| f Prin Key Mrneegs = SEP 291891 = pam ietatay es T owe the following note to the kindness of M. Stanislas Julien :— ‘La maniére dont le mot ego s’exprime dans les différentes conditions est fort curieuse. ‘Un homme ordinaire dira par humilité: yu, le stupide; ti, le frere cadet ; siao-tt, le petit; now-thsai, lesclave. ‘L’empereur dit: siao-tseu, parvus filius; siao-eul, parvus infans. Un prince dit: kowa-jin, exigue virtutis homo; kou, Vorphelin; pou-kou, non bonus. ‘Un magistrat supérieur (un préfet) dit: pen-fou, ma ville du premier ordre. Un magistrat inférieur (sous-préfet): hia-kowan, le magistrat infime. Pen-hiew, ma sous-préfecture ; pi-tchi, la basse charge. ‘Un Tartare parlant & ’empereur : nou-thsai, Vesclave. ‘Un religieux bouddhiste: pin-seng, le pauvre religieux ; siao-seny, le petit religieux. ‘Une femme parlant & son mari: nou-now, esclave-esclave; nou-kia, esclave-maison ; tsien-tsie, la méprisable concubine. ‘Un domestique: do, le domestique. ‘Un fils parlant & son pere: pou-siao, pas semblable (c’est-a-dire, dégénéré). ‘Un vieillard dit: lao-fon, le vieil homme; lao-han, le vieux Chinois ; lao-tchue (vieux-stupide) ; lao-hieow, vieux-pourri. ‘Un religieux : tao-sse ; pin-tao, le pauvre tao ; siao-tao, le petit tao. ‘Une religieuse bouddhiste: pin-ni, la pauvre religieuse; siao-nié, la petite religieuse. ‘Une vieille femme: lao-chin, le vieux corps; /ao-niang, la vieille dame, etc,’ D 2 36 CHAPTER TI. polite ; but we cannot expect that different nations should hit on exactly the same polite speeches, though they may agree in the common sense of orammar. The past tense is indicated in Chinese also by particles meaning ‘already’ or ‘ formerly, but we do not find among them the Annamitic da. The same applies to the future. The system is throughout the same, but the materials are different. Shall we say, therefore, that these languages cannot be proved to be related, because they do not display the same criteria of relationship as French and English, Latin and Greek, Celtic and Sanskrit ? I tried on a former occasion! to explain some of the causes which in nomadic dialects produce a much more rapid shedding of words than in literary lan- guages, and I have since received ample evidence to confirm the views which I then expressed. I was not aware at that time how clearly Schelling, in his Kinlettung in die Philosophie der Mythologie (vol. i. p. 114), had perceived the necessity of change and dialectic variety in all nomadic languages. Speaking of the languages of Southern America, as described by Azara in his voyages (vol. ii.), he says :— Among that population the Guarani is the only language which is understood over a large area, and even this point requires more careful examination. Apart from this, as Azara remarks (and he has not only passed through these countries, but lived in them for years), the language changes from clan to clan, from cottage to cottage, so that often the members of one and the same family only understand each other. Nay, the very power of speech seems sometimes to become extinct. ' Letter to Chevalier Bunsen, on the Turanian Languages. NEW MATERIALS FOR THE SCIENCE OF LANGUAGE. 37 Their voice is never strong or sonorous; they only speak low, never loud, even when they are being killed. They hardly move their lips while speaking, and there is no expression in their face to invite attention. They evidently dislike speaking, and if they see a friend a hundred steps off, they rather run after him than call him. Language, therefore, here hovers on the very edge, and one step more would entirely put an end to it. My excellent friend, Bishop Patteson of Melanesia, of whom it is difficult to say whether we should admire him most as a missionary, or as a scholar, or as a bold mariner,’ met in every small island with a new language, which none but a scholar could trace back to the Melanesian type. ‘What an indication, he writes, ‘of the jealousy and suspicion of their lives, the extraordinary multiplicity of these languages affords! In each generation, for aught I know, they diverge more and more; provincialisms and local words, &c., perpetually introduce new causes for perplexity.’ The northern peninsula of Celebes, of which the chief town is Menado, is inhabited by a race quite distinct from the other people of the island. They are Malays, but have something of the Tatar and something of the Kuropean in their physiognomy. They agree best with some of the inhabitants of the Philippines; and Mr. Wallace, a most accurate ob- server, supposes that they have come from those islands originally by way of the Siaou and Sanguir islands, which are inhabited by an allied race. Their languages show this affinity, differing very much from all those of the rest of Celebes. A proof, how- 1 He was murdered in 1871, a true hero and martyr. 38 CHAPTER TI. ever, of the antiquity of this immigration, and of the low state of civilisation in which they must have existed for long periods, is to be seen in the variety of their languages. In a district about one hundred iniles long by thirty miles wide, not less than ten distinct languages are spoken. Some of them are confined to single villages, others to groups of three or four; and though of course they have a certain family resemblance, they are yet so distinct as to be mutually unintelligible.! Te pi. There are many causes at work to produce dialectic change. In addition to those which I have explained already, I shall mention but one more which has acted very powerfully on the Polynesian languages. It may seem at first sight very insignificant, but as one of the multifarious influences which are at work in nomadic dialects, constantly changing their aspect and multiplying their number, it ought not to be over- looked. It will serve at all events to convince even the most incredulous, how little we know of all the secret springs of language if we confine our researches to a comparison of the classical tongues of India, Greece, Italy, and Germany. The Tahitians,? besides their metaphorical ex- pressions, have another and a more singular mode of displaying their reverence towards their king, by a custom which they term Je pi. They cease to em- _* A.R, Wallace, ‘Man in the Malay Archipelago, Transactions of the Ethnological Society, iii. p. 206. * Hale, l. c. p. 288. NEW MATERIALS FOR THE SCIENCE OF LANGUAGE. og ploy, in the common language, those words which form a part or the whole of the sovereign’s name, or that of one of his near relatives, and invent new terms to supply their place. As all names in Poly- nesian are significant, and as a chief usually has several, it will be seen that this custom must produce a very considerable change in the language. It is true that this change is only temporary, as at the death of the king or chief the new word is dropt, and the original term resumed. But it is hardly to be supposed that after one or two generations the old words should still be remembered and be reinstated. The literary activity of the missionaries also will in future serve to check the influence of this ancient national custom, because words, if once incorporated in the translation of the Bible, in grammars and dictionaries, will acquire a strong persistence and defy the ceremonial loyalty of the natives. Vancouver observes (Voyage, vol. i. p. 185) that at the accession of Otu, which took place between the visit of Cook and his own, no less than forty or fifty of the most common words, which occur in conversation, had been entirely changed. It is not necessary that all the simple words which go to make up a compound name should be changed. The alteration of one is esteemed sufficient. Thus in Po-mare, signifying ‘the night (po) of coughing (mare), only the first word, po, has been dropped, mz being used in its place. So in Ai-mata (eye-eater), the name of another queen, the az (eat) has been altered to amu, and the mata (eye) retained. In Te-arii-na-vaha-roa (the chief with the large mouth), roa alone has been 40 CHAPTER I. changed to macro. It is the same as if, with the accession of Queen Victoria, either the word victory had been tabooed altogether, or only part of it, for instance to7v?, so as to make it high treason to speak during her reign of Tories, this word being always supplied by another; such, for instance, as Liberal- Conservative. The object was clearly to guard against the name of the sovereign being ever used, even by accident, in ordinary conversation, and this object was equally well attained by tabooing even one portion of his name only. But this alteration (as Mr. Hale continues) affects not only the words themselves, but syllables of similar sound in other words. Thus the name of one of the kings being Tu, not only was this word, which means ‘to stand,’ changed to tia, but in the word fetu, star, the last syllable, though having no connection except in sound, with the word tu, underwent the same alteration—star being now fetia; tui, to strike, became tiai; and tu pa pau, a corpse, tia pa pau. So ha, four, having been changed to maha, the word aha, split, has been altered to amaha, and murihd, the name of a month, to murimaha. When the word ai was changed to amu, maraai, the name of a certain wind (in Rarotongan, maranai), became maraamu., The mode of alteration, or the manner of forming new terms, seems to be arbitrary. In many cases, the substitutes are made by changing or dropping some letter or letters of the original word, as hopoi for hapai, to carry in the arms; ene for hono, to mend; au for tau, fit; hio for tio, to look ; ea for ara, path; vau for varu, eight; vea for vera, not, &e. In other cases, the word substituted is one which had before a meaning nearly related to that of the term disused—as tia, straight, upright, is used instead of tu, to stand; pae, part, division, instead of rima, five; piti, together, has replaced rua, two, &c. In some cases, the meaning or origin of the new word is unknown, and it may be a mere invention—as ofai for ohatu, stone ; pape, for vai, water; pohe for mate, dead, at i NEW MATERIALS FOR THE SCIENCE OF LANGUAGE. 41 &c. Some have been adopted from the neighbouring Pau- motuan, as rui, night, from ruko, dark; fene, six, from hene ; avae, moon, from kawake. It is evident that but for the rule by which the old terms are revived on the death of the person in whose name they entered, the language might, in a few centuries, have been completely changed, not, indeed, in its grammar, but in its vocabulary. When such liberties could be taken with language, we need not be surprised that one of the kings of the | Sandwich Islands conceived the idea of inventing an entirely new language. About the year 1800, as Chamisso tells us in his Travels,) King Tameia- meia (only another rendering of Kamehameha) in- vented a new language in honour of the birth of a son. The new words were not related to any roots of the current language, nor derived from them. Even the particles which take the place of grammatical forms and bind a sentence together, were similarly changed. The story goes that some of the influential chiefs who disapproved of this innovation, poisoned the child that had been the innocent cause of it, and that at his death the changes were suppressed which had been introduced at his birth. The old language returned, the new one was forgotten, not so much however that Marini, Chamisso’s authority in this matter, could not mention a few instances of novel words which survived, such as anna, man, for the old kanaka; karavu, woman, for the old waheini; amio, to go, for the old kokine; ja papa, dog, for the old irzo or lio.? 1 Chamisso, Werke, ii. 77. 2 Something of the same kind is mentioned by Dobrizhoffer with regard to the language of the Abipones; History of the Abipones, part ii, chap. 17. 