certainly would do in any language if left to themselves — they are got rid of by means of the process of levelling by analogy. Thus, in some dialects of Welsh the divergence in the words just mentioned is got rid of by forming a new singular taten from the plural, so that the comparatively regular inflection taten, tatws is developed. So also in English we might make the inflection of woman regular by giving it a regular plural womans on the analogy of the vast majority of English nouns. We might make the plural of man itself regular in the same way. But as these two words are of extremely frequent occurrence, it is easy for us to remember them, especially as the whole number of irregular plurals is but small. Different languages tolerate different irregularities. Thus Welsh is very irregular in the formation of its plurals, but it makes up for this by getting rid of all its case-inflections. Finnish, as we have seen, has many cases, but they are, on the whole, very regular. If a language is very regular and simple in one department, we may expect it to be irregular and complex in another. In this way there is a balance of difficulties, although this is often ignored through taking a one-sided view. Thus in English the formal part of the grammar is fairly simple and regular ; but the vocabulary shows the greatest complexity and irregularity, which in the spoken language extends to the form as well as the meanings of the words, as we see in such a group of words as (foutagrasf, fout3 - gra3fik, fa'tografa), where there is not only divergence in stress but also in sound, so that the first and last have very little resemblance to the ear. It is evident that every language in its colloquial form must DIFFICULTIES OF LANGUAGE 60 be adapted to the average capacity of its speakers. Although each language is constructed to a great extent by the philo- sophers and poets of the race, it cannot in the form of it which serves for ordinary intercourse go beyond the capacity of the average mind. Learning a language, therefore, is not in any way analogous to learning mathematics or metaphysics : it does not imply any attempt to enter into higher regions of thought — to commune with a higher mind. On the contrary, as the greater part of all existing languages was evolved by people in a rudimentary state of civilization, it implies the very reverse. Hence, as we shall see hereafter, it is often a positive obstacle to learning a language to be rigorously logical and minutely analytical. From the admission that all languages are in themselves equally difficult, it does not necessarily follow that we are never to apply the word 'difficult' to languages. But it must be understood that when we say that one Oriental language is more difficult than another, we only imply that the external obstacles are greater, or that the structure of the language differs more from that of the average European language. CHAPTER IX GENERAL PRINCIPLES OF METHOD We now come to our main problem — how to overcome these various difficulties. Language Only Partly Rational Before going any further it is important to realize clearly the fact that language is partly rational, partly irrational and arbi- trary. Thus, when a language enlarges its vocabulary by systematically utilizing material words to express abstract ideas, as when it uses such words as spring or source to express the idea of ' origin,' it is rational ; so also when it indicates different grammatical relations between words by the order in which they follow each other. When, on the other hand, language developes such a system of grammatical gender as we find in French and German, or when it allows inflections to become irregular and ambiguous, it is irrational. It is true that we can prove by historical philology that there was once a reason for grammatical gender, and that the inflections that are now irregular and anomalous were once the regular ones, or that at any rate they are the result of regular sound-changes ; but this does not in any way alter the fact that they are now, from a practical point of view, irrational. We might as well argue that the buttons that are still put at the back of men's dress coats are useful because our dress coats were originally coats with long tails which were buttoned up in riding. The arbitrariness of language is most strikingly evident in its vocabulary. The type of a rational word is such a one as cuckoo, which, to those who already know the object it repre- sents, is as self-interpreting and as easily remembered as any gesture or picture. But in all languages the vast majority cf primitive words have no connection with the meanings they 70 GENERAL PRINCIPLES OF METHOD 71 express, and, what is worse still, one sound-group often stands for a variety of ideas, which are sometimes quite disconnected, as in the various meanings of such English words as bear and box. Again, in a rational vocabulary words similar in form would have allied meanings, and similar meanings would be expressed by similar words, but in English such formally almost identical pairs as bit and beat, bed and bad, have nothing in common as regards their meanings, and even such ideas as 'good' are expressed by a variety of distinct words, such as good, well, virtue. The only rational part of the vocabulary is that which forms new words by composition and derivation, and gives words new meanings by means of metaphor, simile, and other processes of the same kind ; but all these processes are often irregular and arbitrary both in their operation and their results. Irrational Combinations in Language : We Cannot Speak by Rule Language is often irrational even in the way it combines words into sentences — in its synthesis. If language were perfectly rational in this respect, we should be able to handle words like the nine digits in arithmetic, and combine them into sentences at pleasure by applying a few simple grammatical rules. In practice, however, we find that a great part of all languages consists of a limited number of natural sentences, only some of which admit of being formed a priori and freely modified by the substitution of other words, as when from have, ink, pen we make up such sentences as / have the ink ; who has the pen ? who has the ink ? he has the ink, and so on. But just as we cannot go on speaking long without using irregular inflections, so also we cannot go on speaking naturally for any length of time without using irregular combinations of words — combinations which cannot be constructed a priori. The sentences which make up natural speech are of two kinds — general sentences, such as those which have just been given, and special sentences or idioms, such as how do yon do ? never mind, which are really on a level with simple words, such as salutation, indifference, and, like them, have to be learnt one by one, in the same way as the iregularities of the grammar. Many of them, indeed, have meanings inconsistent with those of the words of which they are made up. Thus do by itself 72 THE PRACTICAL STUDY OF LANGUAGES never has the meaning it has in how do you do ? and help in the idiomatic expression / could ?iot help being late has the meaning ' prevent,' ' avoid,' which is the exact contrary of its ordinary meaning. Again, even in those cases in which the grammar and dictionary allow us to express an idea by various combinations of words, there is often only one of these combinations in actual use. Those who have had to do Latin prose composition know that the main difficulty of the art consists in having an instinctive knowledge what combinations to avoid. French has a similar character. English and Greek are much freer in this respect, a fact which many foreigners find it difficult to realize. When they ask me such questions as ' can one speak of an " elegant supper " ? ' ' can you say, " he was bad last night " ? ' I always answer that English is a free language, and that there is nothing to prevent any one calling a supper 1 elegant,' although I do not remember ever doing so myself. Nevertheless, English has its limitations as well as other languages. Foreigners' English often presents the curious spectacle of a language constructed on strict grammatical principles, but with hardly a single genuinely English sentence in it. The following extract from the published works of a distinguished French Orientalist who lived many years in England, and wrote most of his books in English, will illustrate this. The writer is Prof. Terrien de la Couperie (The Pre- Chinese Languages, § 235), and he is protesting against the systematic study of phonetics : — ' Another point which requires due consideration is that of pronunciation. The scientific achievements lately obtained in perfection of transcription by several English and German scholars go beyond human looseness. They have reached the high level of the respective idiosyncrasies of the speaker and of the transcriber, above the common average of speech. The activity of man's speaking-organs and also that of his ear- sense, have nowhere the mechanical and permanent precision which their principles and those of the new school of gram- marians imply. Uncultured populations and uneducated men are not naturally bent in the material of their speech to the yoke of steady precision which is only the result of a training in educated social surroundings through several generations. Audition and articulation of language, except in the higher races, seldom arrive together at some sort of perfection in their GENERAL PRINCIPLES OF METHOD 73 effectiveness. For instance, we may quote the well-known fact that the acuity of the ear among the races paying peculiar attention to the colour and pitch of the vowels exists only at the expense of precision in the articulation. ' Tribes in a rude state of culture have a looseness and uncouthness of pronunciation and hearing, which escapes, in its group's fancies or individual distortions, from any unflinching law of regularity. The cases and causes of variance from analogy, relative easing, symbolical strengthening or weakening, scorn anything like a formulated law. The segmentation, dispersion, and migration of tribes grown from a homogeneous linguistic stock in that state of unculture, combined with the complication from the frequent though often unknown super- imposition of races and languages in a similar condition or otherwise, imply large divergences of pronunciation apparently inconsistent with their genuine derivation from common parents. And the efforts at reducing the whole of the divergences to regular and somewhat mechanical equivalence cannot lead otherwise than to numerous confusions and misapprehensions. 'After the disturbance of ideologies, the most important result for all the languages engaged in the struggle, a result produced at the same time by the intermingling of blood, concerns the phonesis.' The Arithmetical Fallacy The 'arithmetical fallacy,' as we may call it, is well illus- trated in the practice of exercise-writing and translation into the foreign language, a subject to which we will return later on. In the well-known methods of Ahn, Ollendorff, and Arnold it is developed into a regular system, intended as a substitute for the ordinary grammar and dictionary method— at least for the beginner. The result is to exclude the really natural and idiomatic combinations, which cannot be formed a priori, and to produce insipid, colourless combinations, which do not stamp themselves on the memory, many of which, indeed, could hardly occur in real life, such as the cat of my aunt is more treacherous than the dog of your uncle \ we speak about your cousin, and your cousin Amelia is loved by Jier uncle and her aunt J my sons have bought the mirrors of the duke \ horses are taller than tigers. At one school where I learnt — or rather made a pretence of learning — Greek on this system, the master 74 THE PRACTICAL STUDY OF LANGUAGES used to reconstruct the materials of the exercises given in our book into new and strange combinations, till at last, with a faint smile on his ascetic countenance, he evolved the following sentence, which I remembered long after I had forgotten all the rest of my Greek — the philosopher pulled the lower jaw of the hen (tou tijz o'naipos senou gnaepos). The results of this method have been well parodied by Burnand in his New Sand- ford and Merton, thus : the merchant is swimming with (avec) the gardener's son, but the Dutchman has the fine gun. Isolated Phenomena of Language : Grammar and Dictionary One result of language being partly rational, partly irrational, is that some of its phenomena can be brought under general rules, some cannot. Thus in English the fact that tree is made into trees when we speak of more than one tree is a general one ; for we can add s in the same way and with the same change of meaning to nearly all other names of things. But the fact that /, r, e, e expresses the idea ' tree,' and not any other idea, is an isolated one ; for, given these sounds, we can- not tell beforehand what the meaning will be, and given the idea ' tree,' we cannot tell beforehand what combination of sounds will express it. This constitutes the whole distinction between grammar and dictionary. Grammar, like all other sciences, deals with what can be brought under general laws, and relegates all the other phenomena of language to that collection of isolated facts which we call the dictionary. It need hardly be said that there is no absolute line of demarcation between the two ; thus the prepo- sitions and many other particles belong both to the grammar and the dictionary. It also follows from our definition that what belongs only to the dictionary in one language may fall — partially, at least — under grammar in another, and vice versa. Thus in that remarkably symmetrical family of languages, the Semitic — of which classical Arabic is the best type — many of the details of the formation of roots and the structure of the primitive vocabulary are rightly included in the grammar. Again, such languages as German and Russian — though in many respects they fall short of the Semitic languages in word- forming power — still have great resources in the way of compo- sition and derivation. In English, on the other hand — which, GENERAL PRINCIPLES OF METHOD 75 from the point of view of the vocabulary, must be regarded as a degenerate language — even such a simple matter as the for- mation of an adjective from a noun is often the business, not of the grammar, but of the dictionary, as in sun, solar, man, human, virile. We see, then, that the existence of grammars and diction- aries is founded on the nature of language itself. The Natural Method But many undeniable abuses in the use of these helps have led some reformers to a revolt not only against the use of grammars and dictionaries, but also against all system and method whatever in learning languages. This revolt against method has further led to an advocacy of the ' natural method ' by which children learn their own language. These enthusiasts forget that the process of learning one's native language is carried on under peculiarly favourable cir- cumstances, which cannot be even approximately reproduced in the later study of foreign languages. In learning our own language, we begin young, and we give our whole time to it. Our minds are perfect blanks, and we come to it with all our faculties fresh and unworn. The fact, too, that we generally learn new words and new ideas simul- taneously, and that the word is often the key to the idea, gives a peculiar vividness and interest to the process of word-learning. But the process has also its disadvantages. It is a very slow process ; and the results are always imperfect. Indeed, so im- perfect is this natural method, that even with the help of school- training and the incessant practice of everyday life, very few ever attain a really thorough mastery of their own language. When we say that any one is ' eloquent,' or that he ' has a good style,' or ' is a good speaker,' or ' can tell a story well,' we hardly mean more than that his command of his own language is rather less imperfect than that of his fellows. If languages were learnt perfectly by the children of each generation, then languages would not change : English children would still speak a language as old at least as ' Anglo-Saxon,' and there would be no such languages as French and Italian. The changes in languages are simply slight mistakes, which in the course of generations completely alter the character of the language. The disadvantages we have to labour under when we learn a 76 THE PRACTICAL STUDY OF LANGUAGES foreign language are evident enough, and the later in life we begin, the more evident these disadvantages become. The power of imitation has greatly decreased, which is especially noticeable in the pronunciation. Not only has the power of imitation decreased, but also the desire to use it : the mind has lost its freshness and susceptibility to new impressions. On the other hand, the mind is formed : it is capable of generalization and abstraction ; it has an immensely wider and more accurate knowledge of the things and ideas represented by words and their combinations ; it has greater powers of concentration and methodical perseverance. And these advan- tages more than compensate the disadvantages we have just mentioned. Nevertheless, there is one disadvantage which turns the scale; that is, the fact that the student has already learnt another language — his own. Hence in learning the new lan- guage he has, as it were, to try to unlearn the other language, to struggle continually against the formidable difficulties caused by cross-associations. When he tries to pronounce a new sound, his tongue tends to slip back into the position for forming the nearest native sound. So also with word-order, grammatical construction generally, and the whole fabric of the language. The fundamental objection, then, to the natural method is that it puts the adult into the position of an infant, which he is no longer capable of utilizing, and, at the same time, does not allow him to make use of his own special advantages. These advantages are, as we have seen, the power of analysis and generalization — in short, the power of using a grammar and dictionary. Residence Abroad The natural method almost necessarily implies a residence in the country where the language is spoken. But residence abroad has also its own linguistic drawbacks. It sounds well to talk of ' picking up a language by ear in the country itself,' but most good linguists "will confess that they learnt nearly everything from books, especially in the beginning of their study of the foreign language, and but little from con- versation. There are, indeed, many obstacles to learning from conversation. In the hurry of talk we are apt to mishear and GENERAL PRINCIPLES OF METHOD 77 forget, so that what we pick up in that way is never reliable. Conversation is really not a means of learning new words and expressions, but only of practice in hearing and reproducing what we have already learnt. In conversation we also have the disadvantage of hearing only the answers to our questions, while we have no means of knowing whether our questions are expressed correctly, for it is very difficult to overhear the natives asking questions which will serve as patterns for our own. Rash reproduction of what we hear casually may land us in vulgar, ludicrously slangy, or otherwise objectionable ex- pressions. The results of picking up a language entirely by ear from the beginning may be seen in uneducated adults who come among a population speaking a strange language : after years of residence in the country they are often unable to utter anything but a few words and phrases. In fact, a residence in the country itself before the elements have been mastered at home is positively injurious, for it forces the learner to improvise incorrect expressions on the spur of the moment ; and these incorrect expressions then tend to become stereotyped by incessant repetition, so that they can scarcely be got rid of. This is specially the case with the equivalents of such particles and phrases as Oh ! to be sure, don't you know. Nor must the learner expect too much from a residence abroad. There are many external obstacles, especially in the case of English-speakers. Thus it is often almost impossible for an Englishman to learn educated colloquial German in the country, because all the Germans want to practise their English upon him ; and, besides, he is often thrown by circumstances almost exclusively among English-speakers in foreign schools and boarding-houses. I heard of one case in which an English boy was at Bonn for a year ; when he came home, he said he had not spoken a single word of German the whole time, not even in the shops. Then there is the difficulty of avoiding confusion of dialects, even if the learner is able to choose his place of residence exclusively from that point of view ; in a University town the professors and students come from all parts of the country, and therefore often speak different dialects. Many people, however, who admit the utility of grammars and dictionaries, are inclined to discard systematic study as soon as they have mastered the elements of the language, 78 THE PRACTICAL STUDY OF LANGUAGES especially if they have an opportunity of pursuing their studies in the country itself. A little reflection ought to convince them that systematic study is almost as necessary at the end of the course as at the beginning. After what has been said about the difficulty, or rather impossibility, of picking up reliable knowledge by hearing — which applies also, though perhaps in a less degree, to cursory reading— it is evident that giving up systematic study means simply giving up learning. After we have once given up systematic study, we cannot be said to learn the language, we only ' keep it up.' Those who wish to derive the fullest benefit from residence in the country itself should, therefore, be guided by the follow- ing principles : (i) prepare yourself thoroughly beforehand; (2) choose a place where you will have an opportunity of hearing a good standard of pronunciation and language generally, as un- mixed as possible ; (3) keep up systematic study till the last. Speaking Foreign Languages at Home There are several substitutes for residence abroad. One is, to converse with foreigners in one's own country. In this way many foreigners get a good knowledge of colloquial French and English. It is evident that the success of this method depends to a great extent on the number of foreigners who come to the learner's country, and on the extent to which they learn the language of the country, the most favourable condi- tions being for the learner to speak the language of a small country much frequented by foreign tourists, as when a Norwegian learns English from tourists of that nationality. But this course has its drawbacks and dangers, which become more and more evident the more the conditions diverge from those sketched above. A tourist, who stays only a short time in the country, preserves his national habits of thought and speech, which are generally those of an educated man ; but foreigners who settle permanently in another country may partially lose their nationality in speech as well as in other respects, and may be bad models from the beginning. The greatest of these drawbacks is, of course, that the sur- roundings are not foreign, so that we miss a good deal of what we should learn spontaneously in the country itself, and what we do learn is learnt under wrong surroundings and associations. Thus instead of learning the words and phrases GENERAL PRINCIPLES OF METHOD 79 associated with the national games and amusements of the foreigner, we hear perhaps the description of a game of cricket or lawn tennis, interlarded, of course, with many English words and phrases. German clerks in our large towns may be heard using such expressions as die bill of lading ist noch nicht da, and, of course, da hatf ich einen kep ( = cab) genommen. Foreigners who have lived long in the country often import even its idioms into their own language. Thus Germans in America in conversation with each other have been heard to say backen Sie nicht aus, ' don't back out (of your promise) ! ' Similar objections apply to the practice of letting children learn languages from foreign nurses and governesses. Of course, the younger the child, the more perfect its imitation of the foreign language. But if this is carried too far, it implies that the child does not learn its own language. Then, again, if young children learn easily, they forget still more easily : in extreme cases a child may learn a little of its own language, then learn a foreign language tolerably well, forgetting its own language in the process ; it then begins to learn its own language again, and forgets the foreign language, the final result being simply to delay its learning of its own language. The results, too, are generally unsatisfactory in many ways : the child learns to speak the little it learns with great fluency, but the pronunciation is not good, nor the construction perfect ; and if there is a large family of children, they soon invent a French or German of their own with a pronunciation made up exclusively of English sounds. Good results are due either to exceptional ability on the part of the child, or to exceptionally favourable circumstances which make the child bilingual from the beginning. Thus when the children of foreign parents settled in England speak the two languages perfectly, this is not a case of learning a foreign language in the ordinary sense of the word, any more than when children are taken abroad by their parents. Natural Aptitude Every one knows that the natural aptitude for learning foreign languages varies greatly in different individuals. It varies in children as well as adults, though perhaps not to the same degree. 80 THE PRACTICAL STUDY OF LANGUAGES Children show different degrees of quickness and accuracy in learning to speak their own language. Gabelentz says, in speaking of children learning their own language (Gab. 65) : ' Some take years to overcome the difficulties of pronunciation and grammar, while for others these difficulties seem scarcely to exist. I could mention German children who, from the very beginning of their attempts to speak, pronounced the gutturals and the consonant-groups of their own language and even foreign words with ease and correctness, and seldom violated the rules of German gender, or the irregularities in the formation of the plural and the conjugation of the verbs. Other children built up independently a language of their own with special laws.' He goes on to mention a child who, of its own accord, developed a system of modifying the vowels of the German words it learnt for symbolic purposes, some- what as in the Semitic languages, and thus constructed a language of its own, in which, for instance, the vowel u was associated with bigness, the vowel * with littleness. This is interesting, as illustrating what we shall have occasion to notice hereafter, that originality of mind does not make a good linguist. In fact, a talent for languages does not imply any higher intellectual development of any kind. The truly original mind seizes instinctively on the most efficient means of expres- sion at its command— that is to say, it prefers to express itself in the language it knows best, which is its own. Such minds avoid learning foreign languages as much as possible. Sweden- borg would no doubt sooner have written in Swedish than in Latin, were it not for his wish to have his books read as widely as possible. As for those who are drawn to the original investigation of the science of language, they do not, as a rule, speak them any better than other people — often worse. We need only mention the bad Latin in which the great founder of comparative philology made his first discoveries known. The considerations to which we were led before, namely, that languages are only partly rational, show that their acqui- sition must be, to a great extent at least, a mechanical process. Mechanical learning does not require originality of mind or a critical spirit. These are, indeed, hindrances rather than helps. What is required is the faculty of observation, quick imitation, adaptiveness to grasp the phenomena of the new language, and memory to retain them. All these qualifications are required in the highest degree in GENERAL PRINCIPLES OF METHOD 81 speaking, ease in which— especially, of course, with the more remote languages — is the greatest test of the born linguist as opposed to the scientific philologist. One of the most perfect types of what the latter would call ' the parrot linguist ' was Palmer the Orientalist ; and it used to be said of him at Cam- bridge that when he talked to Orientals in their own language, he seemed to speak faster than they did. This excessive fluency often blinds the superficial observer to the defectiveness of the imitation, especially in the pronunciation, which in the born linguist of the highest type is always good, but apparently never perfect, unless with the help of phonetic training. It is said that when Palmer talked to the Arabs of the desert, they thought he was an Arab of a different tribe. There is also a lower type of general linguist who cannot speak, but reads a large number of languages, and, perhaps, writes them. This type is the natural result of the combination of a less quick mind with a retentive memory and a natural taste and enthusiasm for polyglot linguistics. Although originality and independence of mind are to some extent anti-linguistic, they are not positive bars to the acquisition of languages. Strength of purpose, based on a conviction of the utility or perhaps the absolute necessity of learning a given language, will work wonders, especially if there is a real love of the study, which does not necessarily imply any special talent. It is difficult to define the opposite extreme of the purely anti-linguistic mind except as the negation of the other extreme, that is, as the result of slowness of mind, want of adaptability and power of imitation, together with shortness of memory. Such an absolutely anti-linguistic mind is the slave of the associations of its own language : when it expresses itself in a foreign language, it tries to do so by translating the native expression of each idea word for word into the foreign language, perhaps grammatically, but regardless of idiom and the genius of the foreign language, as when an Englishman of the old- fashioned John Bull type said to a German uh habe einen groszen grist Sie niederzuklopfen. This anti-linguistic mind is not un- common among grammarians and philologists. It must, of course, be understood that the intellectual qualities which constitute linguistic talent are of a special kind : the quickness must be linguistic quickness, the memory must be a linguistic memory, however much it may extend to other subjects as well. In the same way the adaptability and G \ 82 THE PRACTICAL STUDY OF LANGUAGES sympathy must be linguistic sympathy : the feeling which makes us feel an interest in the individuality of each language — in the way in which it expresses ideas. The linguistic interest, though allied to the literary, is not identical with it — least of all, in the higher developments of the latter. Thus the great linguist Palmer wrote verses with great facility, but these verses had nothing of poetry beyond the mere form, which was itself generally trivial. No phenomenal linguist has ever produced real literature, nor, what is more remarkable, ever made any great contribution to the science of language. National Aptitude There does not seem to be any valid reason for supposing that one nation has more talent for languages than another. The great linguists have not been confined to one country any more than the phenomenally strong men. But nevertheless the observations we have made concerning individuals apply, to some extent, to nations also. In the first place, original and intellectually independent nations which have a long civilization behind them, do not generally take kindly to learning foreign languages. A French- man in a mixed company abroad expects every one to talk French, even if he is the only Frenchman present. Englishmen are less egotistical, but they generally prefer to talk English with foreigners, even if they can speak the foreigner's language better than the foreigner speaks English. The Germans, on the other hand, whose sense of nationality has been of later growth, never speak their own language if they have a chance of speaking a foreign one ; but, as might be expected from the most intellectual nation in Europe, they seldom speak foreign languages really well. The imitative Russian and the supple Oriental seem to be often better linguists than the slower and more independent European. But the Russian aptitude for learning languages has been doubtless much exaggerated. Foreigners who have lived long in the interior of Russia have often assured me that the Russians, as a rule, do not speak foreign languages better than other nations. Tolstoi, too, in one of his novels, remarks of one of his Russian characters that ' he spoke that excellent French which is so seldom heard now.' The fact is, that those GENERAL PRINCIPLES OF METHOD 83 Russians who used to speak perfect French had to pay the price in expatriation and partial oblivion of their own language. We may safely prophesy that as the national life of the Russians developes, they will become worse and worse linguists. Some of the conditions of national linguistic skill are purely external. Belgians, Swiss, Dutchmen, and Danes are better linguists than Englishmen partly because the smallness of their respective countries obliges them to learn other languages. The Russians were obliged to be good linguists, partly because their retarded civilization obliged them to be imitative and adaptive with regard to the older civilizations of Western Europe, partly because the newness and inaccessibility of their own language prevented foreigners from acquiring it. One Method for All However great the differences may be between individuals and between nations as regards ease of learning foreign languages, these differences are differences of degree only. All minds work by the same fundamental psychological laws. No one can learn a language without exerting the faculties of . association and memory. However bad his linguistic memory, however weak his linguistic associations may be, he must have some linguistic memory and be capable of forming some linguistic associations, or he will not be able to learn any language at all — not even his own. The mere fact of his having learnt his own language shows that he is capable of learning other languages as well. That the difference between the dull learner of languages and the born linguist is one of degree only, seems to be con- firmed by the fact that even such a prodigy as Mezzofanti used to learn paradigms by heart like any schoolboy. The only difference was that Mezzofanti learnt them quicker and remem- bered them better, and was more ready in applying them to the grammatical analysis of the texts he read. His memory was so retentive that he could repeat a whole folio page of a Greek Father by heart after reading it through once. These considerations will help us to settle the important question, how far the method of learning languages ought to be the same — that is, of course, the same for all normally and fully developed minds. 84 THE PRACTICAL STUDY OF LANGUAGES If one linguist gives another linguist an account of the method by which he has learnt — or professes to have learnt — a language, the other may agree with him, or may think some other method better. But he may also take an agnostic attitude : he may say that every one has his own method of learning languages, and that it is ■ impossible to setup any general principles. But the facts we have been considering certainly tend to show that even if there is not one absolutely invariable method, there are at least general principles. If in learning languages by whatever conceivable method we must all make use of the same fundamental psychological processes, and if these faculties are present in all minds, differing only in degree, it seems reasonable to assume that all learners will have to travel by the same road, although some will take a longer time for the journey. The comparison of the process of learning languages with a journey is halting in this respect, that most of the learners can hardly be said to reach their destination at all ; that is, they fail to learn the foreign language perfectly. But this, again, is only a question of degree ; for it is doubtful whether even the best linguists learn foreign languages perfectly — unless, of course, they learn it under circumstances in which any one might reasonably be expected to become perfectly bilingual. Thus, as already remarked, Palmer was taken for an Arab, but never for an Arab of the tribe he was among, showing that he did not really speak any one dialect perfectly, but took the Arabs in partly by his amazing volubility and powers of mimicry generally. It must be remembered that he was not only a linguist, but also a powerful mesmerist and a most expert conjurer. All this helped the illusion. It is very difficult to get at the exact truth about these born linguists, most of whom are surrounded with a mist of exagge- ration and fable. Indeed, one does not quite see how such a statement as that such-a-one ' speaks forty languages like a native ' is to be tested. One would first have to collect forty indubitable natives ; then to confront them with the linguist ; and then to make sure that their complimentary criticisms of his speaking were to be taken literally. As it is, such statements are generally made by people who know nothing of the lan- guages in question, and who draw their conclusions solely from the fluency of the speaker, or take his statements on trust. The GENERAL PRINCIPLES OF METHOD 85 achievements of Mezzofanti have certainly been exaggerated in this way. I was told by Prof. Johan Storm, who got his information from a Norwegian who had had an interview with the great linguist, that the current statements about his being able to distinguish the different Norwegian dialects were pure fable, and that he kept his visitor waiting a long time in the antechamber, while he primed himself with a selection of Nor- wegian phrases, which he uttered slowly and with considerable hesitation. It is really not difficult to get, or make, the reputa- tion of speaking a foreign language perfectly. An Englishman travelling in out-of-the way parts of South Germany has only to speak anglicized book German to be taken for a Prussian, and then to go home and tell people he was taken for a German everywhere. But even if we grant that some adults are practically in- capable of learning to speak a remote language with fluency, or even of reading its classics with ease, this does not invalidate our conclusion that all must travel by what is essentially the same road : the fact that the traveller does not reach his desti- nation by one road does not prove that he would have got any further by another road. It is lastly to be observed that the doubts and objections we have had to meet are founded on the results obtained by the antiquated methods of study still generally employed in this country. One of the most important results of the perfection of rational methods will be that differences in natural aptitude will be more and more levelled by systematic training. The same adult who would otherwise be incapable of imitating a single unfamiliar foreign sound, would certainly, if he had been trained in phonetics from his infancy, be able to reproduce every foreign sound with ease and perfect accuracy, and would therefore in this important respect be completely on a level with— or rather, superior to — the most highly gifted linguist trained on the old system. No training will ever make a slow mind or a bad memory equal to the mind and memory of a great linguist : we can never expect that all learners will reach the goal with the same ease and quickness. But perfected methods will reduce these inequalities to a minimum ; and we may reasonably hope that they will bring the goal within the reach of all who are ready to make the necessary sacrifices of time and trouble. 86 THE PRACTICAL STUDY OF LANGUAGES Another consideration is, that nothing will ever make the learning of languages easy : it will always be a difficult and unnatural process — unnatural because it involves constant con- flicts with the associations of the learner's native language. It is not true that ' to learn to speak no matter what language is a thing as natural and easy to a child as learning to fly is to a bird.' This was said by Gouin in praise of his own system, the great merit of which, according to Gouin's disciple Swan, is that ' the stupidest scholar can learn it as easily as the smartest ' because 'all intelligences are sensibly equal ' (Br. G. 29). This last statement is only an extravagant exaggeration of the one- method-for-all principle. The preceding statement may be true, but, unfortunately, learning Gouin's method does not imply knowing the language. The Historical Method With the rise of comparative philology and its great develop- ment during the present century came the historical view of language. It was shown that the irregularities and anomalies of language could be explained by comparison with their older forms as preserved both in the earlier stages of the language itself and in the cognate languages belonging to the same family, and that the further a language is traced back, the more clear and regular does its structure seem to become. Hence it was inferred that the historical treatment of language would also lighten the drudgery of acquiring a practical mastery of its grammar. Although the scientific study of language is impossible without historical method, it is possible to carry the historical view of language too far. The historical study of language degenerates into one-sided antiquarianism when, as is often the case, it concentrates all its energies on the determination of the oldest formations in a language or group of languages, valuing the inflections and other forms of modern languages only in as far as they throw light on those of the older stages. The great defect of antiquarianism is that it ignores the fact that every language and every stage of a language has an individuality of its own. It is not enough to trace the forms of a language back to what we conventionally regard as their original forms ; we must also gain a clear idea of the structure of the language of a given period as an organic whole without GENERAL PRINCIPLES OF METHOD 87 regard to the antiquity of its morphological characteristics or their older forms. From this point of view it is, for instance, of primary importance to know that the modern Scandinavian languages have a passive voice, while the fact that this inflection is of late origin is comparatively unimportant. Again, a know- ledge of the fact that such a plural as feet is exceptional and anomalous, and that the great majority of English plurals are formed by adding -s, is essential to the comprehension of the structure of English, while the historical explanation of the origin of the form feet through foii, fosti, feet, fet, fit does not materially assist that comprehension. It is no doubt interesting to know that such plurals as men, feet, mice were once perfectly regular, and interesting to trace the steps by which they gradually assumed their present forms ; but this does not in the slightest degree modify the fact that these plurals are now isolated forms or irregularities. The difficulty the foreign learner feels in mastering such forms lies in the effort of forming associations supported only by a few words, and directly opposed to those involved in acquiring the regular plurals ; nor is the tendency to expect mans, foots, mouses instead of men, feet, mice and the effort of overcoming this tendency at all affected by the learner's conviction that the forms that are now isolated irregularities were once regular. Mischievous as one-sided antiquarianism is in the scientific study of language, it is still more so in the practical study of language. As we see, the anomalies and irregularities of language retain all their practical difficulty, however much they may be illuminated by the light of history ; and the main result of the application of the historical method is to add to the effort of overcoming the cross-associations involved in the anomalies and irregularities themselves, the further one of mastering a number of theoretical statements and of learning a number of hypothetical forms which afterwards have to be unlearnt. The Crude Form System An extreme development of the historical method is the so-called ' crude form ' system. It is strange that the advocates of this system do not see that the student who has learnt, for instance, the Greek paradigm dnax, dnaktos, etc., by heart has learnt exactly as much as 88 THE PRACTICAL STUDY OF LANGUAGES another who has been first taught that the crude form is anakt, that the nominative is formed by adding -s, and that anakts is then contracted into anaks, dnax, the only difference being that the crude-former not only has to learn the actual forms dnax, dnaktos, but also a variety of hypothetical forms, besides having to make the additional effort of remembering that the forms anakt, anakts, etc., do not exist. So also in Finnish the mere juxtaposition of such forms as nominative singular kiisi, ' hand,' illative kciieen, plural nominative kddet, together with the posses- sive nominative singular kateni, ' my hand,' is enough to give a practical knowledge of the fact that the stem or crude form is kdte, from which the nominative singular kdsi and the nomi- native plural kddet are formed by perfectly regular sound- changes. Putting kdte at the head of the paradigm simply unsettles the learner's associations with the nominative kdsi; and the confusion is made worse, when, as is sometimes the case in Finnish grammars, nouns are given sometimes in their nominative singular, sometimes in the form of the bare stem. If the learner only has two such forms as kdsi, kateni, or kdsi, kddet, he has material enough to enable him to construct the stem together with all the inflectional forms. The Etymological Fallacy- Similar criticisms apply also to the ' etymological fallacy.' The meaning of a word in a given period of a given language is a matter of usage, and the fact of its having had a certain meaning at some earlier period or in some cognate language does not necessarily afford any help in determining, and still less in remembering, its present meaning. Etymological trans- lation should, above all, be avoided in dictionaries. Thus in Old English dictionaries we find g<)>ofta defined as ' one who sits on the same rowing-bench, companion ; ' but the only meaning the word has is the second one, the former being an inference from the etymology of the word. The inference is no doubt correct in as far as it assumes that the word had the other meaning once ; but this does not alter the fact that in the language as known to us this meaning does not occur. Besides, any one can draw the inference for himself; so it is a waste of space first to give the etymology, and then to inter- polate the inference drawn from it among the meanings. Etymological translation often takes the silly form of trans- GENERAL PRINCIPLES OF METHOD 89 lating an Old English word by some obsolete or dialectal word which is assumed — sometimes erroneously— to be etymolo- gically connected with the other word, as when the Old English beam is translated by the Scotch bairn — as if modern English were such a poverty-stricken language that it could not find a word for ' child ' ! Then the German lied and the Old English leo\ is translated lay — a French word which has nothing to do with lied. This practice is carried to an extrava- gant extent in many translations from the Icelandic. On this principle we might translate the German Jener kleine knabe ist ?iichtfaul by yon clean knave is naught foul. It has also been suggested to me that the lines in Faust — Bist du es, Faust, dess stimme mir erklang, Ein furchtsam, weggekriimmter wurm ? ought to be translated — Be'st thoti it, Fist, whose voice to me did clink, A frights ome 'way-y crumpled worm? Comparison with Cognate Languages Even when the historical method does not require the help of hypothetical, non-existent forms, it involves the importation of words from other languages into the text-books of the language which is being studied. Now it is true that, for instance, a knowledge of Latin considerably facilitates the acquisition of Italian and the other Romance languages. But where the connection between the two languages is self-evident, the help of scientific historical philology is not needed : every one sees for himself that padre is connected with patrem, aimer with amare. If the connection is not self-evident, the question arises, Is a knowledge of the etymology of any practical use? How, for instance, can the Latin sitim help us to remember the French soif 7 Why, they have only a single sound in common ! ' That is true,' says the philologist ; ' but when the learner has once mastered the intermediate stages, the connection becomes perfectly clear.' Very likely it does; but when it turns out that these inter- mediate stages involve no less than nine distinct sound-changes, some of them very difficult to understand, we are forced to ask, Is it practical and rational to seek our object in so round- about a way ? So also a knowledge of Sanskrit is a great help oo THE PRACTICAL STUDY OF LANGUAGES in learning Zend ; for