42 CHAPTER f. Nor is this custom of Te pi, a kind of linguistic Tabu, confined to the Malayo-Polynesian dialects. A similar tendency exists in Chinese. Schlegel, in his Sinico-Aryaca,! p. 4, makes the following statement : En Chinois nous retrouvons le méme usage pour la langue Lay a Ale Say P écrite. Le caractere Ry tchou, p.e. désignant une espéce de toile grossiére, est en méme temps le nom particulier de lEmpereur Hien-fung. Depuis, on ne peut plus se servir de ce caractere pour désigner cette espece d’étoffe, mais on doit la at : Yi désigner par le caractére tronqué qe par respect pour le nom sacré du Souverain. Le caractére F{5 pang, un état, fut éliminé de la littérature chinoise pendant tout le temps que regnait la maison du fonda- teur de la dynastie de Han, puisque le nom de ce fondateur était All Fi Liu-pang. I fut remplacé par le caractere [zt] kwoh, qui signifiait primitivement, une principauté; mais, qui, par l'élimination temporaire du mot pang, a recu une signification plus large, tandis que le mot pang est descendu-de son rang supérieur et a pris l’acceptation qu’avait primitivement le caractere kwoh. (Notes and Queries on China and Japan, vol. ill. pp. 179-181.) A similar custom, according to Aymonier, prevails in Cambodja. ‘Si le nom du roi, he writes in his Dictionnaire Franeais-Cambodgien ( 1874), p. 4, ‘est emprunté & un mot du langage usuel, chose trés- commune au Cambodge, ce mot est souvent chargé. Ainsi depuis Ang Duong, le mot duong, qui désignait une petite piéce de monnaie, est remplacé par le mot dom, " Verhandlingen van het Bataviaasch Genootschap, Deel. xxxvi, Batavia, 1872. This subject has been very fully treated by the Rev. Hilderic Friend, ‘Euphemism and Tabu in China,’ Folk-lore Record, vol. iv. NEW MATERIALS FOR THE SCIENCE OF LANGUAGE. 43 Ukuhlonipa. It might, no doubt, be said that a custom such as Te pi is a mere accident, a fancy peculiar to a fanciful race, but far too unimportant to claim any considera- tion from the philosophical student of language. I confess that at first it appeared to myself in the same light, but my attention was lately drawn to the fact, that the same peculiarity, or at least something very like it, exists in the Kafir languages. ‘The Kafir women, as we are told by the Rev. J. W. Appleyard, in his excellent work on the Kafir language,' ‘ have many words peculiar to themselves. This arises from a national custom, called Ukuhlonipa, which forbids their pronouncing any word which may happen to contain a sound similar to one in the names of their nearest male relations. It is perfectly true that the words substituted are at first no more than family idioms, that they would be confined to the gossip of women, and not enter into the conversation of men. But the influence of women on the language of each 1 The Kafir Language, comprising a sketch of its history ; which in- cludes a general classification of South African dialects, ethnographical and geographical ; remarks upon its nature; anda grammar. By the Rey. J. W. Appleyard, Wesleyan missionary in British Kaffraria, King William’s Town : printed for the Wesleyan Missionary Society; sold by Godlonton and White, Graham’s Town, Cape of Good Hope, and by John Mason, 66 Paternoster Row, London. 1850. Appleyard’s remarks on Ukuhlonipa were pointed out to me by the Rev. F, W. Farrar, the author of an excellent work on the Origin of Language. See also Tylor, Early History of Mankind, p. 147, and the Rev. J. L. Dohne, Zulu-Kafir Dictionary, Cape Town, 1857, s.v. hlonipa, to be bashful, to keep at a distance through timidity, to shun approach, to avoid mentioning one’s name, to be respectful. On Ukuhlonipa in Tasmania, see Bonwick, Daily Life in Tasmania, p. 146, 4.4, CHAPTER I. generation is much greater than that of men. We very properly call our language in Germany our mother-tongue, Unsere Muttersprache, for it is from our mothers that we learn it, with all its peculiarities, faults, idioms, accents. Cicero, in his ‘Brutus’ (c. 58), said:—‘It makes a great difference whom we hear at home every day, and with whom we speak as boys, and how our fathers, our tutors, and our mothers speak. We read the letters of Cornelia, the mother of the Gracchi, and it is clear from them that her sons were brought up not in the lap, but, so to say, in the very breath and speech of their mother.’ And again (Rhet. iii. 12), when speaking of his — mother-in-law, Crassus said, ‘When I hear Lelia (for women keep old fashions more readily, because, as they do not hear the conversation of many people, they will always retain what they learned at first) ; but. when I hear her, it is as if I were listening to Plautus and Neevius.’ | But this is not all. Dante ascribed the first at- tempts at using the vulgar tongue in Italy for literary compositions to the silent influence of ladies who did not understand the Latin language. Now this vulgar Italian, before it became the literary language of Italy, held very much the same position there as the so-called Prakrit dialects in India ; and these Prakrit dialects first assumed a literary position in the San- skrit plays where female characters, both high and low, are introduced as speaking Prakrit, instead of the Sanskrit employed by kings, noblemen, and priests. Here, then, we see the language of women, or, if not of women exclusively, at all events of women NEW MATERIALS FOR THE SCIENCE OF LANGUAGE. 45 and domestic servants, gradually entering into the literary idiom, and in later times even supplanting it altogether; for it is from the Prakrit, and not from the literary Sanskrit, that the modern vernaculars of India branched off in course of time. Nor is the simultaneous existence of two such representatives of one and the same language as Sanskrit and Prakrit confined to India. On the contrary, it has been re- marked that several languages divide themselves from the first into two great branches; one showing a more manly, the other a more feminine character ; one richer in consonants, the other richer in vowels ; one more tenacious of the original grammatical ter- minations, the other more inclined to slur over these terminations, and to simplify grammar by the use of circumlocutions. Thus we have Greek in its two dialects, the AMolic and the Ionic, with their sub- divisions, the Doric and Attic. In German we find the High and the Low German; in Celtic, the Goidhelic and Cymric, as in India the Sanskrit and Prakrit ; and it is by no means an unlikely or merely fanciful explanation, that, as Grimm suggested in the case of High and Low German, so likewise in the other Aryan languages, the stern and strict dialects, the Sanskrit, the Molic, the Goidhelic, represent the idiom of the fathers and brothers, used at public assemblies ; while the soft and simpler dialects, the Prakrit, the Ionic, and the Cymric, sprang originally from the domestic idiom of mothers, sisters, and servants at home. But whether the influence of the language of women be admitted on this large scale or not, certain 1t 1s, that through a thousand smaller channels their idioms 46 | CHAPTER I. everywhere find admission into the domestic conver- sation of the whole family, and into the public speeches of their assemblies. The greater the ascendancy of the female element in society, the greater the influence of their language on the language of a family or a clan, a village, or a town. The cases, however, that are mentioned of women speaking a totally different language from the men, cannot be used in confirmation of this view. The Caribe women, for instance, in the Antille Islands,’ spoke a language different from that of their husbands, because the Caribes had killed the whole male population of the Arawakes and married their women ; and something similar seems to have taken place among some of the tribes of Greenland. Yet even these isolated cases show how, among savage races, In a primitive state of society, language may be influenced by what we should call purely accidental causes, and more particularly wherever the system of exogamous marriage is prevalent. But to return to the Kafir language, we find in it clear traces that what may have been originally a mere feminine peculiarity—the result, if you like, of the bashfulness of the Kafir ladies—extended its influence. For, in the same way as the women eschew words which contain a sound similar to the names of their nearest male relatives, the men also of certain Kafir tribes feel a prejudice against employing a word that is similar in sound to the name of one of their former chiefs. Thus, the Amambalu do not use ilanga, the general word for swn, because their first chief’s name was Ulanga, but employ dsota instead. For a similar * Hervas, Catalogo, i. p. 212. * Ibid. i. p. 369. NEW MATERIALS FOR THE SCIENCE OF LANGUAGE. 47 reason, the Amagqunukwebi substitute vmmela for isitshetshe, the general term for knife." Here, then, we may perceive two things: first, the influence which a mere whim, if it once becomes stereotyped, may exercise on the whole character of a language, for we must remember that as every woman had her own male relations, and every tribe its own ancestors, a large number of words must — constantly have been tabooed and supplanted in these African and Polynesian dialects ; secondly, the cu- rious coincidence that two great branches of speech, the Kafir and the Polynesian, should share in common what at first sight would seem a merely accidental idiosyncrasy, a thing that might have been thought of once, but never again. It is perfectly true that such principles as the Te pi and the Ukuhlonipa could never become powerful agents in the literary languages of civilised nations, and that we must not look for traces of their influence either in Sanskrit, Greek, or Latin, as known tous.* But it is for that very reason that the study of what I call Nomad languages, as distinguished from State languages, becomes so in- structive. We see in them what we can no longer expect to see even in the most ancient Sanskrit or Hebrew. We watch the childhood of language with all its childish freaks, and we learn at least this one lesson, that there often is more in real language than is dreamt of in our philosophy. One more testimony in support of these views. ' Appleyard, J. c. p. 70. 2 See Lorédan Larchey, Les Excentricités du Langage: Paris, 1865. 48 : CHAPTER T. Mr. H. W. Bates, in his interesting work, The Natural- ist on the Amazons, writes :— But language is not a sure guide in the filiation of Bra- zilian tribes, seven or eight languages being sometimes spoken on the same river within a distance of 200 or 300 miles, There are certain peculiarities in Indian habits which lead to a quick corruption of language and segregation of dialects. When Indians, men or women, are conversing amongst them- selves, they seem to take pleasure in inventing new modes of pronunciation, or in distorting words. It is amusing to notice how the whole party will laugh when the wit of the circle perpetrates a new slang term, and these new words are very often retained. I have noticed this during long voyages made with Indian crews. When such alterations occur amongst a family or horde, which often live many years without com- munication with the rest of their tribe, the local corruption of language becomes perpetuated. Single hordes belonging to the same tribe, and inhabiting the banks of the same river, thus become, in the course of many years’ isolation, unin- telligible to other hordes, as happens with the Collinas on the Jurtia. I think it, therefore, very probable that the dispo- sition to invent new words and new modes of pronunciation, added to the small population and habits of isolation of hordes and tribes, are the causes of the wonderful diversity of lan- guages in South America.—(Vol, i. pp. 329-30.) As I mostly borrow my materials for the illustra- tion of the general principles of the Science of Language from Greek and Latin, with its Romance offshoots ; from English, with its Continental kith and kin, and from the much-abused, though indispen- sable, Sanskrit, I thought it all the more necessary to guard against the misapprehension that the study of Sanskrit and its cognate dialects could supply us with all that is necessary for our purpose. It can do so as little as an exploration of the tertiary epoch could tell ae NEW MATERIALS FOR THE SCIENCE OF LANGUAGE. 49 us all about the stratification of the earth. But, nevertheless, it can tell us a great deal. By display- ing the minute laws that regulate the changes of each consonant, each vowel, each accent, it disciplines the student, and teaches him respect for every jot and tittle in any, even the most barbarous, dialect he may hereafter have to analyse. By helping us to an under-— standing of that language in which we think, and of others most near and dear to us, it makes us perceive the great importance which the Science of Language has for the Science of Thought. Nay, it shows that the two are inseparable, and that without a proper analysis of human language we shall never arrive at a true knowledge of the human mind. I quote from Leibniz: ‘I believe truly, he says, ‘that languages are the best mirror of the human mind, and that an exact analysis of the signification of words would make us better acquainted than anything else with the operations of the understanding.’ Lie E CHAPTER II. LANGUAGE AND REASON. ANGUAGE has two aspects under which it presents itself to the eye of the student. It has a body and a soul which, though they cannot be separated, can be distinguished and be subjected separately to scientific treatment. I shall treat therefore first, of the body or the out- side of language, its letters, syllables, and words, describing their origin, their formation, and the laws which determine their growth and decay. Here we shall have to deal with some of the most important principles of etymology. After that, I-shall try to investigate what may be called the soul or the inside of language, examining the first concepts that claimed utterance, their com- binations, their ramifications, their growth, their decay, and often their resuscitation. We shall have to deal then with some of the fundamental principles of mythology, both ancient and modern, and try to determine the sway, if any, which the old language exercises on the ever new language, or, as it is generally expressed, which language, as such, exercises over thought. LANGUAGE AND REASON. 51 I am fully aware that this division is liable to some grave objections. To treat of sound as in- dependent of meaning, of thought as independent of words, seem to defy one of the best established prin- ciples of the science of language. Where do we ever meet in reality, [mean in the world such as it is, with articulate sounds—sounds like those that form the body of language, existing by themselves, and inde- pendent of language? No human being utters arti- culate sounds without an object, a purpose, a mean- ing! The endless configurations of sound which are: collected in our dictionaries would have no existence at all, they would be the mere ghost of a language, unless they stood there as the embodiment of thought, as the realisation of ideas. Even the interjections which we use, the cries and screams which are the precursors, or, according to others, the elements, of articulate speech, never exist without meaning. Arti- culate sound is always an utterance, a bringing out of something that is within, a manifestation or revela- tion of something that wants to manifest and to reveal itself. It would be different if language had been in- vented by agreement; if certain wise kings, priests, and philosophers had put their heads together and decreed that certain conceptions should be labelled and ticketed with certain sounds. In that case we might speak of the sound as the outside, of the ideas as the inside of language. * Ait. Br, II.: ‘Manasa va ishité vag vadati, yam hy anya- mand vakam vadaty asurya vai si vag adevagushta, ‘The voice speaks as impelled by the mind; if one utters speech with a different mind or meaning, that is demoniacal speech, not loved by the gods.’ EK 2 52 CHAPTER II. Artificial Language. Why it is impossible to conceive of living human language as having originated in a conventional agree- ment, I have endeavoured to explain before. We should want language in order to arrive at a con- ventional agreement on language. But I should by no means wish to be understood as denying the possibility of framing some language in this artificial manner, after men have once learnt to speak and to reason. It is the fashion to laugh at the idea of an artificial, still more of a universal language. But if this problem were really so absurd, a man like Leibniz would hardly have taken so deep an interest in its solution. That such a language should ever come into practical use, or that the whole earth should in that manner ever be of one language and one speech again, is hard to conceive. But that the problem itself admits of a solution, and of a very perfect solution, cannot be doubted. The Universal Language of Leibniz. As there prevails much misconception on this sub- ject; I shall give a short account of what has been achieved in framing a truly philosophical and there- fore universal language. Leibniz, in a letter to Remond de Montmort, written two years before his death, expressed himself with the greatest confidence on the value of what he calls his Spécieuse générale, and we can hardly doubt that he had then acquired a perfectly clear insight into his ideal of a universal language.1 ‘If he succeeded, * Guhrauer, G. W. Freiherr von Leibnitz, 1846, vol. i. p. 328, LANGUAGE AND REASON. 53 he writes, ‘in stirring up distinguished men to culti- vate the calculus with infinitesimals, it was because he could give palpable proofs of its use; but he had spoken to the Marquis de L’Hopital and others, of his Spécieuse générale, without gaining from them more attention than if he had been telling them of a dream. He ought to be able, he adds, to support his theory by some palpable use; but for that purpose he would have to carry out a part of his Characteristics— no easy matter, particularly circumstanced as he then was, deprived of the conversation of men who would encourage and help him in this work.’ A few months before this letter, Leibniz spoke with perfect assurance of his favourite theory. He admitted the difficulty of inventing and arranging this philosophical language; but he maintained that, if once carried out, it could be acquired by others without a dictionary, and with comparative ease. He should be able to carry it out, he said, if he were younger and less occupied, or if young men of talent were by hisside. A few eminent men might complete the work in five years, and within two years they might bring out the systems of ethics and meta- physics in the form of an incontrovertible calculus. The Philosophical Language of Bishop Wilkins. Leibniz died before he could lay before the world the outlines of his philosophical language, and many even among his admirers have expressed their doubts whether he ever had a clear conception of the nature of such a language. It seems hardly compatible, however, with the character of Leibniz to suppose 54, CHAPTER II. that he should have spoken so confidently, that he should actually have placed this Spécieuse générale on a level with his differential calculus, if it had been a mere dream. It seems more likely that Leibniz was acquainted with a work which, in the second half of the seventeenth century, attracted much attention in England, ‘The Essay towards a Real Character and a Philosophical Language,! by Bishop Wilkins (London, 1668), and that he perceived at once that the scheme there traced out was capable of much greater perfection. This work had been published by the Royal Society, and the author’s name was so well known as one of its founders, that it could hardly have escaped the notice of the Hanoverian philoso- pher, who was in such frequent correspondence with members of that society.? Now, though it has been the fashion to sneer at Bishop Wilkins and his Universal Language, his work seems to me, as far as I can judge, to offer the best solution that has yet been offered of a problem which, if of no practical importance, is of great interest from a purely scientific point of view; and though it is impossible to give an intelligible account of the Bishop’s scheme without entering into particulars which cannot be but tedious, it will help us, I believe, * The work of Bishop Wilkins is analysed and criticised by Lord Monboddo, in the second volume of his Origin and Progress of Language, Kdinburgh, 1774. * This supposition has been confirmed by a passage in which Leibniz actually quotes Bishop Wilkins. See Benfey, Geschichte der Sprach- wissenschaft, p. 249; Trendelenburg, Uber Leibnizens Entwurf einer allgemeinen Characteristik, Berlin, 1856; Monatsberichte der Berliner Akademie, 1860, p. 375; and a note in the French translation of my Lectures by Harris and Perrot, p. 57. LANGUAGE AND REASON, 55 towards a better understanding of real language, if we can acquire a clear idea of what an artificial language would be, and how it would differ from living speech. The primary object of the Bishop was not to invent a new spoken language, though he arrives at that in the end, but to contrive a system of writing or repre- senting our thoughts that should be universally in- telligible. We have, for instance, our numerical figures, which are understood by people speaking different languages, and which, though differently pronounced in different parts of the world, convey everywhere the same idea. We have besides such signs as + plus, — minus, x to be multiplied, + to be divided, = equal, < greater, > smaller, © sun, © moon, @ earth, ¥ Jupiter, ) Saturn, ¢ Mars, ¢ Venus, &c., which are intelligible to mathema- ticians and astronomers all over the world. Now if to every thing and notion,—I quote from Bishop Wilkins (p. 21)—there were assigned a distinct mark, to- gether with some provision to express grammatical derivations and inflexions, this might suffice as to one great end of a real character, namely, the expression of our conceptions by marks, which shall signify things, and not words. And so, likewise, if several distinct words (sounds) were assigned to the names of such things, with certain invariable rules for all such grammatical derivations and inflexions, and such only as are natural and necessary, this would make a much more easy and convenient language than is yet in being. This suggestion, which, as we shall see, is not the one which Bishop Wilkins carried out, has lately been taken up by Don Sinibaldo de Mas, in his Idéographie.’ 1 Idéographie. Mémoire sur la possibilité et la facilité de former une écriture générale au moyen de laquelle tous les peuples puissent s’entendre mutuellement sans que les uns connaissent la langue des 56 CHAPTER II. He gives a list of 2,600 figures, all formed after the pattern of musical notes, and he assigns to each a certain meaning. According to the interval in which the head of such a note is placed, the same sign is to be taken as a noun, an adjective, a verb, or an ad- verb. Thus the same sign might be used to. express love, to love, loving, and lovingly, by simply moving its head on the lines and spaces from f to e, d, and c. Another system of signs is then added to express | gender, number, case, person, tense, mood, and other grammatical categories, and a system of hieroglyphics is thus formed, by which the author succeeds in rendering the first 150 verses of the Aneid. It is perfectly true, as the author remarks, that the diffi- culty of learning his 2,000 signs is nothing in com- parison with learning several languages ; it is perfectly true, also, that nothing can exceed the simplicity of his grammatical notation, which excludes by its very nature everything that is anomalous. The whole grammatical framework consists of thirty-nine signs, whereas, as Don Sinibaldo remarks, we have in French 310 different terminations for the simple tenses of the ten regular conjugations, 1,755 for the thirty-nine irregular conjugations, and 200 for the auxiliary verbs, a sum total of 2,265 terminations, which must be learnt by heart.) It ig perfectly true, again, that few persons would ever use more than 4,000 words, and that by having the same sign used throughout as noun, verb, adjective, and adverb, this autres; écrit par Don Sinibaldo de Mas, Envoyé extraordinaire et Ministre plénipotentiaire de S. M. ©, en Chine. Paris: B. Duprat, 1863. 1 Page 99, LANGUAGE AND REASON. 57 number might still be considerably reduced. There is, however, this fundamental difficulty, that the assign- ment of a certain sign to a certain idea is purely arbitrary in this system, a difficulty which, as we shall now proceed to show, Bishop Wilkins endea- voured to overcome in a very ingenious and truly philosophical way. If these marks or notes (he writes) could be so contrived as to have such a dependence upon, and relation to, one another, as might be suitable to the nature of the things and notions which they represented ; and so, likewise, if the names of things could be so ordered as to contain such a kind of affinity or opposition in their letters and sounds, as might be some way answerable to the nature of the things which they signified; this would yet be a farther advantage superadded, by which, besides the best way of helping the memory by natural method, the understanding likewise would be highly improved; and we should, by learning the character and the names of things, be instructed likewise in their natures, the knowledge of both of which ought to be conjoined." The Bishop, then, undertakes neither more nor less than a classification of all that is or can be known, and he makes this dictionary of notions the basis of a corresponding dictionary of signs, both written and spoken. All this is done with great circumspection, and if we consider that it was undertaken nearly two hundred years ago, and carried out by one man single- handed, we shall be inclined to judge leniently of what may now seem to us antiquated and imperfect in his catalogue raisonné of human knowledge. A careful consideration of his work will show us why this language, which was meant to be permanent, 1 Page 21. 