the languages are so closely allied that whole passages of Zend can be translated into Sanskrit word for word simply by applying the laws of etymological sound- change. But, as we have seen (p. 55), this very closeness is a source of difficulty; so that, instead of wishing to have his associations with Sanskrit strengthened, the learner ought rather to try to forget his Sanskrit as soon as it has helped him over the first difficulties; and consequently he is only exasperated when he finds he cannot look at a paradigm in his Zend grammar without having his mind confused by the constant intrusion of parallel Sanskrit forms. Nor does the beginner in Arabic want to be reminded of Hebrew. It is besides conceivable that the study of Zend or Arabic may be begun without any previous knowledge of the two other languages, in which case the confusions resulting from cross- associations become still more serious. Comparative Philology Sometimes Useful But our scepticism with regard to the help afforded by comparative philology and etymology must not be exaggerated into an unreasoning rejection of it. Cognate forms may be just far enough from one another to make it a matter of doubt whether or not the learner will recognize their affinity; under such circumstances it seems reasonable to give the learner a hint which may perhaps enable him to establish many other similar associations which would otherwise have escaped his notice. Thus, if the learner fails to see that German zehn is cognate with English ten, a state- ment of the correspondence between the initial consonants in the two words will not only help him to remember zehn, but will also enable him to establish an association between German zeit and English {noon)tide, and so with hundreds of other words. But there is alw T ays a danger of going too far; the teacher must be careful not to allow himself to be drawn into an elaborate exposition of Grimm's Law or any other philo- logical generalization until he is quite sure that the practical gain will outweigh the expenditure of time and trouble. In most cases it certainly will not. Fifty years ago, the main laws of Aryan and Romance etymology could be tabulated in a brief space and with delusive simplicity; but nowadays the phonetic changes from Latin to French alone can hardly GENERAL PRINCIPLES OF METHOD 91 be mastered even by specialists, and Grimm's Law has been developed and subdivided into a whole series of laws with endless complications. Under these circumstances, the dream of making comparative philology and etymology a part of ordinary education has to be abandoned. But there is no great harm in occasionally introducing scraps of comparative philology into elementary books, if only the information is correct — which it often is not. Chance Resemblances between Languages Not that this matters much from a practical point of view ; for it often happens that a false etymology is a greater help to the learner than the correct one. Thus every beginner in Greek remembers the meaning of hblos by its similarity in form to the English whole, while its real affinity with the Latin salvits is mastered only by an effort. Such accidental resemblances are instinctively seized on by the beginner as the natural foun- dation of his new vocabulary, and none the less if they appeal only to his sense of the ludicrous or paradoxical, as when Hood says of the French ' they call their mothers mares, and all their daughters fillies.' A Latin primer was once published in which, among other similar suggestions, the learner was told to remember that hasta meant ' spear ' by thinking of the warning not to be hasty with it. This is really more sensible than giving the Sanskrit cognate form. These chance resemblances are especially valuable in learn- ing remote and unconnected languages, where, therefore, there is no scope for comparative philology, and where the new vocabulary is the main difficulty. Thus in Arabic it is some help to note that the first numeral, wahid, begins with the same consonant as the English one, and that the seventh numeral sab' resembles German sieben. So also in learning Finnish we cannot help associating poika, 'son, boy,' with English boy, whether or not we are inclined to believe in any closer connection between the two words. Borrowed Words There are few languages of any degree of culture from which our Western languages have not borrowed to some extent. At any rate, we are generally familiar with some of the proper 92 THE PRACTICAL STUDY OF LANGUAGES names in the language. In the case of Latin, the number of borrowed words is so great that we really know the vocabulary beforehand. English gives us, too, a fairly full vocabulary for Greek also, where affinity with known Latin words is often a further help. Of the remoter languages, Arabic is particularly well represented by borrowed words. When we consider the great difficulty of the Arabic vocabulary, it is a pity that our elementary text-books do not make a systematic use of this link of association. Thus, starting from salaam - Arabic salam, originally meaning ' peace,' we get to the verb sa/im, ' be safe and sound,' whence the fourth form as/am by the regular process of dropping the second vowel and prefixing a, the meaning being ' give oneself up, resign oneself (to the will of God), become a Mahommedan,' whence by equally regular changes the infinitive islam, ' true faith,' and the present active participle muslim, ' true believer,' while in the name Muhammad, ' praised ' or ' praiseworthy,' we have the corresponding passive participle of the second class of verbs, formed by doubling the middle consonant. In this way a few Arabic loan-words can teach us not only a good deal of the vocabulary, but of the grammar as well. So also in Chinese, if we bear in mind that the native forms of Pekin and Nankin, namely pek king and nam king, mean ' north capital ' and ' south capital ' respectively, and that kuarj tung, the native name of Canton and the province in which it is situated, means ' extensive east,' and that the name of the neighbouring province kua-q si means ' extensive west,' we have a memoria technica which helps us to recall the Chinese names of the points of the compass. The associations with borrowed words have this great advan- tage over those with cognate words that the connection between the borrowed word and its original form is generally simple and direct, both in form and meaning. Borrowed words do not generally require any Grimm's Law to explain them. If they are disguised, the disguise is generally a simple one. Thus the fact that Welsh rhwyf, ' an oar,' is the borrowed Latin renins, though not self-evident, is soon made clear by a few of the numerous parallel cases. CHAPTER X SPECIAL PRINCIPLES OF METHOD Rules ; Mechanical Isolation One result of language being only partly rational is that only part of it can be brought under general rules, so that while some linguistic phenomena can be learnt by bringing a number of them under a general statement, others have to be learnt disconnectedly, one by one. With those that can be brought under general statements or rules, the question still remains to be answered for each par- ticular fact of language, Is it worth while referring it to a rule, or is it better to learn it simply as an isolated fact ? The usefulness of a rule depends : (i) on its extent — that is, the number of examples included under it; (2) on its efficiency — that is, the number of exceptions it has to admit, the rule that has the fewest exceptions being the most efficient ; (3) its definiteness, clearness, and simplicity — that is, the ease with which it is learnt and applied, independently of its extent and efficiency. Such a rule as that for the formation of the plural of nouns in English stands high in all three respects : its meaning and scope are definite and clear, and its extent and efficiency make it applicable to every noun in the language, with few exceptions. A still more perfect rule is that of Latin grammar, by which in indirect narration finite verbs are in the subjunctive mood. This example illustrates the fact that syntactical rules are, in the nature of things, more perfect than those which deal only with the forms of words. Many syntactical rules, indeed, hardly admit of exceptions ; when there are exceptions, they are the result of crossing by other syntactical rules, or, at any rate, the exceDtion is one for which a clear reason can be 93 94 THE PRACTICAL STUDY OF LANGUAGES given. Thus in Old English, where the verb is in the subjunc- tive in indirect narration, as in Latin, it is nevertheless put in the indicative in such a sentence as I zvish to say that I am ready to start, because the whole sentence practically means the same as the direct statement / am ready to start, the clause / wish to say being almost an ' empty clause.' The subjunctive is here also avoided because it would imply that the speaker wished to make a false statement. It is evident that our first two criteria balance one another to some extent. If a rule has no exceptions — or none but self-evident and necessary exceptions — it is worth learning, even if it applies to only a few words. If a rule covers a great many words, it may be worth while learning it even if there are a good many exceptions. The exceptions must be in the minority to make a rule worth learning ; if the regular forms are at least twice as numerous as the exceptions, then the rule is generally decidedly worth learning — that is, if a rule is really useful or necessary. Thus, as we shall see hereafter, the genders of nouns in such languages as French and German are better learnt one by one than by rule ; hence it is not worth while to give any rules for gender in these languages except those which practically admit of no exception. All rules of any extent have this great utility, that they tell us how far the analogy of the form we are dealing with extends. Thus suppose a foreigner began the study of English with the word-group hands and feet, or men, women, and children. In either case he would be puzzled by the variety of plural-forms, and would instinctively feel a wish to know whether any one of these methods of forming the plural predominated in the language, and if so, which. The answer, ' the regular way of forming the plural of nouns is shown in the first word ; nearly all English nouns form their plural in this way ; the others are irregular forms which you need not trouble yourself with at present,' gives him the information he wants, enabling him to con- centrate his attention on those forms which he can associate together by bringing them under a simple rule. Where there is greater complexity and irregularity, we may either make our rules correspondingly elaborate, adding long lists of exceptions, or we may content ourselves with giving only those rules fully which are most efficient, and then content ourselves with general statements. Thus in dealing with the complicated noun-plurals in Welsh, we may content ourselves SPECIAL PRINCIPLES OF METHOD 95 with stating that Welsh nouns form their plurals by about a dozen vowel-changes (da/ad, ' sheep,' plural defaid), and by adding various endings (pen, ' head,' plural penau), which are sometimes accompanied by vowel-changes (wab, ' son,' plural meibion). This very general and vague rule may then be sup- plemented by such statements as that the most frequent endings are -au, -iau, -on, -ion ; that the ending -od is used chiefly with names of animals (Ihuynog, ' fox,' plural llwynogod). In Arabic the difficulty of bringing the plurals under simple and definite rules is still greater. Under such circumstances it is safest to err on the side of ignoring rules rather than that of elaborating them. The beginner will find the simple mechanical method of associating each singular form with its plural the most effectual : that is to say, he must repeat such pairs as dafad, defaid, till one form instan- taneously recalls the other. When he has accumulated a stock of examples in this way, he will then be able to derive all the more benefit from learning rules of increasing elaborateness. So also with grammatical gender : the simplest way of learning them is to associate each noun with the definite article or any other word which marks the gender. Thus the learner of German who has learnt to repeat das haus, ' house,' hauser, has a practically exhaustive knowledge of the word. There are other considerations by which the choice between the method of rule and that of mechanical isolation is guided. Such plural formations as those of Welsh and Arabic are very distinctive : they are full and sonorous, and make a strong impression on the ear, so that they have greater individuality, and consequently are easier to discriminate and retain in the memory. But in a language such as German, where the end- ings are more worn away, the inflections have a more abstract character, so that such an ending as -e, though frequently used to form plurals, does not in itself suggest any such idea, because it is used for a great variety of other grammatical pur- poses, besides being in itself of little phonetic weight. Hence German plurals, though simpler than those of Welsh and Arabic, are much more in need of rules to prevent otherwise inevitable confusions. Again, the isolating method does very well with purely formal distinctions such as those of grammatical gender, because these require no thought— nothing but a mechanical asssociation between the noun and certain accompanying words, such as the g6 THE PRACTICAL STUDY OF LANGUAGES definite article. But with syntactical rules such as those for the use of the subjunctive, purely mechanical methods are rarely effective, or, indeed, available : the different construc- tions can only be discriminated by the help of reason and logic. Hence syntactical rules not only tell us how far the analogy of any particular construction extends (cf. p. 94), but they also save us the labour of finding out for ourselves why such a con- struction as the subjunctive mood is used in any particular sentence. So also it would be hopeless to try to master the initial mutations in modern Welsh without knowing the rules which govern their highly abstract and varied syntactical functions. We see, then, that the syntax is the most important part of the grammar, and that it requries a much fuller and more detailed treatment than the accidence. Fortunately, too, syntax lends itself to such a treatment more easily and naturally than accidence does (p. 93). We are now able to answer the general question, Should languages be learnt with or without the help of grammatical rules ? The tendency among reformers now is to revolt against rules, and lay stress on such facts as that ' we learn to speak by pattern rather than by rule' (Paul, 89), and that 'we learn living languages more by imitation than by rules ' (Storm, For- bedret Undervisning, 20). As Storm remarks, ' those who learn such a language as French mainly from grammars are often greeted by Frenchmen with the remark, " What you say is certainly very correct, but it is not French ! " ' But it must be remarked that such results are generally due not to using grammars, but to using the wrong grammars — those which ignore the living language in favour of the old- fashioned literary form of it. No grammar that really restricts itself to modern French can possibly teach anything that is not modern French. It is true that we can often dispense with rules in modern European languages, because they have so much in common grammatically that to a great extent we know their grammar beforehand, just as we do their vocabulary. Thus any one who has learnt the rules for the subjunctive in Latin or German will soon pick up those which govern its use, say, in Italian. Most European languages show a certain similarity in the SPECIAL PRINCIPLES OF METHOD 97 construction of sentences through the great influence Latin has had on their prose. Nevertheless, while admitting the importance of the imitative principle, we must, even from the limited point of view of the modern European languages, add that ' rules are often a great help ' — we may say ' an indispensable help.' Those foreigners who try to learn the English verb without definite rules for its modern use, generally fail to master its delicate syntactical distinctions. This is partly because English lies to some extent off the beaten track of modern European linguistic development. Hence also English learners are at a disadvantage when they learn one of the ordinary European languages : grammatical gender, the subjunctive mood, the accusative case, are all novelties to them, unless they are already familiarized with them from Latin grammar. In remoter languages the necessity of definite rules is felt from the beginning. We can pick up a knowledge of Italian by desultory reading of the Italian New Testament, but this method would break down with such a language as Welsh, although it, too, is an Aryan language : no ordinary learner could be expected to find out for himself the mutations and the different uses of the verb ' to be,' or the principles on which the various equivalents of ' yes ' are formed — without a detailed grammatical analysis all this would be a chaos of apparently arbitrary distinctions. The more unfamiliar the language, the greater the amount of grammatical analysis required, and the more elaborate and detailed it must be. Old English differs considerably from modern English grammatically, and yet I have in my First Steps in Anglo-Saxon been able to give all the grammatical information absolutely required by the beginner in 25 pages, comprising not only accidence and pronunciation, but also syntax and full examples. Even the much fuller grammar in my Anglo-Saxon Primer takes up only 54 pages. Classical Chinese, on the other hand — a language which has no acci- dence whatever, in which nouns have not even a plural, and in which verbs have neither person, tense, nor mood distinguished by form — takes up 84 pages in Gabelentz's Anfangsgriinde der chinesischcn grammatik. And yet the most thorough knowledge of this book will not enable the learner to read a single line of the Chinese classics by himself — so great are the difficulties of H 98 THE PRACTICAL STUDY OF LANGUAGES the grammatical construction in Chinese, which can only be overcome by long-continued and elaborate syntactical training carried on side by side with a careful study of the texts. Analysis and Synthesis Although language is made up of words, we do not speak in words, but in sentences. From a practical, as well as a scien- tific, point of view, the sentence is the unit of language, not the word. From a purely phonetic point of view words do not exist. As I have said in my Primer of Phonetics (p. 42), ' No amount of study of the sounds only of a sentence will enable us to recognize the individual words of which it consists. We may write down every sound, every shade of [phonetic] syn- thesis, but we shall never be able to analyse the sentence into separate words till we know its meaning, and even then we shall find that word-division postulates much thought and com- parison of sentences one with another.' Thus the sound-group (tela) may stand for the single word teller or the two words till her, there being no more pause between the words of a sentence than between the syllables of a word. In French, where word- division is much less clearly marked by stress and other formal criteria than in English, it is still more difficult to mark off the divisions of words by ear only. Thus the title of Darmesteter's well-known popular book on etymology, La vie dts mots, is pro- nounced (lavidemo) with practically equal stress on all the vowels, and nothing to show, as in English, whether the internal consonants form groups with the preceding or the following vowel, so that if we did not know what it meant, we might transcribe it into nomic spelling in half a dozen ways, especially if some unknown proper name entered into it : la vie . . . , I'avis . . . , Vavide et ...... . des mots, . . . d'e maux, Lavy ....... Manx, . . . Desmatix. We see, then, that there are two ways of dealing with languages: (1) the synthetic, which starts from the sentence; (2) the analytic, which starts from the word. From the point of view of the practical study of language the synthetic method implies that the analysis of the language is not carried further than, at the most, cutting it up into sentences, which are grasped and learnt as wholes, instead of being separated into words, and put together like pieces of mosaic, as on the analytic method. SPECIAL PRINCIPLES OF METHOD 99 As the division of sentences into words is an essential preliminary to grammatical study, the synthetic principle is as opposed to grammatical analysis as it is to the analysis of a sentence into words. The great development of analytic methods in modern times is partly the result of our fixed word-division in writing and printing, partly of the increasing elaboration of grammars and dictionaries, and partly of the growth of minute scholar- ship, philology, and etymology. These analytic methods are often carried to a monstrous and almost incredible extreme in the historical and ' scientific ' study of dead languages, as elaborated in Germany, and now being imported into this country. On this system, the words of an Old English or any other text are taken word by word and discussed etymologically, each word being transliterated into the form it assumes, or ought to assume, in the other cognate languages. The result of such a method is that the students learn a good deal about words, but nothing about the language itself, the sense of whose individuality is com- pletely lost amid the chaos of conflicting associations. Paradigms A knowledge of the grammar by no means necessarily implies a knowledge of the language itself : the grammar with its rules and paradigms merely gives the materials for acquiring that knowledge. The schoolboy who has learnt his tiipto, tupteis, tiiptei, t&ptomen, tuptete, tiiptousi by heart has simply established a series of external associations between these six words, an association which is at first so strong that he is unable to get to his tuptei or tiiptousi without repeating all its predecessors in order — an association of which the actual language knows nothing. It is not till such a context as ' the master beats the boy when he does not know his lesson ' lias been learnt in Greek, so as to establish an instantaneous association between thought and sound, that any real know- ledge can be said to be gained. Nor does being able to state that lune in French is feminine necessarily imply a practical knowledge of its gender. When the student has learnt to associats lune with the article la or with the adjectives belle, blanche, he really knows its gender; till then he has simply transferred the 'lune, subst. fern.' of his dictionary to his roo THE PRACTICAL STUDY OF LANGUAGES own memory, and has, after all, only facilitated his reference to a statement, not mastered the fact it involves. In the case of paradigms such as tupto, tupteis, tuptei, there is a certain amount of natural association between the words — although so weak that we can scarcely imagine them ever coming together in one sentence — and this is one of the justifi- cations of the practice of learning paradigms. But there is no natural association between lune and the word 'feminine,' or the letters s.f., or with printing in small capitals — lune — as has been actually proposed as a means of learning the genders, and consequently these associations are useless ; while the simple rule of never repeating lune without a preceding la establishes a natural association, and at the same time gives all the informa- tion contained in the statement that the word is feminine. Learning Lists of Words The worst kind of isolation is to begin the study of a language by learning lists of words by heart : ' if I learn two hundred words a day, I shall have a perfect knowledge of German in a fortnight.' It is conceivable that there may be a period when the learner finds it worth while to sum up his knowledge of the vocabulary of the language he is studying by running over ' kopf head, auge eye, ohr ear,' and so on ; but the beginner is not concerned with isolated words, but with their combinations into natural sentences : it is no use telling him that kopf means ' head ' when he wants to say ' the head ' or to speak of ' heads ; ' nor would even the information contained in der kopf ' head,' plural kopfe, be of any use to him till he had learnt some grammar, which again implies previous text-reading. Detached Sentences ; Context As already remarked, we speak in sentences. But we do not generally speak in detached sentences ; we speak in con- catenations of sentences. In ordinary speech sentences are connected together in the form of a dialogue, which, again, often consists of an alternation of questions and answers. In books sentences are joined together into larger groups called paragraphs, which again form chapters, which again constitute a complete connected text. The relations between sentences and texts are analogous to SPECIAL PRINCIPLES OF METHOD 101 those between words and sentences : both are relations of context. Just as a word apart from its context may be ambiguous both in grammatical form and in meaning — for even in Latin we cannot tell, apart from the context, whether boni is genitive singular or nominative plural — so also, though in a less degree, the grammatical construction or the meaning of a sentence may- be ambiguous when it is detached from its context. Hence, also, the meanings of words are brought out more clearly in connected texts than in detached sentences. These considerations point clearly to the conclusion that the main foundation of the practical study of language should be connected texts, whose study must, of course, be accompanied by grammatical analysis. But in a grammar, the rules must be illustrated and justified by examples, which also serve to strengthen the learner's hold of the rule, and to make it easier for him to recognize the working of the rule in the texts he reads. These examples must in the nature of things be detached words or detached sentences. For this and other reasons we cannot dispense with detached sentences. But we must be careful to employ as far as possible only those sentences which will really bear detaching. Such a sentence, for instance, as the sun rises in the east and sets in the west conveys a perfectly definite and distinct meaning, and requires no further context. In grammars in which the examples are taken from the higher literature we often meet sentences which are almost unintelligible. One of the great weaknesses of the a priori methods of the Ollendorff type is that they involve the substitution of detached sentences for connected texts. But detached sentences are not peculiar to these methods. They are the natural and inevitable result of all methods which make the grammar the centre of instruction instead of the texts. Widgery remarks (p. 47), quoting partly from Perthes : — ' Since the vain attempt to teach a language by means of short, disconnected sentences was introduced into Germany about seventy years ago, there has been a steady rise in the number of hours devoted to Latin, but the results are not better, nay, they are worse. After the French Revolution had given the death-blow to the real use of Latin as a means of io2 THE PRACTICAL STUDY OF LANGUAGES communication, this new method was gradually evolved in the hope of infusing some show of life into the ghostly dilettan- teism of "prose composition," that sickly branch of study kept alive only by the golden sap of prizes and scholarships.' Storm remarks (Forbedret Undervisning, 17) with special reference to modern languages : ' It is but little relief in the study of a difficult grammar to have to ruminate hour after hour dry, detached sentences without a trace of connection, indeed often without intelligible meaning.' He then gives an extract from a manual of French, which he says is in pretty general use in Norway : The more merit one has, the more modest o?ie is. Thy sisters ate apples, and mine ate nuts. Receive, sir, the assurance of my high respect. These (!) threw bombs into the fortress in order to compel the besieged to surrender. Yield to his importunity, if you do ?iot possess enough strength to make a resistance. As he remarks, ' an intelligent pupil will ask, Who are these ? But such unintelligible language has simply a stupe- fying effect on most learners ; the meaning is entirely lost to them ; and how much they retain of the French form it is not difficult to imagine.' In that form of the Ollendorff system developed by Prender- gast in his ' Mastery Series,' each detached sentence is regarded as a bag into which is crammed as much grammatical and lexical information as it will hold. The following are examples of this ' sentence-cramming ' method as applied to French and German : — 1 Pourquoi ne voulez-vous pas me faire le plaisir de passer avec moi demain chez le frere de notre ami dans la rue neuve ? N'avez-vous pas besoin d'aller h. Londres aujourd'hui, avant votre promenade du matin, chez le cordonnier francais, pout faire elargir vos bottines ? Dites au gargon, je vous prie, de m'apporter tous les jours sans faute, a sept heures ou plus tot s'il peut, un pot d'eau chaude, une tasse de cafe au lait, et mes habits bien brosses. Savez-vous comment se nomme cette vieille dame anglaise qui demeure pres du pont neuf, dans la meme maison ou il y a une famille francaise, et un jeune ministre allemand ? J'ai eu pour moins de deux francs dans un grand magazin de Paris 011 tout se vend bon marche, du papier a lettre tres-beau, des plumes me'talliques excellentes, et un joli petit buvard. SPECIAL PRINCIPLES OF METHOD 103 Da er, der junge Freund des reichen Marines, dem Diener den Brief nicht hat geben wollen, so werden Sie mir ihn gleich holen lassen miissen. Wenn der alte und kluge Lehrer uns den guten Rath selbst gegeben hatte, wiirden wir diesen grossen Fehler kaum gemacht haben konnen. Die kleine Freundin der schonen Dame liess sich die neue Kutsche nach der nachsten Station der Eisenbahn schicken, um in derselben zu der Stadt zu fahren. Aber endlich schickte die Alte aus, und bestellte ihre Kiste, weil sie die Absicht hatte, die schon oft vorgehabte lange Reise sobald als nur moglich zu unternehmen. Ich hore, dass das schon gestern Morgen friih erwartete Schiff selbst heute Abend, wie ich glaube, wegen des schlechten Wetters, schwerlich mehr hier ankommen wird. Ein dringender Brief eines kranken Geschaftsfreundes, welcher mir so eben gebracht worden ist, nothigt mich zu einem kurzen Ausflug, um einen keinen Badeplatz an der Nordkiiste von Deutschland zu besuchen.' This attempt to give each word a context without overstepping the boundaries of a single sentence must be pronounced a failure. The sentences are quite as insipid as those of Ahn, and even more unnatural and impossible ; the last sentence is practically nonsense as it stands. The construction of the German sentences is stiff to the last degree ; observe the repe- tition of the ' split article ' construction in two consecutive sentences (the fourth and the fifth). The incessant heaping of epithets — 'the young friend of the rich man, the old and saga- cious teacher' — is alone enough to give an uncolloquial, or rather exaggeratedly literary, character to these sentences (p. 52). Association The psychological foundation of the practical study of lan- guages is the great law of association, to which we have fre- quently had occasion to allude already. The whole process of learning a language is one of forming associations. When we learn our own language, we associate words and sentences with thoughts, ideas, actions, events. The words themselves are associated into groups of various kinds. Thus such words as tree, wood, forest form an associa- tion-group by virtue of their meaning ; the words trees, woods, 104 THE PRACTICAL STUDY OF LANGUAGES forests also constitute a group in another way, namely, by all having the same plural inflection; all six words, lastly, are associated together by forming part of the grammatical group ' nouns.' These groups are independent of any linguistic con- text : even if we never met tree and wood associated together in one sentence, the mind would still pick them out and associate them together by virtue of the meaning and grammatical function they have in common. These groups often cross one another in different ways ; thus wooden by its meaning belongs to the preceding group, but from a grammatical point of view it is outside it, and belongs to the same group as such words as good, green. These associations are unconscious, but none the less real : every speaker of English, even the most uneducated, knows instinctively what a noun is. The sole problem of grammar is to make these unconscious associations into con- scious and analytic ones by defining and analyzing them, and stating them as briefly and clearly as possible by means of a suitable terminology. The function of grammar is, therefore, to sum up the associa- tions by which we all understand and speak our own language as well as any foreign languages we may learn. When we say that certain nouns are feminine in French, we mean that they are associated with certain forms of the definite article and other adjectives, which we call ' feminine,' because these forms are to some extent also associated with the idea of the feminine or female sex. We have seen that the practical way of learning genders is to start, not with the abstract grammatical statement, but with the actual associations themselves. But when we have accumulated in our memory a certain number of direct associations such as that between la and tune, maison,femme, and between le and soleil, garcon, it is a help to have all these associations summed up in a brief statement, the more so as some of the associations connected with gender are complicated and contradictory. Thus the learner of French finds that (la) and (bon) are regularly associated together (la bon itiezo), while (la) is generally associated with (bo) ; he is then puzzled to find the collocation (la bon om). Here the grammar comes in, and saves him the trouble of collecting a large number of examples and comparing them, by informing him that such masculine forms as (bo) assume the feminine form before a word beginning with a vowel. SPECIAL PRINCIPLES OF METHOD 105 Unconscious association is not, as we have seen, necessarily dependent on the actual juxtapositions which occur in language itself : there is a real unconscious association between the forms tupto, tuptcis, tiiptei, between see, saw, seen, and between am, is, are, be, in spite of these four words not having a sound in common ; although, as these forms could hardly occur together in a sentence, the association is not so strong and direct as, for instance, between la and maison in French. But they may easily be associated together in two connected sentences ; and such a dialogue as are you ready ? yes, I am ready, but he is not ; he will soon be ready though, implies a definite association between the four verb-forms that occur in it ; the dialogue would, indeed, be impossible if the second speaker had not a clear feeling that / am stands in the same relation to he is as / see, I h:ar, do to he sees, he hears. There are two association- groups connected with every inflected word : one which con- nects it with all other words having the same inflection, as in the group he sees, he hears, he comes, he is ; another which groups together all the inflections of the same word, as in the group / see, he sees, saw, seen — am, is, was, been, be — tree, trees — man, man's, men, men's. So also there are groups formed by deriva- tion and other formal changes, such as big, bigger, biggest — happy, happily, happiness, unhappy, which again involve such groups as bigger, stronger, less — biggest, least — happiness, good- ness, unselfishness. This is an additional proof of the utility of grammatical paradigms. A paradigm of the Latin declensions is simply a brief summary of these unconscious associations which we have just been describing. A paradigm is useful both as a guide through the mazes of these often conflicting associations, and also as a test of the learner's practical mastery of them. In this way, the being able to repeat a paradigm by heart, useless as it would be to the beginner, is a gain to the more advanced student, for it strengthens and reduces to order associations already partially formed — or, at any rate, prepared — by a natural process. The following are the main axioms of the principles of association : — (1) Present the most frequent and necessary ele- ments first : The first associations are the strongest, because they are the 106 THE PRACTICAL STUDY OF LANGUAGES least disturbed by conflicting associations, because they have the longest time to establish themselves, and because the greater effort required in mastering the first elements fixes them more strongly in the mind. It is evident that in learning a language we should establish the strongest associations with — that is, we should begin with — the commonest and most necessary words, phrases, idioms, and constructions of the ordinary spoken language before proceeding to the vocabulary and style of the higher literature. This principle has been well illustrated in our discussion of the relations of the spoken to the written language (p. 52), where we have also seen its importance in cases where cross- associations arise. When a foreigner learns archaic literary and modern colloquial English simultaneously, he constantly hesi- tates between such forms as he hath and he has, quotha and he said. But if he first forms strong associations with he has and he said exclusively, he can then form weaker secondary associa- tions with he hath and quotha without much fear of their inter- fering with the primary associations. (2) Present like and like together, and then (3) Contrast like with unlike till all sense of effort in the transition ceases : Thus in learning the English noun-plurals, the beginner may have the regular inflections exhibited in a variety of nouns. Then, when these are firmly fixed in the mind, he will have the mutation-plurals, such as men, geese, brought before him in a group, till they also are firmly fixed in the mind. Lastly, the regular and irregular forms may be contrasted in carefully selected natural collocations such as hands and fed, ducks and geese, men and animals, till not only the sense of discontinuity of association is overcome, but a new special association is formed between the contrasted words, so that the one suggests the other, and both in common suggest, and are suggested by, the idea of plurality. It is to be understood that this is not intended as a model way of learning English, but simply as an illustration of how the principle of association works under certain given conditions. So also in teaching German, it is a violation of the principles of association to put before the beginner such a contrast as that between der band, ' volume,' and das band, ' ribbon ; ' these words ought at first to be kept entirely apart and mastered separately, each in its natural SPECIAL PRINCIPLES OF METHOD 107 context. But when they have been learnt in this way, it is not only allowable but advisable to confront them, and call the learner's attention to the difference of gender. Otherwise he might be tempted to transfer the gender of the word he was more familiar with to the less familiar one. (4) Let the associations be as definite as possible : Thus in giving examples of the use of the ablative case in Latin, the grammarian should be careful to choose, as far as possible, sentences containing words whose ablative case is distinct from their dative. So also no text should be published for beginners without full phonetic information in the way of quantity-marks, stress-marks, and so on, in addition to all the helps that can be given by the use of the ordinary marks of punctuation, the use of italics, etc. : if the learner of Latin were taught from the beginning to recognize the distinction between such pairs as labor, ' labour,' and labor, ' I slip,' populus, ' people,' and populus, ' poplar tree,' by eye and ear, instead of having to rely entirely on the context, he would certainly learn to under- stand Latin quicker. The common practice of withholding information of this kind with a view to exercising the learner's intellect and test- ing his knowledge is an example of the violation of this prin- ciple of association. Thus in text-books of Oriental languages it is usual to give transliterations only on the first few pages, not to mark the short vowels (in Arabic) after a certain page, adding them, only when the learner is supposed to want them. But as no one can possibly tell beforehand the weak places in another person's memory, each learner complains that the information he wants is withheld, and that which he does not want is repeated over and over again. Gabelentz, in his Chinese grammars, shows his practical good sense by invariably giving transliterations however frequently the word may occur. It is a pure fallacy to imagine that withholding information and forcing the learner either to guess or waste time in seeking elsewhere for the information withheld adds to his knowledge : on the contrary, it not only puts superfluous mental labour on him, but also weakens his associations, and leads him into inevitable errors, which can be corrected only by still greater and more painful efforts. All examination and testing of knowledge should be reserved till there is reasonable ground for supposing that the learner has a firm grasp of the subject. 108 THE PRACTICAL STUDY OF LANGUAGES (5) Let the associations be direct and concrete not indirect and abstract : The crude form system (p. 87) is an example of the fallacious substitution of indirect for direct associations : the learner has first of all a non-existent crude form or stem presented to him, and then is taught how to deduce from it the actually existing form which ought to have been presented to him at the outset. All associations which involve remembering a certain order — first class, second declension, third conjugation— are indirect, just as calling a certain class of people ' the third estate ' is less direct than calling them ' the commons,' which does not involve any knowledge of what the other two estates are, and what the order of the three estates is. All associations of order should be made direct, as when we call that group of strong verbs to which Old English ceosan belongs the ' choose-class ' instead of expecting the learner to identify it by remembering its order in a series. The objection to going entirely by numbers is not only that it is difficult to remember the order, but also that all numbering is essentially more or less arbitrary, so that there is always a possibility of a variety of orders. Thus in my arrange- ment, the choose-class is the seventh, in Germany it is the second, while the name ' choose-class ' has the double advantage of conveying information instead of being purely negative, and of being entirely independent of any changes of order. Of course, if the order is part of the meaning of the words, the mere repeating of them in their order — which, on this supposition, is always a fixed and definite one — does establish a direct association with the meaning of each separate word, as when we repeat one, two, three . . . , first, second, third . . . , Monday, Tuesday, Wednesday . . . , October, November, December, January . . . , and in a less degree in such groups as north and south, east and west, and men, women, and children, or ladies and gentlemen, because these orders may differ in different languages. Thus in Chinese the order of the four quarters is nam, pek, si, turj, ' south, north, west, east.' The order in such groups is very strict in Chinese, and has great grammatical importance, for any deviation from the fixed order implies change of construction. Thus tsi' niti ', ' son daughter,' means ' sons and daughters ' on the principle of the male preceding the female; if this normal order is inverted, niu ts'{, the combination becomes attribute + noun instead of being co-ordinate : = ' female child.' SPECIAL PRINCIPLES OF METHOD 109 As a further illustration of the distinction between direct and indirect associations, if we wanted to distinguish between the two Mills, we might either state it abstractly by saying ' the elder Mill's christian name was James ; he wrote the History of British India and . . . the younger Mill's christian name was John Stuart ; he wrote a System of Logic and . . .,' or we might simply repeat to ourselves ' James Mill history of British India, John Stuart Mill system of logic . . .' It is evident that the latter method would establish more direct associations, because the association of ideas would be helped by the sounds and organic formation of the words themselves. Such mechanical and external associations are of the greatest help in learning languages. (6) Avoid conflicting associations (cross-associations) : Attempting to teach a language through its nomic form and its phonetic transcription simultaneously would be an example of the violation of this principle. It also involves basing the study of a language at first exclusively on one definite dialect and period. Thus it involves not beginning to read Herodotus till one has a firm grasp of Attic Greek. A striking example of the ill effects of cross-associations is afforded by the gender-lists which figure so prominently in some French and German grammars, especially the older ones. In these grammars we are told that a certain ending is feminine except in four words, which are given together with, perhaps, two words to serve as examples of the regular feminine gender. The result is that a stronger association is established with the irregular than with the regular forms. This is really worse than cross-association, it is ' inverted association.' A further objection to this procedure is that the lists of exceptions must either comprise a large number of rare and useless words, or else be incomplete and comparatively useless, for it is impossible to draw any definite line between rare and frequent, useful and useless words. # It is to be remembered that this axiom is only meant as a protest against unnecessary cross-associations caused by defective methods of teaching or defective statements of the phenomena of language. Those cross-associations which are inevitable — that is to say, which exist already in language itself, are dealt with under (3). no THE PRACTICAL STUDY OF LANGUAGES Memory ; Repetition The next problem is, how best to retain these associations in the memory. As even the strongest associations are liable to be weakened by disuse and lapse of time, the principle of economy is all- important : that is, of giving the learner only such material as he wants at the time or is likely to want within a short period. Thus, if he is to give a certain time to reading nothing but Caesar's Commentaries, in which the verb occurs only in the third person, it is evident that if he is to be provided with a special Latin grammar- for that purpose, it ought to exclude the first and second persons of the verb. In the German grammar I began with the word homung, ' February,' was given as an excep- tion to the rule that nouns in -wig are feminine, and for many years no German word was more familiar to me, except perhaps petschaft, ' seal,' whose acquaintance I made at the same time and in the same way. But to the present day I cannot re- member having met with either of them in any Modern German book, still less of ever having heard them in conversation, hornung being now entirely obsolete except in some German dialects. At last, when I began Middle High German, I met with it for the first time in my life in a poem of Walther von der Vogel- weide, but by this time I had forgotten all about it, and so failed to recognize it, especially as it appeared in the slightly disguised form of hornunc, which, I know not why, made me guess it to mean ' hornet.' I am glad to see that this and other words of a similar character are now often omited from German grammars. » Economy teaches us to begin with as small a vocabulary as p Jssible, and to master that vocabulary thoroughly before pro- c jpding to learn new words. In this, and in many other ways, it Confirms the general principles of directness and simplicity. ^Repetition is essential both for forming associations and retaining them in the memory. It is an additional argument for working as long as possible with a limited vocabulary, for the smaller the vocabulary, the greater chance the different words, forms, and constructions have of being repeated. But there is a point beyond which repetition becomes wasteful — and in two ways. In the first place, the excessive repetition SPECIAL PRINCIPLES OF METHOD in of one detail hinders the due repetition of other details. Secondly, such excessive repetition is wearying to the learner, who is already familiar with the detail in question, and so any further repetition of it causes his attention to flag. This is the great danger of using grammatical illustrations made on the impulse of the moment by the writer instead of being collected from a variety of texts by different authors. In such illus- trations certain words and constructions tend to recur with a frequency of which the writer is unconscious until he revises what he has written from this special point of view. He will then find that in a chapter on the syntax of the numerals he has, perhaps, given one particular numeral five times as often as any other, and has omitted to give any examples at all of some of them, when he