58 CHAPTER II. unchangeable, and universal, would, on the contrary, by its very nature, be constantly shifting. As our knowledge advances, the classification of our notions is constantly remodelled ; nay, in a certain sense, all advancement of learning may be called a corrected classification of our notions. If a plant, classified ac- cording to the system of Linnzeus, or according to that of Bishop Wilkins, has its own peculiar place in their synopsis of knowledge, and its own peculiar sign in their summary of philosophical language, every change in the classification of plants would necessitate a change in the philosophical nomenclature. The whale, for instance, is classified by Bishop Wilkins as a fish, falling under the division of viviparous and oblong. Fishes, in general, are classed first as swbstwnces, then as animate, as sensitive, and lastly as sanguineous, and the sign attached to the whale, by Bishop Wilkins, expresses every one of those differences which mark its place in his system of knowledge. As soon, therefore, as we treat the whale no longer as a fish, but as a mammal, its place is completely shifted, and its sign or name, if retained, would mis- lead us quite as much as the names of rainbow, thunderbolt, sunset, and others, expressive of ancient ideas which we know to be erroneous. This would happen even in strictly scientific subjects. Chemistry, for instance, adopted acid as the tech- nical name of a class of bodies of which those first recognised in science were distinguished by sourness of taste. But as chemical knowledge advanced, it was discovered that there were compounds precisely analogous in essential character, which were not sour, LANGUAGE AND REASON. 59 and consequently acidity was but an accidental quality of some of these bodies, not a necessary or universal character of all. It was thought too late to change the name, and accordingly in all European languages the term acid, or its etymological equivalent, is now applied to rock-crystal, quartz, and flint. In like manner, from a similar misapplication of salt, in scientific use, chemists class the substance of — which junk-bottles, French mirrors, windows, and opera glasses are made, among the salts, while ana- lysts have declared that the essential character, not only of other so-called salts, but of common kitchen salt, the salt of salts, has been mistaken; that salt is not salt, and, accordingly, have excluded that sub- stance from the class of bodies upon which, as their truest representative, it had bestowed its name. The Bishop begins by dividing all things which may be the subjects of language into six classes or genera, which he again subdivides by their several differences. These six classes comprise :— A. TRANSCENDENTAL NOTIONS. B. SUBSTANCES. C. QUANTITIES. D. QUALITIES. Ki. ACTIONS. I’. RELATIONS. In B to F we easily recognise the principal pre- dicaments or categories of logic, the pigeon-holes in which the ancient philosophers thought they could stow away all the ideas that ever entered the human ' Marsh, History of the English Language, p. 211; Liebig, Che- mische Briefe, 4th edit. i. p. 96. 60 CHAPTER II. mind. Under A we meet with a number of more abstract conceptions, such as kind, cause, condition, &e. By subdividing these six classes, the Bishop arrives in the end at forty classes, which, according to him, comprehend everything that can be known or ima- gined, and therefore everything that can possibly claim expression in a language, whether natural or artificial. To begin with the beginning, we find that his transcendental notions refer either to things or to words. Referring to things, we have I. Transcenpentats Generat, such as the notions of kind, cause, differences, end, means, mode. Here, under kind, we should find such notions as being, thing, notion, name, substance, accident, &c. Under notions of cause we meet with author, tool, aim, stuff, &e. II. Transcenpenrat or Mrxep Retation, such as the notions of general quantity, continued quantity, discontinued quantity, quality, whole and part. Under general quantity the notions of greatness and littleness, excess and defect ; under continued quantity those of length, breadth, depth, &c., would find their places. IIT. Transcenpewtat Retations or Actions, such as the notions of simple actions (putting, taking), comparate action (joining, re- peating, &c.), business (preparing, designing, beginning), com- merce (delivering, paying, reckoning), event (gaining, keeping, refreshing), motion (going, leading, meeting). TV. Tur Trayscenventat Notions or Discourse, comprehending all that is commonly comprehended under grammar and logic : ideas such as noun, verb, particle, prose, verse, letter, syllogism, question, affirmative, negative, and many more. After these general notions, which constitute the first four classes, but before what we should call the categories, the Bishop admits two independent classes of transcendental notions, one for God, the other for the World, neither of which, as he says, can be treated as predicaments, because they are not capable of any subordinate species. LANGUAGE AND REASON. 61 V. The fifth class, therefore, consists entirely of the idea of (rop. : VI. The sixth class comprehends the Wort or universe, divided into spiritual and corporeal, and embracing such Beton as spirit, angel, soul, heaven, planet, earth, land, &c. After this we arrive at the five categories, subdivided into thirty-four subaltern genera, which, together with the six classes of transcendental notions, complete, in the end, his forty genera. The Bishop begins with substance, the first difference of which he makes to be inanimate, and distinguishes by the name of VII. Exemenz, as his seventh genus. Of this there are several differences, fire, air, water, earth, each comprehending a number of minor species. Next comes supstance anrmatz, divided into vegetative and sensitive. ‘The vegetative again he subdivides into imperfect, such as minerals, and perfect, such as plants. The imperfect vegetative he subdivides into VIII. Sronz, and IX. Merat. Stone he subdivides by six differences, which, as he tells us, is the usual number of differences that he finds under every genus; and under each of these differences he enume- rates several species, which seldom exceed the number of nine under any one. Having thus gone through the imperfect vegetative, he comes to the perfect, or plant, which he says is a tribe so numerous and various, that he confesses he found a great deal of trouble in dividing and arranging it. It is in fact a botanical clagsi- fication, not based on scientific distinctions like that adopted by Linnzeus, but on the more tangible differences in the out- ward form of plants. It is interesting, if for nothing else, at least for the rich native nomenclature of all kinds of ee shrubs, and trees, which it contains. The herb he defines to be a minute and tender plant, and he has arranged it according to its leaves, in which way con- sidered, it makes his 62 CHAPTER II. X. Class, Lear-nerss. Considered according to its flowers, it makes his XI. Class, or Frowsr-Herss. Considered according to its seed-vessels, it makes his XII. Class, or Srep-nerss. Kach of these classes is divided by a certain number of differences, and under each difference numerous species are enumerated and arranged. All other plants being woody, and being larger and firmer than the herb, are divided into XIII. Survss, and XIV. Trees. Having thus exhausted the vegetable kingdom, the Bishop proceeds to the animal or sensitive, as he calls it, this being the second member of his division of animate substance. This kingdom he divides into XV. Exsancurneovus. XVI., XVIL, XVIII. Sancurnzovs, namely Fisn, Birp, and Beast. Having thus considered the general nature of vegetables and animals, he proceeds to consider the parts of both, some of which are peculiar to particular plants and animals, and constitute his XIX. Genus, Pecuuiar Parts ; while others are general, and constitute his XX. Genus, Generar Parts. Having thus exhausted the category of substances, he goes through the remaining categories of quantity, quality, action, and relation, which, together with the preceding classes, are represented in the following table, the skeleton, in fact, of the whole body of human knowledge. LANGUAGE AND REASON. 63 General; namely, those universal notions, whether belonging more properly to : GENERAL. I. rnings called TRANSCENDENTAL | Retain MIXED. II. RELATION oF Action, III. Words ; Discourse. IV. Special ; denoting either Pe V. Creature ; namely, such things as were either created or concreated by God, not excluding several of those notions which are framed by the minds of men, considered either { Collectively ; Worup. VI. Distributively ; according to the several kinds of beings, whether such as do belong to Substance. Pea ine ; Exiement, VII. Animate; considered according to their several Species ; whether Vegetative ; Imperfect ; as Minerals } a See Herp, considered ( LEAF. X. according to { FLowrr. XI. Perfect; as Plant eco XIIL,- | Reowee XI. XI. Tree, XIV. EXSANGUINEOUS, XV, Sensitive | Fisu. XVI. Sanguineous { Brrp. XVII. ( Beast, XVIII, PECULIAR. XIX, Parts Vek XX, Accident, MaGnirupe, X XI. Quantity ; Space, XXII. Measure. XXITl. NATURAL PowER. XXIV. Hasit. XXV. Quality ; MANNERS. X XVI. SENSIBLE QUALITY. XXVII. Sickness. XXVIII. SPIRITUAL, X XIX. SG CoRPOREAL, XXX. * 3 Morton, XXXI. OPERATION. XXXII. CECONOMICAL, XX XIII. Private | Bossesstos XXXIV. PROVISIONS, XX XV. ip Bs . 7 sy) Relation ; whether more /Civir. XXXVI. | JUDICIAL. XX XVII. Public Miuitary. XX XVIII. NAVAL. XXXIX. ECCLESIASTICAL, XL, The Bishop is far from claiming any great merit for his survey of human knowledge, and he admits most fully its many defects. No single individual could have mastered such a subject, which would baffle even the united efforts of learned societies, Yet 64, CHAPTER II. such as it is, and with all its imperfections, increased by the destruction of great part of his manuscript in the fire of London, it may give us some idea of what the genius of a Leibniz would have put in its place, if he had ever matured the idea which was from his earliest youth stirring in his brain. Having completed, in forty chapters, his philoso- phical dictionary of knowledge, Bishop Wilkins pro- ceeds to compose a philosophical grammar, according to which these ideas are to be formed into complex propositions and discourses. He then proceeds, in the fourth part of his work, to the framing of the language, which is to represent all possible notions, according as. they have been previously arranged. He begins with the written language or Real Cha- racter, as he calls it, because it expresses things, and not sounds, as the common characters do. It is, therefore, to be intelligible to people who speak dif- ferent languages, and to be read without, as yet, being pronounced at all. It were to be wished, he says, that characters could be found bearing some resem- blance to the things expressed by them; also, that the sounds of a language should have some resemblance to their objects. This, however, being impossible, he begins by contriving arbitrary marks for his forty genera. The next thing to be done is to mark the differences under each genus, This is done by affix- ing little lines at the left end of the character, forming with the character angles of different kinds, that is, right, obtuse, or acute, above or below; each of these affixes, according to its position, denoting the first, second, third, and following difference under the LANGUAGE AND REASON. 