might just as well have utilized his sentences to give each of the more important numerals a fair proportion of examples. The various devices of artificial memory or memoria technica are of even less use in language than in other branches 01 study. The whole business of learning languages consists in establishing associations, which can often be effected only by long-continued effort. It is therefore a waste of energy to take on oneself the additional burden of the extraneous associations by which an artificial memory is built up. Such devices as printing feminine nouns in a dictionary in capitals are liable to similar objections, and are quite super- fluous (p. ioo). Of course, if extraneous associations come unsought, they should — and, indeed, inevitably will — be utilized, as in the cases already discussed under the head of ' accidental resem- blances ' (p. 91). But most of these are not strictly parallel to memoria technica — at least, not those in which the association between the two words is direct, as in hblos = whole, and does not require the introduction of a third element. Some of the methods recommended under the head of ' nomic pronunciation ' (p. 34) have also a resemblance to memoria technica, but they are simply cases of the modification of the materials of existing associations. Learning by heart should not be attempted till the piece has been thoroughly studied from all sides. To learn a piece by heart before it has been properly studied and grammatically ii2 THE PRACTICAL STUDY OF LANGUAGES analyzed is often rather injurious than otherwise, as it tends to take away the sense of interest and freshness, and to deaden and blunt the observing faculties. Besides, by the time the piece has been thoroughly studied, the knowledge implied in learning by heart will have come of itself if the learner has a fairly good memory. If he has not, learning by heart is simply a waste of time. If he cannot retain in his memory even a short, simple poem in his own language, he cannot be expected to learn by heart in a foreign language ; and if he can learn his own language by imitation and reproduction after a pattern without learning by heart, he can do the same with a foreign language. Interest Memory depends also on attention, and this partly on the interest taken in the subject. If we take no interest either in the language itself or the text we are reading, our attention inevitably flags. The genuine linguist, on the other hand, is only stimulated all the more by difficulties. Oriental languages are more difficult than European languages, but they have the charm of remoteness and complete novelty, and stimulate curiosity and interest to the highest degree, so that in learning them we endure drudgery which would seem intolerable if spent on a comparatively insipid Romance language, which we half know beforehand. But we must be careful not to confuse interest in the litera- ture with interest in the language. An absorbing interest in what we are reading, speaking, or hearing, so far from helping us to remember and observe the phenomena of the language, has the opposite effect. If the reader is ' panting to arrive at the thrilling denouement ' of a sensational novel, he is certainly not in the mood for observing niceties of syntax. Another difficulty is that the unfamiliar is what is interesting, while all sound principles of linguistic study tell us that we ought to begin with the expression of those ideas and the descriptions of those things and circumstances which are most familiar to us, or will be so when we have acquired the language. In learning French we ought to begin with what is common to both France and England, French and English life, and when we pass beyond English associations, to be initiated gradually into French ones : we do not wish to SPECIAL PRINCIPLES OF METHOD 113 accompany Jules Verne into the heart of Africa. Nor will reading about exciting adventures of Englishmen in New Guinea give a foreigner a good vocabulary for a visit to London. Then, again, all reading that is profitable from a linguistic point of view must at first be very slow, and interrupted by incessant repetitions ; and no text can be very interesting under these conditions. If the learner is interested in the language itself, that is enough. If he has a strong motive for learning the language as quickly as possible as a means to an end, or simply because he wants to get through the drudgery as quickly as possible, he will regard those texts as most satisfactory which bring him to the goal with the greatest ease and quickness ; that is, he will prefer texts in which the meanings of words and their constructions unfold themselves easily from a simple context of progressive difficulty, in which there is repetition enough to help the memory, and yet variety enough to keep the attention on the alert. He will prefer such texts as long as they are not ostentatiously trivial and vulgar, to more interest- ing ones with which he feels he is not making the same linguistic progress. If he has to choose between an anecdote of a Lacedemonian and an Athenian, a fable about a fox and a goat, a funny story about a red rose, ' Twinkle, twinkle, little star,' and a description of the furniture of a drawing-room, he may possibly choose the latter for a variety of reasons : because he knows the anecdotes already, because he does not care for poetry, but mainly because he thinks the description of a drawing-room may teach him some words which he cannot find explained in his dictionary, and which may be useful to him when he visits the country itself. The Gouin-method is a good instance of the ' interest-fallacy.' According to Gouin himself, his series-method was first sug- gested to him by observing a nephew of his, who, after seeing a mill for the first time, began to play at being a miller, talking all the time to himself, ' First I fill the sack with corn — then I put it on my back and carry it to the mill . . . the water falls on the mill-wheel, and the wheel goes round- the wheel turns the millstone — the millstone grinds the corn,' and so on. Gouin fails to see that there is a wide difference between taking a lively interest in a novelty and being interested in the 1 ii4 THE PRACTICAL STUDY OF LANGUAGES vocabulary connected with the object after it has ceased to be a novelty. Even while the child was playing at being a miller, its interest was not in the words, but in what the words ex- pressed : the attitude of its mind was that of the absorbed novel- reader. We know how soon the child's mind tires of any one object of interest ; and we may be sure that if a year afterwards M. Gouin's nephew had had to go through the same mill-series in a foreign language, the old interest would not have been forthcoming, and the youth would perhaps have declined to take part in any series in which tin soldiers and a popgun did not figure. If the old interest had been forthcoming, it would have been as much a hindrance to mastering the details of the foreign language as in the case of the novel-reader. Besides, all children are not equally interested in the construction of a mill, even when it is a novelty; and certainly some of the series, such as that which gives a detailed description of opening and shutting a door — ' I walk towards the door, I approach the door, I approach nearer, I approach nearer still, I put out my arm, I take hold of the handle ' — are as uninterest- ing as they are useless. As I have indicated already, the only safe concessions that can be made to interest are negative : be dull and common- place, but not too much so. Thus, although repetition is essential, there are some kinds of repetition which are so wearisome to the learner that they can hardly be used in teaching, in spite of certain special advantages they possess. I mean such methods as that of repeating a long Latin speech in oratio obliqtia in order to show the accompanying changes of construction, or of con- jugating a whole sentence through a variety of moods and tenses. It is strange that Gouin, who attaches so much impor- tance to stimulating the pupil's interest in the subject-matter, should advocate teaching the verb by means of such repetitions as these : ' To-day the postman will come before we have breakfast — while we are at breakfast — after we have had break- fast. Yesterday the postman came before we had breakfast . . . to-morrow the postman will come before we have break- fast . . .' Such methods should only be used occasionally in the grammar, not made a standing feature of the method. SPECIAL PRINCIPLES OF METHOD 115 Relations between Texts, Grammar, and Vocabulary We have seen that the traditional division of the materials and apparatus for the practical study of languages into (1) Connected texts — the reader, (2) Grammar, (3) Dictionary, vocabulary, is founded on the nature of things. We now have to consider the relations between these three. We have already come — either expressly or tacitly — to the following conclusions on this subject : that the beginner's grammar ought to deal only with the inflections and construc- tions which actually occur in the texts he is reading, and that the dictionary — if a dictionary is used at all — ought to take the form of a special glossary to those texts. My Anglo-Saxon Primer is a simple typical example of this threefold division in a single book. Deducing Grammar from Texts. — But some reformers go further than this. Some of them go so far as to abolish grammar altogether, at least in the elementary stages, and train the pupils to deduce the laws of the language — the rules of grammar — from the texts they are reading. An obvious objection to this plan is the time it would take. The most practical way of collecting materials for grammatical investigations is to write each quotation on a separate slip of paper, adding the necessary headings, arid then to sort the slips under these headings. I am told that the great English lexico- graphers of the present day look down with contempt on any- thing less than a ton of such materials ; but I am sure that by the time the boys had sorted a hundredweight or so of slips, they would have had enough of it ; and by the time the master had gone over the work of a biggish class of boys, he would have had enough of it too, and would perhaps welcome the suggestion of one of the German reformers, namely, that of using copybooks with printed headings and blank spaces to enter the quotations in. But even when all the boys' mistakes had been corrected, the material would still be defective, and would require to be supplemented from other texts. To make a long story short — the master would find it best in the end to n6 THE PRACTICAL STUDY OF LANGUAGES do the work himself; and at last, perhaps, a happy thought would dawn on him : Why not print the whole thing ? The book would be useful to other teachers, and it might pay. When the book was published, the author would discover to his astonishment that the result was nothing more or less than an ordinary grammar. These considerations show that this method would be a failure, if carried out on any large scale. It would involve great waste of time and effort as compared with the ordinary grammatical methods. And there would be a sense of unreality about it : teachers and pupils alike would feel that they were only playing at grammar — pretending that they had to make their own grammar, while they knew perfectly well that the work had been done for them long ago, and that the results were accessible in hundreds of grammars of every degree of elaborateness. This method of ' inventional grammar ' would be highly useful as an occasional stimulus and exercise for the pupils, but there its legitimate sphere of usefulness would end. Such inventional methods — of which Spencer's Inventional Geometry is a good type — have often been tried in various branches of education. There is certainly something plausible in the idea of making the learner's progress consist in finding out by himself the solution of a series of problems of progressive difficulty and perplexity, till at last he stands on the highest pinnacle of knowledge with the proud consciousness of having arrived there entirely by his own efforts. But although these inventional methods excite great interest at first in the minds of the more gifted pupils, those who are less original and slower in mind instinctively rebel against them, and all, sooner or later, get tired of their sham originality. As regards the difficulty of the problems or other work involved in inventional methods, if we look at the question from the point of view of the average learner, we have to face this dilemma : if the work really requires much thought or originality of mind, it will be too difficult for them, or, at any rate, will cause them to make so many mistakes that the labour of establishing correct associations will be far greater than it is worth ; if, on the other hand, the work is so easy as not to tax the intellectual powers of the pupils, it will cease to excite their interest. But there is a method allied to the inventional which may SPECIAL PRINCIPLES OF METHOD 117 form an integral part of a systematic course of linguistic study — that is, the method of inductive grammar. It is only the fully developed mind of the adult that can plunge straight into the study of the grammar of a foreign language. A less developed mind, one which is less used to dealing with general and abstract statements, requires to start with something more individual and concrete. There is, besides, as we shall see hereafter, a pre-grammatical stage in every progressive course of linguistic study — whether for children or adults — in which no grammar is taught, but only the materials on which grammar is based, that is, sentences and short texts. In the case of very young children, the pre-grammatical stage is, indeed, the only one suitable for their intelligences. Now, although the grammar is rightly banished from this stage, it is possible to familiarize the pupils with some of its principles almost from the beginning, that is, as soon as they have read or heard enough to furnish a few examples of some grammatical category. Thus, as soon as they have met with three or four examples of a certain case or other inflection, the teacher calls their attention to this category by writing the words containing that inflection on the blackboard, and making them see what these words have in common, as far as is possible without using any technical or abstract terminology. In the same way he can collect together on the blackboard from the texts already read the scattered words which make up such a paradigm as / am, you are, he is. When this has gone on for some time, the teacher may expect the pupils to find out for themselves what grammatical category a word belongs to. This, then, is the deductive method of teaching grammar, or rather of preparing for the systematic study of grammar. It is capable of various stages of development, according to the mental development of the pupils, according as the grammatical categories are left undefined, or are stated explicitly in more or less technical and abstract language. As already remarked, there will be no harm in varying the course of inductive grammar by an occasional application of the inventive method — letting the pupils find out some of the categories by them- selves — although, for the reasons already given, this ought not to be made an integral part of the course. After all, the main thing is that the texts and the grammar should be intimately associated, and studied as much as possible simultaneously — the exact order is generally of less importance. n8 THE PRACTICAL STUDY OF LANGUAGES Stages of Progressive Method : Irregularities I will now give a general sketch of a rationally progressive method of linguistic study on the principles already discussed. The complete course may be divided into five stages : ( x ) the mechanical ; (2) the grammatical ; (3) the idiomatic and lexical (dealing with the vocabulary of the colloquial language) ; (4) the literary ; (5) the archaic. (1) The first stage, the mechanical, begins with a thorough mastery of the pronunciation of the language which is being learnt, which presupposes a general practical knowledge of phonetics based on the sounds of the learner's own language. Every sentence must be practised till it runs glibly off the tongue without effort or hesitation. Even with a thorough preliminary training in phonetics, this will take long practice at first, until the learner is familiar with the organic basis of the language. The result will be that at first everything will practically be learnt by heart. Hence the importance of care- fully choosing the most instructive words and sentences for these phonetic exercises, and of associating every word and sentence with its exact meaning — in the case of sentences by means of idiomatic translations, together with a translation of each word separately. As the energies of the pupil will be mainly taken up by phonetic difficulties — especially if his previous phonetic training is either defective or altogether wanting — there will be no time for grammatical analysis. Even the analysis into separate words need not be carried farther than the translation of the ' full words,' the meaning of the form-words — the prepositions and other particles — being left to be gathered from the context. Such idioms as how do you do ? in which words are used in special meanings which they do not otherwise have, might also be left unanalyzed — partially at least. But it would be advis- able, perhaps, to exclude such idioms from the first stage. To omit word for word translation altogether would be carrying the mechanical principle to an irrational extreme : we do not wish our pupils to fall into the error of the student who was being examined in the Greek Testament, and after translating neos oinos correctly as ' new wine,' was asked ' which is new and which is wine,' whereupon, suspecting a trap, and distrusting SPECIAL PRINCIPLES OF METHOD 119 the similarity between the Greek and the English forms of the words, he answered ' n'eos wine, obws new.' Irregularities. — The phonetic exercises should, as already implied, include some of the most necessary and frequent elements of the grammar and vocabulary, and, perhaps, some of the most indispensable idioms. In this way many of the irregularities could be learnt in this stage, and they would be learnt without effort, for the learner would not know that they were irregularities. Thus to a foreign beginner who has not learnt any English grammar, the regular singular feat and the irregular plural feet are on exactly the same level - they are purely phonetic difficulties, the difficulties being identical in both — and it is not till he learns the grammar that such a collocation as hands and feet causes any hesitation through associations which tend to make him change it into hands and foots. It is indeed possible thai foots is more difficult to him Xbanfeet — of course on phonetic grounds. This would be the solution of what from a strictly grammatical point of view is an insoluble dilemma. The dilemma is this : the irregular forms are the most frequent, and should therefore be learnt first ; but in the grammar the irregular forms must necessarily be subordinated to the regular ones. The answer is, as we see, that irregularities are psychological, not mechanical difficulties, and should therefore be mastered in the mechanical stage. When some progress has been made in the first stage, the learner may be allowed to read short texts of the simplest character — still without any grammatical analysis. The time spent on the first stage will depend on the confor- mation of the learner's mind. If his mind is mature and quick to grasp general principles, he will remain in this stage only as long as is necessary to give him a thorough command of the pronunciation, which, again, will depend partly on his natural aptitude for phonetics, partly on the degree of training he has had in practical phonetics. In the case of immature or slow minds the first stage may be indefinitely prolonged. The more it is prolonged, the more will phonetic considerations be subordinated to those of grammatical structure and the acquisition of a useful vocabu- lary, so that the texts will become longer and more varied, and iso THE PRACTICAL STUDY OF LANGUAGES the method of grammatical induction will be more and more applied. (2) The grammatical stage. It is evident from what has been said that the transition from the first to the second stage may be either quick and abrupt or slow and gradual, and that the two stages may overlap in various ways. This second stage presupposes a thorough mastery of the pronunciation and the acquisition of a certain amount of materials for grammatical study in the shape of words, sentences, and texts whose meanings are known. What further preparation for grammatical analysis has been made will depend on the length and character of the first stage In this stage the texts will be so chosen as to embody the different grammatical categories in progressive order of difficulty as far as is compatible with employing genuine texts which reproduce the actual language. The texts will naturally become longer and less simple in style and subject, and will embody a more and more extensive vocabulary. But as the vocabulary is in this stage entirely subordinated to the grammar, there will be no attempt to develope the vocabulary systemati- cally. It will be taken into account only from the negative point of view of keeping out rare and superfluous elements, and using as small a vocabulary as is consistent with general efficiency. In most cases the grammatical training will consist in a gradual expansion of the deductive method, till the learner is able to read with profit a grammar founded on the texts he is studying together with those he has learnt in the first stage. When he has gone through his first grammar, he will begin again at the beginning and revise all the texts in the first stage from a grammatical point of view. The study of grammar is not confined to the second stage, but is necessarily continued through all the following stages. At the end of the second stage the learner will be able to read a general grammar — one that takes its material from the whole of the language, not merely from the texts already read — but this grammar will necessarily deal only with the modern colloquial language. The student will not be able to read a grammar that includes the literary language till he is in the fourth stage, and for historical grammar he will have to wait till he has finished the fifth stage. SPECIAL PRINCIPLES OF METHOD 121 The historical study of grammar lies outside the domain of the practical study of languages. Even if we admit, with Storm and the majority of German linguists, that the study of historical grammar and comparative philology ' is practical in a higher sense, because it facilitates the comprehension and acquisition of the facts,' we cannot admit that it is an essential part of the practical study. We only have to ask ourselves the question whether three years spent in the exclusively practical study of a language, or the same time spent partly in practical, partly in theoretical studies such as historical grammar, would yield the better results. We cannot hesitate in answering that the latter method would be a failure as compared with the former, if only because it would not allow time for acquiring the necessary practical knowledge of the older periods of the language. If we extended the period to five years, the disparity as regards practical results would not be so glaring, but the advantage would still be on the side of the purely practical course. (3) In the idiomatic and lexical stage the idioms will be learnt systematically, partly from reading idiomatic texts, partly from a phraseology in which the idioms will be classed under psychological categories, as will be explained hereafter. At the end of this stage the learner will have acquired a thorough command of a limited number of words and phrases and idioms expressing the most necessary ideas. His voca- bulary will not be large— perhaps not more than three thousand words — but he will command it with ease and certainty. Those who learn a language through its literature often have as wide a vocabulary as the natives, but have no command of the elementary phraseology : they know words, but do not know how to combine them, except from a purely grammatical point of view. They are, indeed, often unable to describe the simplest mechanical operations, such as ' tie in a knot,' ' turn up the gas,' or express such ideas as ' make haste ' or ' what is the matter ? ' As Storm remarks (Forbedret Undervisning, 22), there are hundreds of expressions in French, which, although they occur incessantly in conversation, are seldom or never taught in the ordinary school-books because they cannot be brought under the conventional rules of grammar. Hence even those who have learnt French for years do not know that, for instance, the French for ' it is kind of you ' is Jest aimable a vous, not de vous, and that ' it smokes here ' cannot be i22 THE PRACTICAL STUDY OF LANGUAGES translated by il fume iri, which means ' he smokes,' but only by cafume id. ' Very few have the gift of being able to learn such expressions from books. The material afforded by literature, even in that form of it which approximates most closely to the colloquial language, namely, novels and comedies, is such a medley, so varied, and so mixed, and often so difficult, that one expression drives out the other ; the reader has enough to do to understand the contents, and has not time to concentrate himself on the separate expressions. The great art is, not to learn everything, but to take note of the special expressions that one really requires ; but this is an art which only very few are capable of.' All this points to the necessity of a systematic study of the vocabulary and phraseology of the language, which should begin in this stage, and be carried on in the next stage as well, where it will have the further use of helping to prevent con- fusion between the colloquial and the literary language. It must be understood that the study of the phraseology is only a part of the study of the vocabulary, as given in an ideological dictionary, as explained hereafter. The learner should begin with phrases and idioms, and then study the whole of his vocabulary from the ideological point of view. (4) The literary stage. As our ideal student advances, he will be able to choose his texts with greater freedom and with less subordination of matter to form, till at last he is able to enter on the fourth stage, and begin to read the actual literature unmodified and uncurtailed, beginning, of course, with ordinary prose, and proceeding gradually to the higher prose literature and to poetry. There is no reason why some literary texts of exceptional simplicity should not be read in the previous stage. In fact, simple poetry might be read almost from the beginning, for the metrical form is generally an effectual bar to any cross-asso- ciations with the divergent forms of colloquial prose. The greatest danger of confusion is with the antiquated or artificial colloquial style of the drama. In the course of this stage the learner will begin to acquire the nomic spelling of such unphonetically written languages as French and English. In dealing with less unphonetic lan- guages, the nomic spelling may be begun earlier. With others the nomic spelling will be used almost from the beginning. SPECIAL PRINCIPLES OF METHOD 123 The learner will henceforth be able to dispense with the phonetic transcription altogether, except when he wishes to refresh his memory for purposes of conversation. (5) The archaic stage presupposes a thorough mastery of the modern literary language in its most important branches, as far, at least, as understanding it goes. In proceeding to the older literature of such a language as English, he may either work his way back through Milton to Shakespeare and Spenser, or he may begin at once with Old English (Anglo-Saxon), and work his way down through Chaucer to the modern period. The choice between these two main lines of study and the details of the study will, of course, depend on what his objects are — especially on whether his interests are purely linguistic, or whether he means to use his knowledge of the language as a key to literary, historical, or other non-linguistic studies and investigations. CHAPTER XI GRAMMAR Grammar, like all the other divisions of the study of language, has to deal with the antithesis between form and meaning. Accidence and Syntax The fact that in language there is generally a divergence between form and meaning — as when the idea of plurality is expressed by a variety of forms, and sometimes by none at all (trees, men, sheep), or when the same form is used to express distinct grammatical functions (he sees the trees) — makes it not only possible, but in many cases desirable, to treat grammatical form and grammatical meaning apart. That part of grammar which concerns itself simply with forms, and ignores the meanings of the grammatical forms as far as possible, is called accidence or ' forms ' (German formenlehre) ; that which concentrates its attention on the meanings of grammatical forms is called syntax. Thus under accidence an English grammar describes, among other details, those of the formation of the plural of nouns — how some add -s, some -es, while others mark the plural by vowel-change, and so on. In the syntax, on the other hand, the grammar ignores such formal distinctions as are not accompanied by corresponding distinctions of meaning, or rather takes them for granted, and considers only the different meanings and grammatical functions of noun-plurals in general. The business of syntax is, therefore, to explain the meaning and function of grammatical forms, especially the various ways in which words are joined together to make sentences. As the form of a sentence depends partly on the order of its words, word-order is an important part of syntax, especially when it serves to make such distinctions as 124 GRAMMAR 125 in the English, the man saw the fox first, and the fox saiv th • man first. In fact, word-order is the most abstract part of syntax, just as word-order is the most abstract grammatical form. In accordance with its etymology, syntax is by some gram- marians regarded entirely from this latter point of view, so that it is by them identified with the analysis of sentences, the mean- ing of grammatical forms being included under accidence. Thus the peculiar meaning of the plural inflection in such words as sands, kads, waters of the Nile, would by such grammarians be discussed under accidence, on the ground that accidence deals with isolated words, syntax only with combinations of words into sentences. Although the application of grammatical terms cannot be allowed to depend on their etymology, yet, as we cannot avoid saying something about the meaning of grammatical forms under accidence — if only to discriminate between such inflections as trees, Johris, comes — it is often convenient to clear off this part of the grammar under accidence, especially if the variations of meaning are only slight, or else so great that they cannot be brought under general rules. The whole question is, after all, one of convenience. The separation of meaning from form is a pure matter of conveni- ence, and is not founded on any logical necessity, but only on a defect of language as it is, for in an ideally perfect language form and meaning would be one — there would be no irregu- larities, no isolated phenomena, no dictionary, and what is now dictionary and grammar would be all syntax. Even in lan- guages as they exist form and meaning are inseparable, so that the separation of accidence and syntax must always be a more or less arbitrary one, which may vary in different languages, quite apart from any questions of convenience. Formal and Logical Syntax The duality of form and meaning allows us to study syntax from two points of view. Formal syntax starts from the grammatical forms, and explains their uses ; logical syntax starts from the grammatical categories expressed in language generally, and describes the different forms by which they are expressed, as when we describe the different ways in which predication is expressed — by a single verb, by the verb to be with an adjective 126 THE PRACTICAL STUDY OF LANGUAGES or noun, and so on. So also in logical syntax the two con- structions maris disobedience and the disobedience of man would be treated of under the same head, while in formal syntax the one would go under 'inflections of nouns,' the other under ' prepositions.' It is evidently the first business of syntax to deal with the phenomena of language from the formal point of view, reserving logical groupings till all the grammatical forms have had their functions explained. G. v. der Gabelentz seems to have been the first to insist on the distinction between formal and logical— or, as he calls them, ' analytic ' and ' synthetic ' grammar (Gab. 86, 90). In his larger Chinese grammar he has tried to carry out the dis- tinction in detail Grammar and Dictionary We have seen that grammar deals with those phenomena of language which can be brought under general rules, while the dictionary deals with isolated phenomena — especially with the meanings of separate words. But not of all words. It is clear that while the meaning of such a word as man or house belongs to the dictionary, that of such a word as of in the disobedience of man belongs to the grammar, for it has exactly the same function as the -s of the genitive case : it cannot, indeed, be said to have any meaning of its own at all. From the point of view of the practical study of languages, such a question as whether or not the prepositions are to be treated of in the grammar as well as the dictionary, and the further question whether all of them, or only some of them, are to be included in the grammar, must be answered by showing whether or not the acquisition of the language will be facilitated thereby ; and this will depend on the structure of each language. Accidence and Syntax Taught Together We have seen that there is no real necessity for the separa- tion of accidence and syntax. Although practical convenience often seems to call for a separation, there may be circumstances under which it is desirable to treat forms and their grammatical functions and meanings together. GRAMMAR 127 In Beyer and Passy's Elementarbuch des gesprochenen Franzb- sisch this principle has been carried out consistently. Thus, under ' definite article ' first the forms are given (la metr, le peer, dy peer), and then under the heading ' gebrauch ' (use) the syntax of the definite article is given. In dealing with the verb, the forms are first given in a lump, the periphrastic forms as well as the inflected being given, and then the ' gebrauch.' But this arrangement is only a compromise : it simply amounts to giving a chapter of accidence and a chapter of syntax alter- nately, instead of printing all the chapters on accidence together, and then giving the chapters on syntax together. In going through such a book as this, one feels doubtful whether it is not after all more convenient to have the accidence all together, so as to facilitate reference to the paradigms and other sources of information, instead of having to search through the whole grammar for them. In my First Steps in Anglo-Saxon I have also tried the experiment of teaching accidence and syntax together. So far from subordinating syntax to accidence, I have in some cases advocated teaching syntax first, and for the following reasons, as stated in the preface to the book : ' Inflections may be recog- nized in two ways : by their form — as when we know that a noun is in the dative plural by its ending in -um ; and by their function — as when we infer from a word expressing more than one person and standing in the indirect-object relation that it is a noun in the dative plural. Of these two methods of parsing — the formal and the syntactical — sometimes one is easier, sometimes the other. There is therefore every reason why elementary syntax should be learnt simultaneously with acci- dence. It seems irrational to oblige a beginner to recognize such a grammatical category as the subjunctive mood solely by irregular and perplexing inflections, when such a simple rule as " it is always used in indirect narration " may enable him to recognize a large number of subjunctives with mathematical certainty.' In accordance with these principles, I have in the grammar to First Steps in Ang/o-Saxonb\ended accidence and syntax together more closely than in Passy's book. Thus under ' cases ' I first describe the formal peculiarities of each case, and then describe its functions. One of the advantages of this arrangement is that the syntactical examples serve to imprint the formal details more firmly on the learner's memory, being, in fact, chosen i28 THE PRACTICAL STUDY OF LANGUAGES partly for that purpose. In beginning the verbs, I confine myself at first to the indicative mood together with the infinitive and participles. Then, when I have given a general sketch of the different classes of verbs from this limited point of view, I go on to describe the forms of the subjunctive mood, and how they differ from those of the indicative, which takes up only half a page ; I then devote two pages to stating the chief rules for the use of the subjunctive, with examples. In this way the danger of confusing the forms of the subjunctive with those of the indicative — and in Old English these two moods are especially liable to be confused — is reduced to a minimum. Stages of Grammatical Analysis In this book I have also tried to do justice to another im- portant principle of practical grammar, namely, that grammatical analysis has two stages, one of recognition or identification, and another of reproduction or construction. As I say in the pre- face, ' The first requisite is to understand written texts, which involves only the power of recognizing grammatical forms, not of constructing them, as in the further stage of writing or speaking the language. Thus in beginning the second text in the present book, a learner in the first stage is expected to find out for himself that manna is in the genitive plural, and that craftum is in the dative plural, and to infer from the ending -ne in hwelcne that craft is masculine. He will then be able to infer with tolerable certainty from what he has learnt in the grammar that the plural of crceft is crceftas, but this inference belongs really to the second stage : a learner in the first stage is only expected to recognize the inflection of crceftas when he meets it. The first object, therefore, of a simplified grammar is to give what is necessary to enable a beginner to recognize the grammatical forms in the texts he is about to read. . . . The first thing is to explain the general structure of the language — that in Old English, for instance, nouns have three genders, that the gender is partly grammatical, that nouns have four cases — and to state those general rules which admit of no exception, such as that all nouns in -a are masculine, and that compound nouns follow the gender of their last element. Those irregular forms which are of very frequent occurrence — such as the in- flections of the definite article — must of course be learnt by heart at once, the learner relying on their incessant repetition GRAMMAR 139 to fix them in his memory. Less frequent irregularities need not be included in the grammar at all, their explana- tion being relegated to a note [to the texts]. ... In deal- ing with the strong verbs, it will be seen that after giving a general account of their formation and a few general rules — such as that in the preterite the second person singular always has the same vowel as the plural — I content myself with giving the typical forms of each verb in a note to the passage where it first occurs. ... In some cases where there is more than one form, but without there being any great complex ity or irregularity, I steer a middle course : I mention the various forms, but without giving any rules for their use. Thus I merely say that most strong neuters take -u in the plural or else remain unchanged. ... In the grammar I have been careful to group parallel forms together as much as possible. Thus under " cases " I give the inflections of nouns, the definite article, and the personal pronouns all together, so that, for instance, the learner may make them, her stepping-stones to yam, hire, yTe/r, and afterwards to the corresponding strong adjective inflections. The occasional paradigms are in most cases not intended to be learnt by heart, but serve only to sum up the scattered information already given.' I then go on to say, ' All these principles are those which are carried out — consciously or unconsciously — by most linguists. An experienced linguist in attacking a new language begins with the shortest grammar he can find. He first takes a general bird's-eye view of the language, finds out what are its special difficulties, what has to be brought under general rules, what to learn detail by detail, what to put off till a later stage. The rash beginner who starts with a big grammar forgets two-thirds of it soon after he begins independentlreading. Such a grammar as the one in the present work simply attempts to give him the really useful residue which, when once learnt, is not and cannot be forgotten.' Grammar Learnt Unconsciously We have already seen how the first or mechanical stage of learning a language, being the pre-grammatical stage, may be utilized to convey a good deal of grammatical information not directly through rules, but indirectly through examples, so that when the learner comes to the rule, he finds that he knows it K 130 THE PRACTICAL STUDY OF LANGUAGES already, or, at any rate, has advanced half way towards knowing it — a result which is a special help in mastering irregularities (p. 119). Thus in the grammar to First Steps I give under the phonology, among other examples of the vowels, twa handa, ' two hands,' twegen pet, ' two feet,' twegen menn, ' two men,' so that when the learner comes to the numerals, he finds that the paradigm twegen twa twa V ^ , tw&m twegra offers hardly any new difficulties; for he finds the above examples repeated with a reference to the place where they occur, but without any translation, together with mid twam handum as an example of the dative, whose ending -m is already familiar to him, from the nouns and adjectives. The only remaining form twegra is sufficiently illustrated by the parallel genitives yre'ora and yritigra, of which 1 proceed at once to give examples in sentences. The form twegra is only added to complete the paradigm, as it does not occur in the texts in First Steps, for which reason no special example is given of it. It will be seen that after the learner has gone through such a book as First Steps, in which the grammar is kept strictly within the limits of the recognition-stage, a great part of his grammatical knowledge will be unconscious instead of analytic and systematic. Thus he will know a good many individual forms of strong verbs, but will know nothing of the distinctions of class. Thus he may know that brecan has preterite brae and preterite participle brocen, but he has not learnt to refer it to the bear-class, although he may have noticed the parallelism between brcec, brocen and bear, boren, and may have strengthened this association by remembering the further parallel stdan, steel, stolen. In this way he will be well prepared for the classifica- tion of the strong verbs. A few weeks' work at the Anglo-Saxon Primer, which is constructed on the rigorous grammar-and- glossary historical method — though otherwise made as simple and easy as possible — will then, as I have said in the preface to First Steps, enable him to ' systematize his knowledge and round it off, and he will proceed to the elements of historical and comparative grammar with all the more zest through not having had them crammed into him prematurely.' GRAMMAR 131 Evils of the Separation of Syntax from Accidence The evils of the separation of syntax from accidence are well shown in the way in which the dead languages are taught in schools. Boys are made to learn paradigms by heart, and are then set to read the classical authors with the help of a dic- tionary before they have acquired any real knowledge of the meanings of the inflections they are expected to recognize in their texts — much as if they were taught the names of tools without being taught their uses. Thus in learning Greek they are taught to recognize the optative mood entirely by its form without having any idea of its meanings and functions as dis- tinguished from those of the subjunctive, of which, indeed, they come to regard it as an arbitrary and unmeaning variation ; to which may be added that their ideas about the meaning and function of the subjunctive mood itself are vague enough. When they are afterwards made to learn the rules of syntax, they are unable to apply these rules to what they are reading, and in most cases the possibility of doing so never enters their minds : they prefer to go on as before, and to guess at the meaning from the context without paying any regard to the moods. It is not very long ago that the rules of Greek syntax were learnt in Latin — an effectual bar to any intelligent appli- cation of them. Examples It is now generally admitted that a grammatical rule without an example is of no practical use : it is an abstraction which is incapable of entering into any direct associations with anything in the language itself.. The example, on the other hand, is concrete : it can be imprinted firmly on the memory by the mere force of the mechanical associations involved in carefully reading it and carefully pronouncing it aloud ; while, on the other hand, it is logically associated with the rule, which it explains, illustrates, and justifies. The example serves also as a standard or pattern by which the learner can recognize other examples of the rule as they occur in his reading. The example is thus a link between these other examples and the rule itself. Many of the older grammarians, while expending much 132 THE PRACTICAL STUDY OF LANGUAGES thought and care on elaborating their statement of the rules, considered the choice of examples as of subordinate impor- tance. They forgot that the first object of grammatical study is not the acquisition of rules, but of a practical command of the language itself; so that instead of the examples being intended solely to illustrate the rules, the true relation is almost the reverse : the rules are mere stepping-stones to the under- standing of the examples ; so when the latter are once thoroughly understood, the rules become superfluous and may be forgotten. These considerations have led some reformers to advocate putting the example before the rule, the idea being that the learner is thereby led to study the example carefully and then deduce the rule for himself, and finally compare his deduction with the rule as formulated in the grammar. This is the old inventional fallacy (p. 116) over again. Experience shows that when the learner knows that the work of deducing the rule from the examples has been already done for him, he naturally declines to do it again, so that, if the rule is put after the example, he simply reads the rule first, and then returns to the example. If, however, he prefers to read the example first, there is nothing to prevent him from doing so, whether it precedes the rule or not. Most learners prefer to read the rule first in order to know what the example is about, and what to look for in it — for a sentence may be, in itself, an example of a dozen rules of grammar — and if they do not understand the rule, they then read the example and return to the rule again, and when they finally understand the rule, they concentrate their attention on the example. We may say, in short, that the order of rule and example is of no importance compared with their mutual relations. The number of examples depends partly on the nature of the rule, partly on the scope and size of the grammar. Some rules hardly require any example at all through being practically of universal application, or self-evident, or because they are of no intrinsic importance, and are added only for the sake of completeness. But it is a safe principle never to take for granted that a rule does not require an example : if adding a few words in parentheses will make the statement or rule any easier to grasp, or prevent some misunderstanding that the GRAMMAR 133 writer never thought of, they certainly ought to be added. If they are superfluous, no harm is done. Besides, what is super- fluous to one reader may be helpful or even necessary to another. German writers often exasperate the reader by giving half a page of examples of some pet truism that requires only two words to illustrate and prove it, and then make a series of abstract generalizations expressed in unfamiliar and arbitrarily defined terminology without any help in the way of example, so that they often become unintelligible even to their own countrymen. If every rule is to have an example, it follows that a com- pound rule ought to have an example of each division of the rule. Thus, such a rule as that ' verbs expressing joy, desire, memory govern the genitive ' requires at least three examples. But in such a case as this many short grammars would give only one, on the mechanical principle that each paragraph is to have only one example. Even in the shortest grammar space may generally be found for a full number of examples by omitting some of the irrelevant matter of which such ill- planned books are generally full. If there is not room for more than one example to those rules which really seem to require it, additional examples to those rules that most require it may be given in a separate book. A good example must fulfil two conditions : (1) It must illustrate and confirm the rule unambiguously. Thus, as already remarked (p. 107), examples of the use of the ablative in Latin should, if possible, be forms which cannot be taken for datives. (2) The example must be intelligible as it stands, without any further context. If the example is a sentence or is contained in a sentence, the sentence should be one which will bear isolation from the context. In dealing with separate words, it is often a great help to the learner to give them in natural groups such as hands and feet \ buy and sell, past, prese?it, and future, dead or alive, neither here nor there. The more concrete a word is, the better it will bear isolation. It is mere waste of space to give bare lists of prepositions, conjunctions, and other form-words in an elementary practical grammar. It need scarcely be said that the examples must be in the language with which the grammar deals. Thus no one would 134 THE PRACTICAL STUDY OF LANGUAGES think of illustrating a rule of Spanish syntax by a Portuguese example. But it is almost as great an absurdity to illustrate rules of modern English syntax by examples taken from Shake- speare, except in special cases where the earlier constructions have been imitated by modern writers. All of this would, however, be quite out of place in a practical grammar for beginners. Carrying this principle a little further, we must be careful that our examples in an elementary grammar do not contain any specially difficult or rare words or irregularities of con- struction which do not directly illustrate the rule. Examples made up extempore for the purpose of illustrating a rule are not so good as those which have been collected from a variety of writings. There is, first of all, the danger of monotonous repetition of words, ideas, and constructions. In the effort to frame collocations of words to illustrate some rule, the grammarian is apt to produce unnatural, trivial, or other- wise objectionable sentences, such as the golden sun shines brightly \ the happy children of our teacher sing sweetly enough from their book of hymns, both taken from an English grammar of some repute in its time. Every example ought to be explained — even in the phono- logy. The translation of a new word not only gives a useful piece of information, but serves also to identify the word. But the explanation need not necessarily take the form of translation. There is one objection to translating the examples in a grammar : the learner is tempted to read them carelessly, and so not get all the benefit that would result from a con- scientious analysis of them. In my First Steps in Anglo-Saxon I have therefore tried the experiment of putting the explanation of the examples in the grammar on the same footing as the words in the texts, as far as possible, so as to oblige the learner to read the examples with the same care as the texts themselves. At the beginning of the grammar each example is translated in full. When a word or word-group or sentence is repeated in the grammar, it is not translated, but the learner is referred back to the place where it is translated ; and after the first few pages each new word in the examples is explained in the notes at the end of the book. Hence the reader is obliged to study each example carefully, and with constant comparison of what GRAMMAR J 35 he has already learnt, while at the same time he has every inducement to learn thoroughly every page before proceeding to another. Paradigms The paradigms and tabulations in an elementary grammar ought to be regarded mainly as summings up of what has already been learnt indirectly or in the form of scattered details (p. 129). The principle of combining words into groups rather than presenting them singly (p. 133) should be carried out in paradigms for beginners as far as possible. Thus, in dealing with an inflectional language, nouns should be accompanied by the definite article or some similar word, adjectives by a noun whose gender is known. In Old English and German the weak inflection of adjectives should be exhibited in its natural surroundings, that is, with the definite article as well as a noun. This, I believe, is generally done in the German grammars used in England. To take an example from Old English, it is evident that such a collocation as \one godan witan must strengthen the associations by which the learner has already learnt to recognize \one and witan as accusative masculine singulars, so that the weak inflection godan offers no new difficulty, while the preceding definite article always reminds him of the syntactical conditions on which weak adjective- inflection depends. From this point of view the Middle-Age grammarians with their incessant hie, haee, hoc were really more practical than their successors. It was not the old grammarians' fault that Latin had no article ; and they certainly took the best substitute they could find. So also were the old-fashioned French grammars when they indicated the subjunctive mood by the addition of que: que je sois, que tu sois ... I need hardly say that I follow this precedent in my First Steps : indicative ic wear]>, subjunctive \I33M 3HV XS3H 3HX 33V 3SnV039 'AVON/IS What is that which you and every living person have seen, but can never see again ? 'AVrm, such as can you tell me what the right titne is ? which is hardly an idiom, but an ordinary normal sentence, what is the time ? being on a level with what is the hour? which is the way ? Familiarity of Subject The subject of the texts ought to be in harmony with the language they are intended to teach, both as regards place and time : an English reading-book for French learners ought to deal with scenes of modern English life rather than with Lacedemonians. Nor would the English learner of French care to have French adaptations of insipid and antiquated English children's stories put before him. But when the learner has got a firm hold of the foreign language, it is instructive for him to read descriptions of his own country written in the foreign language, for although such descriptions are not a preparation for either a stay in the foreign country or a study of its literature, they have the advantage of dealing with objects and ideas with which he is familiar, so that the resulting associations, though less directly useful, are more definite and distinct. Descriptions and defini- tions of familiar objects and ideas are peculiarly instructive. In dealing with languages embodied in old-established literatures such as English and French, and still more with 176 THE PRACTICAL STUDY OF LANGUAGES dead languages such as Latin, there is often a great difficulty in rinding texts which are at once genuinely national in character and at the same time simple in matter and style. The difficulty is that highly developed literatures are apt to be too rhetorical or too ornate, too epigrammatic or too cynical, and, generally speaking, wanting in naivety. Hence the foreigner in search of simple texts is apt unconsciously to select old-fashioned pieces, which, while fulfilling the require- ments of simplicity of language and familiarity of subject, do not fulfil them with reference to the present day, the life and language being those of past generations. Such a book as Vietor and Dorr's Englisches Lescbuch is pleasant reading to an English adult, precisely because it brings back half-faded associations of childhood and traditions of the eighteenth century, but for that very reason is in many respects a mislead- ing guide for Germans who wish to learn to understand English life and language as they now are. There is no fault to be found with such texts as ' London (in 1880),' but it is difficult to see how the choice of such a nonsense rhyme as ' A diller, a dollar, a ten o'clock scholar, What makes you come so soon ? You used to come at ten o'clock, But now you come at noon,' can be justified, for a dollar is not an English coin, and the word dollar in the above rhyme is not connected with the context, and when the learner looks up diller in the glossary, he gets full information about its pronunciation, but is dis- gusted to find that he has taken all this trouble about a word which is only a ' scherzwort.' Even when perfectly suitable modern texts exist, the diffi- culties of copyright come in. Hence in my Elementarbuch, being in want of a short story, I took The Gypsy Party by Thomas Hood (in his Hood's Own), and modernized the language and, to some extent, the incidents, changing the title to The Picnic, the original title being quite an antiquated expression, which many English people would not understand. In this way I have combined the advantages of good matter and modern language. The story of The Old Chapel in my Primer of Spoken English was obtained in the same way. Although the language of both these pieces as given by me is purely modern, the spirit of them is not so. But, on the other hand, if every TEXTS; THE READING-BOOK 177 reading-book had to be perfectly up to date, we should have to write new ones every five years or so, and they would then embody many very transitory elements, confined perhaps to a limited sphere. But, fortunately, there is a certain foundation of English style and phraseology which is even older than the nineteenth century ; there are whole pages even of such writers as Swift and Arbuthnot which, with a very little alteration, are good colloquial English of the present day — in fact, it is only by its being so good that we know that it is not Present English. Simplicity of Language Simplicity of language demands, in the first place, that the texts should be colloquial rather than literary : that they should be written in short sentences, not in long and com- plicated periods, that they should be as free as possible from metaphors and other figures of speech. But colloquial tendencies must not be pushed to an extreme. It must be remembered that in the beginning we do not advocate col- loquialism so much for its own sake as because, as a general rule, a colloquial style fulfils certain requirements better than a literary one. But when colloquialism developes into abrupt, elliptical, disconnected dialogues full of unnecessary idioms and slang, it becomes almost as unsuited for elementary practical purposes as the opposite extreme. Again, many considerations point to descriptions of nature as the best texts to begin with; but such descriptions cannot be colloquial in the strict sense of the word, for continuous descriptions con- stitute not a dialogue but a monologue, which would be hardly possible in real life. Such texts are, in fact, almost as much literary as colloquial, but they give the literary style simplified to the utmost degree in the direction of the spoken language. Hence, too, as already observed, there is no harm in giving at an early period pieces of simple poetry. For there are many poems whose language is so simple and free from archaisms that it diverges but slightly from colloquial speech as regards vocabulary and grammatical structure, while the marked character given by the metre and diction serve to diminish the danger of cross-associations with the colloquial language. The little poem, Past and Present, given at the end of my Elementarbuch, is an example of this. In its thirty- N 178 THE PRACTICAL STUDY OF LANGUAGES two lines the only uncolloquial features I notice are morn for morning, bear away for carry away, 'tis for it's, he instead of it (said of the sun), and the compound lily-cup, together with some trifling divergences in word-order. Note, on the other hand, the pure colloquialism, I'm for / am. We might also substitute ifs for 'tis without injury to the metre. Variety The great advantage of natural, idiomatic texts over artificial ' methods ' or ' series ' is that they do justice to every feature of the language, if only representative pieces of the three great classes of texts are chosen. The artificial systems, on the other hand, tend to cause incessant repetition of certain grammatical constructions, certain elements of the vocabulary, certain com- binations of words to the almost total exclusion of others which are equally, or perhaps even more, essential. Thus the Ollen- dorff and Ahn methods result in the total exclusion of idioms, even the most necessary ; and Gouin's ' series ' deal only with concrete and objective words, and almost entirely exclude the abstract and subjective elements of the language, so that he is obliged to supplement his objective series with a subjective course — or, rather, to promise such a supplement, for, as might be imagined, he soon found the task far beyond his strength. In its present form the Gouin method is incapable of teaching the pupil to say, ' I think so,' or ' I would rather not do it,' or, indeed, to express anything that falls under the categories of emotion or intellect. As Brekke remarks (Brekke, Gouin, 44), the series method results in the most astounding grammatical limitations : only principal sentences, verb only in the first or third person, only assertive sentences (no interrogative or negative sentences), everything in the present tense, and so on. Gradation of Difficulties After what has been said, there can be little doubt as to the true principles of the gradation of texts. The simplest in grammatical structure are descriptive pieces, in which the verb can be restricted to the present tense and the third person. The practical value of this restriction will, of course, depend on whether the language is highly inflected or not. Even in TEXTS; THE READING-BOOK 179 English it would save the learner some difficulties, such as those of the preterites of strong verbs. With such a language as Chinese it would have no grammatical meaning at all. But descriptive texts have the further advantage of affording the clearest, most definite, and most connected and continuous context. With these, therefore, the beginning should be made. They agree with Gouin's series in giving mainly the concrete and objective elements of the vocabulary. Our texts would differ, however, materially from them in giving only the really useful combinations. The grammatical forms which are wanting in the descriptive texts are supplied by colloquial dialogues, which, in their highest and freest development, are the most difficult of all. The maximum of variety is attained by that mixture of description, narrative and dialogue, which is exemplified in a novel or short story. This kind of text has the advantage of being infinitely elastic, so that it admits of almost as great simplification as a purely descriptive text, from which it then differs only in giving greater variety of grammatical con- struction, vocabulary, and idioms. This, then, should be the central type of text : it is at the same time a preparation both for reading and speaking the language. Interest The remarks already made on the question of interest (p. 112) will, I think, be confirmed by a consideration of the different ways of studying texts. To be interested in a thing, we must be in the mood for it, and the thing itself must be a novelty. The learner who is struggling with the combined difficulties of pronunciation and grammatical analysis, together with all the difficulties caused by an unfamiliar vocabulary, is not in the mood to appreciate jokes or national humour, which, even if not already stale to him, or uncongenial through national pre- judices of his own, will certainly lose their novelty by the time he has learnt to pronounce them and to parse their linguistic embodiment. There are many passages in my Elcmentai-biich and Primer of Spoken English which hardly ever fail to elicit signs of amusement from English readers, but I have seldom known any of my foreign pupils show the slightest signs of appreciation of them from this point of view. And even if it were advisable to make use of sensational 180 THE PRACTICAL STUDY OF LANGUAGES narratives of shipwreck, piracy, murders, and apparitions, our fundamental principles of slow reading and incessant repetition would soon take the interest out of them. The teacher would not fail to hear the remark frequently made by those who begin the study of a foreign language with this kind of litera- ture, ' I should like to read this book in a translation — I cannot remember the plot of the story when I only read twenty lines a day.' But if learners are often callous to the literary or humorous merits of their texts, they are, on the other hand, very ready to criticize their defects. Young children, in particular, have a great dislike to being condescended to, and being offered what is aggressively babyish, or too obviously intended to serve moral and pedagogic interests. Often, indeed, they prefer the other extreme : they like to have glimpses of something just a little beyond them. We all dislike unnecessary triviality. This is why I based my descriptions in the Ekmentarbuch mainly on popular scientific and sociological works (p. 166), although this involves some slight deviations from the principle of familiarity. But even when I introduce details out of the life of the Middle Ages or any other unfamiliar scene, I take care to describe them in language which recalls ideas familiar to the modern reader. Another insuperable obstacle to making texts positively interesting and not merely non- trivial or non-objectionable, is the variety of tastes. Descriptions of nature are soothing and pleasing to some minds even if a little commonplace. To others even the most eloquent and imaginative descriptions of nature are as tedious and depressing as nature itself is to them : those who do not care to hear about ' The new soft -fallen mask Of snow upon the mountains and the moors ' cannot be expected to be interested in a matter-of-fact descrip- tion of atmospheric or marine phenomena. Storm objects to my descriptions of nature that ' the children have enough of them at school,' and Passy says that ' no French boys would have the patience to go through them.' These divergences of taste depend partly on nationality and changes in public taste. A certain style of literature goes out of fashion in one country, and is then introduced into another, TEXTS; THE READING-BOOK 181 where it is welcomed as a novelty, just as extinct German philosophies find a sleepy home elsewhere. Hence it is pos- sible that the ' goody ' stories in Passy's Elementarbuch may be acceptable to German children, although they certainly were not so to English children even thirty years ago, at which period they had already become old-fashioned in this country. I am certain that such a piece as (1 ekol bqisonjftr) in Passy's book (nr. 42) would provoke lively antagonism in most English readers, not on account of the sentiments conveyed in it, but of the manner in which the moral lesson is put forward. Nor can I believe that school-children care to read descriptions of schoolrooms and of pedagogues swaggering before a big black board chalk and duster in hand, such as Passy gives in his first piece (la klaas). I do not make these remarks with any inten- tion of depreciating this valuable book, but simply as an illus- tration of the impossibility of making linguistic texts permanently interesting to the majority of learners. It is with texts and selections as with pronunciation : every one likes his own best. I find, too, as regards my own books, the Elementarbuch and the Primer of Spoken English, that every reader has different tastes. Besides avoiding triviality and over-childishness and naivety, it is evident that the texts should be of moderate medium length — neither as long as three-volume novels on the one hand, nor as short as proverbs on the other (p. 172). As to monotony, the principles of variety and gradation already discussed will fully obviate that. Literary Texts The language of purely literary texts is generally inconsistent with our principles of selection. It is tolerably sure to be more or less archaic from a strictly colloquial point of view, or to contain unnecessary words and phrases, or to be accompanied by complications of grammatical structure, or vagueness of context. But if a literary piece is exceptionally suitable for any linguistic purpose, or seems to fit in well with the context, or to illustrate it and make things clearer, there can be no very strong objection to admitting it, if the divergences from the colloquial standard are not too marked or such as to cause linguistic confusion. Many of these divergences can, indeed, often be removed 1 82 THE PRACTICAL STUDY OF LANGUAGES without injury to the general character of the piece, and this should always be done when practicable. Useful texts may be constructed by retelling the story of some literary composition in simple language. Epic and narra- tive poems may be dealt with in this way in languages other- wise wanting in prose texts, such as many dead languages. Thus one of the texts in my First Steps is a simple prose para- phrase of the epic poem of Beowulf, which in its metrical form bristles with obscurities and difficulties. I here give a specimen, first of the poem itself, and then of the corresponding portion of my own paraphrase into simple Old English prose : — ' Sbwa a dryhtguman dreamum lifdon eadigllce, obbast an ongann firene fremman, feond on helle. Wses se grimma. giest Grendel haten, masre mearcstapa, se be moras heold, fenn and fasten. Fifelcynnes eard wansselig wer weardode hwlle, sibban him Scieppend forscrifen hasfde in Caines cynne, bone cwealm gewrasc ece Dryhten, bass be he Abel slog. Ne gefeah he basre fashbe, ac he hine feorr forwrsec metod for by mane manncynne fram ; banon untydras ealle onwocon : eotenas and ielfe and orcneas, swelce gigantas ba wib Gode wunnon lange brage ; he him }>ass lean forgeald ! Gewat ba neosian, sij>ban niht becom, hean huses, hu hit Hringdene sefter beorbege gebun hsefdon. Fand ba bairinne sebelinga gedryht swefan sefter symble ; sorge ne cubon wansceaft wera. Wiht unfaslo grimm and graadig gearo sona wass, reoc and re^e, and on reste genam britig begna ; banon eft gewat hube hremig to ham faran, mid bsere waelfylle wica neosan.' ' On bisse blisse burhwunode HrSbgar cyning and his menn lange tid, obbaet him feond onseege^ wearb. Past wass unfselu wiht, Grendel hatte. Se bade on \ asm mearclande, and haefde TEXTS; THE READING-BOOK 183 him faesten geworht on fennum, onmiddan ]>£em sweartum mdrum. Sume menn cwsedon bast Grende waere of Calnes cynne. Forbaim, ba Cain ofslog Abel his brobor, ba wear b him se aelmihtiga gram, and nine on wraecsib asende, and het hine on westenne wunian, feorr mancynne. Panon onwocon ealle unfaele wihta, dweorgas, and ielfe, and eotenas, be wib God wunnon. Pa ne mihte Grendel bolian bast he ielce daege blisse gehierde on Heorote, and he self ute wunode on piestrum. fa on niht aefter ))im gebeorscipe, ba ba menn slepon on jJaare healle, ba wearf> se reba Grendel sona gearo : he him on ungearwe on bestael, ba hie him nanes yfeles ne wendon, and hira britig genam, and mid him ferede ham td his faestenne, baere herehyj^e fasgniende ; forbaam hit waes his beaw baet he hlaf ne ast, ne waster ne dranc, ac aet manna lichaman and hira blod dranc.' But such paraphrases must be into a simple, colloquial style of language, as far as possible. Such books as Lamb's Tales from Shakespeare are useless for our purposes because written in an artificial archaic style. In dealing with dead languages we cannot be so fastidious, especially with one that has only a limited prose literature, such as Old English. Thus in the above paraphrase of Beo- wulf I have made no attempt to keep to the language of one peirod, but wherever I have found a suitable model or pattern for any portion of it, I have followed it, whether it comes from Alfred or ^Elfric, or from the early or later parts of the Chronicle. Condensed Treatises Advanced students of a language often feel the want of a knowledge of the vocabulary of some special technical or scien- tific subject which they cannot expect to pick up by ordinary general reading — such subjects as commerce, gardening, management of a sailing-boat, cycling, trigonometry, chemistry, electricity. A full logical dictionary would, of course, give some informa- tion as to the vocabulary of such subjects, but necessarily in a very concise form, especially in the wider branches of know- ledge. For mastering the vocabulary of these, it would be 1 84 THE PRACTICAL STUDY OF LANGUAGES desirable to have condensed special treatises resembling the science primers and practical guides with which we are familiar, but differing essentially from them in strictly subordinating actual information to explanation and illustration of the special vocabulary and terminology of the subject in question. This might be extended to more general subjects. Thus we might have a series of ideal condensed histories of different periods with typical battles, sieges, sea-fights, insurrections, trials for treason, embassies and so on, the information — which may be imaginary — being only just as much as will suffice to give a certain number of examples of the terminology required. Subordination to Form ; Grammatical Texts We have seen that the general character of a text determines to some extent the character of its vocabulary and grammatical forms — that, for instance, in descriptions and narratives the verbs may be exclusively in the third person. In Caesar's Commentaries even the dialogues have their verbs in the third person through being put in indirect narration. If, then, for any pedagogic purpose we wanted a text of this description, it would be perfectly easy to make one without doing any vio- lence to the genius of the language. It is a different matter when we try to write a text under formal limitations which do not naturally follow from the general character of the text. Even so apparently natural and reasonable a restriction as using only the present tense in descriptions of nature might cause embarrassment, although it is partially founded on the character of the text. Thus the very first verb in my descriptive texts in the Elementarbuch is a preterite (used). Nor must it be forgotten that even if all the verbs are made present in meaning, we cannot avoid the pre- terite in clauses of rejected condition, as in if it were implying 1 it is not.' English has so many monosyllabic words that it is quite possible to write long texts in words of one syllable ; and this has often been done from a mistaken idea that such texts facilitate learning to read. But when we consider that such a restriction allows us to mention only a single season of the year, only three out of the twelve months, and not a single day of the week, it is evident that such texts must be hampered by many unnatural omissions and awkward circumlocutions. TEXTS; THE READING-BOOK 185 In China, whore all the words of the language are monosyl- labic, a book written under much more embarrassing restrictions is still used as a primer for teaching boys to read and write. This is the famous Book of a Thousand Characters (ts'icn tsf wen). The origin of this book is a curious story. It is said that one of the emperors summoned the best scholar of the time, and gave him a thousand slips of paper, each with a different cha- racter — that is, a different word — written on it, and told him to arrange them so as to make sense. The scholar solved the problem in a single night ; but in the morning his hair had turned white. The peculiar difficulty of the task lay, of course, in the restriction that no word was to be used twice over — not even the commonest particle. The result was a text that was never really intelligible throughout, not even with the help of the many commentaries that have been written on it. Such a task could not have been even attempted in any other language but Chinese, which, at a pinch, can dispense entirely with auxiliaries or particles of any kind, and express every grammatical relation by mere position. Other Oriental languages can show long poems written entirely to illustrate grammatical and lexical forms. Thus in Sanskrit there is an epic written for the express purpose of giving examples of verb-forms found in the grammars but non- existent in the literature. The artificiality lies here not in restriction, but simply in finding a connected context for a certain number of words. But comparatively easy as the task is, we cannot believe that the result can be anything but insuf- ferably tedious. In Europe such texts are constructed in a less ambitious spirit. Our Ahns and Ollendorff's do not write poems ; they do not even try to write consecutive prose. Franke's Phrases de tons les jours contains excellent materials, but has the same defect of want of continuity. As Storm remarks (Forbedret Unde?~visning, p. 26), ' we have here a good selection of idiomatic material, but not a single actual conversation, nor any arrange- ment according to the grammar. The material is as discon- nected as in the ordinary manuals. We meet, for instance, every minute pronouns without being able to see who is referred to, and questions without answers. It is of little use having good material, if it cannot be assimilated. When the sense is interrupted every moment and the context becomes uninte 1 - ligible, it is impossible to adapt oneself to the situation, and 1 86 THE PRACTICAL STUDY OF LANGUAGES feel at home in the surroundings.' I may add that Franke probably thought that the arrangement of his idioms under logical categories would be enough to associate them together in the learner's mind. But this seems not to be the case ; as they stand, Franke's idioms are of no use except as a summary of what has already been learnt from connected texts. And this was the main object of my colloquial sentences in the Ekmentarbuch, although at the same time I was fully alive to the advisability of making the sentences as connected as possible. But I soon saw that to carry this out fully would require much more space than I could afford. In a full thesaurus or in a primer which dealt only with the commonest words there would, I think, be little difficulty in making the ex- amples form continuous dialogues or narratives or descriptions of some length. The continuity would, of course, be logical, not formal — that is, not according to grammatical categories. Storm himself has, in his .Dialogues Franfais, 1 attempted to construct connected texts for systematic practice in the chief rules of grammar, so arranged that the rules are learnt more by unconscious imitation than by deliberate grammatical analysis. The author has taken the principle of beginning with the spoken language literally by giving his texts in the form of dialogues. But it must be remarked that the book is not intended for beginners, but for grown-up students who have already worked at French for two or three years. From this point of view the choice of dialogues instead of more e ementary forms of texts is fully justified, and, perhaps, to some extent, the complete absence of any phonetic transcription. Considered from the purely grammatical point of view, it must be admitted that these dialogues have been adapted to their purpose with great skill. But it must also be admitted that this subordination of matter to form has made many of them Irather trivial and uninteresting in themselves. But the dialogues are frequently interspersed with little anecdotes and occasional literary pieces (cf. p. 181), among which we find the well-known passage from Moliere about M. Jourdain speaking prose without knowing it. One danger of |\vriting texts for a certain purpose is the tendency to spin them out indefinitely by heaping up illustra- tions and dwelling too long on one rule. Even with the most 1 There is an authorized English edition by G. Macdonald under the title of French Dialogues by Joh. Storm. TEXTS; THE READING-BOOK 187 rigid limitations the attempt adequately to embody all the rules of grammar in such texts would probably result in a book of impracticable length. Storm himself seems to feel this difficulty, for he often interrupts his dialogues to give groups of detached proverbs, phrases, and idioms, which have not even the logical connection of Franke's sentences, being associated solely by grammatical considerations. I will now give a few examples of Storm's texts : — ' II. L'article partitif. Avez-vous du vin ? * Je n'ai pas de vin, mais le marchand de vin en a. 2 A-t-il du vin rouge, du vin blanc, de bon vin, 3 de mauvais vin? II n'a pas de mauvais vin, il n'en a que de bon. Quel vin desirez-vous, du rouge ou du blanc ? Donnez-moi du rouge. Ce n'est pas du vin, c'est du vinaigre. J'en ai d'autre ; j'en ai de meilleur ; en voici. Voila. du vrais bordeaux, et du meilleur. Vous avez de si bon vin, que je vous en demanderai encore. Cela fait du bien. Un peu plus de vin ne vous fera pas de mal. II me faut peu de vin et beaucoup d'eau. Vous mettez trop d'eau dans votre vin ; mettez moins d'eau et plus de vin. II n'y a pas de vin ; moi du moins je n'en ai pas. II n'y a plus de vin. Je n'ai plus de vin. II n'y en a plus ? II n'en reste plus. Si, il en reste encore. En avez-vous ? Oui, j'en ai. En voulez-vous ? Ddsirez-vous encore du vin ? En voulez-vous encore ? Oui, donnez-m'en encore un peu. Garcon, encore du vin, s'il vous plait. Encore un peu de vin, s'il vous plait. Encore un verre de vin, s'il vous plait. Mais vous avez un verre de vin devant vous. Pardon, il y a bien un verre [a vin], mais pas de vin. Un peu plus de vin, monsieur ? 1 All the texts have translations in parallel columns. 2 This is Ollendorffian. * In a note the author tells us that the colloquial form is du bon vin. 1 88 THE PRACTICAL STUDY OF LANGUAGES Merci. [Je ne veux] 1 plus de vin. Monsieur n'en veut pas davantage ? Pas davantage, je vous remercie. Du vin, mon ami ? Merci. Pas de vin. Je ne bois pas de vin. Je ne veux pas de vin. II y a un verre de trop, 6tez-le. ' Avez-vous du pain ? Non, je n'ai pas de pain, mais le boulanger en a (il y en a chez le boulanger). ' Allons diner. Garcon, la carte, s'il vous plait. Quel potage de'sirent ces messieurs ? I Messieurs, en dinant, je vais vous raconter une petite histoire. L' autre jour un Anglais, assis dans un restaurant, criait a rue-tete, a plusieurs reprises : " Garcon ! plus de soupe ! Garcon ! plus de soupe ! Garcon ; n'entendez-vous pas ? plus de soupe ! " Le garcon repond d'abord : " Bien, monsieur." A la fin il dit : " Monsieur, j'entends tres bien ; vous ne desirez plus de soupe ; aussi ne vous en servirai-je plus." L' Anglais, tres-e'tonne, s'ecrie : " Mais c'est justement plus de soupe que je veux." " Ah," dit le garcon, " c'est different ; alors il fallait vous expliquer plus clairement. Si vous m'aviez dit que vous desiriez encore du potage, je vous en aurais servi tout de suite." Notre Anglais, honteux et confus, s'est remis h prendre des lecons de francais. ' II n'y a pas de regie sans exception. II n'est point de roses sans epines. II n'y a pas de fumee sans feu. Ne'cessite' n'a point de loi. Faire de ne'cessite vertu. A bon entendeur peu de paroles. A sotte question point de reponse. Ventre aflame" n'a point d'oreilles. On prend plus de mouches avec du miel qu'avec du vinaigre. 1 These additions in [ ] might be relegated to notes, as they confuse the learner, who ought to have only one form presented to him at a time — that is, in this case, Merci. Plus de vin. TEXTS; THE READING-BOOK 189 II n'a ni feu ni lieu. II n'a ni foi ni loi. Cela n'a ni rime ni raison.' The difficulty of constructing grammatical texts may depend on the nature of the language. The more highly inflectional a language is, the more easily it seems to lend itself to such a priori construction ; while, on the other hand, the complexity of its forms is an additional inducement to make such texts. It will therefore be worth while to notice a Finnish analogue to Storm's book, intended to teach Finnish to Swedish-speaking natives of Finland — Kallio's Finsk Elementarbok} The plan of this book differs, however, widely from that of Storm's Dialogues. It is divided into four parts: (1) texts, (2) vocabularies, (3) general index to vocabularies (4) grammar. Each text has its own vocabulary, in which the meanings of the words are explained in the order in which they occur in the text. The index to these vocabularies is an alphabetic list of all the words in the texts, each word having a reference to the number of the text where it occurs first, which is, of course, also the number of the corresponding vocabulary, no further information being given. The object of this is to induce the learner to master as thoroughly as he can the vocabulary of each piece before going on to the next, so as to save himself the trouble of looking up the word in the index and then referring to the vocabulary there indicated. This is one of the weak points of the book : the learner ought to be referred not to a dry list of isolated words, but directly to the text itself, so that he can take in the context. As it is, if he wishes to compare the context of the first appearance of a word, he has to make three different references — two to lists of words, and another to the text itself. Notes like those in my First Steps in Anglo-Saxon are simpler and more effective than these short glossaries, which neither give full information nor are convenient to refer to through not being alphabetic. It would really be simpler to do away with the special vocabularies, and have an alphabetical glossary, and nothing else. The learner cannot be expected to remember every word at once — least of all in a strange language like Finnish — so that 1 I know it only in its fourth edition, in which it has undergone some modifications by another hand. i 9 o THE PRACTICAL STUDY OF LANGUAGES practically he is obliged to look up many words at least three times, besides occasional references to the grammar. There is, in short, too much to and fro work — there is more turning over of pages than with the ordinary grammar and dictionary method. The grammar begins with two pages of introduction dealing with the phonology. After that it is divided into numbered sections, each number referring to that of the piece in which the grammatical rules given in that section are exemplified. Thus the first section (p. 139) gives part of a verb-paradigm with analysis and rules : — ' (mind) mene-11 I go (sind) mene-t thou goest (me) mene-mme we go (te or Te) mene-tte ye go or you go Stem : mene. Personal endings : -;/, -/, -mme, -tie. Rule 1. In an inflected Finnish word we distinguish stem and ending. Rule 2. By adding different endings to the stem we get different inflections of the word. Rule 3. The subjects mind, sind, me, te (or Te) can be omitted if there is no emphasis on them.' All this seems rather dry and unnecessarily pedantic ; but it must be remembered that the book is intended for teaching children in classes, not for self-instruction. The corresponding text (p. 1) and its glossary (p. 73) are as follows : — 1. Puheharj otitis. 1 Hyvaa paivaa ! Kuinka te voitte ? — Kylla me hyvin voimme. Kuinka sinavoit? — Kiitoksia, hyvin mina voin. — Mihin menet? — Mina tulen teille ja toivon, etta te huomenna 1 tulette meille. — Kiitoksia ! Kylla me tulemme, jos mina voin hyvin huo- menna. — Toivon, etta Te voitte hyvin. Hy vasti nyt ! Hy vasti, hyvasti ! — Siis tulette huomenna ? — Kylla me tulemme. 1. puheharj otitis, speaking-practice, kuinka, how ? hyvdd pdivdd, good day ! te, Te, ye, you; teille, to you. 1 Pronounced huomena. TEXTS; THE READING-BOOK 191 vein, be (ill or well). tukn, come. kyllii, certainly. ja, and. me, we ; meille, to us. toivon, hope. hyvin, well. ettti, that. si/; if, thou. huomenna, to-morrow. kiitoksia, thanks. Jois, if. mind, I. hyvasti, good-bye ! mi kin, whither ? Tytf, now, menen, go «'«, so, therefore.' We cannot expect much of a text produced under such conditions, but it is certainly a great advance on Ahn and the rest of them. At any rate, it is connected. The average length of the later pieces is about a page. The following are translations of some of the headings, to which I have occasion- ally added the first sentence or two of the piece: Father (Father often goes away early in the morning. Sometimes he comes home late in the evening . . .) ; the Neighbours ; the Poor Woman (Yonder is a cottage. The cottage is old and bad. There dwells a poor woman . . .) ; the Gardiner ; What do we buy and sell ? ; Journey abroad ; On the Ice (Near us there is a skating-rink) ; the Eagle's Nest ; the Months ; the Lighthouse ; Norway ; Wolves ; a Fairy-story ; Kalevala (the Kalevala tells of the life of our ancestors. There are fifty cantos in it. The chief personages are . . . ). It will be seen that the texts are of a very varied character, only a few of them being in the dialogue form. Some of them are necessarily rather trivial, especially the earlier ones, and sometimes the constructions are a little unnatural, through the necessity of avoiding certain difficult forms, such as most of the infinitive and participle constructions, which are the great difficulty of the language. There is a second part, on the same plan as the first, in which the rest of the grammar is worked out in the same way, accidence and syntax being kept abreast throughout. On the whole Kallio's Finsk Elementarbok gives as good a compromise between free texts and grammatical texts as could reasonably be expected. But the general question still remains, Which method will yield best results within a given time — that of progressive grammatical texts, or of free texts accompanied by a complete grammar founded on the texts ? These considerations bring us face to face with the problem, How are we to bridge over the gulf between grammar and 192 THE PRACTICAL STUDY OF LANGUAGES reader ? The dilemma is this : If the texts are perfectly free and natural, they cannot be brought into any definite relation to the grammar. If the learner reads a sufficient number of systematically varied texts, he may depend on finding examples of all, or nearly all, the rules of grammar ; but the examples will occur practically at haphazard without any natural grouping and without any regularity of reoccurrence. Thus in a descrip- tive text all that we can promise a priori, from a grammatical point of view, is that the verbs shall be in the third person present, while from a logical point of view we can determine with definiteness and certainty what concrete or other categories shall be represented. The other horn of the dilemma is that if we try to make our texts embody certain definite grammatical categories, the texts cease to be natural : they become either trivial, tedious, and long-winded, or else they become more or less monstrosities, or, finally, they are broken up into detached sentences. Storm, as we see, openly adopts the detached sentence method ; and this is better than giving texts which are outwardly connected, while in reality their sentences are detached. We may, then, repeat our question in a different form : Is it really worth while trying to construct grammatical texts ? Is it not simpler to rely on natural texts on the one hand, and detached sentences on the other ? If we resign ourselves to this compromise, we shall find that detached sentences are the real bridge across the gulf between texts and grammar. The bridge is constructed by taking the detached sentences used as examples in the grammar from the texts the learner is either reading at the moment or is about to read. This is the method I adopt in my A?iglo-Saxon Primer : all the examples in the syntax are taken from the texts which follow. In this book I have, like Storm, supplemented the texts and the grammar by adding a selection of detached sen- tences, arranged so as to illustrate the different grammatical categories. In First Steps in Anglo-Saxon I generally follow the same principle, though here, not being bound to adhere strictly to the texts, I frequently modify the sentences in the grammar which are taken from the texts, so as to make them more instructive for the immediate purpose I have in view. This method acts well either way, whether the learner begins with the texts or with the grammar. In the first case, he remembers the context of his sentence when he meets it in the TEXTS; THE READING-BOOK 193 grammar, so that it is no longer isolated to him. In the second case, when he meets his sentence in the texts, he sees more easily what grammatical rule it illustrates. By going through grammar and text alternately several times, both these advan- tages may be secured. It is evident that the question of the relation between texts and grammar cannot arise till the systematic study of grammar has been begun. Kallio's book introduces grammatical analysis, with its stems and cases, at the very beginning. In my First Steps in Ajiglo-Saxon I utilize the section on pronunciation as a preparation for the grammar, but, as Old English in the nature of things is not learnt by very young or linguistically untrained beginners, it was not worth while giving much space to lengthening and systematically developing the pre-gram- matical stage. If this were done, the gulf between texts and grammar would be partially bridged over beforehand : when the learner came to the grammatical stage, he would not only be better able to understand the detached sentences in his grammar, but would also be familiar with many of them indi- vidually. Another important result of the development of the pre- grammatical stage would be that very elementary and conse- quently trivial and unnatural grammatical texts would no longer be needed at all, and grammatical texts generally would perhaps become superfluous. CHAPTER XIV RELATIONS BETWEEN DIFFERENT LANGUAGES ; TRANSLATION It is evident that any general plan of study cannot be applied to any one language without certain modifications of detail. There are, moreover, further modifications of detail dependent on the special relations between the language to be learnt and the student's own language. Thus German offers certain special difficulties to an Englishman, other special difficulties to a Frenchman, not only in pronunciation, but also in gram- mar, vocabulary, and phraseology. But all these special relations are governed by the same general laws of association as the ideal general plan of study itself. Begin with a knowledge of one's own language. — The first preparation for the study of a foreign language is the acquisition of a thorough knowledge of the peculiarities of one's own language. We have already seen that the first re- quisite for acquiring foreign pronunciations is a practical know- ledge of the sounds of one's own language. So also the first requisite for understanding the grammatical structure of another language is a thorough knowledge of the grammatical structure of one's own language. This is one of the reasons why we should from the first be taught to regard the grammar of our own language from the point of view of general grammar. Just as in the study of the phonetics of a foreign language we are often surprised to find that the germ of an unfamiliar sound exists already in our own pronunciation, so also a systematic study of English grammar enables an English learner to point out analogies to unfamiliar foreign constructions which might otherwise escape his notice. Thus in Welsh and many other languages, adjunct-words or modifiers follow the word they 194 RELATIONS BETWEEN LANGUAGES 195 modify instead of preceding it as in English. Thus the Welsh gwr givellt, ' straw man, effigy,' means literally ' man straw.' But in English an adjunct-group consisting of a preposition and a noun follows the same order as in Welsh, and we only have to think of man (of) straw with the of dropped to get the Welsh order, so that by degrees we can feel at home even in such complicated series as llonaid llwy de llaelh, ' tea-spoonful of milk,' literally ' fullness spoon tea milk.' It is as important in grammar as in phonetics to have a clear idea of the defects and anomalies of one's own language; which, again, points to the importance of regarding the gram- matical structure of our own language from a general linguistic point of view, as is done in my New English Grammar. Thus in English we have a group of defective verbs, such as may, can, which have no infinitives or participles ; so that when we have occasion to use them in the functions of infinitives or participles, we have to substitute other words for them : / can come, I shall be able to come, I have not been able to come. Other English verbs are defective in other ways ; thus we cannot transfer he used to go there every year to the present without a complete change of construction : he goes there every year, or he is in the habit of going there every year. A distinct conscious- ness of these defects in English helps the English learner to get over the hesitation he feels when in speaking foreign languages he has to use such a construction as / shall can . . . , especially in a language such as German, which expresses these ideas with etymolog'ically allied word (ich kann, infinitive konnen). So also, as already observed, the use of English up in pack tip is contrary to that of most other languages, in which ' pack up' suggests the idea of unpacking, these languages generally expressing the idea of our pack up by ' pack in,' or some such construction, unless they use separate words. So if the English learner has once learnt to recognize that his native use of up in such constructions as tie up, pack up, shut tip, is illogical, or, at any rate, contrary to the genius of other languages, he will be more ready to accept their divergent constructions. Difficulties also arise from the opposite reason, namely, that the native language is more normal and rational or simpler than the foreign language. Thus the English speaker is apt to feel impatient of the distinctions of grammatical gender in most other European languages. He learns easily enough to associate feminine definite articles with feminine nouns, and so i 9 6 THE PRACTICAL STUDY OF LANGUAGES on, but is continually liable to relapse into calling a tree or a house ' it ' instead of ' he ' or ' she,' even when he knows the gender. That son frcre in French should mean ' her brother ' as well as ' his brother ' seems confusing and irrational to him : he feels it ought to be sa frere. It sounds even more absurd to him to talk of women as le beau sexe: he feels the adjective ought to be feminine. The only remedy for these and similar wrong associations is to regard the matter from a rigidly formal and mechanical point of view— to suspend the reasoning power, or, in some cases, to divert it into purely grammatical as opposed to logical channels. The utilization of resemblances between the two languages — whether the result of affinity or accident — has already been discussed (p. 89). Cross-associations. — We have already seen (p. 55) that the closer the connection between two languages, the greater chances there are of confusions arising from cross-associations. But cross-associations extend far beyond the limits of comparative philology, and may occur between any two languages however remote they may be from one another genealogically. But in such cases they are confined mainly to the syntax and phrase- ology and general structure of the two languages. But it is evident that if two languages have any general principles of structure in common, there must be a tendency to level differ- ences of detail. Thus the main principles of word-order are the same in English and Chinese, so that the English learner is tempted in cases where the order differs to make the Chinese words follow the English order. If the word-orders of two languages follow fundamentally distinct principles, there is less effort required to keep up differences in detail. So also the general similarity in structure and word-order in English and French, and also in phraseology — all of which is the result partly of independent development, partly of borrowing — makes an English speaker more apt to introduce English constructions, word-order, and phraseology into French than into German, which, although more directly cognate to English, is neverthe- less further removed in general structure. Distrust of Similarity. — A linguist who has learnt a cer- tain number of foreign languages of different families and RELATIONS BETWEEN LANGUAGES 197 different morphological structure, and has found certain con- structions, idioms, or developments of word-meanings of his own language uniformly rejected by these foreign languages, gets at last quite instinctively into the habit of mistrusting the associations of his own language on these points, so that instead of feeling inclined to translate such a sentence as ask him to come literally, saying, for instance, in German, fragen Sie ihn zu kommen, he gets into the habit of always expecting something quite different. Hence when he comes to a language such as Arabic, in which, as in English, the meaning 'interrogate' developes into that of ' request,' he resists the tendency to reproduce this usage as strongly as he was once carried away by it. So also it is a surprise to an English linguist to find the Chinese kien, ' see,' used in the sense of ' visit.' The French apprendre par cornr we do not distrust because we at once con- jecture that the English learn by heart is simply a translation of it ; but we should not venture to transfer it to any other language. Cross-associations between two foreign languages. — It is evident that the more foreign languages we learn, the greater our liability to form cross-associations. We not only form associations between our own language and a foreign language, but between the foreign languages themselves, especially between the one we are learning and the one last learnt. If an Englishman, after learning to speak Welsh fluently, were to go to Egypt and begin Arabic there, he would find that, in spite of the total want of affinity between the two languages, he would be constantly substituting Welsh for Arabic words in his attempts at conversation. This influence of the language last learnt implies not only that the language last learnt has been recently acquired, but also that it has been acquired with an effort, so that the resulting associations are strongly impressed on the mind in such a way as to be easily called forth by the slightest external stimulus ; if our English- man had been familiar with Welsh from his childhood so as to be perfectly bilingual, the influence of cross-associations derived from Welsh would be no stronger than cross-associations derived from English. But although want of affinity is no safeguard against cross- associations, there is, of course, much greater danger of confusion when the foreign languages are cognate. Every comparative philologist knows this by experience. 