65 genus, these differences being, as we saw, regularly numbered in his philosophical dictionary. The third and last thing to be done is to express the species under each difference. This is done by affixing the like marks to the other end of the cha- racter, denoting the species under each difference, as they are Prrnered 3 in the dictionary. In this manner all the several notions of things | which are the subject of language, can be represented by real characters. But besides a complete dic- tionary, a grammatical framework, too, is wanted before the problem of an artificial language can ‘be considered as solved. In natural languages the gram- matical articulation consists either in separate par- ticles or in modifications in the body of a word, to whatever cause such modifications may be ascribed. Bishop Wilkins supplies the former by marks denoting particles, these marks being circular figures, dots, and little crooked lines, or virgulee, disposed in a certain manner. The latter, the grammatical terminations, are expressed by hooks or loops, affixed to either end of the character above or below, from which we learn whether the thing intended is to be considered as a noun, or an adjective, or an adverb; whether it be taken in an active or passive sense, in the plural or singular number. In this manner, everything that can be expressed in ordinary grammars, the gender, number, and cases of nouns, the tenses and moods of verbs, pronouns, articles, prepositions, conjunc- tions, and interjections, are all rendered with a precision unsurpassed, nay unequalled, by any living language. | iat: . F 66 CHAPTER II. Having thus shaped all his materials, the Bishop proceeds to give the Lord’s Prayer and the Creed. written in what he calls his Real Character; and it must be confessed by every unprejudiced person that with some attention and practice these specimens are perfectly intelligible. Hitherto, however, we have only arrived at a written language. In order to translate this written into a spoken language, the Bishop has expressed his forty genera or classes by such sounds as ba, be, bi, da, de. di, ga, ge, gt, all compositions of vowels, with one or other of the best sounding consonants. The differences under each of these genera he expresses by adding to the syllable denoting the genus one of the following consonants, b, d, g, p, t, ¢, Z, 8, n, according to the order in which the differences were ranked before in the tables under each genus, b expressing the first difference, d the second, and so on. The species is then expressed by putting after the consonant which stands for the difference one of the seven. vowels, or, if more be wanted, the diphthongs. Thus we get the following radicals, corresponding to the general table of notions, as given above: if A General 5 i Ba Gi, ein Ae Relation Mixed . Ba III. ’ Relation of Action. Be AVE Discourse : Bi V. God ° ° ° Da VI. World . : : Da VII. Element : 2 De Vid Stone. , : Di IX. Metal . ; ; Do LANGUAGE AND REASON. The differences of the first genus would be ex- pressed by, Leaf Flower Seed-vessel Shrub Tree : Hxsanguineous Fish | Bird Beast Peculiar . General . | Magnitude Herbs } Space Measure. : Natural Power Habit Manners Quality, sensible Sickness Spiritual Corporeal Motion . Operation (Economical . Possessions Provisions Civil Judicial . Military . Naval | Ecclesiastical. Bab, bad, bag, bap, bat, bac, baz, bas, ban. The species of the first difference of the first genus would be expressed by. Ga Ga Ge Gi Go Baba, baba, babe, babi, babo, babs, baby, babyi, babys. Fe 2 68 CHAPTER II. According to the system of Bishop Wilkins, as ex- plained before, baba would mean being, baba thing, babe notion, babi name, babo substance, babe quantity, baby action, baby? relation. For instance, if De signify element, he says, then Deb must signify the first difference, which, according to my tables, is fire; and Deba will denote the first species, which is flame. Det will be the fifth difference under that genus, which is appearing meteor; Deta the first species, viz. rainbow; Deta the second, viz. halo. Thus if 7% signify the genus of Sensible Quality, then Tid must denote the second difference, which comprehends colours, and Zida must signify the second species under that difference, viz. redness, &c. The principal grammatical variations, laid down in the philosophical grammar, are likewise expressed by certain letters. If the word, he writes, is an adjec- tive, which, according to his method, is always’ de- rived from a substantive, the derivation is made by the change of the radical consonant into another consonant, or by adding a vowel to it. Thus, if Da signifies God, dua must signify divine ; if De signifies element, then due must signify elementary; if Do signifies stone, then duo must signify stony. In like manner voices and numbers and such-like accidents of words are formed, particles receive their phonetic representatives ; and in the end, all his materials being shaped, a complete grammatical translation of the Lord’s Prayer is given by the Bishop in his own newly-invented philosophical language. I hardly know whether the account here given of LANGUAGE AND REASON, 69 the artificial language invented by Bishop Wilkins will be intelligible, for, in spite of the length to which it has run, many points had to be omitted which would have placed the ingenious conceptions of its author in a much brighter light. My object was chiefly to show that to people acquainted with a real language, the invention of an artificial language is by no means an impossibility, nay, that such an artificial language might be much more perfect, more regular, more easy to learn, than any of the spoken tongues of man. The number of radicals in the Bishop’s language amounts to not quite 3,000, and these, by a judicious contrivance, are sufficient to express every possible idea. Thus the same radical, as we saw, expresses with certain slight modifications, noun, adjective, and verb. Again, if Da is once known to signify God, then zda must signify that which is opposed to God, namely, idol. If dab be spirit, odab will be body; if dad be heaven, odad will be hell. Again, if saba is king, sava is royalty, salba is reigning, swmba to be governed, &c. Volaptk, Pasilingua, etc. It must be clear from these extracts how totally different in character and purpose were these schemes of a universal, because philosophical, language from the schemes lately put forward under such names as Volapiik, Pasilingua, Lingvo Esperanto, &. The propounders of these systems have a purely practical purpose. They take one or more languages as they find them, try to remove all irregularities, and by simplifying both grammar and dictionary, to facilitate 70 CHAPTER II. the acquirement of an easy means of communication. Such experiments are quite unobjectionable, and, if properly conducted, may in time lead to something like a telegraphic language for the whole world. But they have nothing in common with the ideas of Descartes, Wilkins, and Leibniz. Reason and Language Inseparable. Let us now resume the thread of our argument. We saw that in an artificial language, the whole system of our notions, once established, may be matched to a system of phonetic exponents; but we maintain, until we are taught the contrary, that no real language was ever made in this manner.! There never was an independent array of deter- minate conceptions waiting to be matched with an independent array of articulate sounds. As a matter of fact, we never meet with articulate sounds except as wedded to determinate ideas ; nor do we ever meet with determinate ideas except as bodied forth in articulate sounds. Thisis a point of some importance on which there ought not to be any doubt or haze, and I therefore declare my conviction, whether right or wrong, as explicitly as possible, that thought, in one sense of the word, i.e. in the sense of reasoning, is impossible without language or without signs. After what I stated in my former lectures, I shall not be understood as here denying the reality of thought or mental activity in animals. Animals and infants who are without language, are alike without reason ; but ‘See an important letter of Descartes on the same subject in his (Huvres completes, ed. Cousin, v. 61; quoted in the French translation of my Lectures. LANGUAGE AND REASON. yA: the difference between animal and infant is, that the infant possesses the healthy germs of speech and reason, only not yet developed into actual speech and actual reason, whereas the animal has no such germs or faculties, capable of development in its present state of existence. We must concede to animals ‘sensation, perception, memory, will, and judgment, but we cannot allow to them a trace of what the Greek called (dgos, i.e. reason, literally, gathering, a word which most rightly and naturally expresses in Greek both speech and reason. Animals were called by the Greek dloga, whether in the sense of without reason, or in the sense of speechless. Ldgos is derived from légevn, which, like Latin /egere, means, originally, to gather. Hence, katdlogos, a catalogue, a gathering, a list; collectio, a collection. In Homer,? légevn is hardly ever used in the same sense of saying, speaking or meaning, but always in the sense of gathering, or, more properly, of telling, for to tell is the German zdhlen, and means originally to count, to cast up. Ldgos, used in the sense of reason, meant originally, like the English tale, or the German Zahl,® gathering ; for reason, ‘though it penetrates into the depths of the sea and earth, elevates our thoughts as high as the stars, and leads us through the vast spaces and large rooms of this mighty fabric, + is nothing more or less than the gathering up of the single by 1 Cf, Farrar, p. 125; Heyse, p. 41. 2 Od, xiv. 197: ov Te Stamphéaiu A€ywv Euad Kndea Ovpov. Ulysses says he should never finish if he were to tell the sorrows of his heart, i. e. if he were to count or record them, not simply if he were to speak of them. 3 Hxrod. v. 8, the tale, i.e. the number of the bricks. ' Locke, On the Understanding, iv. 17, 9. Cae CHAPTER II. means of the general.' To sum up, as Kant says, it is the office of the senses to perceive, and the office of the understanding to think; but to think is to unite different conceptions in one act of conscious- ness.2_ The Latin intelligo, i.e. inter-ligo, for inter- lego, expresses most graphically the interlacing of the general and the single, which is the peculiar province of the intellect. Expressions like cogitare, i.e. co- agitare, or to comprehend, rest on similar metaphors. But Ldgos used in the sense of word, means likewise a gathering, for every word, or, at least, every name is based on the same process; it represents the gathering of single impressions under one general conception. As we cannot tell or count quantities without numbers, we cannot tell or recount things without words. There are tribes, we are told, that have no numerals beyond four. Should we say that they do not know if they have five children instead of four? They certainly do, as much as a cat knows that she has five kittens, and will look for the fifth, if it has been taken away from her. But if they have no numerals beyond four, they cannot reason beyond four. They would not know, as little as children know it, that two and three make five, but only that two and three make many. * This, too, is well put by Locke (iii. 3, 20) in his terse and homely language : ‘ I would say that all the great business of genera and species, and their essences, amounts to no more but this: that men making abstract ideas, and settling them in their minds, with names annexed to them, do thereby enable themselves to consider things, and discourse of them, as it were, 7 bundles, for the easier and readier improvement and communication of their knowledge, which would advance but slowly were their words and thoughts confined only to particulars.’ 2 Kant, Proleg. p, 60. LANGUAGE AND REASON. 73 Formation of Names. Man could not name a tree, or an animal, or a river, or any object whatever in which he took an interest, without discovering first some general quality that seemed at the time the most characteristic of the object to be named ;? or, to borrow an expression of Thomas Aquinas (I. P. 9. 