198 THE PRACTICAL STUDY OF LANGUAGES Safeguards against confusion. — The great safeguard against confusions between different languages is, of course, to learn each language separately, and bring one's study of it to some definite conclusion before beginning another language. But a good deal of help might be afforded by systematic summaries of the conflicting associations — the confusions and divergences — in each pair of languages. Thinking in the Foreign Language ; Not Translating The remedy usually prescribed is to ' learn to think in the foreign language.' But we cannot think in a foreign language till we have a thorough and ready knowledge of it ; so that this advice — sound as it is in itself — does not alter the fact that when we begin to learn a new language we cannot help thinking in our own language. Thinking in the language implies that each idea is associated directly with its expression in the foreign language instead of being associated first with the native expression, which is then translated into the foreign language. This has led many into the fallacy that if we were only to get rid of translation in teaching a foreign language, substituting pictures or gestures, we should get rid of the cross-associations of our own language. But these cross-associations are independent of translation. They arise simply from the fact that each idea that comes into our minds instantly suggests the native expression of it, whether the words are uttered or not : and however strongly we may stamp the foreign expression on our memories, the native one will always be stronger. This is proved by the well-known fact that in moments of great excitement, we invariably fall back on our native language or dialect. Even if we admit that transla- tion strengthens such cross-associations, we cannot admit that it is the cause of them. If it were, how can we explain those confusions between two foreign languages which we have been considering ? It is not even necessary that we should be very familiar with the language last learnt to cause confusion with the one we are learning : all that is necessary to establish cross- associations is that we should have made an effort to learn the former one. TRANSLATION 199 Translation from the Foreign Language Translation is of two kinds : from the foreign language into our own language, and into the foreign language from our own language. The great practical difference between them is that the latter presupposes a thorough knowledge of the foreign language. Translation from the foreign language stands on quite a different footing. It does not imply any previous knowledge of the word or sentence translated, and is at the same time the most obvious and convenient way of explaining its meaning. But some reformers wish to exclude even this kind of transla- tion from the beginning. The Picture-method. — This revolt may be seen carried to its extreme in the plan of teaching the vocabulary of a language by means of pictures. This old idea met with a warm champion in Franke (Fr. S. 34). He argues that the ' translation-method ' involves a complicated psychological process : by this method a German learns the* meaning of the French word chapeau- by first associating it with its German equivalent hut, and then associating Imt with the idea ' hat ; ' but show him the word chapeau in connection with a picture of a hat, and he will be able to establish a direct association between the word and the idea. Such reasoning involves the fallacy that a psychological process must necessarily be difficult because it is complicated. The fact is that to a German the word hut and the idea ' hat ' are so intimately connected that the one suggests the other instantaneously and without effort. Again, the picture gives us only part of the ideas associated with the word chapeau ; the shape and size of a hat varies, and is, besides, a secondary matter compared with the fact that a hat is meant to protect the head from the weather. Now the great advantage of a word as opposed to a picture is that it is practically an epitome of this whole group of ideas, and the equation chapeau — hut enables a German to transfer bodily such a group of ideas from his own to the foreign word. This the picture cannot do \ for even if we ignore everything but the shape of the hat, we must either give pictures of every conceivable shape of hat — tall, hard felt, soft felt, clerical, sailor, cocked, etc. — or else risk implying that chapeau means ' tall hat,' not 'hat in general.' 200 THE PRACTICAL STUDY OF LANGUAGES The picture method is, besides, very limited in its applica- tion. Pictures and diagrams are often useful, and sometimes almost indispensable, but in other cases they are either in- adequate or useless, or absolutely impracticable, as in dealing with abstract ideas. Explanation in the Foreign Language. — A less ex- treme view is that translation should be used only as a crutch for the beginner, to enable him to grasp the meaning of the foreign words and sentences, and should then be thrown away, the new foreign words being henceforth explained in the foreign language itself. Several advantages are claimed for this method by its adherents. The only incontestable one is that it affords additional practice in the foreign language. The other advantage claimed for it as well as the picture-method, that it diminishes the risk of cross-associations between the two languages, is, as we have seen, of more theoretical than practical importance. We find as a matter of fact that cross- associations cannot be got rid of by ignoring them : on the contrary, they have an awkward habit of cropping up when we least expect them. We cannot get rid of them for the simple reason that every idea is indissolubly associated with some word or phrase in our own language. The main argument against explaining in the foreign lan- guage is that as long as we are learning the foreign language it is our first business to have it explained to us as clearly and unambiguously as possible. Therefore all explanations ought to be in the language we know — that is our own — not in the one we do not know. Again, definitions, like pictures, may be ambiguous : if I define a hat as ' a covering for the head,' the learner may think I mean a cap, or a bonnet, or a hood, or a helmet. Or the definition, like the picture, may be too precise. Thus, if I define a hat as ' a cylindrical head-covering with a brim,' or show the learner a picture of such a hat, he may think I mean to restrict the meaning of the foreign word to ' tall hat.' It is further evident that a misleading or obscure definition will not be made clearer by being expressed in a partially unknown language. But translation or paraphrase in the foreign language may occasionally have its advantages for the more advanced student. Nor can there be any objection to it in cases where we can rely with certainty on the learner understanding it perfectly : TRANSLATION 201 even if it does no other good, it will at least, as remarked before, serve to give him practice in the foreign language. There is one application of such translation which is directly useful and instructive. When the advanced student comes to read the literature itself, he will derive much benefit from having the more out-of-the-way words and phrases translated into the corresponding simpler forms in the same language. This will teach him to discriminate clearly between what is general, modern, and colloquial on the one hand, and what is excep- tional, archaic, or purely literary on the other hand. And this advantage would be lost if the translations were into his native language only. It would, indeed, be useful, not only for practical, but also for critical and philological purposes, to have complete idio- matic translations of older standard works of literature into the modern language — to have, for instance, a translation of Shake- speare into Modern English prose, a translation as literal as the divergences of the two periods would allow. Such a translation would be more useful in many ways than a com- mentary, which, however lengthy, can never be made exhaustive. Translation makes knowledge more exact. — Trans- lation from the foreign into the native language has other and higher uses than that of being a temporary link between the foreign word and its meaning. When the learner has once clearly grasped the meaning of all the words in a phrase by means of translation, and has also grasped the meaning of the whole phrase, it is well that he should put aside his explanation- crutches for a time, and learn to associate the phrase directly with its meaning, without thinking of the corresponding phrase in his native language more than he can help. He can then begin to think in the foreign language — ' to live himself into it,' as the Germans say. But, as Storm remarks (Forbedret Undervisning, 29), ' the living oneself into the foreign language has also its dangers. One easily accustoms oneself to a partial understanding ; one does not form a definite idea of the special shade of meaning, because one has not thought of corresponding ex- pressions in the native language. It is not till one can translate the word, that one has complete mastery over it, so that one not only understands it, but can use it.' In fact, translation has much the same function in the vocabulary as grammatical 202 THE PRACTICAL STUDY OF LANGUAGES rules and parsing have in construction : it tells us how far we can go in our unconscious or half-conscious associations. Thus, when an Englishman hears a Frenchman say in French, I ask myself (je me denmnde) what this means,' he feels that this makes perfectly good sense as it stands, being, indeed, a possible English expression of incredulity or astonishment. But when he has it translated into its exact English equivalent, ' I wonder what it means,' he sees that what he assumed to be an exceptionally strong expression is a mere expletive, and that he was quite wrong in translating it mentally word for word. Again, without this translation he would be at a loss to find the French equivalent of the English ' I wonder . . .' In this way translation is a most valuable means of testing the accuracy and correcting the mistakes in our unconsciously and mechanically formed associations between our ideas and their expressions in the foreign language. Three Stages in translation. — We may distinguish three stages in the use of translation. In the first stage translation is used only as a means of conveying information to the learner : we translate the foreign words and phrases into our language simply because this is the most convenient and at the same time the most efficient guide to their meaning. In the second stage translation is reduced to a minimum, the meaning being gathered mainly from the context — with, perhaps, occa- sional explanations in the foreign language itself. In the third stage the divergences between the two languages will be brought face to face by means of free idiomatic translation. To these we may perhaps add a fourth stage, in which the student has so complete and methodical a knowledge of the relations between his own and the foreign language that he can translate from the one to the other with ease and accuracy. Translation into the Foreign Language ; Exercises As already remarked, translation into the foreign language presupposes — or ought to presuppose — a thorough knowledge of the foreign language. If the arithmetical fallacy were true — if sentences could be constructed a priori by combining words according to certain definite rules — then all that would be required for translating TRANSLATION; EXERCISES 203 into a foreign language would be a knowledge of the grammar and the possession of a good dictionary. This is the fallacy on which the old practice of writing exercises was based. In its crudest form the exercise-method consists in giving the beginner half a dozen words and a few rules, and then giving him detached sentences embodying these rules for translation from and into the foreign language either viva-voce or in writing or in both- and this from the very beginning. Thus Ahn's New Practical and Easy Method of learning the German Language, after a page or two on pronunciation, begins thus : — Masc. der Vater, the father; Fem. die Mutter, the mother ; Neut. das Buch, the book. gut, good ; grosz, tall, big ; klein, small, little ; ist, is. Der Vater ist gut. Die Mutter ist gut. Das Buch ist gut. Ist der Vater grosz ? Ist die Mutter klein.? 1st das Buch gut ? 2. The father is tall. The mother is little. The book is good. Is the father good ? Is the mother tall ? Is the book small ? ' After sixty or seventy pages the learner has only got as far as the following sentences : — 129. ' bitten, to beg, to pray, to entreat; der Krieg, the war. Do you know of what I am speaking, of what I am thinking ? It is not the same street through which we passed this morning, the same house where we have been yesterday. Are you speaking of the war ? Yes, we are speaking of it. Are you thinking of the concert? We are not thinking of it. Are you satisfied with that ring ? I am. Why do you not come up ? Tell your brother that I shall come down immediately. Come in, my friends. I beg of you to come in. Shall you go to-night to the play ? We shall not. Do you know where that gentleman lives, who he is, and where he is going ? We do not.' Of the more advanced use of exercises as a supplement to 2o 4 THE PRACTICAL STUDY OF LANGUAGES the detailed study of the grammar, the following are examples from Bernays' German Exercises ; — 'R. 23. a. Of the spoon Loffel; of the broom Besen ; to the father Voter. We have a governor Gouvernetir. Of the tea and to the coffee. b. Of the chain-of-mountains Gebirge ; of the evil Uebel ; to the seal Siegel ; of the knife Messer; of the young-lady Fraukin ; to the little-man Manncften (R. 15). c. Of the assessor Assessor ; to the author Autor.' 'R. 161. His cattle is run away, and his pigeons are flown Vieh laufeii Taiibe fliegen (ir.) away ; nevertheless he has worked on, as if (R. 304) nothing dennoch arbeiten fort had (R. 164) happened. — The labourers have run after the vor-fallen knecht horses. — They have run in-imitation-of the rope-dancers. — nach Seil-tdnzer Their pond is fished out; now they have done fishing. — He fischen aus has jumped after me.' springen (ir.) It is instructive to compare these examples of Ahn's and Bernays' methods. It is evident that the impossible task of translating into an unknown or only partially known language can be accomplished only under restrictions which make it either an evasion or a failure. In the first place, translation from one language into another ought to imply as a matter of course that what is translated has a meaning— that it is, if not a complete text, at least a sentence with an independent meaning of its own worth stating — and, of course, that it as well as the translation is grammatically and idiomatically correct. But such groups of words as of the tea and to the coffee and they have run in imitation of the rope- dancers are neither of them fit objects of translation from Eng- lish, the first because it cannot have any meaning, the second because it is not English. Nor is it enough that the texts or sentences should fulfil the negative conditions of making sense and being expressed correctly in both languages ; it 1.; also necessary that they should express something useful, something TRANSLATION; EXERCISES 205 worth saying, even if it were only a trivial dialogue between a traveller and a waiter at a restaurant. But although it is con- ceivable that any one of the sentences quoted from Ahn might occur in real life, yet taken as a whole they are impossible : instead of the first exercises introducing the learner to sentences and constructions which will help him to understand and express what he is most likely to meet with first, they give him a string of disconnected ideas which he might never have occasion to hear expressed or to express himself, even if he lived for years in the country where the language is spoken. Again, although the sentence about the ' rope-dancers ' is not English, yet the result of the translation will certainly be a fairly good German sentence, if a perfectly useless one. But this result is only obtained by giving so many helps in the way of glosses and direct references to the rules of the grammar that the work of translation becomes almost as great a farce as if the learner were set to copy from a book first the English original and then the complete German translation. In fact, such a process would in most respects be a more instructive and improving one ; for the learner would have the advantage of being able to compare the two languages in their correct idiomatic forms. We have also to realize what is meant by making mistakes in our exercises and correcting them afterwards. It means the laborious formation of a number of false associations which must be unlearnt before the labour of forming the correct ones can be begun. Even when no positive errors are made, the writing of exercises which require any thought must produce vague and hesitating, instead of the clear and instantaneous associations which constitute a real practical command of a language. And yet this process of going out of one's way to make mistakes, and then laboriously correcting them, is almost the only way of learning languages — at least, of learning grammar — that some people can conceive. I remember, when I first went up to Oxford as an undergraduate, I told my tutor that I was rusty about some point of Greek grammar ; so he said, ' You had better do a paper on it.' I could not help thinking even then that strengthening one's false associations by ' doing a paper ' was a curious preliminary to getting rid of them. It must also be remembered that the knowledge and con- viction that a certain linguistic combination is erroneous does 206 THE PRACTICAL STUDY OF LANGUAGES not necessarily get rid of the false association itself, for that is a matter of habit, not only of conviction. Thus, if in speaking German I once get into the habit of making ' bread ' masculine instead of neuter, even when I am told that brod is neuter, I am still liable to fall back into saying — as I once heard an Englishman say — (haaban zij kainsn vaisbroud) through pure force of habit. Getting rid of this habit may imply that I must repeat das brod at least as often as I formerly repeated der brod. There was once a professor who taught some Oriental language by correspondence. One of his pupils — a middle- aged military man — after going through a course, asked to be allowed to go through it again, so as to perfect the knowledge already gained before going any further. When he did so, he made exactly the same mistakes over again. He then asked to be allowed to go over the same course for the third time. The professor, who seems to have been a good-natured fellow, was inclined to grant this request, but was dissuaded by his wife. As we see, the only way to avoid the necessity of making mistakes is either practically to do the work for the learner by giving him a more or less complete word-for-word translation ; or to make the exercises so easy that they cost no effort, and afford no real practice at all, so that they slip through the mind without making any impression, these very easy exercises being at the same time necessarily unidiomatic and consequently of little or no use when learnt. These facts are now generally recognized among reformers. This is, indeed, the one point on which there is the greatest unanimity among them, namely, that everything of the nature of exercise-writing ought to be abolished, not only in the beginning but throughout the whole course. Free Composition ; Question and Answer There is also a general agreement among reformers that the place of exercises and translations into the foreign language should be taken by free composition in the foreign language on subjects taken from the texts already studied, so that the compositions are reproductions of what is already known. Continental reformers also make great use of a system of question and answer carried on in the foreign language by the FREE COMPOSITION 207 teacher and pupils, the former asking the questions, the latter answering them, or the teacher telling one pupil to ask a certain question of another pupil. The subjects of the questions are, of course, taken from the texts which the pupils have just been reading. Thus even a short sentence such as we can easily see that the earth is round by watching a ship sailing out to sea can be made the subject of a number of questions, such as what is the earth ? or what is the earth like ? or what shape is the earth ? \ how can we see that it is this shape ? or how can we see that the earth is round? | what can we see by watching a ship sailing out to sea ? Of course, if any unfamiliar word, such as shape, is used in the questions, it must be explained, unless its meaning is quite clear from the context. This method of question and answer is older than is commonly supposed. As I have several times drawn on Bernays for examples of bad methods, it is a pleasure to me to be able to quote the following remarks from the introduction to his Gcrmati Reader : — ' I have always found it very advantageous to my pupils, both in private lessons and classes, to let them translate back again into German. For this purpose I make use of the third section, generally beginning with this kind of exercise about the time the student has reached nearly the end of the first section, proceeding at the same time with the construing of German into English. When the learner is thoroughly master of a piece, however short, I question him on it in German, and receive his answers in the same language. By this means, his ear becomes familiarized with the pronunciation of another person without the aid of the eye, while he insensibly acquires the habit of speaking German himself. Take, for instance, the first short anecdote, page 119 [I have given this very anecdote on p. 167] ; I ask : — Question. Wer schricb an die Griechen ? Answer. Alexander. Q. Was that Alexander? A. Er schricb. Q. A?i wen schrieb er ? A. A?i die Griechen. Q. Was schrieb er an die Griechen ? A. Dasz sie ihnfiir einen Gott erkennen soil fen. Q. Fur was sollten sie ihn erkennen ? A. Fur einen Gott, 208 THE PRACTICAL STUDY OF LANGUAGES Q. Wen sollten siefiir eine?i Gott erkennen ? A. Alexander. Q. Wer sollten ihn fur einen Gott erkennen ? A. Die Gricchen. 1 This exercise may be continued and varied to any extent, if directed by any person capable of conversing in German, provided he is sufficiently familiar with the grammar to correct the mistakes of the student.' I do not know when this preface was first published — cer- tainly before 1856, the date of the seventh edition of the Reader. Dr. A. Bernays, who was professor of German language and literature in King's College, London, and was, I believe, more successful in the combination of language and literature than is always the case, began to publish his helps for the study of German about 1830. Although he was under the full influence of the methods of detached sentences and exercise-writing which attained their most extravagant development about his time, his books contain many good ideas. It is strange he did not see the absurdity of teaching his pupils to converse in German about Alexander and the Lacedemonians. The purely oral exercises of question and answer in the foreign language should precede any attempts at written repro- duction of what has been learnt, partly on the general ground that the fixed associations of the ear should precede the secondary and perhaps variable associations of the written form of the language, partly because of the facility and quick- ness with which they can be worked. They have the further advantage of training the pupils both to understand what is said, and reproduce it with accuracy and ease. They are, in fact, the best possible substitute for a phonetic method, although they will be ten times more efficient if preceded by systematic training in phonetics. They are also in the highest degree stimulating to the pupils, and develope quickness, presence of mind, and the power of observation. This reproductive or ' imitative ' method has the great advantage of being progressive. The questions and answers may be exact literal reproductions of what has been learnt, or they may be free paraphrases of it. The questions may also embody new words, which, again, may be expressly pointed out, and explained, either beforehand or afterwards, or left to be inferred from the context. VISUALIZING 209 So also with the written compositions. At first the pupils will simply be expected to write down from memory the subject of what they have been studying. Then they may be set to write an essay on a subject analogous to that of the text they have been studying. In this way the written compositions become gradually more and more independent of the texts, and more and more general in their subjects, as the learner's command of the language is widened, till at last he is able to express himself both in speaking and writing on any ordinary topic. Visualizing By visualizing we understand the establishment of a direct association of the words and sentences of the foreign language with the ideas they express by means of a direct appeal to the sense of sight. This can be effected in three principal ways, namely, by — (1) Object-lessons — the presentive or object-method: 'here is a piece of chalk,' ' this is called a black board,' ' this is my nose.' (2) Models, pictures, diagrams — the representive or pictorial method. (3) Gestures, mimicry — the dramatic method. It is also possible to establish direct associations independent of the help of a second language by appealing to the other senses. Thus the teacher may illustrate ' cock ' or ' fowl ' not only by exhibiting a picture of the bird, and by the dramatic method of flapping his arms and raising himself on tiptoe, but also by an imitation of its crow. So also the pupils may be invited to taste sugar, salt, tartaric acid, and alum in connection with a study of the foreign words expressing the accompanying sensations of taste. But the visual impressions are evidently the only ones of which any extended use can be made. Of the purely visualizing methods, it is evident that the first two are best suited for words expressing concrete ideas, the last for words expressing phenomena and actions. But they are all limited in their application. And of those associations which can be established by visual means, many are, as we have seen, vague and ambiguous as compared with those established by means of translation. It is so even with the object-method. Thus a cube of boxwood may just as well p 210 THE PRACTICAL STUDY OF LANGUAGES suggest the idea of ' wood ' as that of ' cube * ' a piece of sugar may suggest the idea of ' sugar in general,' or it may suggest the narrower ideas of ' loaf sugar ' or ' white sugar ' — which, again, may be subdivided into ' cane sugar ' and ' beet-root sugar ' — or ' lump of sugar.' It ought also to suggest the idea of ' lump ' or ' piece ' in the abstract ; but even if a piece of chalk, a piece of coal, and a piece of bread were exhibited together with the lump of sugar, it is by no means certain that the class would grasp what was meant. Pictures are even more liable to be misunderstood. Let us suppose two pictures, one of a human head, the other of a railway station, with numbers and dotted lines leading to various parts of the pictures, these numbers referring to an accompanying vocabulary of the foreign words expressing the ideas supposed to be excited in the learner's mind by the con- templation of these pictures — a contemplation which, if he is not much interested in the subject of either picture, he will perhaps enter on only with a certain effort and without much attention to details. His first difficulty will be a mechanical one ; as the pictures are shaded in some parts, he sometimes cannot see clearly where the dotted lines lead to : a certain line may point to the pupil of the eye, or the iris, or the cornea, or it may indicate the bridge of the nose. Then comes the old difficulty of determining the degree of generality of the ideas called forth by the pictures : is the eye, or only the pupil of the eye, meant ? is the number to be taken literally as indi- cating the dial of the clock, or does it imply ' clock ' generally ? does the other number refer to the railway carriage as a whole, or only to its roof ? In fact, did I not possess enough know- ledge of the foreign language to know that whatever wagon may mean, it does not mean ' roof,' I should find these pictures most misleading guides. Gestures are equally liable to be misunderstood. The argument that the substitution of visualizing methods for translation prevents cross-associations is, as we have seen, a fallacy. If we did not use a phonetic notation, there would indeed be something in the argument that visualizing methods enable us to save the learner the confusions that result from letting him see words written in an unphonetic spelling. But no teacher who has once used the phonetic method will ever think of wasting time over such an inefficient method of teaching pronunciation. CHAPTER XV CONVERSATION Conversation in a foreign language may be regarded from two very different points of view: (i) as an end in itself, and (2) as a means of learning the language and testing the pupil's knowledge of it. But there is, of course, no reason why the second process should not be regarded as being at the same time a preparation for the first. The difficulty and, at the same time, the utility of conver- sation, is due to the quickness and presence of mind that it requires. What we speak we have to know perfectly ; and we must have it ready at a moment's notice. Even the elementary question-and-answer method described above requires that the pupil should have thoroughly mastered the little he knows. Hence every speaker's knowledge must be definitely limited within comparatively narrow boundaries. Even in our own language we can only speak one special form of it. It is true that our knowledge is not confined to the spoken language, but extends to the literary language, and even to the archaic literary language. But although we know the literary language well enough to be able to read it with perfect ease, and per- haps to write it in its modern form, we cannot speak it for any length of time without the risk of continual relapses into the colloquial. The language of a few generations back we can neither speak nor write. We speak our own colloquial language without hesitation in spite of the confusing associations of the written language, because our associations with the former are by far the stronger; and the only way to acquire a colloquial style of speech in a foreign language is to make our associations with the spoken language stronger than those with the written language — by beginning with the spoken language and confining ourselves 211 212 THE PRACTICAL STUDY OF LANGUAGES exclusively to it till we can handle it with ease and certainty. I We have seen also that this is the rational method, whether we I wish to learn the language for conversational purposes or not. / This does not by any means imply that the beginner should learn dialogues about railway travelling and life in hotels, but merely that his elementary training should not be such as to unfit him for doing so hereafter, if he has occasion to travel abroad or converse with foreigners. Phrase-books We will now consider what ought to be the character of the dialogue- or phrase-books intended for this special purpose. As regards the ordinary phrase-books, the want of phonetic notation is alone enough to make them useless. But they are often quite as defective in their idioms. Not only is there a want of system in selecting the really useful and necessary idioms, and rejecting or subordinating the others, but the idioms and phrases given are often incorrect from the point of view of ordinary speech, being archaic, or literary, or vulgar, or, what is worse still, the result of mistranslation or over-literal translation of some foreign idiom. Most phrase-book writers fail to reproduce the natural spoken language, partly from want of preparatory training in the practical study of languages, partly from fear of being thought vulgar, but also from pretentiousness and conceit, which leads them into a spurious literary style, so that their dialogues read like extracts from badly written novels. Thus Franke remarks that German grammars and phrase-books for foreigners generally give eilen Sie ! \ dieses ist mein Bruder instead of the colloquially idiomatic beeilen Sie sick or mac/ien Sie schncll \ das (kier) ist mein Bruder. I find in English phrase-books such fossils as may 1 have the pleasure of drinking wine with you, Miss? | your health, Sir ! together with dinner-table comments such as this beef is delicious : it melts in the mouth \ I love fat. In some of these books a wife is still a good lady. On these principles learned foreigners might address impudent cabmen with zounds, sirrah / or even 'sdeath ! Some compilers of phrase-books seem to forget that we no longer cut our pens or snuff our candles. Storm quotes from Otto's French conversation- grammar la servante nettoie la chambre, which, he says, would make the same impression on a Frenchman as the handmaiden CONVERSATION 213 cleanses the chamber would on an Englishman ; the correct modern form is la bofine fait la chambre. It is interesting to compare the modern phrase-books or 1 parleurs ' with a much older specimen of this kind of litera- ture — the Hermeneumata or Interpretamenta of the Greek Julius Pollux, who was born 150 a.d. and died in 208, and was pro- fessor of literature at Athens under the reign of Commodus. His book was intended primarily for the use of Greeks who wished to learn to speak Latin. The following extracts are from A. Boucherie's edition in Notices et cxtraits dcs mamiscrits de la Bibliotheque JVationale, Tome xxiii., scconde partie (1872), with most of the peculiarities of the orthography unaltered (the manuscript is a ninth-century copy of an older text), but with the text written continuously instead of in two columns of mostly single words, one column Greek, the other Latin, as in the original : ' Epide or5 pollous epithu- Quoniam video multos cupi- mountas ellenisti dialegesthai entes graece disputare et latinae, kai r5maeisti, mete eukerSs neque facile posse propter diffi- dinasthai, dia ten diskerian kai cultatem et multitudinem ver- poluplethian ton rematon, ouk borum, non peperci hoc facere ephisamen touto poiesai, eina en ut in tribus libris interpretoriis trisin bibliois ermeneumatikois omnia verba conscribam. panta ta remata sungrapsomai. Arkomai graphin apo alpha Incipio scribere ab alpha eos 6. usque o.' I will now give the Latin text only : — ' Bona fortuna, dii propitii. Preceptor, have. Quoniam volo et vaide cupio loqui graece et latinae, rogo te, magister, doce me. Ego faciam, si me adtendas. Adtendo diligentur. Quoniam ergo video te hujus rei, hoc est, ejus interpreta- tionis quae dicitur latinae, cupientem, demonstrabo tibi, fili, quoniam non est cujuslibet hominis deprehendere, sed docti et ingeniosi esse doctrinam. Propter hoc etiam tibi magis, qui nescis nihil disputare, exponam. Opus ergo tibi est quae praecipio : auditus, memoria, sensus ; usus cotidianos artificem facit. Hoc tibi, si praesteteris mercedes, potes discere. Duo ergo sunt personae quae disputant, ego et tu : tu es qui interrogas, ego respondeo. Ante omnia ergo lege clare, diserte. 214 THE PRACTICAL STUDY OF LANGUAGES Libenter te vidi | Et ego te. Quis pulsat ostium ? | A Caio ad Lucium. Si hie est, nuntia. Venit a Caio. | Roga ilium. Quid est, puer? j Omnia recte, etiam domine. Misit tibi epistolam signatam. | Da legam. Scripsit mihi de negotio. Vade, puer, et nuntia quoniam venio. Date mini calciamenta ; adfer aquam ad faciem ; da subar- male, cinge me ; da togam, operi me ; da penulam et annulos. Quid stas, sodalis ? Tolle quae opus sunt, et veni mecum ; festino ad amicum antiquum, senatorem populi Romani, qui a Romulo deducit genus, a Trojanis Aeneadarum.' It will be seen from these specimens that Pollux' dialogues are, on the whole, neither better nor worse than most modern ones. Although generally simple and to the point, the pedant has certainly triumphed over the practical linguist in the last sentence, of which Ollendorff himself need not be ashamed. The succeeding dialogues deal in the same way with such subjects as going to the shops, taking a bath, dining. The writer often falls into the mistake of bare enumeration of words belonging to the same category without adding anything to differentiate them (p. 173), as in praecide cervum et gallinam et leporem et colliculos. Indeed, he soon tires altogether of such trivial compositions, which afford him no opportunity of displaying his learning and rhetorical skill, and his book degenerates into a mere vocabulary of words arranged roughly under categories. He begins with the names of divinities, then goes on to the signs of the zodiac, enumerates the constellations and stars, then gives words relating to the atmosphere and its phenomena, the winds, seasons, divisions of time, terms relating to medicine, navigation, civil government, military organization, agriculture, names of trees, edifices, relationship, serpents (!), parts of a city, the different trades and professions, and so on. If we compare the Hermetieumata of Pollux with the Colloquy of Archbishop ^Elfric and his disciple yElfric Bata, composed about 1000, we cannot hesitate to give the palm to our own countrymen. The full title of the work is Colloquium ad pueros linguae Latinae locutione exercendos, ab sElfrico primum compi- lation, et deinde ab sElfrico Bata, ejus discipulo, auctu?n, Lati?ie et Saxonice. These dialogues are not only good from a pedagogic point of view, but have intrinsic merits of their own. They are inspired by a liberal and humane spirit, and are full of graphic descriptions and incidents. Accordingly, CONVERSATION 215 in my First Steps in Old English I have taken the Old English interlinear version of them, and made it into an idiomatic Old English text. The following are specimens of the original Latin : — 1 Nos pueri rogamus te, magister, ut doceas nos loqui Lati- aliter recte, quia idiote sumus, et corrupte loquimtir. Quid vultis loqui ? Quid curamus quid loquamur, nisi recta locutio sit et utilis, non anilis aut turpis ? Vultis flagellari in discendo ? Carius est nobis flagellari pro doctrina quam nescire ; sed scimus te mansuetum esse, et nolle inferre plagas nobis, nisi cogaris a nobis. Interrogo te quid mihi loqueris ? quid habes operis ? Professus sum monachum, et psallam omni die septem sinaxes cum fratribus, et occupatus sum lectionibus et cantu ; sed tamen vellem interim discere sermocinari Latina lingua. Quid sciunt isti tui socii ? Alii sunt aratores, alii opiliones, quidam bubulci, quidam venatores, alii piscatores, alii aucupes, quidam mercatores, quidam sutores, quidam salinatores, quidam pistores loci. Quid dicis tu, arator ; quomodo exerces opus tuum ? O mi domine, nimium laboro. Exeo diluculo, minando boves ad campum, et jungo eos ad aratrum. Non est tam aspera hiemps ut audeam latere domi, pre timore domini mei ; sed junctis bobus et confirmato vomere et cultro aratro, omni die debeo arare integrum agrum aut plus. Habes aliquem socium ? Habeo quendam puerum minantem boves cum stimulo, qui etiam modo raucus est pre frigore et clamatione.' Neither Pollux nor the Archbishop need fear a comparison with the following extract from Waddy's English Echo (10th edition, 1877), in which I have enclosed superfluous words and word-groups in ( ), so as to enable the reader better to realize the Gladstonian verbosity of the language : — ' A little bread, (if you) please. — Will you be good enough to pass (me) the salt ? I do not think this soup is sufficiently seasoned. — My husband is so very fond of salt and (of) sugar. I tell him sometimes that if he eats so much sugar he will get shockingly stout. — Do not terrify me in that way ; I should be horrified if I thought I was likely to be a fat man. — I remember when you were in Germany you were very slim and 216 THE PRACTICAL STUDY OF LANGUAGES agile. 1 Our friends gave you the nickname of the active Englishman. — I do not think I could run or leap as I used to (do) then. I have not tried anything in that way since I became sedate so many years ago. — Why, then, we must now call you the lazy Englishman. — (Oh dear !) that is worse than ever; I hate to be thought slow and torpid. — May I have the pleasure (, Madam,) of assisting you to some wine ? — Thank you (, Sir,) that is Sherry ; if you will allow me I will trouble you for some of the Bucellas. It is in the other decanter. — Adolph, you and I must drink a bumper (with each other) in memory of bygone times. — Your very good health, my old friend ! — And yours, and that of your wife and family ! — Will you let me send you some of this grouse, or would you prefer some venison ? —Thank you, I will wait for a piece of your old English roast beef, of which I have heard so much. — I am sorry (that) we have none to-day. But there is a very nice shoulder of mutton. — I will ask you for some of that macaroni pudding. — Charlotte, you are drinking with your mouth full, that is very unlady-like indeed. John, remove these things and bring us the dessert and wine. — Try some port, Adolph; I think you will approve of it ; or would you prefer claret ? You always had the reputation of being a good judge of wines, 2 so I will accept your recommendation. — Emily, here are your walnuts ; I will break them open 3 for you. — Charles ! Charles ! you are cracking those nuts with your teeth ; you should use the nut-crackers. — I have not got any, mama. — There is a pair just under the edge of your plate ; what a careless little fellow you are. — Now, my children, go with your mama. May this little gentleman, Charley, stay with us once? Yes, if he will behave himself very well. — We will rejoin you, my dear, directly. Now, Adolph, draw your chair nearer the fire and fill your glass again. — These pears are very large and fine. Do you grow them yourself ? — No, I have no garden here of any size. My wife got them in Covent Garden Market. — Really, I must felicitate 4 you upon having so beautiful and amiable a wife. I am delighted with her. — You will like her better the more you know of her. Her beauty is her least recommendation. She is my greatest treasure.' This could never have been genuine English conversation, 1 = active; the word agile is put to avoid the repetition of active; but this is literary, not colloquial. 2 = wine. 3 break open = crack. * — congratulate. CONVERSATION 217 not even in i860, about which time Storm supposes this book was first published, although it must be admitted that the matter of these dialogues is generally good. Strange to say, Jeaffreson and Boensel's English Dialogues (1891) are even more stilted, though in a different way. Take such a sentence as this : ' The practical results of science during that period are doubt- less unsatisfactory, but still I think they have been underrated. If Nature was not interrogated as she has been since, there was notwithstanding a considerable improvement in the mechanical arts, so far as they affect the daily comforts of life. Besides, we cannot deny that the foundations of chemistry and optics were laid in that period. The art of navigation, even before the invention of the compass, was greatly improved, and you must not forget that the printing-press belongs to what are called the middle-ages, though coming within a few years of their expiration. There are many other points I might urge in favour of a more respectful attitude towards the science of the middle-ages, but I must leave you now, as I have an appointment.' This is preaching, not talking. But even in the lighter dialogues there is a constant substitution of unnecessary artificialities for plain colloquialisms, such as your very humble servant for /, together with archaisms such as how is that, pray ? and, what is strange in a book for foreigners, unnecessary insertion of foreign expressions such as nuit blanche, auf wiedersehen ! As a further illustration of the difficulty of getting really idiomatic conversational language, I may refer to Spoken English : Everyday Talk, by E. A. True, which was originally a translation of Franke's Phrases de tons les jours. How mis- leading a guide it is, may be seen from the following examples : ' Since the month of August we have constantly had fine weather (= we have had constant fine weather since August). Please, put a fire in my room. He resembles his eldest brother. Put the lamp on the piano ; it is too much exposed here and might be upset. Mary, pick up your hat, it is lying on the floor, and then don't let your toys lie about everywhere. Remove these things from the table ; they are in my way. On his way to Paris he must needs pass through this place.' Prof. Jespersen, who re-edited this work, has got rid of most of the worst faults of the original, but it is hopeless to 218 THE PRACTICAL STUDY OF LANGUAGES try to make idiomatic dialogues by translation; one of the languages must sacrifice its individuality. In short, it should be clearly understood that to write plain colloquial dialogues is a task of great difficulty, and should not be undertaken without preparation— quite apart from a knowledge of phonetics. No one but a native should ever attempt it. In the present dearth of really useful and reliable phrase- books, the safest way of learning conversational idioms is to read novels and comedies, selecting those idioms which seem most useful and passing over the rest. But this is a slow and difficult process, and requires a peculiar linguistic tact and a special kind of memory to produce good results (p. 122). It is also difficult for the foreigner to know whether the idioms are really modern colloquialisms such as he can safely assimilate without fear of falling into archaisms on the one hand, and vulgarity or slanginess on the other. Plays are, on the whole, not so good as novels for this purpose, as the stage in most countries seems to develope a traditional and conventional colloquial style of its own. The difficulty of learning the spoken language from literature lies deeper than this : it is the result of the literary being necessarily anti-commonplace. We do not go to literature to find a photograph of our everyday life and talk : we seek the flavour of originality and divergence from the associations of ordinary life. Even when humorous literature is founded on descriptions of the commonplace, the result is seldom anything that the foreign learner can assimilate with advantage. We must, therefore, have books specially written for this purpose, and for no other. The best book of the kind that has hitherto appeared is Storm's edition of Bennett's Norwegian Phrase-book. 1 But as it is mainly intended for tourists who know little or nothing of the language, it does not claim to be anything but a rough guide, the idioms being arranged alpha- betically for ready reference. 1 This book has since been re-edited by another hand, and the phonetic transcription has been discarded. CHAPTER XVI LITERATURE ; LITERARY COMPOSITION In a rationally progressive method of learning languages the approach to the literature of the foreign language will be made gradually. The learner, as he advances, will be able to choose his texts with greater freedom and with less subordination of matter to form, till at last he is able to read with profit the actual literature itself, unmodified and uncurtailed. As there are gradations in the approach to the literature, so also there will be gradations in the study of the literature itself. The student will, of course, begin with modern prose in its simplest form and that which approximates most closely to the spoken language. In fact, some of this simpler literature will perhaps be already familiar to him in the pre-literature stage (p. 181). He will then proceed gradually to the higher rheto- rical and imaginative prose, and then to archaic prose and poetry. This procedure is quite opposed to the older method of not only introducing the learner to the literature of the language before he has mastered its vocabulary and grammar, but ot making its classics the vehicle of elementary instruction. This is much as if a music-teacher were to give his pupils classical sonatas to learn the notes on instead of beginning with scales and exercises. Even in Latin there is no necessity for begin- ning with such an author as Virgil, whose literary merits no beginner can be expected to appreciate ; in a modern language there is no excuse whatever for such a course. Besides, when the classics of a language are ground into pupils who have neither knowledge enough of the language to appreciate their stylistic merits, nor maturity of mind and taste to appreciate their ideas, the result is often to create a disgust for literature generally. 219 220 THE PRACTICAL STUDY OF LANGUAGES Composition The ambitious student who aspires to original composition in the higher literary style of the foreign language must work his way up cautiously and slowly, beginning with more or less free reproduction of what he has read, at first in an almost colloquial style, then in a higher literary diction. The simplest and most spontaneous form of literary com- position is letter-writing, which is at the same time the most useful. Letter-writing is, indeed, the only form of literary composition that most people ever attain to, even in their own language. Even in letter-writing we must firmly resist the temptation to imitate ornate literary models in the foreign language. I have already remarked that a too early introduction to the masterpieces of literature often inspires the learner with a dis- like for literature generally. But it may have the opposite effect of rousing too much enthusiasm — a too burning desire to emulate the example of the great masters of literary style. There was once a firm of German merchants at Hamburg who had a native clerk to do the foreign correspondence, which he did very well on the whole, his only failing being a weakness for fine writing, the result of assiduous devotion to the works of Byron and Bulwer Lytton. It happened in course of time that the firm received a consignment of leeches from England. The passage was rough, and as the leech is a delicate insect, the result was that the German clerk wrote a letter to the English firm in which he said, 'Dear Sirs, we beg leave to acknowledge receipt of the consignment of leeches as per invoice, but regret to be obliged to add that the greater part of them have gone to glory.' In many Oriental languages there would be no incongruity in such a mixture of styles ; for in them the inflation and artificiality of the literary language often goes to such a length that all sense of congruity and proportion is lost, and the style is valued according to the degree of its departure from the plain colloquial style. This is why many Hindoos who have a perfect knowledge of English for speaking purposes, become entirely un-English when they write even an ordinary letter, not by making mistakes in grammar, but by using words and idioms in unappropriate surroundings. A Hindoo clerk in LITERATURE; LITERARY COMPOSITION 221 whose charge a pony had been left by his employer, wrote thus to him : ' I have the honour to report that the little horse, since your honour's departure, has assumed a devil-may-care attitude, and has become violently obstreperous. This morning, at 6 a.m., the said little horse eloped from my custody, but, with the favour of heaven, he may return.' In a well-known biography written by a Hindu, a description of the confusion caused in the house by some calamity calls forth the ejaculation, ' Here was a pretty kettle of fish ! ' In the same book it is said of the hero of the biography that ' in his youth he was filamen- tous, but he afterwards became plump as a partridge.' Next to the imitation of unsuitable models, the greatest snare in composition in a foreign language is originality of style. We can be original in our own language only, although even there we cannot deviate far from the beaten track : in a foreign language we must adhere rigidly to our models. This is why original writers are seldom good linguists : they know instinc- tively that their own language is the only instrument of thought they can hope to handle freely, and so they have no induce- ment to try to master any other. However wilful the deviations of a native may be from rule and tradition in his attempts to frame new expressions for new ideas, or to express more forcibly the old ones, the result will always have a native flavour — it may be uncouth or obscure English or French, but it will always be unmistakeably English or French as the case may be. But if we try to be original in a foreign language, there is always a danger of our originality assuming a native form. The result will be a language which is incorrect in grammatical construction or in phraseology, not merely incon- grous, as in the examples last given — that is to say, it will use forms which are not only non-existent, but which no native writer could possibly have evolved. A foreigner's style may of course have a quaint and apparently original character simply through being tinged with reminiscences of his own language, not through any originality of mind in himself. Dialectal influences may have the same effect. Thus much of the supposed originality of Carlyle's style is the result of the influence of the Scotch dialect. When he speaks of newly built suburban houses as ' the human dog-hutches of the period,' an English reader is struck by the picturesqueness of the word ' dog-hutch,' which he thinks is an original creation 222 THE PRACTICAL STUDY OF LANGUAGES of Carlyle's, perhaps the result of his supposed imitation of German; but this picturesqueness was quite unintentional on Carlyle's part, being simply the result of his ignorance of the correct expression ' dog-kennel.' It is, of course, often difficult or even impossible to distinguish these two factors. The passage already quoted from Terrien de la Couperie (p. 72) is a specimen of the style into which a foreigner may insensibly lapse who lets himself go under the impression that he has a perfect mastery of the language. CHAPTER XVII DEAD LANGUAGES It need hardly be said that the study of dead languages should be based on the same general principles as that of living languages, with, of course, such special modifications as experience and common-sense show to be advisable. It must, in the first place, be realized that there is no essential difference in structure between a dead and a living language. The dead Latin and the living Italian differ widely in structure ; but this is not because all dead languages are inflectional, all living languages analytical. On the contrary, we need not look far to find in Russian a language whose inflections are quite as complicated as those of Latin. In Hebrew, on the other hand, we have a dead language of comparatively simple structure, the simplicity being, like that of Italian, mainly the result of phonetic change. External Difficulties As the study of dead languages is subject to various external difficulties of its own, resulting from their being no longer spoken and being preserved in a limited number of texts which are sometimes fragmentary and often obscure, it is even more necessary than in the case of living languages to remove as many of the other external difficulties as possible. Hence texts in dead languages should be printed with all the helps that transliteration, quantity- and stress-marks and other phonetic diacritics, can afford, not to speak of punctuation, quotation-marks, capital letters or other marks to indicate proper names. If these are found helpful with modern languages — if we do not like to dispense with them even in our own language — they must be still more useful in such 223 224 THE PRACTICAL STUDY OF LANGUAGES languages as Sanskrit, which, as it is, is printed not only with- out most of these helps, but without even the help of word- division ! Normalizing Some dead languages have been handed down to us in a more or less fixed conventional orthography like those of the modern European languages. Other languages died out with- out ever having developed a fixed orthography, so that in these languages the spelling may vary not only from generation to generation, but also from manuscript to manuscript. Old English, Middle English, Old French, Old Irish afford examples of this extreme, while Sanskrit is an example of the other. In Sanskrit the orthography is so absolutely fixed that even the archaic language of the sacred hymns of the Rig- Veda is written with the spelling of the much later classical Sanskrit, in spite of the fact that this later spelling often does violence to the metre, as when the Vedic suria, aria are written su?ya, arya. The worst of these fixed traditional orthographies is that they are generally much more modern than the language they profess to represent, so that they may be really quite unauthori- tative. Thus the present Sanskrit devanagari alphabet and its whole system of orthography was not evolved till long after Sanskrit had ceased to be a living language. This is also the case with the orthographies of Greek and Latin, which are full of misleading spellings, the result of mistakes and confusions of comparatively modern times, as is soon seen by comparing the traditional spelling with that of the few texts — mostly inscriptions — preserved in contemporary documents. Such spellings as Virgilius and some of the details of Greek accentuation are simply modern monstrosities. It is evident that an unfixed orthography such as that of Middle English and the Greek dialectal inscriptions, which attempts — however imperfectly — to do justice to the peculiarities of its period and locality without regard to tradition, is far more valuable for purposes of scientific research than any fixed orthography. It must not be imagined, however, that the chaotic ortho- graphies of such languages as Middle English and Old French give a really faithful picture of the languages themselves any more than the fixed orthographies do ; for language is as little DEAD LANGUAGES 225 lawless on the one hand as it is unchangeable on the other. The varying spellings of one and the same word in the unfixed orthographies are distortions of the truth just as much as the other extreme of writing all the words in a certain period of the language just as they were written several centuries before. The value of these variations — these hesitating attempts to do justice to imperfectly understood distinctions of sound — lies in the varied evidence they afford us by which we are often able to determine with certainty the one sound or grammatical form which lies behind them. Thus when we find an uneducated Englishman sometimes writing father for farther and farther for father, and sometimes writing the correct spellings, we cannot resist the conclusion that in his pronunciation the two words have the same sound (faa-Ss). The various early Modern English spellings lif lyf, life, lyfe, lyffie, etc., point to one single form just as much as the Modern English life does. Hence a normalized orthography gives, as a whole, a truer representation of a language than an unfixed orthography does, although the latter is a great help in correcting the former. For the practical study of languages normalized orthographies are indispensable, for the practical learner cannot afford to waste his time and labour on forming conflicting associations with divergent spellings the value of whose evidence he is not yet able to appreciate. It matters little to him whether or not the spelling of a particular word that is adopted in his normalized texts is the best one or not ; even if there are downright mis- takes in the normalized spelling, it is still worth his while to use it, if it materially assists his mastery of the language. When he has once learnt the language, he can easily correct any errors of this kind, and the divergent spellings will cause but little confusion. If he is studying the language for scientific philological purposes, or if he intends to read manuscripts and original documents, it will be necessary for him to make a thorough study of them. Even those who do not intend to make investigations of this kind will still derive benefit from such a study because of the training it affords in habits of linguistic observation. Pronunciation The practical exigencies of teaching make the adoption of some system or other of pronunciation a necessity in dead as Q 226 THE PRACTICAL STUDY OF LANGUAGES well as living languages ; and where the facts of pronunciation are fairly well established, it is often just as easy to pronounce correctly as incorrectly. Thus in Latin it is just as easy to pronounce non, Caesar correctly (noon, kaisar) as in the English way (non, sijzs). Even if we make the vowel of non into the diphthong (ou), that is at least better than making the word rhyme with on. So also if we distinguish Greek nomos from nomos by treating the accent as a mark of strong stress, we do not do full justice to what was probably the actual distinction, but we certainly get as near to it as is practicable from our point of view, while at the same time — and this is the really important consideration — we greatly strengthen our hold of the distinction between the two words. Even when the correct pronunciation offers difficulties, it is generally worth while to make some efforts to overcome them, without, of course, allowing this to take too much time from the general study of the language. For they may be diffi- culties which will confront the learner in some modern language. In this way, pronouncing a dead language with theoretical correctness may be a valuable help to the learning of living languages. Thus when the learner has once mastered the diffi- culty of pronouncing double consonants in Latin in such words as collo compared with cold, appellare — where he must be careful to double the unstressed/ as well as the stressed / — they will cause him no difficulty in Italian, Swedish, or any other modern language. So also the English learner will find that pronouncing such a word as Latin non with a pure long monophthongal close (o) will greatly improve his pronunciation of almost every foreign language, living or dead. If the correct pronunciation cannot be ascertained, or if its acquisition takes up too much time, the learner may, for the sake of distinctiveness, adopt a schematic, nomic pronunciation (p. 34), which he can, of course, afterwards modify or discard by the light of further knowledge without any practical difficulty. For practical purposes it is specially important to make distinctions of pronunciation in two cases : (1) when the distinctions of pronunciation are significant (p. 5) ; and (2) when they affect the metre of the verse. Thus in Latin it is absolutely necessary from both points of view to pronounce a real double (1) in collo, for cold not only has a different meaning, but also a different function in verse. It makes no difference, DEAD LANGUAGES 227 on the other hand, from either point of view whether we pro- nounce Latin v as (v) or (w). In pronouncing Chaucer we must pronounce the weak e in shoures swote, or else destroy the metre, so that, while we are about it, we may as well restore his genuine pronunciation throughout. With Shakespeare there is no general metrical necessity for a change, so there is no practical inconvenience in reading him with the modern pronunciation. Teaching through the Literature The evil effects of teaching through the literature are even greater in dead than in living languages, for in dead languages every natural obscurity is increased tenfold by our unfamiliarity with ancient circumstances and trains of thought. Such a language as Latin ought to be taught by means of the simplest possible texts, from which every literary complexity or excep- tional form has been carefully weeded. Even after the learner has begun the study of the literature itself, he should not be allowed to look at such authors as Virgil, Tacitus, or Juvenal till he is able to read simple prose and poetry with perfect ease. In modern languages this principle amounts practically to beginning with the spoken language. But as we do not learn Latin to speak it, there is no necessity that the texts should be strictly colloquial in character : all we require of them is that they should imitate the simplicity, definiteness, and directness of the colloquial language — or, at least, that they should not be unnecessarily literary, rhetorical, and artificial. As these requirements are rarely to be found in the actual literatures, it is often advisable to make special texts for our purpose by simplifying and abridging or paraphrasing literary texts suitable in matter but not in style (p. 182). Cross-associations with Modern Languages It is in one sense incorrect to call such languages as Greek and Latin dead languages, for Modern Italian and Romaic or Modern Greek are simply classical Latin and classical Greek which by gradual and perfectly continuous changes have developed into their present form without any change of place. We do not call Italian ' Modern Latin ' on the analogy of 228 THE PRACTICAL STUDY OF LANGUAGES Modern Greek, simply because we find it more convenient to give distinct names to Italian, French, Spanish, and the other dialects of modern Latin. The fact that the speakers of Italian and Modern Greek are probably not the descendants of those who spoke the ancient forms of these languages does not alter the fact that Latin and Italian, for instance, differs only in degree, not in kind, just as the Latin of Tacitus differs from that of Ennius. The only languages which can be called really dead are such as Accadian and Hebrew, which have left no direct living descendants. From a practical point of view, however, we are fully justified in calling such a language as Latin a dead language, differing essentially from a living language (i) in being no longer accessible to direct observation, and (2) in being no longer capable of producing literary works — in having a limited and definitely completed literary development. But from this point of view the earlier stages of a modern lan- guage are also dead. The deadness of such a language as that of Shakespeare differs only in degree from that of Latin. The language of Shakespeare is no more accessible to direct obser- vation than that of Virgil. In fact, as regards its phonology, it is perhaps even less so. We cannot speak, and it is doubtful whether any of us can write Shakespearian English. And from a literary point of view the Tudor period is as finished and shut off from the present period as any ancient literature. If we go a little further back, no one will deny that the Middle English of Chaucer and the Old English of Alfred are dead languages. Old English, indeed, is so remote from Modern English that the learner is often inclined to regard it as a dialect of German. The divergence between these periods is very gradual. Shakespeare's language was perceptibly easier to the middle of the eighteenth century than it is to our period, and although he was taken less seriously, he was more generally read in the first half of the eighteenth century than at any subsequent period. Chaucer, again, was nearer to the Tudor period than the Tudor period is to us, and was connected by insensible gradations with Old English. Hence there is a danger of confusion and cross-association to the foreigner who attempts a historical study of a language. Even a native is not exempt from this danger. The difference is that while the foreigner is apt to import Shakespearisjns DEAD LANGUAGES 229 into his Modern English conversation, the Englishman is more apt to misunderstand Shakespeare through giving Present English meanings to his words and phrases. This suggests two cautions. The first is : Do not work back from Modern to Old English through Middle English ! There is much less risk of confusion if the student, after mastering Modern English thoroughly, goes straight to the other extreme, and masters Old English before making himself acquainted with the intermediate Middle and Early Modern (Tudor) periods. So also the student of German should begin his historical study of the language not with Middle, but with Old High German. In the historical study of French it is still more important to begin with the real Old French, not with fifteenth-century Parisian texts or Anglo- Norman ones. The second caution is : Do not compare the different periods more than you can help ! All the comparisons that are of any use will suggest themselves spontaneously, together with a large number of misleading ones, which unfortunately will be con- firmed by the etymological translations in the- learner's text- books. When the learner first meets with Old English sona, the important thing is not to tell him that it is cognate with Modern English soon, but that it does not mean 'soon,' but ' forthwith, immediately,' just as presently in Edinburgh English does not mean ' after an interval,' but ' at once.' So also Old English smcel does not mean ' small ' (which is expressed by lytel), but ' narrow : ' Norway is a smcel, not a small country. Some well-meaning people, misled by one-sided antiquarian- ism and Freeman's monomania about the continuity of the English language — the only philological generalization he ever seems to have grasped — are incapable of seeing these self- evident facts. They protest that ' we must not isolate Old English from Modern English' — that is to say, that it is wrong to protect ourselves from confusions and mistakes by refraining from comparisons which encourage the formation of cross-associations. The isolation is only a temporary one : when the older periods have once been learnt separately, then is the time to do full justice to the fact of historical continuity. 230 THE PRACTICAL STUDY OF LANGUAGES Dead Methods in Modern Languages There is another fallacy which requires a brief notice. There is a certain school of educationalists who have a strong conviction of the great value of the study of Greek and Latin as a means of training the mind : many of them, indeed, when reminded of the fact that the majority of those who learn these languages at school never acquire even an elementary practical knowledge of them, reply that this really does not matter much, as they still get the benefit of the mental training. Those who hold these views also urge the convenience of the study of dead languages which is the result of not having to attend to pronunciation, and having to deal only with a limited literature which has been thoroughly worked up for educational purposes. Many of them further believe that the present methods of teaching modern languages have the contrary effect of weaken- ing the mind and making it more superficial. Some of them think this is inherent in the nature of modern languages. Others, more liberal-minded, think that the fault lies in the methods of instruction. They argue that if modern languages were taught like dead languages, they would have the same beneficial effect on the mind. Hence instead of assimilating the study of dead to that of modern languages, we are advised to reverse the process. These views are often further combined with antiquarian and etymological fallacies. Thus I was once told by an American pupil of the late Professor Zupitza, of Berlin, that it was a mistake to suppose that Zupitza was not interested in Modern English literature; that, in fact, he had been lecturing on Shelley's Prometheus Unbotmd, but in a more scientific spirit than a purely literary specialist. It turned out that this superi- ority consisted in his making his pupils translate the beginning of the drama into Anglo-Saxon so as duly to impress on them the continuity of the language ! There is something very unreal about this ' dead-alive ' philology. Some — often insignificant — modern text is taken, and elaborately commented upon with a long historical and critical introduction, and elaborate notes are added on points of grammar, etymology, style, and perhaps metre, the editor not being conscious of the absurdity of teaching metre without a previous knowledge of phonetics on the part of the students. DEAD LANGUAGES 231 The text is hardly ever genuinely colloquial, and is often anti- quated, so that the method practically means teaching one language by means of another language. It is not that this kind of study is necessarily objectionable in itself; but it is not the thing to begin with — it should come at the very end, not at the beginning of the course. The reader may be reminded once more that the question whether the study of dead or of modern languages affords the best training for the mind is one which has nothing to do with the question, which is the best way of learning languages. The only question we have to deal with is whether the exten- sion of the old methods of studying dead languages to the study of living languages would make the acquisition of the latter easier. Our answer to this question must be an un- hesitating negative. CHAPTER XVIII ORIENTAL LANGUAGES The study of the Oriental and of the other remoter languages has many analogies with that of dead languages. In the first place, the inaccessibility of these languages, and the difficulty of obtaining native teachers, generally obliges the beginner to approach them from the point of view of the study of dead languages. Secondly, in these languages the true colloquial element is generally even more inaccessible than in European languages, and the divergence between it and the written language is nearly always much greater : classical and vulgar Arabic, written and spoken Japanese, are distinct, mutually unintelligible lan- guages, which have to be treated in separate handbooks and grammars. Even the books which profess to deal with the colloquial form of these languages often give only an approxi- mation to the true colloquial. Thus Green's Practical Arabic Grammar, which ' was originally undertaken to meet the require- ments of English officers in Egypt,' gives a language which is a mixture of classical with modern Egyptian and Syrian Arabic, containing forms which would be quite unintelligible to an uneducated Egyptian, such as haza, ' this' (classical hafta), the learner's confusion being further increased by the occasional insertion of texts in the classical form. Yet this book — which is mainly on the Ahn plan —has lived through at least three editions (third edition 1893). Under these circumstances the learner is often obliged to master a dead form of the language as the only stepping-stone to its colloquial form. When the colloquial language is split up into a number of local dialects which are often practically independent languages, the unity of the literary language is certainly an inducement to beginning with it ; thus classical or 232 ORIENTAL LANGUAGES 233 literary Arabic is the only link by which the different c vulgar ' Arabic dialects of Syria, Egypt, Marocco, etc., can be realized as variations of one language. But when we have really practical guides to the genuinely colloquial forms of each living dialect or language, the only rational course will be to begin with one definite modern dialect, and then work back to the literary language. To learn classical Arabic as a preparation for modern Egyptian Arabic, or written as a preparation for spoken Japanese, would then be as absurd as to learn Latin as a preparation for the practical study of modern Italian. Adherence to Native Methods One of the greatest external hindrances to the study of Oriental languages is the adherence to the native methods of exposition and the native terminology in each language. It is evident that a method which suits an Oriental may not suit a European. Indeed, we may go a step further, and say that a method which suits the one is tolerably certain not to suit the other. To the Oriental ' time is no object,' for he can give his life to his one object of study — the literary form of his own language — which, besides, he already knows to some extent. To him, writing and learning grammar merely means writing and analyzing something that he is already partially familiar with. He learns to read his own crabbed and defective alphabet with comparative ease, not only because he has plenty of time to give to the study, but also because the solution of each orthographic riddle is more or less known to him before- hand. For the same reason in his grammars and dictionaries he can find his way through an abstract and complicated arrangement which baffles the foreign learner, to whom the matter is as unfamiliar as the form. The difficulties of terminology are alone a serious obstacle. Thus, in Arabic the unhappy beginner is expected from the first to remember the three short vowels by their Arabic names fatha, kasra, damma, and has, besides, to remember a number of other technical terms relating to orthography and pronun- ciation which are not clearly explained to him, and even then are difficult to understand and remember — and all this in addition to having to learn a new alphabet. When the learner has at last mastered sixteen or more pages of orthographic 234 THE PRACTICAL STUDY OF LANGUAGES absurdities, he has not learnt a single fact about the language itself. The details of Hebrew orthography — which take up nearly twenty-two pages in the very brief grammar of Strack — are even more repulsive and irritating to any one used to a rational method of learning languages. One cannot blame the scribes who evolved these preposterous orthographic complexities for they knew no better ; although it is a pity that when the Arabs borrowed their system of writing from a Christian people, they did not adopt the Coptic instead of the Syriac alphabet. But there is no reason why European learners should be hampered with them just at the time when they require to be able to give their undivided attention to the very real difficulties that encounter them in the language itself. So also it is excusable in the Chinese that they regard the addition of a stopped consonant to a vowel as a kind of tone, because the peculiar character of their writing made it possible for them to dispense with any minute analysis of sounds ; but it is nevertheless annoying to look up a Chinese word or ' character ' in a dictionary, and then to be told merely that it has the ' entering ' tone (jip fir)), or, in plain English, that it ends in some one of the stopped consonants /, k, p ; the result being that unless we know how the word is pronounced in those modern Chinese dialects which still keep the final stops, and are able to check their often conflicting evidence by know- ing the pronunciation of some word given as a rhyme to our word, we are left in an unpleasant state of uncertainty as to its pronunciation. In many Oriental languages the same difficulties of unneces- sary technicality and confused statements follow us through the grammar and dictionary. Everywhere a new terminology and new arrangements, which have to be learnt over again in each language. Hence even Sanskrit, which in itself is not more difficult than Latin, and whose alphabet is remarkably rational and phonetic in spite of its complexity, was at first considered un- attainable by Europeans. Texts The want of texts suitable for beginners is as keenly felt in Oriental as in dead languages. There are few of them that can show such a collection of comparatively simple and colloquial ORIENTAL LANGUAGES 235 texts as those contained in TJie Thousand and One Nights, which, however, have the disadvantage of being neither classical Arabic on the one hand nor fully modern on the other. So great is the dearth of simple texts in Chinese that Summers, in the chrestomathy to his Handbook of the Chinese Language, actually includes a translation into Chinese of some of yEsop's fables by an Englishman ! In fact, almost the only texts that are even approximately colloquial in Oriental languages are those which have been taken down from dictation by European scholars. Spitta's Contes arabes modernes are an excellent specimen of such work, although unfortunately they are written down so badly from a phonetic point of view as to be misleading to those who have not an independent knowledge of Egyptian Arabic. CHAPTER XIX GENERAL CONSIDERATIONS Time and Effort In learning a language we must advance steadily at a certain speed — neither too fast nor too slow. Hurried reading either of text or grammar results in the learner forgetting half of what he reads, or in his forming vague instead of definite associations with what he does remember. The nearer the language is to those which he already knows, the greater the temptation to hurry. Thus, the beginner in Italian, finding that when he has once learnt to recognize a certain number of disguised particles such as anche, ' also,' />ero, 1 therefore,' he can often guess at the general meaning of whole paragraphs, gets into the habit of superficial reading, until by a succession of gross blunders he is obliged to confess to himself that he has been confusing di, ' of,' with da, ' by,' that he does not know one pronoun from another, and so on. He then sees that he has simply thrown away so many months, and that he must either give up Italian, or else begin again at the beginning, and go through the tedious drudgery of unlearning false associa- tions and going through the elements of the language again after the study has lost the charms of novelty. In fact, in dealing with such a language as Italian, it requires a determined effort on the part of the learner to read as slowly and carefully as he ought. Here we see one of the indirect uses of phonetics : if the learner tries conscientiously to do justice to the pronunciation— if only to the extent of dis- tinguishing the close and open vowels, and sounding the double consonants distinctly— he will have to read each sentence so often that there will be no fear of anything in it escaping his attention. 236 GENERAL CONSIDERATIONS 237 If, on the other hand, the language is remote and unfamiliar, there is no temptation to quick reading, which is, indeed, im- possible at first. Even if the language is not presented to the learner in an unfamiliar alphabet, the unfamiliarity of the vocabulary will enforce a slow progress. The progress, indeed, may be so slow that the learner is unable to keep up the sense of continuity : by the time he comes to another example of some word or construction, he has forgotten the former one. Hence most Oriental languages cannot be learnt by merely reading at the rate of an hour a day : a slow learner might go on at this rate for ten years without making any real progress. Such languages must be studied intensively, with a concentra- tion of effort. Thus it is more economical in the end to give four hours a day for a year to such a language than to spread the same number of hours over three years. Sir Thomas Wade used to tell his pupils that they ought to give eight hours a day to their Chinese ; but this was addressed to those who were qualifying themselves to serve as interpreters, and therefore had to learn a variety of subjects which would be superfluous to the purely literary or philological student. His method, too, was an imperfect one. Under any circumstances most learners would do well to reduce these eight hours to six ; for when tired brain and irritated nerves make the attention flag, the associations necessarily become weaker, and the discrimination of minute points becomes almost impossible. If the student is perfectly free, and his sole object is to learn the foreign language with the maximum of thoroughness in as short a time as possible, he must work at it continuously every day as long as he feels that he is getting his full value out of his time and labour. How many hours this means will depend on the idiosyncrasies of the learner — on the degree of his interest and enthusiasm, the strength of his motive in learning, his surroundings, his health, and lastly on his intellectual capa- bilities. At first, too, his progress will depend much on his method of study, and on the character of his helps. Most learners of remote languages waste many years through using bad methods and bad books, although enthusiasm and per- severance will always triumph in the end. If the time given to the study of a language is to be utilized to the utmost, a certain portion of it will be set aside for repetition. 238 THE PRACTICAL STUDY OF LANGUAGES The thorough student whose memory is not exceptionally quick should always read over again every day what he read the day before. After a month or so, when he has come to a convenient halting-place, he should then go over everything again, so as to pick up those threads of association which have been dropped through the slowness of his progress. He should then read for another month, and then revise his month's reading in the same way. At the end of, say, six months, he should then revise the whole. The more difficult and remote the language, the oftener this process should be repeated. In fact, each text should be gone through over and over again till the learner feels that he is getting no more good out of it — that he must strengthen and freshen his associations with the words in it by meeting them in new texts and in different contexts. Results ; Stages and Degrees of Knowledge Perfect Knowledge. — A perfect knowledge of a living foreign language would imply the power of conversing on ordinary topics with such fluency and correctness as not to be taken for a foreigner, together with that of writing a letter correctly on any familiar subject, and of course being able to read what is written in any branch of general literature. To expect more than this would almost be to expect the foreigner to know the language better than an educated native : by perfect knowledge from a practical point of view we mean a knowledge which puts the foreigner on a level with the average native in all ordinary affairs of life. It need hardly be said that this ideal is seldom attained purely by systematic study. Such a mastery of a foreign language is generally the result of quite exceptional linguistic talent — aided, however, in most cases by some kind of syste- matic grammatical study — or of favourable circumstances. If the circumstances are so favourable as to result in the learner partially or wholly forgetting his own language, the victory cannot be said to be a fair one. This leads us to the question, Is it possible to be truly and per- fectly bilingual ? The answer is, Yes, it is just barely possible. But generally what at first sight seems perfect bilingualism is not really perfect : one of the languages has to suffer. Even GENERAL CONSIDERATIONS 239 when a practically perfect command of two such languages as French and English is kept up by alternate residence in the two countries, the respective speakers of the two languages will generally find that there is something queer, something foreign in the pronunciation of one of the two languages — perhaps in both of them. Where the pronunciation is not perfect, the construction may be theoretically perfect, but is seldom practically so. When bilingualism is the result of living on a linguistic border — as that between England and Wales — the children often learn to speak the two languages with apparently equal ease. But then the languages they learn are themselves already mixed. The children on the Welsh border speak neither pure English nor pure Welsh — they speak anglicized Welsh or celticized English. When our methods of studying languages are perfected, it is probable that perfect bi-, and even poly-lingualism will become more common, though it will be attainable only by those who have special gifts. Thorough Knowledge. — The ordinary learner can aim only at what is called a thorough knowledge of the foreign language. A thorough, all-round knowledge implies speaking with moderate fluency and sufficient accuracy of pronunciation to insure intelligibility, and, of course, the power of understand- ing the natives, and sufficient command of the grammatical structure of the language to avoid grammatical errors, a know- ledge of the necessary idioms, and being able to write a letter and read the literature. All this without implying the being taken for a native. Even this degree of knowledge is not common in this country, and where it exists, is generally the result of infinite expenditure of time and perseverance. Generally this thorough knowledge is one-sided. It often applies only to the written language, a sound critical knowledge of which is often accompanied by complete inability to speak. Polyglot or ' parrot ' linguists may be divided into two main classes : (1) those who can speak their languages — or the majority of them — fluently, and (2) those who can only read them. The former alone fully deserve the appellation of ' born linguists ; ' with the latter, the acquisition of many languages is rather the result of concentrated patience and enthusiasm aided by a good memory than of any special talent. 240 THE PRACTICAL STUDY OF LANGUAGES Elementary Knowledge. — A sound elementary know- ledge implies only the power of reading at sight any simple prose text. Thus, if the learner could translate a page of Grimm's Mahrchen which he had not seen before with moderate accuracy and without any great omissions, we might say that he had an elementary knowledge of German. This knowledge might be attained in about six months of moderate work — an hour a day — by an English adult of average linguistic intelligence, working with good books and with a good method. With a remoter language, offering, how- ever, no special external difficulties, such as Modern Arabic in a Roman transliteration, or Finnish, more time would be required — perhaps a year. With external difficulties, such as those caused by the Sanskrit or Arabic alphabet, the time would have to be largely increased : three years would be a short time in which to learn to read simple texts in Arabic or Sanskrit in their national alphabets. Chinese or cuneiform writing would require still longer time, both as regards number of years and number of hours of daily work. Many learners would require a teacher to reach this standard of progress, especially in the remoter languages ; some, however, advance more rapidly by themselves. Some may take twice as much time, and fail after all. Elementary Theoretical Knowledge.