18, art. 9. ed. 2.), Nomina — non sequuntur modum essendi, qui est in rebus, sed modum essendi, secundum quod in cognitione nostra est. In the lowest stage of language, an imitation of * the neighing of the horse would have been sufficient to call or recall the horse. Savage tribes are great mimics, and imitate the cries of animals with wonder- ful success. But this is not yet language. There are cockatoos who, when they see cocks and hens, will begin to cackle as if to inform us of what they see. This is not the way in which the words of our languages were formed. There is no trace of neighing in the Aryan names for horse. In naming the horse, the quality that struck the mind of the Aryan man as the most prominent was its swiftness. Hence from the root as,* to be sharp or swift (which we have in Latin acus, needle, and in the French diminutive aiguille, in acuo, I sharpen, in acer, quick, sharp, shrewd, in acrymony and even in ’cute), was derived asva, the runner, the horse. This asva appears in * This point has been well discussed by Dr. Otto Caspari, Die Sprache als psychischer Entwickelungsgrund : Berlin, 1864. 2 La Science de Langage, par Alfred Gilly: Paris, 1868. 5 Of. Sk. Asu, quick, w«vs, dewey, point, and other derivatives given by Curtius, Griechische Etymologie, i.101. The Latin catus, sharp, has been derived from Sk. so (syati), to whet. 74 CHAPTER II. Lithuanian as aszva (mare), in Latin as ekvus, i.e. equus, in Greek as itxxos,! or tazos, in Old Saxon as chu. Many a name might have been given to the horse besides the one here mentioned; but, whatever name was given, it could only be formed by laying hold of the horse by means of some general quality, and by thus arranging the horse, together with other objects, under some general category. Many names might have been given to wheat. It might have been called eared, nutritious, graceful, waving, golden, * the child of the earth, &c. But it was called simply the white, the white colour of its grain seeming to distinguish it best from those plants with which otherwise it had the greatest similarity. For this is one of the secrets of onomatopoésis, or name-poetry, that each name should express, not the most impor- tant or specific quality, but that which strikes our fancy,” and seems most useful for the purpose of making other people understand what we mean. If we adopted the language of Locke, we should say that men were guided by wit rather than by judg- ment, in the formation of names. Wit, he says, lies most in the assemblage of ideas, and putting those together with quickness and variety, wherein can be found any resemblance or congruity, thereby to make up pleasant pictures, and agreeable visions, in the fancy: judgment, on the contrary, lies quite on the other side, in separating carefully, one from another, ideas wherein can be found the least difference, 1 Etym. Magn. p. 474, 12, texos onpaive tov immov. Curtius, G. FL. ii. 49. 2 Pott, Htym. F. ii. 139. LANGUAGE AND REASON. 7) thereby to avoid being misled by similitude, and by affinity, to take one thing for another.t. While the names given to things according to Bishop Wilkins’ philosophical method would all be founded on judg- ment, those given by the early framers of language repose chiefly on wit or fancy. Thus wheat was called the white plant, hwaiteis in Gothic, in A.S. hwéte, in Lithuanian kwetys, in English wheat, and all - these words point to the Sanskrit sve ta, 1. e. white, the Gothic hweits, the A. S. hwit. In Sanskrit, sveta, white, is not applhed to wheat (which is called go- dhima, the smoke or waves of the earth), but it is applied to many other herbs and weeds, and as a compound (svetasunga, white-awned) it entered into the name of barley. In Sanskrit, silver is counted as white, and called sveta, and the feminine sveti, was once a name of the dawn, just as the French aube, dawn, which was originally alba. We arrive at the same result whatever words we examine ; they always express a general quality, supposed to be peculiar to the object to which they are attached. In some cases this is quite clear, in others it has to be brought out by minute etymological research. To those who approach these etymological re- searches with any preconceived opinions, it must be a frequent source of disappointment, when they have traced a word through all its stages back to its first starting-point, to find in the end, or rather in the beginning, nothing but roots of the most general powers, meaning to go, to move, to run, to do. But on closer consideration, this, instead of being dis- 1 Locke, On the Human Understanding, ii. 11, 2. 76 CHAPTER II. appointing, should rather increase our admiration for the wonderful powers of language, man being able out of these vague and pale conceptions to produce names expressive of the minutest shades of thought and feeling. It was by a poetical fiat that the Greek probata, which originally meant no more than things walking forward, became in time the name of cattle, and particularly of sheep. In Sanskrit, sarit, mean- ing goer, from sar, to go, became the name of river; sara, meaning the same, what runs or goes, was used for sap, but not fer river. Thus dru, in Sanskrit, means to run, dravat, quick; but drapsa is re- stricted to the sense of a drop, gutta. The Latin evum, Meaning going, from 2, to go, became the name of time, age ; and its derivative wviternus, or eternus, was made to express eternity. Thus in French, meubles means literally anything that is moveable, but 1t became the name of chairs, tables, and ward- robes. In ancient Greek dloga, without reason, was used for brute animals in general. In modern Greek dlogon has become the name for horse Viande, originally vivenda,? the English vwiands, that on which one lives, came to mean meat. SMrwmentum, lit. what serves for food, from frui, means in Latin corn in general; froment in French is wheat. Jumentwmin Latin means a beast of burden ; ywment 1 addoyov, horse, occurs as early as 1198 in the Syllabus Gree. Membr. ed. Trinchera, p. 334: nat 76 ddroyév pou 70 patio, 70 88 -Aoyév pov TO Badioy, et equum meum nigrum, badium vero. * «La viande estoit un peu de poirée,’ dit Vauteur de la Vie d’ Isabella, sceur de Saint-Louis. ‘On ne pouvoit mie assez trouver viandes aux hommes et aux chevaux, rapporte la chronique de Saint-Denis.’ Michel Bréal, De la Méthode comparative, 1864, p. 15. ee ee, ey, 4 eS ee LANGUAGE AND REASON. re in French is a mare. A table, the Latin tabula, is originally what stands, or that on which things can be placed or stood ; it now means what dictionaries define as ‘a horizontal surface raised above the ground, used for meals and other purposes. The French tableau, picture, again goes back’ to the Latin tabula, a thing stood up, exhibited, and at last to the root std of stare, to stand. A stable, the Latin stabulwm, comes’ from the same root, but it was applied to the stand- ing-place of animals, to stalls or sheds. That on which a thing stands or rests is called its base, and basis in Greek meant originally no more than going, the base being conceived as ground on which it is safe to walk. What can be more general than facies, originally the make or shape of a thing, then the face? Yet the same expression is repeated in modern languages, feature being evidently a mere corruption of factuwra, the make. On the same principle the moon was called wna, i.e. losna, or lucina, the shin- ing; the lightning, fulmen from fulgere, the bright ; the stars stella, i.e. sterule, the Sanskrit staras, from st7rz, to strew, the strewers of light. All these etymologies may seem very unsatisfactory, vague, uninteresting, yet, if we reflect for a moment, we shall see that in no other way but this could the mind, or the gathering power of man, have compre- hended the endless variety of nature ! under a limited number of categories or names. What Bunsen called * Cf. Sankara on Vedanta-Sttra, 1, 3,28 (Muir, Sanskrit Texts, iii. 67), Akritibhis ka sabdanam sambandho na vyaktibhik, vyaktinam anantyat sambandhagrahandnupapatteh. ‘The relation of words is with the genera, not with individuals; for, as indi- viduals are endless, it would be impossible to lay hold of relations.’ 78 CHAPTER It. ‘the first poesy of mankind, the creation of words, is no doubt very different from the sensational poetry of later days; yet its very poverty and simplicity render it all the more valuable in the eyes of historians and philosophers. For of this first poetry, simple as it is, or of this first philosophy in all its childishness, man only is capable. He is capable of it because he can gather the single under the general; he is capable of it because he has the faculty of speech ; he is capable of it—we need not fear the tautology—because he is man. No Speech without Reason. Without speech no reason, without reason no speech. It is curious to observe the unwillingness with which many philosophers admit this, and the attempts they make to escape from this conclusion, all owing to the very influence of language which, in most modern dialects, has produced two words, one for language, the other for reason; thus leading the speaker to suppose that there is a substantial difference between the two, and not a mere formal difference... Thus Brown says: ‘To be without language, spoken or written, is almost to be without thought.’* But he qualifies this almost by what follows : ‘That man can reason without language of any kind, and conse- quently without general terms—though the opposite opinion is maintained by many very eminent philoso- 1 In Dutch there is no difference between sede, oratio, and rede, ratio, though Siegenbeek, in his authorised grammar of the Dutch language, 1804, tries to distinguish between rede, speech, and reden, reason, cause. Redeloos is irrational, redelijk, rational, reasonable, the German redlich; rvedenaar, an orator. 2. Works, i. p. 475, LANGUAGE AND REASON. 79 phers—seems to me not to admit of any reasonable doubt, or, if it required any proof, to be sufficiently shown by the very invention of language which in- volves these general terms, and still more sensibly by the conduct of the uninstructed deaf and dumb !—to which also the evident marks of reasoning in the other animals—of reasoning which I cannot but think as unquestionable as the instincts that mingle with it—may be said to furnish a very striking addi- tional argument from analogy.’ Deaf and Dumb People. The wninstructed deaf and dumb, however, have never given any signs of reason, in the true sense of the word, though to a certain extent all the deaf and dumb people that live in the society of other men catch something of the rational behaviour of their neighbours.2, When instructed, the deaf and dumb certainly acquire general ideas even, without being able in every case to utter distinctly the phonetic exponents or embodiments of these ideas which we call words. But this is no objection to our general argument. The deaf and dumb are taught by those who possess both these general ideas and their phonetic embodi- 1 Works, ii. p. 446. * Un médecin célébre de Vinstitution des sourds-muets, Itart, nous a dépeint I’état intellectuel et moral des hommes qu’un mutisme con- génital laissait réduits & leur propre expérience. Non-suelement ils subissent une véritable rétrogradation intellectuelle et morale qui les reporte en quelque sorte aux premiers temps des sociétés; mais leur esprit, formé en partie aux notions qui nous parviennent par les sens, ne saurait se développer.’ Claude Bernard, ‘ Exposé des Faits et du Principe de la Physiologie moderne,’ Revue ethnographique, 1869, p. 253. 80 ‘CHAPTER II. ments, elaborated by successive generations of rational men. They are taught to think the thoughts of others, and if they cannot pronounce their words, they lay hold of these thoughts by other signs, and particularly by signs that appeal to their sense of sight, in the same manner as words appeal to our sense of hearing. These signs, however, are not the sions of things or their conceptions, as words are: they are the signs of signs, just as written language is not an image of our thoughts, but an image of the phonetic embodiment of thought.. Alphabetical writ- ing is the image of the sound of language, hieroglyphic writing the image of language or thought. One of the highest authorities on the teaching of deaf and dumb people, Samuel Heinicke (1729-90), the founder of the German system of education of the deaf and dumb, says, ‘the deaf and dumb must be educated in order to be able to think in concepts, and that in sounding and articulated words of our lan- guage, if he is to learn from us, to understand us, and equally to communicate with us. The thinking of the deaf and dumb without teaching, if one may call so the irregular concatenation of his dark represen- tations, moves only in the sphere of sensuous intu- itions, and its forms and his language are rude and often very uncertain words, framed by himself, imi- tating external impressions, and rendering received impressions. We do not think in written, but in articulated and sounding words. The written word is the representation of the articulated word for the sense of sight, and is taken as an expression of thought only on the supposition of language. It is impossible LANGUAGE AND REASON. 