— Such an ele- mentary knowledge — modest as it seems — is more than sufficient for the purposes of theoretical linguistics and comparative philology. Every one who begins comparative philology is struck first by the limited range of the vocabulary it deals with, and secondly by the fact that a large number of the words quoted do not occur at all in the ordinary literature of the language : half the Greek words seem to come out of Hesychius and Suidas, half the Latin words out of Festus and Varro. In short, lecturing on comparative Greek grammar does not necessarily imply any practical knowledge of Greek : it is enough to have a general knowledge of the structure of Greek so as to be able to avoid mistakes in quoting the few hundred words that reappear over and over again in the comparisons on which Aryan comparative philology is founded. Thus we arrive at a still lower stage of knowledge, which we may call the elementary theoretical knowledge of a language. This kind of knowledge implies only the power of translating GENERAL CONSIDERATIONS 241 certain strictly limited texts which have been already learnt, it being understood that the texts altogether are long enough to give examples of the main features of the structure of the language in its simpler form. This is the kind of knowledge that would be acquired by going through my Anglo-Saxon Primer; that is, the result of learning about fifty pages of gram- mar, and thirty-five pages of texts. A remoter language would, of course, require more grammar and, perhaps, more texts. As this kind of study necessarily presupposes a rather high standard of intellect and a certain enthusiasm for the subject, it need not require more than a month, or even less, according to the nearness of the language. With a remoter language more time would be required, and, as already remarked, longer primers. In Gabelentz' Anfangsgriinde der chinesischen Gra?n- matik the grammar of the classical language takes up eighty-four pages, thirteen of which, however, are given to the explanation of the system of writing, the bulk of the grammar being also increased by everything being given twice over, in the Chinese character and in transcription. The classical Chinese texts at the end, which are accompanied by transcriptions and translations, take up only thirteen pages. General Stages of Knowledge. — In all practical study of languages there are two main stages, one in which everything is strange to us — in which we feel uncomfortable and not at home — the second, in which the main features of the language are familiar, and we begin to know what to expect, and feel instinctively whatever is contrary to the genius of the language. These two stages occur in every branch of the study. Thus in learning the peculiar word-intonation of such a language as Swedish, our own imitation of it sounds at first strange and affected, and we feel as if we were making ourselves objects of ridicule, however correct our imitation may be ; but after a time we have the exactly opposite feeling : we feel that our own intonation is more or less in harmony with that of the natives, and when we hear our own countrymen speaking the language with English intonation, it sounds as strange to us as to the natives. This is a proof that we have learnt to appreciate the native intonation with our ears at any rate. So also when the particles of Old Greek or classical Chinese seem useless encumbrances to the learner, he ought to acknow- ledge that this is a proof of his not having a real knowledge of R 242 THE PRACTICAL STUDY OF LANGUAGES these languages ; when the absence or misuse of one of the particles jars on his linguistic sense, then he may boast that he has really begun to ' live himself into the language.' Epitomes and Note-books In the case of languages and dialects which have hitherto been little studied, there may be a want of the necessary helps, so that the student may have to make part, at least, of the grammar for himself, and may have to make his own dictionary as he goes along. But this is an extreme case. And such independent work is neither an essential element of any practical method of learning languages nor even a supplement to it : it is simply filling up a gap in the materials on which the method is founded. Even when there is a complete grammar, it may be advisable for the learner to make a special abstract of it for his own use. But this, again, is merely filling up a gap in the materials ; for such an abstract ought to have been already provided for the use of other learners as well ; and if we accept the principle of one method for all, there can be no particular object in each learner making a special epitome of grammar for his own use. Indeed, we may ask, How can a beginner know with certainty before- hand what parts of the grammar he will require ? Of course, if his grammar is manifestly unpractical and impossible to work with, then he must do his best ; but it is much better for him to have it done for him by some one who knows the language. So also the use of note-books, in which the learner writes out the words as they occur in his texts with particulars of meaning, inflection, gender, or construction, may be regarded as supply- ing a defect in the text-book — as supplying the want of notes such as those in my First Steps in Anglo-Saxon. But it may also be regarded simply as a means of strengthen- ing the learner's associations. Thus the learner looks up the German word haus in his glossary, and finds that it is a neuter noun with plural hauser, so in order to fix this knowledge more firmly in his memory, he enters in his note-book das hazis neut. 1 house] plur. hauser, or something of the kind. The great danger of this method is that it tends to distract the learner's attention from the texts to the isolated word. Instead of thinking of the word in connection with its natural GENERAL CONSIDERATIONS 243 ' context, the learner gets into the habit of thinking of it as an isolated abstraction — he sees it as a mental picture of the entry in his note-book. This does not matter so much with concrete words of definite and simple meaning, but when it comes to entering abstract words which have hardly any palpable or definable meaning apart from their context, it becomes a pure waste of time. Even in the case of a word which can be easily isolated in thought, it is much better for the learner to read on till he finds such a word as haiis associated with the neuter article, and then to repeat the whole context till it is fixed in his mind, and then to read on till he comes to the plural Muser, and make himself realize in like manner the meaning of the form by association with its context. If he meets the plural before the singular, it does not matter much : he will not have to wait long for the singular. It is much worse when the learner enters in his note-book further particulars about the etymology and history of the word, giving, perhaps, the cognate forms of a German or Old-English word in Gothic and Icelandic, with an occasional Sanskrit root. This is mere madness from a practical point of view. It is the very antipodes of the principle of making the texts the centre of study. The Subject-matter of the Texts We often hear and read complaints about text-editions that the treatment is too exclusively grammatical, linguistic, philological ; that the subject-matter — the realien — is not commented upon, that the social life of the speakers of the language is not described, that no attempts are made to rouse the learner's interest in the literary merits of the texts or their historical value, and so on. It is certainly desirable that the learner should understand the subject-matter of what he is reading. But, on the other hand, it is equally desirable that the texts put before the beginner should deal as far as possible only with topics with which he is already familiar. If they must deal with subjects that are unfamiliar to the learner, they ought themselves to give the required definition or explanation, not of course directly, but indirectly, through the context. Any further information about the subject-matter in the earlier stages is therefore superfluous. 244 THE PRACTICAL STUDY OF LANGUAGES In fact, the question can hardly be said to arise at all till we come to the study of the literature itself. Even here, the explanations of and comments on the subject-matter should be limited to what is absolutely necessary for making the meanings of the words clear — that is, to what is really useful from the point of view of the practical study of languages. Any further elaboration of comment and illustration is irrelevant from our point of view. If the learner in the course of his linguistic reading comes on an allusion to Manichseism, or feels a great curiosity as to whether Shakespeare was really a Free- mason, or what Milton's diet was, there is no harm in his looking these subjects up in a cyclopedia or biographical dictionary, but this has nothing to do with the question what is the best way of learning languages. Even a sketch of the history and literature of the language has nothing to do with this question, although, of course, not even the most narrow-minded linguist would quarrel with his teacher for giving him this information. Teaching Children The most important difference in the classes of learners of languages is that which depends on age. Within childhood itself, again, there are different stages. The different subjects which make up a child's education must be begun at different ages, partly because there is not time to carry them on all abreast, and partly because of the natural gradation and dependence of the different branches one on another. It is of the greatest importance that the succession of studies should correspond with — or, at least, not go directly against— the progressive development of the child's mind. These considerations, combined with the conclusions we have already arrived at as to the right method of learning languages, point to the following order in a child's study of languages : — The foundation of all study of language must be laid by that of the native language. Correct and clear pronunciation of it should be insisted upon from the beginning. The reading- lessons should be made the centre of instruction as soon as possible. The first elements of phonetics and of grammatical analysis should be deduced from them. Great attention should be paid to word-meanings. There is no reason why children should not be taught almost from the beginning of their reading GENERAL CONSIDERATIONS 245 to group the words they meet with into logical as well as grammatical categories — of course, with as little terminology and abstract definition as possible. The same principles apply also — with some necessary modifications — even to the pre-reading stages of education. Phonetics, of course, should be begun in the nursery. The time will come when ignorance of practical phonetics will be held to disqualify a nurse as much as any other form of incapacity. If the infant's attempts to speak were guided into the channel of systematic all-round phonetic drill, it would on entering into school-life be already a thorough practical phonetician : all it would have to learn would be the use of a phonetic notation. The pronunciation of foreign languages would then offer no initial difficulties whatever : it would simply be a question of remembering what particular sounds occurred in the foreign language, and associating them with the symbols of the phonetic alphabet for that language. The reading-books in the native language should at first be mainly in simple prose, with only occasional pieces of simple poetry. They would, of course, be entirely in phonetic spelling on a Broad Romic basis, and with accurate marking of stress and intonation. The further development of the study of the native language would consist in widening the vocabulary, and providing reading of a higher character, and at the same time making the linguistic analysis — both grammatical and psychological — more conscious and more abstract, and framing it more and more into definite rules. The next great step will be that of emerging from the mono- lingual into the bilingual condition. The first foreign language must, of course, be one which admits of being grasped concretely in all the details required; that is, it must be a living, not a dead language. French seems to satisfy our requirements best on the whole. It might be begun at ten. After two years, German may be begun — at twelve. The only dead languages that children ought to have any- thing to do with are the earlier stages of their own language. For reasons already stated (p. 229), I think English children ought to begin with Old English. German and Old English will afford mutual help. On the whole, it would be best to postpone beginning Old English till the elements of German 246 THE PRACTICAL STUDY OF LANGUAGES are fixed in the memory — that is, till the age of fourteen. It is not necessary that much should be read of Old English literature. After a year of Old English, the learner may go on to Chaucer, and then work his way rapidly down to Tudor English. If Latin is to be studied at all at school, it ought not under any circumstances to be begun before the age of sixteen. Greek should be put on a level with Hebrew, Arabic, Russian, Chinese, and other languages, which, in spite of their great intrinsic interest and the importance of the literature they embody, have no necessary direct connection with modern European culture ; in other words, Greek should be regarded as a linguistic specialism to be entered upon, if at all, at the University. As regards literary culture, the schoolboy who has learnt something of the classics of English, French, and German literature will have as much of that kind of culture as is good for him — perhaps too much. These are the main features of a linguistic course for children. To show a different scheme — though founded on similar principles — I quote the following passage from Widgery's Teaching of Languages in Schools (p. 10), as embodying the opinions of a liberal-minded and progressive practical school- teacher who was at the same time well versed in the literature of his subject : — 1 With regard to the study of English, I venture to propose the following : ' Increase the reading-lessons in it ; let them be mainly in modern prose. Teach the very first elements of phonetics and grammar purely inductively; pay special attention to the vocabulary, grouping the words which children meet in their reader under psychologic and grammatical categories. At ten, or earlier, begin to work backwards, say to the age of Anne. With Shakespeare, their attention should be directed to his variations from modern usage, and the beginnings of a sense of the development of language made. At eleven, we might start French, reading at the same time a little Chaucer. Between twelve and thirteen, we might just touch Old English by means of a short Reader with the text on one side, and the necessary grammar on the other; some slight knowledge of the laws of language should be introduced, analogy and the regular changes of sound at least being fully illustrated. The child of twelve GENERAL CONSIDERATIONS 247 and a half is now fit to begin German. After a year's study, bifurcation must come in ; the future classical student could begin Latin at fourteen and gradually drop French, begin Greek at sixteen and devote his time to the classics. The student of the modern languages could now begin a scientific study of his three, keeping English always in the centre.' It will be seen that the main point on which I differ from Widgery is that I would rather begin the study of the older stages of English at once with Old English itself, while he prefers to work gradually backwards. I should advocate great caution in introducing children to classics such as Milton, for which their minds are hardly matured enough. As already remarked, I think Greek — and perhaps Latin too — ought to be excluded altogether from schools. This would obviate the ridiculous bifurcation into a classical and a commercial side. The phrase ' classical education ' has no longer any meaning ; learning Greek and Latin is neither education nor a preparation for it. The future man of science or scholarship wants modern languages as much as the future merchant. What remote or dead languages he or the practical man may require will depend entirely on the details of their pursuits. I would also keep all scientific, theoretical, historical study of languages in schools within very narrow limits, and draw the materials for it ex- clusively from the native language and from French and German. Methods for Adults ; Self-instruction The methods of linguistic study by adults are more varied than with children, for the aims and conditions of study are more varied. The adult can specialize, and he can devote the whole of his time to one language, thus making up by intensive- ness of study for what he has lost in quickness and adaptability of mind. With a Native Teacher. — Another important factor is that the adult can be self-taught. Even if he has a foreign teacher, he may still be self-taught. In fact, he must be so, unless his teacher is a skilled phonetician with a good method. This, it need hardly be said, is rarely the case. So that the teacher is simply a more or less passive object of observation and experiment to the learner, provided, of course, that the latter has had the necessary phonetic and linguistic training. 248 THE PRACTICAL STUDY OF LANGUAGES If the teacher is naturally intelligent, the learner will find it worth his while to try to interest him in the improved phonetic methods. If he succeeds, the gain on both sides will fully repay the time and trouble spent on it. If the teacher is prejudiced against phonetics, and persistently withholds information about the natural colloquial pronunciation and idioms, the only method is, not to let him see the printed page, but to get the information required entirely by putting such questions as ' what do you say when you meet people — when you go away — when you do not understand what is said to you — when you want some one to pass the salt ? ' and so on, and writing down the answers phonetically. But as this is difficult and slow work, especially at the beginning, it is often better to get a dialogue- or phrase-book, and look through it beforehand, so as to get what information one can from the nomic spelling, and then read over the English translation to the native, without letting him see the book, and ask him how he would say that in his own language ; in this way the danger of his simply reading aloud the perhaps incorrect colloquialisms of the book in an artificially distinct pronunciation will be averted, and if he gives the same idioms as the book, the nomic spelling of the latter will be sure to give some help in dis- tinguishing the sounds. If the learner finds he cannot get clear ideas of the sounds by hearing them in connected sentences, he should draw up tentative lists of words containing the elementary sounds — as far as he can ascertain them from his grammar and pronouncing dictionary and other helps — in combinations which present the least difficulty to him. These words must be practised diligently by first listening to the teacher's pronunciation while he repeats each word at least three times, and then trying to imitate. When each word has been gone over in this way, the teacher should read the whole list over several times. At first the learner should confine himself mainly to careful listening, till the sounds are definitely fixed in his ear, so that even if he is unable to pronounce a certain sound during the lesson, he is often able to reproduce it successfully when he practises it by himself. As regards phonetic notation, it is often most convenient at first to improvise a system of diacritic marks — dots, circles, etc., over and below the letters — not attempting to form a com- plete system of transliteration till the sounds are better known. GENERAL CONSIDERATIONS 249 At first, while the student is still unfamiliar with the grammar, he will be able to read only a very short piece — less than ten lines — every day, the pronunciation of which can therefore be studied with some care. Each day's portion should be read over by the teacher and then by the learner, first in very short groups of words ; as soon as enough of these groups have been read to make up a complete sentence, the whole sentence should be repeated, and similarly with the paragraph. The preceding day's reading should be repeated every day before going further. The learner who has not had a phonetic training should often exercise himself in repeating short sentences after the teacher without looking at the book. This will train his ear, and make him less liable to be misled by unphonetic spelling. Start with Definite Knowledge ; with a Translation. — In self-study, without any help from a teacher, the first requisite is to start with a definite and exact knowledge of every sentence in the texts. This should always be aimed at under all circumstances ; but it is doubly important when the learner has to depend on his own vigilance in detecting any mistakes he may have fallen into. It is a fallacy to regard the texts as puzzles to be solved by the help of grammars and dictionaries, thereby forming vague and often false associations which have to be modified or unlearnt. The beginner should from the first provide himself with a fairly literal translation, unless, of course, he is working with such a book as my First Steps in Ajiglo-Saxon, where everything is explained without a continuous translation. At the beginning he should make a point of reading each sentence in the translation before he begins to read and analyze the corresponding passage in the original. The dictionary should be referred to only when the informa- tion so gained is indispensable, or at least instructive, as in determining the nominative case of a noun, the infinitive of an irregular verb, or the original meaning of a word used figura- tively in some special construction. Of course, as soon as the translation has made the sense of the passage clear, it should be put aside, and every word and construction should be care- fully analyzed, as far as the student's knowledge will allow. When some progress has been made, the student should occa- sionally practise himself in making out the sense of a passage 250 THE PRACTICAL STUDY OF LANGUAGES with the help of grammar and dictionary only ; but this must be regarded as a test of knowledge and a stimulus, not as a method of study. The current prejudice against the use of translations is founded on the erroneous assumption that the learner requires to be trained in guessing and unravelling the meaning of sentences, just as any one who is learning to shoot has to make many unsuccessful attempts before he learns to hit the mark with any certainty. It is assumed that a learner who has not been set to guess at the meaning of sentences will never acquire the power of reading without a translation. But there is an essential difference between reading and shooting. While the art of shooting can be acquired only by a series of unsuccessful efforts, a knowledge of the meaning of sentences can be obtained without guessing, that is, by the use of a translation. Not till this knowledge is obtained, is it possible to analyze intelligently. The objection to schoolboys using ' cribs' is a purely practical one, namely, that they are apt to learn the crib by heart instead of comparing it with the original, while the exclusive use of a dictionary forces them both to study the texts themselves, and to do a certain amount of grammatical analysis. It need hardly be said that an intelligent teacher will have no difficulty in testing the soundness of their analysis, whether they have used a translation or not. It is, of course, most satisfactory if the boys can be taught without either a dictionary or a trans- lation, the latter being the special resort of the self-taught adult. Pronunciation. — The learner should start with a definite pronunciation, which may, however, be only a nomic pro- nunciation (p. 34). He should, if possible, read everything aloud, and get into the habit of listening to and criticizing his own utterance as if it were that of a stranger. The Grammar. — Of those principles which are of general application, many are especially important to the self-taught learner, such as that of beginning with a general survey of the language, so as to know beforehand where the difficulties lie, and the degree of attention to be given to each group of linguistic phenomena. Another important general principle is that of beginning with a short grammar. But every short grammar is not suited for self-taught GENERAL CONSIDERATIONS 251 beginners. As I have remarked in the preface to my Anglo- Saxon Reader, many of the elementary grammars and other text-books published abroad ' are intended as companions to the author's lectures, so that he naturally does not care to put his book into such a form as will make his lectures superfluous ; hence such books are generally not suited for self-instruction.' To this I may add that even when they profess to be fitted for purposes of self-instruction, they are often not so, through the author being unconsciously under the influence of the traditional methods. In going through his grammar for the first time, the student should without hesitation cut out all superfluities : he should draw his pen through all comparisons with cognate languages, all archaisms in the paradigms or lists. Even when he is entirely unacquainted with the language, common sense will often enable him to distinguish between the really indispensable and the superfluous. Thus, suppose he is in the third declension in Latin, and has come to the words with accusatives in -im instead of -em ; a little consideration of the meanings of the words amussis, buris, ravis, sitis, tussis, and vis will show that the chances are much against his meeting the first three during the first year of his study of the language, and, in short, that the only words in the list that can possibly occur with any frequency are sitis and vis. These therefore he will learn, and ignore the rest, at least for the present. In going over rules for gender, lists of derivative syllables, and so on, those rules should be singled out which are easily grasped and remembered, which include the largest number of important words and have the fewest exceptions, while those which apply only to a few words or are weakened by numerous exceptions should be passed over. The student should content himself at first with obtain- ing general ideas of the structure of the language, and should never forget that even the most accurate and exhaustive know- ledge of the grammar is in itself only a step towards a real knowledge of the language. The same principles should be followed in studying the syntax as well as the forms. All syntactical rules which are common to language generally, or apply to the native as well as the foreign language, should be passed over. At first the student should confine his attention to those rules which are absolutely necessary for the comprehension of the structure of the language, leaving the others to a later stage. 25 2 THE PRACTICAL STUDY OF LANGUAGES The first time the piece for the day is read over, after its meaning has been learnt from the translation, it should be studied analytically, till the learner understands the meaning and construction of every word, as far as his then state of knowledge will allow him. In revising the same piece the day after, its sentences should be read over and studied more as wholes, analysis being now subordinated to synthesis. Each sentence should be read over and over again till it can be re- peated without hesitation and without looking at the book. This method gives all the advantages of learning long passages by heart without taking so much time. Of course, if the learner has so good a memory that with a little more trouble he can learn by heart whole paragraphs and pages, by all means let him do so. Attention should at first be concentrated mainly on the particles and commonest words together with the general principles of the syntax. Unless these subjects are thoroughly mastered at the outset, the learner will get into the habit of disregarding them, and will then never acquire them properly, however much practice he may have in reading and speaking. This knowledge and that of the vocabulary require distinct methods of study : the former can be acquired only by careful and repeated study of a very limited portion of the literature, while the latter demands an extensive, and therefore necessarily rapid and less careful reading of all the representative branches of it. It is evident, therefore, that no attempt at acquiring the general vocabulary of the language should be made till the particles and commonest words are fully mastered. Of course, as the learner advances, he will be able to read with greater ease and rapidity. Nevertheless he should always set apart a portion of his time every day for slow and careful reading with frequent repetition, and continue this practice up to the very end of his course. The beginning of a new work on a new subject should also be read with special care till the more important elements of its special vocabu- lary have been well learnt, after which it can be read more cursorily. Careful study of the grammar should be carried on con- currently with text-reading, and this should go on during the whole course of study : there should be no idea of getting up the grammar at one stroke, and then throwing it aside. In GENERAL CONSIDERATIONS 253 reading, special attention should be paid to those words and constructions which bear on that section of the grammar which is being studied at the time. This is an additional reason for frequent repetition of what has been read during the last few months. Thus, if a certain text has been read while the learner is studying the syntax of noun-inflections, he will necessarily neglect the syntax of the verb, especially if he has not yet studied that part of the grammar in detail. Hence, when he has come to the syntax of the verb and has mastered it fairly well, he ought to go over the text he read before, partly to get fresh examples of the syntactical rules he has just been learning, partly to perfect his knowledge of the text by means of his newly acquired syntac- tical knowledge. For those words and constructions which offer special difficulties the learner may collect further illustrations from the texts either on slips or in his note-book. The use of note-books and collections generally should, however, not be carried so far as to interfere with the study of the texts themselves ; nor is anything gained in itself by removing words from their natural context to the isolation of the note-book. The learner should always bear in mind that there is no short cut to the knowledge of a language, and that this knowledge can only be obtained by persevering study of the language itself as embodied in the actual literature, and that the whole machinery of grammar, dictionary, and note-book is merely a preparation and an aid for this text-study, not a substitute for it. No plan of study can be a sound one, in which reading the texts them- selves does not take up, on an average, two-thirds of the whole time. It need, therefore, hardly be said that the less time given to composition and elegant translation — not to speak of exercise- writing — the better. Idiomatic translation from the foreign language is beneficial in many ways, but should not be attempted too early. When a firm grasp of the language has been attained, it will be time to contrast its characteristic features with those of the native language by means of such translation. In the intermediate stage between this and the very beginning, the student should learn new words and phrases as far as possible by associating them directly with the ideas they express rather than through the medium of his own language. 254 THE PRACTICAL STUDY OF LANGUAGES Translation into the foreign language, and, in a less degree, original composition in it without direct imitation of any known text, is a task of great difficulty, even when a tolerably full command of the foreign language has been attained. Indeed, most students of modern languages should not attempt anything more than a mastery of the ordinary forms of letter-writing. In dead languages of limited literature, all attempts at translation into them or original composition in them must deal with subjects and styles for which patterns can be found in the literature of the language. Thus the only kind of translation into Gothic that could be attempted would be from such books as the Pilgrim's Progress. No study requires more judgment and common-sense than the practical study of languages. The various capacities of different learners also make it difficult to lay down general rules. The three requisites — sympathetic insight into the structure of language, ear for sounds and power of reproducing them, together withsa good memory — are generally combined in different proportions. Almost total want of the two first may also be combined with high intelligence and power of dealing with abstractions. Such learners often show a deceptive quickness in learning the grammar, to which their progress in the practical command of the language by no means corresponds. So varied are the capacities and circumstances of different learners and their aims and ideals of thoroughness, that it is important to cultivate a sound and independent judg- ment on questions of method, so as to avoid being led astray by preconceived theories, and to acquire the power of profiting by experience, and modifying the plan of study accordingly. Especial judgment is necessary in settling the amount of time to be given to each day's work. Each extreme is equally hurtful. If the learner hurries over his piece of reading, he will himself feel that he has carried away only a blurred mass of associations which are soon forgotten. If, on the other hand, he studies too elaborately, sits too long over his work, and revises too often or at too frequent intervals, his powers of observation become blunted, and at last he feels that his reading makes hardly any impression on his mind, and that he gets nothing more out of it. He should, therefore, abstain from GENERAL CONSIDERATIONS 255 all attempts at exhaustive analysis, and content himself with acquiring as many new associations and new ideas as can be firmly fixed in his mind by one or two repetitions, while at the same time he seizes every opportunity of confirming earlier associations. He must also remember that by the mere process of careful reading he is acquiring a number of unconscious associations, many of which he will be able to analyze con- sciously hereafter, while many he will not find analyzed in any grammar, some of them, indeed, practically defying all analysis. CHAPTER XX ORIGINAL INVESTIGATION We have hitherto confined ourselves to that study of languages which deals with a language that has been already reduced to writing and analyzed grammatically. We have also generally assumed the existence of dictionaries and reading-books and of texts for further practice in the language. This kind of study does not, therefore, require any originality or independence of judgment on the part of the learner beyond that of selecting his helps and forming a plan of study suitable to his special needs and idiosyncrasies, and this only when he is self-taught. Decipherment The task of the investigator who aims at reducing an un- written language to writing, and then analyzing it grammatically and lexically, is a very different one. It calls not only for original research, but also for enterprise, tact, and perseverance of a higher kind than is required in the more plodding work of learning a language by means of helps already provided by others. The unfamiliarity of the language may be of any degree. It makes a good deal of difference whether the language is isolated in its affinities or is cognate with some other accessible language or group of languages. In the latter case the study of the new language may mean little more than the investiga- tion of a new dialect of a known language. But even the investigation of the spoken form of one's own language is really to some extent the investigation of a new language, especially if the real features of the spoken language are concealed by an unphonetic traditional orthography. Thus the investigation of the complicated phenomena of gradation 256 ORIGINAL INVESTIGATION 257 in spoken English — such distinctions as those between (kaen, Saet, wil) and their weak forms (kan, Sat, 1) — is practically the study of an unknown language, for the written language generally ignores not only the details, but the very principle itself of gradation. So also with the colloquial elision of the ' mute e ' (3) in French. Wherever we have to construct a new system of phonetic notation, there we may be said practically to have to deal with a more or less new language. It may happen that the language has been written down by its speakers, so that we have written texts to start with, but that nothing further has been done in the way of linguistic analysis, or that more remains to be done. The most important cases of this kind are those of dead languages made known to us by inscriptions and other docu- ments whose traditional reading has been lost. Such are the inscriptions in the cuneiform writing which from the valley of the Euphrates spread over the adjacent countries, and was used to write a variety of languages ; first, the Sumerian or Accadian language spoken by the Finno-Tartaric founders of the old Babylonian or Chaldean civilization and the neigh- bouring nations of the same stock, and then of the Semitic Babylonians and Assyrians, and lastly of the Aryan Persians and Cypriote Greeks, together with some other languages. The hieroglyphs of Egypt seem also to be of Babylonian origin ; in their oldest forms they preserve the pictorial foundation of the cuneiform writing. Fresh riddles are afforded by the Hittite inscriptions and those of Southern Arabia, together with the much later hieroglyphs of Central America. With all unknown texts the method of decipherment is the same in its general principles, however much it may vary under different conditions. In the first place, no decipherment is possible without some known quantity. Shelley tells us of the youth in his poem of A las tor that ' His wandering step, Obedient to high thoughts, has visited The awful ruins of the days of old : Athens, and Tyre, and Balbec, and the waste Where stood Jerusalem, the fallen towers Of Babylon, the eternal pyramids, 258 THE PRACTICAL STUDY OF LANGUAGES Memphis and Thebes, and whatsoe'er of strange Sculptured on alabaster obelisk, Or jasper tomb, or mutilated sphynx, Dark Ethiopia in her desert hills Conceals. Among the ruined temples there, Stupendous columns, and wild images Of more than man, where marble daemons watch The Zodiac's brazen mystery, and dead men Hang their mute thoughts on the mute walls around, He lingered, poring on memorials Of the world's youth, through the long burning day Gazed on those speechless shapes, nor, when the moon Filled the mysterious halls with floating shades, Suspended he that task, but ever gazed ^.nd gazed, till meaning on his vacant mind Flashed like strong inspiration, and he saw The thrilling secrets of the birth of time.' But this is a description of what may be, not of what is : with our present faculties we must go to work in a slower and more methodical way — we must have something to start from. It is the want of this starting-point, this known quantity, which still baffles us in the Etruscan inscriptions. The Etruscan alphabet offers none of the formidable difficulties of the cuneiform and the Egyptian writing, being, indeed, almost as easy to read as the Greek alphabet from which it is derived ; but the key to unlock the meaning of the words has not yet been found. We know that certain words are numerals, but we do not know in what order to take them, nor have we yet found any solid basis of comparison with other languages. Until a bilingual inscription is found into which some known language enters, it does not seem likely that we shall advance further than guesses at a few isolated words. In the case of the cuneiform inscriptions, the known quantities were certain proper names in the Persian inscriptions, which led to the discovery of a genitive plural ending, the language itself being practically known beforehand through being an Aryan language closely allied to Sanskrit and practically almost identical with Zend. In the Persian inscriptions the com- plicated syllabic writing of the Babylonians had been simplified into a comparatively easy system which had nearly emerged into the simplicity of the Phenician alphabet with its single letter for each consonant. If we had had only the original Sumerian inscriptions to go upon, the problem would have been ORIGINAL INVESTIGATION 259 hopeless. As it was, the numerous bilingual and trilingual in- scriptions into which cuneiform Old Persian entered soon enabled the decipherers to read the Semitic inscriptions first in the Assyrian and then in the more difficult and archaic Babylonian writing. But the great difficulties of the task could hardly have been surmounted if Assyrian and Babylonian Semitic had not been practically little more than dialects of classical Arabic and Hebrew. When the Semitic inscriptions had once been read, the bilingual inscriptions in Semitic and Sumerian made the discovery of the latter language and its cognates a matter of certainty. So also with modern texts. Even if the strange language is presented to us in Roman transcription and in a connected text of some length, we can do nothing in the way of decipher- ing it till we either know what known languages to compare it with, or have a translation, or, at any rate, know what the text is about. If we know, for instance, that it is a translation of the Gospels or of the Prayer-book, the decipherment is only a question of time. Help afforded by Comparative Philology Of all helps, that of comparative philology is the most uncertain and the most liable to mislead unless severely con- trolled by a critical and scientific habit of mind. One page of translation is worth any number of comparisons with other languages and conjectural etymologies. All that a comparison of a certain word with some other word in a known language can do is to give hints which may lead to the discovery of its true meaning. First there is the difficulty of proving that the similarity is anything but accidental. If the two languages are only distantly connected, the trained philologist distrusts any great similarity. It takes very little theoretical divergence to make languages mutually unintelligible. A speaker of High German who does not know Platt-Deutsch can hardly understand a word of Dutch, nor can a Dutchman understand Frisian, nor a Frisian an Englishman, although Frisian is more closely related to English than to any other Germanic language. Even two dialects of the same language may be mutually unintelligible. Philological dilettantes who have learnt to 260 THE PRACTICAL STUDY OF LANGUAGES pick out similarities between cognate languages and to ignore the differences, often rush into the most extravagant statements about the similarities between languages. Thus they look into a Dutch book, and boast that with nothing but a know- ledge of German and English to help them, they can read the language at sight, not considering that a great many of the words they recognize by their forms have quite different mean- ings from what they have in German and English, and that some of the resemblances may be accidental and misleading — that, for instance, of has nothing to do with English of but has the meaning ' or ; ' os, plural ossen, does not mean ' horse,' but ' ox,' and so on. When Leibnitz first noticed the agreements in vocabulary between Persian and German which are the result of both being Aryan languages, he was so carried away by his dis- covery that he ventured on the astounding assertion that the similarity between the two languages was so great that an educated German could understand whole strophes of Persian poetry. The simple answer to this is, that if a North German cannot understand more than a word here and there of a South German dialect, it is not likely that he should understand more of a language which is not only non-Germanic, but belongs to the most remote subdivision of the Aryan family. There are certainly some very remarkable resemblances between English or German on the one hand, and modern Persian on the other, some of which are due to real affinity, but these few similarities are not enough to counterbalance the divergences in the rest of the Aryan portion of the vocabulary, together with the fact that about half the vocabulary is Arabic. The comparative philologist, of course, ignores the latter element, but to the practical linguist a modern Persian word of Arabic origin is just as much a Persian word as one of Aryan origin ; and the decipherer has to approach his problems from the practical point of view. The decipherer will then always distrust great similarity. The degree of average similarity that he expects will depend on the closeness of affinity between the two languages. Thus, if he is comparing a Germanic with a Slavonic language, he will expect on the whole greater divergence between them than between the Germanic language and any other Germanic language. So also if he is comparing the Germanic languages as a whole with the Finnic languages, he will expect the general divergence ORIGINAL INVESTIGATION 261 between the two families — the Germanic and the Finnic — to be greater than that between the most distant members of the Aryan family or the most distant members of the Finnic family. If he meets in Finnish such words as hmingas, ' king,' which are almost identical in form and meaning with words in the Germanic languages, while the corresponding cognate words in such closely allied Aryan languages as the Slavonic are mostly so different that their affinity with the Germanic words requires elaborate proof, he at once assumes that most of such agreements are the result of borrowing on one side or the other — in this case on the Finnish side. It need hardly be said that borrowed words are among the most valuable aids to the decipherer, just as they are to the practical linguist. Although their identity may be occasionally disguised by changes of form and meaning, their evidence is generally of a much more direct character than that of cognate words, and if the borrowing is of recent date, there is no reason why there should be any divergence at all in meaning or any but a slight divergence in form. Etymological affinity, on the other hand, proves nothing — it only raises expectations which may be fulfilled or not. Thus, if in learning German I come on the word fusz, I guess, on the analogy of the identity of the form and meaning of German hand with that of the English hand, that it is cognate with the "English, foot and has the same meaning. But if I were to go on to assume on similar grounds that German kopf is not only cognate with English cup, but has the same meaning, I should find myself as much hampered in my attempts at decipherment by this correct etymological identification as by any incorrect one. Of course, when I once find out that kopf means ' head,' my knowledge of the changes of meanings in other languages would enable me to prove that this change of meaning is quite a natural one. I might, indeed, have been prepared at the outset for this change of meaning ; but experience shows that such flights of a prioj'i imagination may lead to results which are as baseless as they are plausible. Hence we find that while the older school of cuneiform investigators made great use of comparisons with Hebrew and the other Semitic languages, and thereby obtained many valuable hints towards decipherment, the later scholars have got their best and most reliable results by the comparison of parallel uses of words in the texts themselves, so that the 262 THE PRACTICAL STUDY OF LANGUAGES testimony of comparative philology has now only a secondary weight : the evidence of the texts settles the etymology, the etymology does not settle the meaning of the text. Decipherment a Practical Problem We see, then, that decipherment is more a practical than a theoretically scientific problem, and that even when it calls in the help of comparative philology, its methods still are more allied to those of the practical linguist than of the comparative philologist. Gabelentz has some interesting remarks on this subject (Gab. 76) ; in speaking of the method of dealing with languages in which we have only texts in a known alphabet to start with, without any help beyond a translation or general knowledge of the contents, he says — ' It might seem that in dealing with texts in a foreign language we should have to rely from the beginning entirely on learned investigations. But this is not the case : here, as elsewhere, a purely naive attitude (naives verhalten) is best at first. Let the student read a few pages, aloud if possible, in order to help the memory by the ear as well. In doing so, he need not trouble himself particularly about the correctness of the pronunciation, taking care only to distinguish whatever is written differently [p. 34]. He will soon notice that words, perhaps also word-stems and word-forms, repeat themselves, and perhaps occasionally discovers their meaning. In this way the instinct of analysis gradually asserts itself, the text talks to us, and we learn to understand it better page by page. Any one who learns by heart easily, and only wants to get a practical command of the language, will probably attain his object quicker in this way than if he conscientiously set to work to make collections (collectanien) like a philological investigator.' Gabelentz then goes on to recommend this method of learning languages as a useful training for every linguist and philologist ; he says — 1 If he chooses some remote, but not too difficult language, such as one of the Bantu family, a Malay, Polynesian, Mela- nesian, or a Ural-altaic language, he can be certain of success, even if he has no previous acquaintance with any of the languages belonging to the family. He will at the same time receive a quantity of entirely new scientific ideas, his sagacity ORIGINAL INVESTIGATION 263 will be exercised, and after a short spell of certainly rather dry labour, he will have the enjoyment of self-gained know- ledge which increases hour by hour. And skill in this method of investigation is immensely increased by practice. We all know that a talent for languages does not always imply scientific capacity. But he who is trained in this school may expect that his scientific judgment will increase in the same proportion as his linguistic knowledge, for in such work theoretical speculation comes of itself.' Work of this kind might well form part of the training in an ideal seminary of comparative philology in its wider sense in which it is equivalent to the German ' allgemeine sprachwissen- schaft.' All Text-reading Implies Originality The methods of studying languages we have just been con- sidering all imply a considerable amount of originality. But there is no absolute line of demarcation in this respect between the most difficult decipherment of an unknown inscription and the reading of the easiest text in the most familiar language. All free text-reading implies a certain amount of independent thought. Even in reading our own language we may at any moment come upon unknown words and obscure or ambiguous constructions without being able to get help from any dictionary or grammar. Text-editing : Original Research In deciphering a new text in a known language, as when a philologist copies and edits a hitherto unpublished manuscript text, the originality reduces itself to explaining such meanings of words and phrases and grammatical constructions as cannot be found in the existing dictionaries and grammars and other helps. If the text is a literal translation whose original is accessible, the originality may be reduced almost to a vanishing- point. A good deal of what is grandiloquently called ' original research ' is purely mechanical work, requiring almost less originality than the routine of a bank clerk., The ' researcher ' looks through a catalogue of manuscripts, and finds, say, a Treatise on the Seven Deadly Sins in the Kentish dialect of the fifteenth century, or a fragment of a translation of the French romance of The Adventures of Sir Arthur and tfie Green Lady, 264 THE PRACTICAL STUDY OF LANGUAGES which his professor assures him has never been published. Our student copies it by the help of a facsimile of the handwriting of the manuscript, translates it with the help of the Latin or French original, and then publishes the text with a glossary and introduction, two-thirds of which perhaps is written by his professor. On the strength of this original research he is then himself made a professor — a professor who never in the whole course of a long and laborious career shows the slightest glim- mering of originality. The evils of the German system which requires, if not the reality, at least the semblance of originality from every candi- date for a doctor's degree are manifold and self-evident. Any measure that would stop this over-production would be welcome. Some kind of tax on useless and superfluous literature is much needed. Investigations of Unwritten Speech The investigation of unwritten forms of speech requires much higher qualities than publishing a manuscript text. Phonetic training, quickness of observation, presence of mind, are here essential. Equally important is the power of recording one's observa- tions in phonetic writing. I remember a young foreign philo- logist showing me his notes of the pronunciation of some Turkish dialect, written in a phonetic notation he had hastily improvised on a French basis, the result being that he had to confess that he was quite unable to remember what sounds his symbols stood for. He was no phonetician, and made no pretence of a knowledge of phonetics. But there are many who profess to be phoneticians, and are almost equally helpless when they have to face the difficulties of having to write down a dialect for the first time. Such work requires not only accuracy, but quickness. For really good dialectal work, a phonetic shorthand will in future be regarded as indispensable. But however well-equipped with theoretical and practical knowledge the investigator may be, and however much practice he may have had, the phonetic analysis of a new dialect and the writing down of its sounds must always be a slow and difficult process at first. Let the beginner be under no delusions ORIGINAL INVESTIGATION 265 on this point : no one can write a language down straight off under such circumstances. All who profess to be able to do so deceive themselves. Even familiar and easy distinctions such as that of close and open vowels are often confusing in a new language : what is, relatively speaking, the close sound may be a little opener than in other languages, so that the observer perhaps writes it down roughly and tentatively as (t) ; but when he finds a still opener sound which is, however, not so open as the (a?) in English man, he sees that he has made a phonetic shifting or ' verschiebung,' so that what he at first wrote (e) must be written (e), and what he at first wrote (ae) must be written (e). Again, many beginners, in listening to the speech of a native, will often confidently assert that the pronunciation of a certain sound varies from word to word ; and when the natives assure them that they are mistaken, they only regard it as a proof of their own superior acuteness of hearing; they do not know that it is one of the surest signs of unfamiliarity with the sound in question : the impressions of the ear sway, as it were, from side to side in the vain attempt to identify the unfamiliar sound with some familiar one ; and when the sound is thoroughly mastered, this fluctuation ceases. This leads to another consideration. We cannot write down a strange language or dialect till we know it practically. We can only write down what is familiar to us. Till the elements of the language are familiar to us we can only take rough provisional notes. The only way to describe the formation of the strange sounds is to describe the organic actions by which we imitate them. Experimental phonetics may some day alter this, but at present it is practically the way in which we have to analyze strange sounds. It is the same with the idioms and constructions of the language : at first we can only write down what we have assimilated ourselves. Help from Natives. — An evident objection to this method is that unless our knowledge is perfect, our description of the new form of speech cannot be perfect. As the only people who have this perfect knowledge are the speakers themselves, a second method suggests itself, that of training an intelligent native to do the work under the supervision of the foreign investigator. 266 THE PRACTICAL STUDY OF LANGUAGES The investigator should first try to find some speaker of the dialect who is intelligent enough to be able to learn to use a phonetic notation, and sufficiently interested in the subject to take the trouble of writing down the tales, songs, riddles, or other traditional pieces he remembers, or making dialogues and other texts on topics suggested by the investigator. A young village schoolmaster will often prove the best help in this way. At first the native will write his texts in a mixture of dialectal and literary forms. The investigator, by comparing parallel passages and noting apparent inconsistencies, will soon find these out, and by degrees will be able to train the native to write phonetically. The first prejudices once overcome, the latter will soon take a pleasure and pride in being as purely colloquial as possible, and will perhaps feel something of the charm of exploring a country which is at once strange and familiar. This last method always implies practically a combination of it with the preceding one. Compared with this combined method of learning the language oneself, and training a native to write it down, all other methods are mere makeshifts. Questioning. — The one that is most frequently employed is, perhaps, that of questioning. If the investigator has to deal with the language of illiterate savages, of whose language he is ignorant, he has, of course, to begin with gesture. Here he will meet with the difficulties already discussed under ' visualizing ' (p. 209). If he points to his mouth, he may get the word for 4 mouth,' but he is just as likely to get ' bite, eat, teeth, lip.' If he is certain that it is ' teeth ' and not lips, he is still in doubt whether the plural or the singular is meant. Prolonged questioning is apt to tire the intelligent European schoolmaster, still more the flighty barbarian. Many a traveller who has attracted crowds of dusky natives into his tent by dis- playing his stores, has soon found himself alone when he begins to ask them questions about their language and religious views. If the natives have a keener sense of humour than of the obligations of veracity, they may revenge themselves by giving misleading answers. A missionary who had been in the South Seas was once observed to burst into repeated fits of laughter while reading what professed to be a list of numerals in a Polynesian language with which he was familiar. He explained that, knowing that the speakers of this language could not count ORIGINAL INVESTIGATION 267 beyond twenty, he was at first surprised and interested to find the numerals given as high as ten thousand, but found that all the higher numbers were simply strings of words of the most ludicrous and improper associations. The European peasant of the north is apt to turn sulky, if questioned beyond a certain point. A story is told of a Swedish dialectologist who, wishing to know what was the preterite of the verb die in a certain Swedish dialect, asked one of the natives whether he said / died or / dew. The only answer he could get was, ' When we are dead, we don't speak.' Still more hopeless is the method of investigating a dialect by means of correspondence, although it must be confessed that Ellis by this means accumulated a vast mass of information about the English dialects which would otherwise have been lost. But such information cannot under ordinary circumstances be reliable. It must be remembered also that the information given in the fifth part of Ellis's Early English Pronunciation, although originally obtained by correspondence, was in many cases checked by personal interviews with his authorities. The only possible way of dealing satisfactorily with a whole body of dialects such as the English, is to have a school of phonetics at some real University which will attract speakers of the different dialects, each of whom will pursue his investiga- tions under his teacher on a uniform plan. Collecting Materials Every one has his own methods of literary work, and every investigator of unwritten dialects has his own way of collecting his material. But this does not prove that there are not certain broad principles of general application. A method which makes elaborate and carefully digested collections useless to others cannot be a sound one. Handwriting. — The first requisite is to write clearly. Business men and those who keep secretaries can afford the luxury and distinction of writing an illegible hand. Scientific investigators, who seldom do good work after starting a secre- tary, and who often have to read nothing but their own hand- writing for months together, must learn to write. 268 THE PRACTICAL STUDY OF LANGUAGES Their writing should be as small as is convenient for their sight, and as compact as possible, the letters being nearly up- right and close together, without more separation between the words than is necessary. They should get into the habit of always leaving a margin — which may be marked by folding the paper — of writing in paragraphs and on one side of the paper only, so as to be able to make additions on the back — unless they are writing in a book, and then only when saving of space is essential. If they can write evenly without ruled lines, their handwriting will be all the better. Such a system of writing as my Current Shorthand will be found to add greatly to the speed and ease of literary work, as it can both be written quicker and packed into a smaller space than ordinary longhand, even if written without any special contractions. Notes that are to be kept for any length of time should never be written in lead pencil, but in ink, or, if that is inconvenient, with a solid ink pencil. If it is desired to make the writing with the latter specially distinct, the paper should be slightly moistened, and the pencil carefully sharpened. For rapid and continuous work several of these pencils should be kept in readiness. Form. — As regards the form of the paper on which the collections are written, they may take the form of slips, of loose sheets, or of bound books. For the first preliminary rough work of collecting isolated words or quotations for a dictionary or grammar, slips are the most convenient. The proper average size of a small slip is about four and a half inches by two, or less for a fine writer ; a short-sighted writer will find three and a quarter inches by one and a quarter large enough ; he will be able to write the Lord's Prayer many times over in Current Shorthand on one side of such a slip. If the slips are required only for temporary use, and a great number is required, they may be of thin paper. If the collection of slips is intended to be more permanent and for reference, they should be of stiffer paper, the best for continuous use being evenly cut thin cards. They can be kept in boxes like cigar-boxes, or, better, in shallow trays with divisions. ORIGINAL INVESTIGATION 269 For purposes of further division and classification, stiff cards with ' tabs ' or square pieces in the middle projecting above the general level may be inserted at intervals among the slips, so that the eye can see the letters or other index-marks on the tabs at a glance. If accuracy of reference is essential, the slips should be kept in their boxes in the exact order in which they were first written, which of course will be the order in which the words or quotations they contain occurred in the text which was excerpted ; and then each reference should be verified before the slips are sorted into an alphabetic or any other order which interferes with their original order. The heading — the word, the grammatical categoiy, etc. — should be written at the top left-hand corner of the slip, the reference — name of text, number of page or paragraph and line — at the right-hand lower corner, the quotation itself between. The more mechanically the work of slip-writing is done, the better. It is a good rule never to put two entries on one slip. There is no occasion to economize in paper. The great advantage of slips is that they require no previous calculation of space ; a collection of slips is perfectly elastic. Their other advantage is the speed with which they can be written, as they involve no reference, no turning over of pages. But the sorting of them is a most wearisome and irritating, to some minds intolerable, drudgery ; nor can the sorting be left to others unless it is purely mechanical. When sorted, they are easily misplaced or lost. Altogether, they are difficult to handle and refer to, each of the hundred slips which make up perhaps only a page of print being practically on a page by itself. For less mechanical or more comprehensive collections loose sheets, nearly square, so as to allow room for a margin, are very convenient, five inches by four being the medium size for a fine shorthand-writer. The deductions from the materials collected on slips may be summarized on such sheets. If there is any doubt about the sequence of ideas, or any probability of additional matter coming in, each sheet may be restricted to what would be a paragraph in a book. Indeed, this will be found the most convenient way of preparing a book for press, as the paragraphs can be rearranged at pleasure till the book is complete, and can then be transcribed into longhand. 270 THE PRACTICAL STUDY OF LANGUAGES The sheets can be kept in boxes or drawers, or can be kept together with indiarubber bands passing over sheets of stiff cardboard at both ends of the pile of sheets. Larger sheets can be kept in portfolios, subdivisions being made by keeping each lesser group inside a folded sheet of thicker paper. Bound books, according to Gabelentz, are suited only for travellers, who cannot afford to risk the loss of slips or sheets. The advantages of books are not only that the leaves do not get lost or mislaid, but also the perfect facility of reference, which, again, is increased by the greater amount of matter that can be included in one page. But they postulate that we know beforehand how much space will be required, or else involve a great waste of space, which, of course, diminishes the ease of reference. Then, too, the order of the quotations or the categories under which they are put cannot be altered without causing confusion and waste of space. The bigger the book, the greater the waste of time in making entries. Some, how- ever, still prefer, even in doing a glossary, to enter the head- words in a book — or paste the headings from a printed dictionary along the margin — with ample blank spaces, to going through the drudgery of sorting slips which can never be made easy to refer to. Some make their slips more easy of reference by pasting them into a book. For this purpose any strongly bound printed book will do. It is only necessary to cut out every sixth leaf or so in order to allow for the thickness of the slips. The work of pasting down the slips is irksome, but it may be shortened by simply running two lines of paste down the page, and then putting down the slips without having to apply the paste to each separately. But the result is always untidy and wanting in compactness. A practical shorthand-writer would hardly hesitate between this method and the slower but far more satisfactory one of simply copying his slips into a book in shorthand. It is never worth while to interleave printed books such as dictionaries unless we are certain of having to make numerous entries into them. Otherwise it is better to write on the margin, and, if necessary, insert a sheet occasionally, either loose or pasted. Many have noticed that interleaved books in ORIGINAL INVESTIGATION 271 libraries often have a few entries on the first few sheets, which are then left completely blank. For many special purposes it is convenient to make one's own manuscript books by fastening together sheets of folded paper either by stitching, or by simply making holes and putting a piece of thin string through, by which the paper is less liable to be torn ; the back may be strengthened by pasting on a narrow strip of cloth. Principles of Collecting Collect everything at first. — In working at a text, for the first few pages one should, as a general rule, collect every- thing. If the collections are for a dictionary, every word should be noted, with, of course, such exceptions as common sense dictates, such as the conjunction mid. After a time, the text should be read over again cursorily, and the method of collection for the future carefully considered and methodically planned out. Collect mechanically at first. — The work should also be done mechanically at first. No a priori generalizations should be allowed to interfere with the first aim of the investigator, that is, gathering together enough material to form the basis of sound deductions. When a certain definite amount of material has been collected, or when the most important texts have been gone through, the investigator may well pause and review his gains from a higher and freer point of view, lest prolonged drudgery and impracti- cable ideas of fullness or exhaustivity of collection lead him into working on a scale which will make it impossible for him to complete his enterprise within reasonable time, until at last he sinks into a monomaniac machine incapable of any higher work. It should never be forgotten that it is much easier to heap up material than to utilize it. It is easy for the dictionary-compiler to brag of the tons of material, the millions of slips that have been collected for him, but when it comes to sorting these slips according to the meanings of the words, and weighing the evidence of each, he often wishes he had started with a ton 272 THE PRACTICAL STUDY OF LANGUAGES or two less. Let us, then, take warning by Browning's gram- marian — ' That low man seeks a little thing to do, Sees it and does it : This high man, with a great thing to pursue, Dies ere he knows it. J Let us be low men — at least at first. Classification. — We have arrived, then, at the stage of a logical classification of a moderate and reasonable amount of material which, without pretending to be exhaustive, may be relied on for giving a sufficient number of examples to illustrate the general principles we are investigating. The ideal for a dictionary would be to have enough examples of the rarer words and rather more than enough of the very common words. If our classification is partly arbitrary — as in an alphabetic dictionary — then the classification of the materials will be partly mechanical. The logical classification will then consist mainly in arranging the quotations under the meanings expressed by the head-words. If the quotations are intended to form the foundation of a syntax, then the classification will probably be purely logical from the beginning. In all logical classifications the investigator should proceed cautiously at first, so as not to start with prejudices or hastily formed generalizations. The material should at first be sorted only into definitely marked off main groups which interfere with one another as little as possible. At first, the language should be explained as far as possible out of itself. There should be no comparison with cognate languages till this has been done. Otherwise the investigator runs the risk of importing into the language tendencies which do not really belong to it, and so missing, perhaps, some explanation which would otherwise be obvious to him. CHAPTER XXI MIND-TRAINING; CLASSICAL AND MODERN LANGUAGES We will now examine the grounds of the widely spread assumption that ancient languages — that is, Greek and Latin — are more perfect and more highly developed in structure than modern ones, and that consequently their study is a better training for the mind. Now that the question of subordinating ancient to modern languages — even to the exclusion of Greek from the general scheme of education — is being earnestly discussed, and is winning more and more adherents, the statement of the innate superiority of ancient languages is incessantly repeated by the advocates of Greek and Latin. Some of them, indeed, go so far as to hint that the study of modern languages is not only useless as an instrument of intellectual training, but is even positively injurious, as tending to create a superficial turn of mind. A thorough examination of the reasons of these assumptions, and of the real distinctions between ancient and modern languages, will lead us to the very opposite conclusions in every respect. We shall see that the arguments of the supporters of ancient languages are based on an erroneous idea of the nature of language, which has been further supported by the one- sidedly historical method which has hitherto prevailed in philo- logical investigation. The assertion of the higher development of ancient languages may be reduced to the more precise one that ancient languages have a more copious grammar than modern ones. The com- parison of the two extremes, Greek and English, has often prompted the remark that English has 'no grammar.' And, indeed, if we compare the numerous cases and declensions in 273 T 274 THE PRACTICAL STUDY OF LANGUAGES Greek and Latin with the English genitive and plural -s, and the interminable complexities of the Greek verb in all its voices, moods, tenses, numbers, and persons with the few endings which make up the inflections of the English verb, we are apt to accept the statement as a matter of course. But even the most superficial observer cannot but be struck by the evident contradiction implied in the generally admitted fact that English is one of the most expressive and concise languages that have ever existed, and that ideas can be ex- pressed in it with as much facility and accuracy as in Greek and Latin. Thus the idea, or rather ideas, expressed by the Latin amat can be expressed with the same brevity by the English he loves, which, like amat, consists of only two syllables, and with greater precision, for the English form denotes not only the person, but also the gender. Again, the Latin amabit has not only a corresponding he will love in English, but also a shorter dissyllable form he'll love. The supposed superiority of the Latin over the English forms consists in the former being able to express their meaning with one, while the latter require two or more words. But the distinction is really a purely graphic one. The actual spoken language admits no division into words, its lowest unit being the sentence, within whose limits there is no division or pause of any kind. Historically considered, moreover, the Latin amat is really two words, as much as its English representative, the final t being originally a pronoun signifying ' he,' ' she,' or ' it/ and it is only reasons of practical convenience that prevent us from writing am at or ama t as two, and heloves as one word. It may, of course, be urged that the ^ of loves is historically the same ending as the Latin -t, and consequently that he loves is really three words, but this does not apply to / love. Besides, these historical views lie outside of the practical ques- tion, and the s of loves is in English practically nothing but a fossilized archaism. The really essential difference between amat and he loves is that in the former the pronominal element is expressed by a suffix, in the latter by a prefix. The end of a word being more hurried over and slurred than the beginning, it naturally follows that in those languages which express general relations, such as the persons of verbs or the cases of nouns, by means of suffixed words or syllables, these elements will be much more liable to various phonetic changes and shortenings, which will MIND-TRAINING 275 vary greatly according to the sounds which precede them, as when the s of the nominative in Greek is preserved in dnax but dropped in daimon, lengthening the preceding vowel. Hence have arisen the varied and complex inflections of the ancient languages. English, on the other hand, prefers to denote general relations by prefixes, which are not liable to be modified, or incorporated into the root-word. The practical result in writing is that most English modifiers can be written as separate words, and re- garded as such, even when their meanings are quite as abstract as those of the inflections of the old languages. The preposition of, for instance, in of a man is quite as abstract as the is in Latin hominis, and, like it, is absolutely unmeaning when separated from its noun, although the accident of its being written as an independent word blinds us to the fact. The real functional distinction between the two is that while of is always perfectly distinct and recognizable in all its combinations, the Latin -is is both ambiguous in itself, being used to express a number of other cases as well, and is only one of a large number of means of denoting the same case, as may be seen by comparing the endings of hominis, mensae, dominl, do?nus, which have not a single sound in common. What must strike an impartial observer is the waste of power involved in employing so many forms, most of which have at the same time a number of other vague and contradictory meanings, to express an idea which in a modern language like English can be expressed by a single unambiguous word. By the side of their useless complexity of inflection, ancient languages show a remarkable vagueness of thought, both in grammar and vocabulary. Compare the extreme vagueness of the meanings of the half-a-dozen cases in Greek and Latin with the precision of the numerous English prepositions which corre- spond to them. The same want of differentiation is shown in the vocabulary as well. Even in those cases in which an ancient language has a considerable number of words to express a given group of abstract ideas, it often happens that each single word runs through and exhausts the whole series of ideas, so that nothing at all is gained by the fullness of the vocabulary. In such a case a modern language utilizes each word to express a definite idea. The traditional character of ancient languages often leads them into downright absurdities, such as the use of grammatical 276 THE PRACTICAL STUDY OF LANGUAGES genders, which, strange to say, are still retained more or less in all the Aryan languages of Europe with the exception of English. In fact, there can be no question that the highly inflected Aryan languages are in many respects far more irrational than those which stand on a lower scale of development, and that such a language as English owes its superiority as a means of expression in a great degree to developments which have many analogies to those of non-Aryan and even savage languages. The statement that English has little or no grammar simply means that the grammatical structure of English is so regular and transparent that a very moderate amount of analysis is enough to enable the learner to find his way through it. But regularity and symmetry are by no means inconsistent with complexity, and, as a matter of fact, English is one of the most complex languages that has ever existed. If grammar be defined as the expression of general relations, whether that expression be effected by suffixes or prefixes, by inflections or prepositions and auxiliaries, then English has the most copious grammar of any in the world. The difference between the complexity of an ancient and of a modern language is that that of the former is to a great extent unmeaning and useless, while that of the latter implies a correspondingly full and minute analysis of the ideas expressed by it. Of course, it must not be forgotten that all languages are extremely defective, if compared with an ideal standard, and that consequently the difference between them can only be one of degree ; but if those languages are the most rational which express ideas most clearly, simply, and regularly, there can be no question of the superiority of the modern languages in ration- ality, and consequently as a means of intellectual training also. If, on the other hand, the mechanical acquisition of irrational distinctions of form, and familiarizing oneself with vague and loose expressions of thought, is the best training for the mind, then there can be no question of the superiority of ancient languages. It cannot be denied that the defects of ancient languages are compensated by many real advantages, although these advan- tages have nothing to do with intellectual training. One superiority of most ancient languages is the simplicity, clear- ness, and sonorousness of their phonetic structure. The very vagueness of their meanings, again, although in itself a serious defect, brings with it great freshness, freedom, and picturesque- MIND-TRAINING 277 ness of metaphor, which, together with their fullness of sound, eminently fit them for poetry and oratory, and for aesthetic purposes generally. The assimilations, contractions, and other phonetic changes of modern languages not only diminish their harmony and fullness of sound, but also make them indistinct by diminishing the individuality of the older distinctions, or even, as is so often the case in English, by confounding originally distinct words under one common form. But even these defects do not affect the value of modern languages as instruments of intellectual training. The defects of the inflectional languages are most clearly shown in those cases in which an inflectional system has been retained after it has been made superfluous by the development of prefixes, auxiliaries, etc., and a fixed word-order. Modern German is a marked example of such a transition language. Although it has adopted the fixed word-order of a modern language, and makes an extensive use of auxiliaries, prepositions, etc., it still retains many of the old inflections, together with the three grammatical genders. The result is that while in some cases the old inflections still express an independent meaning, as when the distinction between the English in and into is expressed by in with the dative and accusative respectively, in others they are superfluous, the idea being already fully expressed by means analogous to those employed by such a language as English. Such distinctions, for instance, as those between guter and gutc in ein guter mann and der gutc mann are really quite useless, being fully expressed by the ein and the der. Again, in the old languages the distinctions of grammatical gender, together with the laws of concord, allowed the separation of adjectives from the nouns to which they belong, which, although of little use for purposes of expression, yet added greatly to the harmony and picturesqueness of the language by causing variety, and especially by preventing the repetition of the same heavy endings close together ; but in German, with its fixed word-order, they are almost useless, and, indeed, the agreement between adjective and noun is abandoned when the adjective stands predicatively — curiously enough, in the only position in which it would be of any use — although it is superfluously retained in the attribu- tive position of the adjective. Such a language as Swedish, on the other hand, with its simplicity, its clearness and harmony of phonetic structure, and 278 THE PRACTICAL STUDY OF LANGUAGES its few, but clear, simple, symmetrical inflections, really combines, to a great extent, the advantages of ancient and modern languages. German has also the antique clearness and sonor- ousness of sound, in which it is infinitely superior to English and French, which certainly carry off the palm for simplicity and precision, English, again, being unquestionably foremost in many-sidedness and power. In comparing the ancient languages among themselves, it must be borne in mind that Greek, owing to the greater intellectual activity of those who spoke it, and the consequent necessity of precision and many-sidedness of expression, is in many respects more modern in structure than Latin. The excessive use of the article and the heaping of particles in Greek are characteristic contrasts with the Latin usages. It must further be remembered that archaism of structure by no means implies that the language is a dead one. We have in Russian an example of a living language of great literary, social, and political importance, which vies in inflectional complexity with Latin and Greek ; and it is a question whether a study of it would not prove as good a practical training in the use of an inflectional language as that of the classical languages. Of course, if modern languages are to be studied at all, they must be studied properly. The superficial study of modern languages certainly tends to deteriorate the mind, just as every other superficial study does, but it is equally possible to study dead languages superficially, as also in a narrow and unscientific spirit. APPENDIX BIBLIOGRAPHY The following notices do not aim at any fullness, but are merely intended as a first guide to those who wish to pursue the subject further. Many of the books here given themselves contain more or less full bibliographies. General Works Paul, H. : Principien der Sprachgeschichte. Halle, 2 i886. [Purely theoretical and abstract.] Von der Gabelentz, G. : die Sprachwissenschaft. Leipzig, 1891. [Deals with the practical as well as the theoretical study of languages.] Storm, J. : Englische Philologie. Leipzig, 2 i8g2. [Purely practical ; gives incidentally much information about general phonetics, together with many details about other languages besides English.] Periodicals Maitre phon^tique, la ; organe de 1' Association phone'tique des professeurs de langues vivantes, edited by P. Passy. Address : Fonetik, Bourg-la-Reine, France. Phonetische Studien, Zeitschrift fiir wissenschaftliche und praktische Phonetik, edited by W. Vietor, I -VI. Marburg, 1888-93. Continued in the form of an Appendix to die Neueren Sprachen, Marburg, 1894, foil. [This publication, like the previous one, is not confined to phonetics, but contains articles of a wider linguistic interest, both theoretical and practical.] Modern Language Quarterly of Language and Literature, edited by H. Frank Heath, London. Phonetics Sweet, H. : Primer of Phonetics. Oxford, 1890. Sketch of Phonetics in Primer of Historical English Gratnmar. Oxford, 1893. Grandgent, C. H. : German and English Sounds. Boston, U.S.A., 1893. [Very good diagrams of the tongue-positions.] Sievers, E. : Grundziige der Phonetik. Leipzig, 4 i893. [Deals 280 APPENDIX ■with phonetics mainly from the point of view of comparative philology.] Vietor, W. : Elemente der Phonetik. Leipzig, 3 i 893. Jespersen, O. : Fonetik. Copenhagen, 1897. [Not yet com- pleted.] Passy, P. : Les sons du Francais. Paris, 4 i897. Michaelis et Passy : Dictionnaire Phone'tique de la langue Franchise, 1897. Hempl, G. : German orthography and phonology. Boston, U.S. A., 1897. General Method Widgery, W. H. : The Teaching of Languages in Schools. London, 1888. Vietor, W. : Der Sprachunterricht muss umkehren ! Heilbronn, 2 i886. Frarike, F. : Die praktische Spracherlernung. Heilbronn, 1884. Klinghardt, H. : Ein Jahr Erfahrungen mit der Neuen Methode. Marburg, 1888. Drei Weitere Jahre Erfahrungen mit der Neuen Methode. Marburg, 1892. Storm, J. : Om en forbedret Undervisning i Levende Sprog (Universitets- og Skoleannaler II.). Christiania. Brekke, K. : Gouins Methode (Univ.- og Skoleannaler, 1894). Christiania. Western, A. : Om Undervisningen i Nyere Sprog. Christiania, 1885. Text-books Sweet, H. : Elementarbuch des gesprochenen Englisch. Oxford, 3 i89i. Primer of Spoken English. Oxford, 2 i895. Passy, P. : Ele'mans d'Anglais parld. Paris, 2 i887. Brekke, K. : Laerebog i Engelsk for folkskolen. Christiania, 1892. Beyer und Passy: Elementarbuch des gesprochenen Franzo- sisch. Cothen, 1893. Passy, J. : Chrestomathie Franchise. Paris, 1897. Pranke, F. : Phrases de tous des jours. Leipzig, 6 i893. Storm, F. : Dialogues Franc, ais. Copenhagen, 1887. Sweet, H. : Anglo-Saxon Primer. Oxford, 7 i893. First Steps in Anglo-Saxon. Oxford, 1897. 10ale Stu&fes in EnoUsb Edited by ALBERT S. COOK, Professor in Yale University I. The Foreign Sources of Modern English Ver- sification. By Charlton M. Lewis, Ph.D. 8vo. 104 pp. Paper, 50 cents. II. jElfric. A New Study of his Life and Writings. By Caroline Louisa White, Ph.D. 8vo. 220 pp. Paper, $1.50. III. The Life of St. Cecilia. From MS. Ashmole43 and MS. Cotton Tiberius E. VII, with Introduction, Variants, and Glossary. By Berth A Ellen Love well, Ph.D. 8vo. 139 pp. Paper, $1.00. IV. Dryden's Dramatic Theory and Practice. By Margaret Sherwood, Ph.D. i6mo. no pp. 50 cents. V. Studies in Jonson's Comedy. By Elisabeth Woodbridge, Ph.D. 8vo. 102 pp. Paper, 50 cents. VI. A Glossary of the West-Saxon Gospels. Latin-West Saxon and West Saxon-Latin. By Mattie Austice Harris, Ph.D. 8vo. in pp. Paper, $1.50. IN PREPARATION. A Guide to the Middle English Metrical Romances. Dealing with English and Germanic Legends, and with the Cycles of Char- lemagne and of Arthur. By Anna Hunt Billings, Ph.D. The Legend of St. Andrew. An Old English Epic, translated into Blank Verse, with an Introduction. By Robert Kilburn Root. Classical Mythology in Milton. By Charles Grosvenor Osgood, Jr. HENRY HOLT & CO., NEW YORK s CQ $ Recent French Books. ?[ Bound in cloth, unless otherwise stated. Prices net, gy postage 8$ additional. 3 BALZAC'S CURE DE TOURS, AND OTHER STORIES. Le Cure* dc •5*J Tours, Les Proscrits, El Verdugo, Z. Marcas, and La Messe de TAthde. -♦»* Edited by Prof. F. M. Warren of Adelbert. xiv + 267 pp. i6mo. 75c. .^ BEVIER'S FRENCH GRAMMAR. With exercises by Dr. Thomas Logik. Qln use by classes in Harvard, Johns Hopkins, the Universities of Michi- gan, Iowa, Chicago, etc. i2mo. 341 pp. $1.00. CANFIELD'S FRENCH LYRICS. The best lyrics since the Renaissance, ^J with particular attention to the XlXth century. Over 50 poets are rep- ^ resented by some 220 poems. Selected and edited by Prof. A. G. Can- g< field of the University of Kansas, xxii + 382 pp. i6mo. $1.00. $3 COPPEE'S ON REND L'ARGENT. A novel giving vivid pictures of the ^ Paris of to-day. Edited by T. B. Bronson of the Lawrenceville (N.J.) school. With eight illustrations. 184 pp. Narrow i6mo. 60c. COPPEE AND DE MAUPASSANT: TALES. Coppee-Lc Morceaudu ^ Pain, Deux Pities, Un Vieux de la Vieille, Les Vices du Capitaine, C/j Scenario, La Robe Blanche, La Remplacant, Un Enterrement Dra- Jv matique, etc. De Maupassant — La Peur, La Main, Garcon, Un Bock, «£ En Voyage, Apparition, Les Idees da Colonel, etc. Edited by Prof. © A. G. Cameron of Princeton. With portraits. xlviii+ 188 pp r 75c. 5 GASC'S FRENCH AND ENGLISH DICTIONARIES. Qj Library Dictionary. 956 pp. (3 columns). 8vo. $4.00. Student's Dictionary. 1186 pp. 12010. $1.50 retail, postpaid. Pocket Dictionary. 647 pp. (3 columns). 12 mo. $1.00. HUGO : SCENES DE VOYAGE. De Paris a Aix-la-Chapelle. Selections itomLeRhm. Edited by T. B. Bronson. xvi-f-277PP. i6mo. 85c. LOTI : SELECTIONS. The stories Chagrin d'un vieux forcat and Viande de boucherie, and representative episodes, from Le Mariage de Loti, Le Roman d'un Spahi, Mon frere Yves, Pecheur d'Island, Madame Chrysantheme, Japoneries d'automne, and Au Maroc. Edited by Prof. A.G.Cameron. Authorized Edition, vii + 185 pp. i6mo. 70c. MERAS & STERN'S FIRST LESSONS IN FRENCH. By Prof. B. nj Meras and Director S. M. Stern of Stern's School of Languages, -1^ New York. 321 pp. 121110. $1.00. *«« MERAS & STERN'S GRAMMAIRE FRANfAISE. The method of this S» book is quite unusual. 312 pp. i2mo. $1 25. 5 MUSSET'S HISTOiRE D'UN MERLE BLANC. Edited by Misses © Williams and Cointat of Smith, iv + sopp. i6mo. Bds. 30c. ^3 OHNET'S LA F1LLE DU DEPUTE. By the author of Le Maitre des SJ Forges. A story giving excellent practice in French conversation. •*»» Edited by Geo. A. D. Beck, Licencie es Lettres. 204 pp. i6mo. 50c. Q SAINTE-BEUVE: SEVEN OF THE CAUSER1ES DU LUNDI. Qu'est- ce qu'un Classique; Grande Epoque de la Prose; Pens^es de Pascal; .^ La Fontaine; Les Mdmoires de Saint-Simon; Madame de Maintenon; "CJ etc. Edited by Prof. G. M. Harper of Princeton, li -{- 176 pp. 75c. ££ SAINT-PIERRE : PAUL ET VIRGINIE. Edited by Prof. L. Oscab CJ Kuhns of Wesleyan (Conn.). x-f-i6opp. i6mo. 50c. £> TOEPFFERS LA BIBLIOTHEQUE DE MON ONCLE. A humorous tale of Geneva. Edited by R. L. Taylor of Yale, xx -f- 201 pp. 50c. VOCABULARY EDITIONS. Erckmann-Chatrian, Le Conscrit de 1813, and Mme. Therese (each 55c); Fcuillet's Le Roman d'un Jeune try Homme Pauvre (55c.) ; MeVimeVs Colomba (50c.) ; Sand's La Mare au "i . Diable (40c.) ; Labiche et Delacour's Poudre aux Yeux (Bds., 30c.). - I HENRY HOLT & CO. 3. , %2S£5BS W "> '99 O 4 Recent German Texts. Bound in cloth unless otherwise indicated. SUDBRMANN'S FRAU SORGE. A Romance. With introduction and notes by Prof. Gustavb Grukner of Yale. With portrait, xx + 268 pp. »g i6mo. 80c, net. '% HAUPTMANN'S DIE VERSUNKENE GLOCKE. Ein deutsches Mar- ft, chen-Drama. With introduction and notes by Thos. S. Baker, Asso- TjJ ciate in the Johns Hopkins University. xviii + 2ospp. i6mo. 80c, net. 5 " Probably the most remarkable play since Goethe's ' Faust.' "—Pro/. H. C. G. Brandt 0/ Hamilton College. rg SCHILLER'S HISTORY OF THE THIRTY YEARS' WAR. Selections 6 portraying the careers and characters of Gustavus Adolphus and Wal- "{g lenstein. Edited, with introduction, notes, and map, by Prof. Arthur *■ H. Palmer of Yale, xxxvii + 202 pp. i6mo. 80c., net. © SCHILLER'S WILHELM TELL. Edited, with introduction, notes, and •O a full vocabulary, by Prof. A. H. Palmer of Yale. Illustrated, lxxvi *H + 300 pp. i6mo. 80c, net. (Without vocabulary 60c.) ^LESSINO'S MINNA VON BARNHELM. Edited, with introduction and notes, by A. B. Nichols of Harvard. With 12 illustrations by Chodowiecki. XXX + 163PP. i6mo. 60c, net. Ss^LESSING'S MINNA VON BARNHELM. With an introduction and V, notes by Prof. William D. Whitney of Yale. New vocabulary CQ edition. 191 pp. i6mo. 60c, net. C! GOETHE'S DICHTUNQ UND WAHRHEIT. Selections from Books I- Q IX. Edited, with introduction and notes, by Prof. H. C. G. Vnjn Jage- ^< mann of Harvard. xiv + 373PP- i6mo. $1.12, net. S\ GOETHE'S GOTZ VON BERLICHINGEN. Edited, with introduction •i*, and notes, by Prof. F. P. Goodrich of Williams, xli -f- 170 pp. i6mo. j?N 70c, net. OOETHE'S HERMANN UND DOROTHEA. Edited, with introduction .^ and notes, by Prof. Calvin Thomas of Columbia. New edition with ^ vocabulary, xxii + 150 pp. Bds. 40c, net. 2 JUNQ-STILLING'S LEBENSQESCHICHTE. With introduction and C vocabulary by Sigmon M. Stern, xxvi -f- 285 pp. 12020. $1.20, net. C A book giving intimate glimpses of Goethe. 0> STERN'S AUS DEUTSCHEN MEISTERWERKEN (Nibelungrn, Parci. [J val, Gudrun, Tristan und Isolde). Erzahlt von Sigmon M. Stern. With a full vocabulary. xxvii-l-225 pp. i6mo. $1.20, net. 1-. A simple version of these great German legends. •^ BAUMBACH'S SOMMERMARCHEN. Eight stories. With introduc ^ tion, notes, and vocabulary by Dr. Edward Meyer of Western Re. Q) serve University, vl + 142 pp. i6mo. Bds. 35c, net. Q ROSEQGER'S DIE SCHRIFTEN DES WALDSCHULMEISTERS. An ^* authorized abridgment. Edited, with introduction and notes, by Prof. 2j Lawrence Fossler of the University of Nebraska. With two poems J^ by Baumbach and frontispiece. xii-f-is8pp. Bds. 40c, net. g; REOENTS* GERMAN AND FRENCH POEMS FOR MEMORIZING. Prescribed by the Examinations Department of the University of the State of New York. Revised 1900. 98 pp. i2tno. Bds. 20c., net. Descriptive List 0/ the Publishers'' Modern Language Books free. HFNRY HOI T <% TO 29W S st23dSt.,HewYork nClNrvi nULl OL K^.KJ . 378 Wabash. Ave., Chicago VI, IOOO WHITNEY'S (W. D.) COMPENDIOUS GERMAN AND ENGLISH DICTIONARY. Cheaper Edition. 538 + 362 pp. l2mo. %\.yo, retail. The standard American German dictionary. It defines some 60,000 words in such a manner as to give com- mand over a vastly larger number by (1) Placing the original meaning first and exhibiting the development of secondary meanings; (2) Marking historical correspondences between German and English words (a particularly helpful device); and (3) Indicating etymologies, particularly the derivation of Ger- man words from one another. GASC'S (F. E. A.) STUDENTS* DICTIONARY OF THE FRENCH AND ENGLISH LANGUAGES. Cheaper Edi- tion. 1 1 86 pp. i2mo. $1.50, retail. It is full, modern, idiomatic, and accurate. It defines thou- sands of French words found in no other French and English dictionary. Space for these has been made by the judicious omission of words alike in both languages, by giving only one of two or three English renderings almost exactly synonymous, and by various devices in arrangement. All the shades of sig- nificance a word assumes in various connections (those of recent origin in particular) are carefully canvassed, and its force in each one fully explained. Words that have many idiomatic applica- tions get an unusual proportion of space. GASC'S LIBRARY FRENCH AND ENGLISH DICTION- ARY. 956 pp. (3 columns.) 8vo. $4.00, net. An expansion of the author's Students' Dictionary. It com- pares favorably, both for completeness and convenience, with any French and English dictionary extant. GASC'S POCKET DICTIONARY OF THE FRENCH AND ENGLISH LANGUAGES. 647 pp. i8mo. $1.00, net. This edition differs from the Students' Dictionary chiefly in being in much smaller type. The number of words treated, and of idiomatic uses noted, is somewhat smaller. BELLOWS' (J.) FRENCH AND ENGLISH DICTIONARY 605 pp. i2mo. $1.00, net. Of the Pocket Edition (roan, $2.55 net, and morocco, $3.10 net) Dr. Oliver Wendell Holmes says : " I was fairly astonished that such an atom of a book could be such a cyclopaedia of phrases." UCX1DV H(~\\ T Rt CC\ 29 West 23d St., New York. nClNtVI HULl Oi V-W., 378 Wabash Ave., Chicago- in -99 "fdo not know where else, within the limits, to find so delightful a selection of noble poems."— Prof. Thomas R. Price of Columbia. PANCOAST'S STANDARD ENGLISH POEMS From Spenser to Tennyson. Selected and edited by HENRY S. Pancoast, author of An Introduction to English Litera- ture, etc. 749 pp. i6mo. $1.50, net. Some 250 complete poems, besides selections from such long poems as "The Faerie Queene," "Childe Harold's Pilgrim- age," etc. There are 19 pages of Ballads, 33 of Spenser, 22 of Elizabethan Songs and Lyrics, 16 of Elizabethan Sonnets, 51 of Seven- teenth Century Songs, 51 of verse from Dryden to Thomson, 277 of verse from Thomson to Tennyson, and 100 of Victorian versa, 164 of Notes (chiefly biographical and appreciative), and an index of titles. New York Tribune : " We believe it will be received cordially by all lovers of poetry, wheiher elementary students or not. Basing his selections on the individual excellence and historic importance of the poems, the editor has not allowed his fidelity to the latter test to overrule his taste, and there is very little matter in the book which is historically significant alone. First and last, this is an anthology of the best poetry." Prof. Henry A. Beers of Yale, author of "English Romanticism in the Eighteenth Century," etc.: " The collection seems to me in gen- eral made with excellent judgment, and the notes are sensible, help- ful, and not too weitldufig." Prof. Albert S. Cook of Yale : "A thoroughly good selection, and the notes are judicious, so far as I have examined." Prof. William Hand Browne of Johns Hopkins: "The scope is amply wide, and the selections as judicious as was possible under the limitations. The notes, judging from a hasty glance, seem full and clear." Prof. Charles W. Kent of the University of Virginia : " Contains nearly all the poems I would wish in such a volume and very few that I would readily dispense with." Prof. James M. Dixon of Washington University: "It is just such a handy volume as can be made, by a sympathetic teacher, a companion to the scholar for life." HcNKY ri(JLl & CO., 3$WabashAve.f Chicago i 1900 THOMAS'S PRACTICAL GERMAN GRAMMAR. For Schools and Colleges. By Calvin Thomas, Professor la Columbia College. 422 pp. iamo. $1.12 net. Part I. is a drill-book for beginners, with abundant exercises which are colloquial, connected, and interesting. Part II. is a reference-grammar for students of literature. " This work is in my opinion an easy first in the line of guides for elementary instruction in German. Its terse clearness of statement, its wise grouping of items in the order of their rela- tive importance for the learner, and its refreshingly sane and idiomatic exercises all tally with the prefatory statement of the author's aim." — S. W. Cutting, Professor in the University of Chicago. " Part I. is a remarkably successful attempt to give all that and only that which the beginner must have in order to read the language with profit. The sentences for drill, also, are far more interesting than in the common run of elementary manuals, — surely no ' mechanical Ollendorfian preparations ' here. I have used it constantly since it first appeared — it stands well that final test of every text-book, class-room use. In short, I find it just what it professes to be, a Practical German Grammar." — Walter R. Newton, Phillips Academy, Andover, Mass. " I have tested it for two years in high-school work as well as in private work. It is equally well adapted to the needs of pupils of average high-school age as to the needs of students of mature age. With it more rapid results can be obtained than with any other grammar." — Carl F. Lange, High School, Manistee. "The Thomas's German Grammar I consider the best work- ing grammar I have had in my department, and it is a delight to use it." — Marie F. Kapp, Professor in Smith College. " For the study of modern German the book has no equal." — W. H. Perkinson, Professor in the University of Virginia. " The best working German Grammar in the English lan- guage. ... It is based on sound pedagogical principles ; it pre- sents living German ; it will create and cultivate in the student a proper Sprachgefiihl." — Charles B. Wilson, Professor in the Uni- versity of Iowa, in The School Review. A circular of the above book containing /urther comments and an extended list 0/ representative Schools and Colleges using it mailed free on application. HENRY HOLT & CO. 379 Wabash Ave. ^Chicago UC SOUTHERN REGIONAL LIBRARY FACILITY AA 001 351 783 CENTRAL UNIVERSITY LIBRARY University of California, San Diego DATE DUE DEC 1 tQ7c APR 2 1 1983 a 39 UCSD Libr. mm mm 'iwm