81 to think in writing, without some whispering support of articulation, because writing absent from sight, is not representable in the soul.’ Locke. The same supposition that it is possible to reason without signs, that we can form mental conceptions, hay, even mental propositions, without words, runs through the whole of Locke's philosophy! He maintains over and over again, that words are signs added to our conceptions, and added arbitrarily. He imagines a state In which man, though possessed of a great variety of thoughts, and such from which others, as well as himself, might receive profit and delight, was unable to make these thoughts appear. The comfort and advantage of society, how- ever, not being to be had without communication of thoughts, it was necessary that man should find out some external sen- sible signs, whereby those invisible ideas of which his thoughts are made up might be made known to others. For this purpose, nothing was so fit, either for plenty or quickness, as those articulate sounds, which, with so much ease and variety, he found himself able to make. Thus we may conceive how words, which were by nature so well adapted to that purpose, came to be made use of by men as the signs of their ideas ; not by any natural connexion there is between particular arti- culate sounds and certain ideas; for then there would be but one language amongst all men; but by a voluntary compo- sition, whereby such a word is made arbitrarily the mark of such an idea. Locke admits, indeed, that it is almost unavoidable, in treating of mental propositions, to make use of words. ‘Most men, if not all, he says (and who are they that are here exempted ?) ‘in their thinking ' Locke, On the Human Understanding, iii, 2, 1. IT. G 82 CHAPTER II. and reasoning within themselves, make use of words, instead of ideas, at least when the subject of their meditation contains in it complex ideas.’ But this is in reality an altogether different question; it is the question whether, after our notions have once been realised in words, it is possible to use words without reasoning, and not whether it is possible to reason without words. Thisis clear from the instances given by Locke. Some confused or obscure notions (he says) have served their turns; and many who talk very much of religion and conscience, of church and faith, of power and right, of obstruc- tions and humours, melancholy and choler, would, perhaps, have little left in their thoughts and meditations, if one should desire them to think only of the things themselves, and lay by those words, with which they so often confound others, and not seldom themselves also. * In all this there is, no doubt, great truth; yet, strictly speaking, it is as impossible to use words without thought, as to think without words. Even those who talk vaguely about religion, conscience, &c. have at least a vague notion of the meaning of the words they use; and if they ceased to connect any ideas, however incomplete and false, with the words they utter, they could no longer be said to speak, but only to make noises. The same holds good if we in- vert our proposition. It is possible, without language, to see, to perceive, to stare at, to dream about things; but, without words, not even such simple ideas as white or black can for a moment be realised. We cannot be careful enough in the use of our words. If reasoning is used synonymously with 1 Locke, 7, ce. iv. 5, 4. 2 Thid. LANGUAGE AND REASON. 83 knowing or thinking, with mental activity in gene- ral, it is clear that we cannot deny it either to the uninstructed deaf and dumb, or to infants and ani- mals.‘ A child knows as certainly before it can speak the difference between sweet and bitter (i.e. that sweet is not bitter), as it knows afterwards (when it comes to speak) that wormwood and sugar- plums are not the same thing.2 A child receives the sensation of sweetness; it enjoys it, it recollects it, it desires it again; but it does not know what sweet is; it is absorbed in its sensations, its plea- sures, its recollections ; it cannot look at them from above, it cannot reason on them, it cannot tell of them.? This is well expressed by Schelling. Without language (he says) it is impossible to conceive philosophical, nay, even any human consciousness; and hence the foundations of language could not have been laid con- sciously. Nevertheless, the more we analyse language, the more clearly we see that it transcends in depth the most con- scious productions of the mind. It is with language as with all organic beings; we imagine they spring into being blindly, and yet we cannot deny the intentional wisdom in the forma- tion of every one of them.‘ Hegel speaks more simply and more boldly. ‘It is in names,’ he says, ‘that we think.’ é ‘ Amusement philosophique sur le Langage des Bestes, par le Pere Bougeant: Paris, 1739. pe loockes J.-c.jin 2.15, * *A child certainly knows that a stranger is not its mother; that its sucking-bottle is not the rod, long before he knows that it is impossible for the same thing to be and not to be.’—Locke, On the Human Under- standing, iv. 7, 9. * Hinleitung in die Philosophie der Mythologie, p. 52; Pott, Etymolo- gische Forschungen, ii. 261. ° Carritre, Die Kunst im Zusammenhang der Culturentwickelung, i AE G 2 84, - CHAPTER II. The Sound of Words has no independent existence. It may be possible, however, by another kind of argument, less metaphysical perhaps, but more con- vincing, to show clearly that reason cannot become real without speech. Let us take any word, for instance, expervment. It is derived from experior. . Perior, like Greek perdn,' would mean to go through. Perittus is aman who has gone through many things ; periculum, something to go through, a danger. Hz- perior is to go through and come out (the Sanskrit, vyutpad); hence experience and experiment. The Gothic jaran, the English to fare, are the same words as perdn; hence the German Lrfahrung, experience, and Gefahr, periculum ; Wohlfahrt, welfare, the Greek euporia. As long then as the word experiment ex- presses this more or less general idea, it has a real existence. But take the mere sound, and change only the accent, and we get experiment, and this is nothing. Change one vowel or one consonant, ea- porvment or esperiment, and we have mere noises, what Heraclitus would call a mere psdphos, but no words. Character, with the accent on the first syllable, has a meaning in English, but none in German or French ; chardcter, with the accent on the second syllable, has a meaning in German, but none in Eng- lish or French; charactére, with the accent on the last, has a meaning in French, but none in English or German. It matters not whether the sound is arti- culate or not; articulate sound without meaning is even more unreal than inarticulate sound. If, then, + Curtius, G. E. i. 237. LANGUAGE AND REASON. 85 these articulate sounds, or what we may call the body of language, exist nowhere, have no independent reality, what follows? I think it follows that this so-called body of language could never have been taken up anywhere by itself, and added to our con- ceptions from without; from which it would follow again that our conceptions, which are now always clothed in the garment of language, could never have existed in a naked state. This would be per- fectly correct reasoning, if applied to anything else ; nor do I see that it can be objected to as bearing on thought and language. If we never find skins except as the teguments of animals, we may safely conclude that skins cannot exist without animals. If colour cannot exist by itself (amav ydp xpépua év odpare), it follows that neither can anything that is coloured exist without colour. A colouring substance may be added or removed; but colour without some substance, however ethereal, is, in rerwm naturd, as impossible as substance without colour, or as substance without form or weight. Granting, however, to the fullest extent, the one and indivisible character of language and thought, agreeing even with the Polynesians, who express thinking by speaking in the stomach,’ we may yet, I think, for scientific purposes, claim the same liberty which is claimed in so many sciences, namely, the liberty of treating separately what in the nature of things cannot be separated. Though colour cannot be separated from some ethereal substance, yet the science of optics treats of light and colour as if they 1 Farrar, p. 125. 86 CHAPTER II. existed by themselves. The geometrician reasons on lines without taking cognisance of their breadth, of planes without considering their depth, of bodies without thinking of their weight. It is the same in language, and though I consider the identity of lan- guage and reason as one of the fundamental principles of our science, I think it will be most useful to begin, as it were, by dissecting the dead body of language, by anatomising its phonetic structure, without any reference to its function, and then to proceed to a consideration of language in the fulness of life, and to watch its energies, both in what we call its growth and its decay. CHAPTER III. THE ALPHABET. EK proceed now to dissect the body of lan- guage. In doing this we treat language as a mere corpse, not caring whether it ever had any life or meaning, but simply trying to find out what it is made of, how sounds are produced, how impressions are made upon our ear, and how they can be clas- sified. In order to do this it is not sufficient to examine our alphabet, such as it is, though no doubt the alphabet, if arranged according to scientific prin- ciples, may very properly be called the table of the elements of language. Greek Classification of Letters. But what do we learn from our AB C? what even, if we are told that £ is a guttural tenuis, s a dental sibilant, m a labial nasal, y a palatal liquid? These are names which are borrowed from Greek and Latin grammars. They expressed more or less happily the ideas which the scholars of Athens and Alexandria had formed of the nature of certain letters. But these ideas were by no means always correct, and, as translated into our grammatical phraseology they have frequently lost their original meaning. Our 88 CHAPTER ITI. modern grammarians speak of tenuis and media, but they define tenuis not as a bare or thin letter, so called originally in opposition to the aspirated con- sonants which in Greek were spoken of as thick, rough or shaggy (dacv), but on the contrary as the hardest and strongest articulation ; nor are they always aware that the mediw or middle letters were originally so called because, as pronounced at Alexandria, they seemed to stand halfway between the bare and the rough letters, i.e. the aspirates, being pronounced with less breath than the aspirates, with more than the tenues.! Plato's division of letters, as given in his Cratylus, is very much that which we still profess to follow. He speaks of voiced letters (@wrijevta, vo- cales), our vowels; and of voiceless letters (dpwva), our consonants, or mutes. But he divides the latter into two classes: first, those which are voiceless, but not soundless (dovievta pev ov, od pévTor ye apOoyya), afterwards called semi-vowels (u(pwva); and secondly, the real mutes, both voiceless and soundless, i.e. all consonants, except the semi-vowels (a@oyya).? In ' Scholion to Dionysius Thrax, in Anecdota Bekk. p. 810: Swvnrina épyava Tpia eiolv, 7 yA@ooa, of ddd6vTEs, TA XEIAN. Tots pev ody dxpos xeiAeot maAovpévois expwveira [TO mw], wate oxeddv pnde dAlyov Tt mvevpa TapekBaive* dvovyopevwy 5€ TOV xELAéwy Tavu Kal TVEdpaToS ToAAov éftdvTos, expwvetTa TO P* TO Se B, Expwvovpevov bpoiws Tois Gkpos THY XELAEwY, TovTéoTL TEpl TOV avTOoY TéTOV TOs TpoAExOEcat Tav pwvntinav Opyavwv, ore navy dvwye TA yxElAn ws TO >, OvTE mavu mAEl ws TO 7, GAAA peony Tiva SiéLodov TO Tvevuare TEperopevws Sidwov, «.7.A. See Rudolph von Raumer, Sprachwissenschaftliche Schriften, p. 102, who shows that the Scholion was written before 730 A.D.; Curtius, Griechische Etymologie, ii. p. 30. It is clear that the scho- liast speaks of the pronunciation of his own time, when the aspirates had become mere spirants, and when the mediz, too, approached to that pronunciation which they have in modern Greek. - 2 Raumer, l. c. p. 100. THE ALPHABET. 89 later times, the scheme adopted by Greek gram- marians is as follows:— I. Phonéenta, vocales, voiced, vowels. Il. Symphona, consonantes. II. 1. Hémtphona, semi-vocales, half-voiced, ], m,n, r, 8; or, Hygrd, liquide, fluid, Pett: II. 2. Aphona, mute. a. Psild, tenues (hard, surd); b. Mésa, medize (soft, sonant) ; ¢. Daséa, aspirate. Keats Ds reget lh ch, th, ph. The Pratisakhyas. Another classification of letters, more perfect, be- cause deduced from a language (the Sanskrit) at a time when it was not yet reduced to writing, but carefully watched, and preserved by oral tradition, is to be found in the so-called Pratisakhyas, works on phonetics, belonging to different schools in which the ancient texts of the Veda were handed down from generation to generation with an accuracy far ex- ceeding that of the most painstaking copyists of MSS. Some of these works have lately been published and translated, and may be consulted by those who take an interest in these matters.! 1 Pratisikhya du Rig-V eda, par M. Ad. Regnier, in the Journal asiatique. Paris, 1856-58. Text und Uebersetzung des Pritisikhya, oder der dltesten Phonetik und Grammatik, in M. M.’s edition of the Rig-Veda. Leipzig, 1856. Das Vagasanéyi-Pratisakhyam, published by Prof. A. Weber, in Indische Studien, vol. iv. Berlin, 1858. The Atharva-Veda Pratisikhya, by W. D. Whitney. New- haven, 1862. The same distinguished scholar has published an edition 90 CHAPTER III. Modern Phoneticians and Elocutionists. Of late years the whole subject of phonetics has been taken up with increased ardour by scientific men, and assaults have been made from three dif- ferent points by different armies, philologists, physio- logists, and mathematicians. The best philological treatises I can recommend (without mentioning earlier works, such as a very excellent treatise by Bishop Wilkins, 1688),! are the essays published from time to time by Mr. Melville Bell,? Mr. Alexander John Ellis, and Mr. Sweet. Other works by R. von Raumer,* F. H. du Bois-Reymond,’ Lepsius,® Thau- sing,’ may be consulted with advantage in their of the Pratisakhya of the Taittiriya-Veda. A similar work for the Saimaveda, under the title of Riktantra-vydkarana, has been discovered and published by Dr. Burnell; Mangalore, 1879. * Republished in Techmer’s Zeitschrift fiir Allgemeine Sprachwissen- schaft, vol. iv. p. 339. * A New Elucidation of the Principles of Speech and Elocution, by Alex. Melville, 1849. The same author has published several other works on phonetics, and has prepared an alphabet which is to indicate the physiological character of each letter, so as really to deserve the name of ‘ Visible Speech,’ a name too freely granted to the ancient Sys- tems of writing. See Visible Speech, a New Fuct, demonstrated by A. Melville Bell. 1865, and 1867. Lectures on Phonetics, delivered at Oxford, 1885. 3 Primer of Phonetics, 1890. * Gesammelte Sprachwissenschaftliche Schriften, von Rudolph von Raumer. Frankfort, 1863. (Chiefly on classical and Teutonic lan- guages. ) > Kadmus, oder Allgemeine Alphabetik, von F. H. du Bois-Reymond. Berlin, 1862. (Containing papers published as early as 1811, and full of ingenious and original observations. ) ° Lepsius, Standard Alphabet, second edition, 1863. (On the subject in general, but particularly useful for African languages. ) " Das natiirliche Lautsystem der menschlichen Sprache, von Dr. M. Thausing. Leipzig, 1863. (With special reference to the teaching of deaf and dumb persons.) THE ALPHABET. 91 respective spheres. The Physiological works which I found most useful and intelligible to a reader not professionally devoted to these studies were Miller's ‘Handbook of Physiology, Briicke’s ‘ Grundziige der Physiologie und Systematik der Sprachlaute’ (Wien, 1856), Funke’s ‘Lehrbuch der Physiologie, and Czermak’s articles in the ‘Sitzungsberichte der k.k. Akademie der Wissenschaften zu Wien.’ ! Among works on mathematics and acoustics, I have consulted Sir John Herschel’s ‘Treatise on Sound,’ in the ‘ Encyclopzedia Metropolitana ;’ Pro- fessor Willis’s paper ‘On the Vowel Sounds and on Reed Organ-Pipes,’ read before the Cambridge Phy- siological Society in 1828 and 1829; but chiefly Professor Helmholtz’s classical work ‘Die Lehre von den Tonempfindungen’ (Braunschweig, 1863), a work giving the results of the most minute scien- tific researches in a clear, classical, and truly popular form, so seldom to be found in scientific books. The whole subject of Phonetics has lately been treated in the most exhaustive and masterly manner by Dr. Techmer in the first volume of his Interna- zionale Zeitschrift fiir Allgemeine Sprachwissenschaft Leipzig, 1884. Spelling Reformers. I ought not to omit to mention here the valuable services rendered by those who, for nearly fifty years, have been labouring in England to turn the results of scientific research to practical use, in de- 1 See also Populdre physiologische Vorirdge, von J. N. Czermak : Wien, 1869. 92 CHAPTER III. vising and propagating a new system of ‘Brief Writing and True Spelling, best known under the name of the Phonetic Reform. I am far from under- rating the difficulties that stand in the way of such a reform, and I am not so sanguine as to indulge in any hopes of seeing it carried for the next three or four generations. But I feel convinced of the truth and reasonableness of the principles on which that reform rests, and as the innate regard for truth and reason, however dormant or timid at times, has always proved irresistible in the end, enabling men to part with all they hold most dear and sacred, whether corn-laws, or Stuart dynasties, or Papal legates, or heathen idols, I doubt not but that the effete and corrupt orthography will follow in their train. Nations have before now changed their nu- merical figures, their letters, their chronology, their weights and measures; and though Mr. Pitman may not live to see the results of his persevering and disinterested exertions, it requires no prophetic power to perceive that what at present is pooh- poohed by the many, will make its way in the end, unless met by arguments stronger than those hitherto levelled at the ‘Fonetic Nuz.’ One argument which might be supposed to weigh with the student of language, viz. the obscuration of the etymological structure of words, I cannot consider as very for- midable. The pronunciation of languages changes according to fixed laws, the spelling has changed in the most arbitrary manner, so that if our spelling followed strictly and unswervingly the pronunciation of words, it would in reality be of greater help to the THE ALPHABET, 93 critical student of language than the present uncer- tain and unscientific mode of writing.! Although considerable progress has thus been made in the analysis of the human voice, the difficulties inherent in the subject have been increased rather than diminished by the profound and laborious re- searches carried on independently by physiologists, students of acoustics, and philologists. The human voice opens a field of observation in which these three sciences meet, and to neglect the results ob- tained by any one of them is entirely to deprive the study of Phonetics of its scientific character. The substance of speech or sound has to be analysed by the mathematician and the experimental philosopher ; the organs or instruments of speech have to be ex- amined by the anatomist ; and the history of speech, the actual varieties of sound which have become typi- fied in language, fall to the province of the student of language, and likewise of the practical elocutionist. Under these circumstances it is absolutely necessary that students should co-operate in order to bring these scattered researches to a successful termination ; and I take this opportunity of expressing my obligation to Dr. Rolleston, our late Professor of Physiology, Mr. G. Griffith, Deputy-Professor of Experimental Philosophy, Mr. A. J. Ellis, and others, for their kind- ness in helping me through difficulties which, but for their assistance, I should not have been able to over- come without much loss of time. * See an article of mine in the Fortnightly Review for April, 1876, ‘On Spelling,’ reprinted separately by Mr. Pitman, London, 1880. 94, CHAPTER III. The Voice. What can seem simpler than the A BC, and yet what is more difficult when we come to examine it? Where do we find an exact definition of vowel and consonant, and how they differ from each other? The vowels, we are told, are simple emissions of the voice, the consonants cannot be articulated except with the assistance of vowels. If this were so, letters such as s, f, 7, could not be classed as consonants, for there is no difficulty in pronouncing these without the assist- ance of a real vowel. Czermak, on the contrary, calls these letters consonants in quite a different sense. He would reserve the name of consonant (Mvtlauter) for all sonant letters, nay even for vowels, while he looks upon the surd consonants as the only true Selbstlauter, because they are not accompanied by voice.’ Again, what is the difference between a, 7,2? What is the difference between a tenuis and media, surd and sonant, hard and soft consonants, a difference almost incomprehensible to certain races; for instance, the Mohawks and the inhabitants of Saxony ? What we hear may be divided, first of all, into Noises and Tones. Noises, such as the rustling of leaves, the jarring of doors, or the clap of thunder, are produced by irregular impulses imparted to the air. Tones, such as we hear from tuning-forks, strings, flutes, organ pipes, are produced by regular periodical (isochronous) vibrations of elastic air. That tone, musical tone, or tone in its simplest form, is produced by tension, and ceases after the sounding 1 Physiologische Vortrdge, p. 107. — ee ee eee ee THE ALPHABET. 95 body has recovered from that tension, seems to have been vaguely known to the early framers of language, for the Greek tonos, tone, is derived from a root tan, meaning to stretch, to extend. Pythagoras! knew more than this. He knew that when chords of the same quality and the same tension are to sound a fundamental note, its octave, its fifth, and its fourth, their respective lengths must be like 1 to 2, 2 to 8, and 3 to 4. Strength, Pitch, and Quality. When we hear a single note, the impression we receive seems very simple, yet it is in reality very complicated. We can distinguish in each note— 1. Its strength or loudness, 2. Its height or pitch, 3. Its quality, or, as it is sometimes called, timbre : in German Tonfarbe, i.e. colour of tone. Strength or loudness depends upon the amplitude of the excursions of the vibrating particles of air which produce the wave. Height or pitch depends on the length of time that each particle requires to perform an excursion, i.e. on the number of vibrations executed in a given time. If, for instance, the pendulum of a clock, which oscillates once in each second, were to mark smaller portions of time, it would cause musical tones to be heard. Sixteen double oscillations in one se- cond would be sufficient to bring out tone, though its pitch would be so low as to be hardly perceptible. For practical purposes, the lowest tone we hear is produced by 30 double vibrations in one second, the * Helmholtz, Linleitung, p. 2. 96 CHAPTER III. highest by 4,000. Between these two lie the usual seven octaves of our musical instruments. It is said to be possible, however, to produce perceptible mu- sical tones through 11 octaves, beginning with 16 and ending with 38,000 double vibrations in one second, though here the lower notes are mere hums, the upper notes mere clinks. The A’ of our tuningforks, as fixed in 1859 by a decree of the French ministry, requires 437-5 double, or 875 single! vibrations in one second. In Germany the A’ tuning-fork makes 440 double vibrations in one second. It is clear that beyond the lowest and the highest tones perceptible to our ears, there is a progress ad infinitum, musical notes as real as those which we hear, yet be- yond the reach of owr sensuous perception. It is the same with the other senses. We can perceive the movement of the pendulum, but we cannot perceive the slower movement of the hand on the watch. We can perceive the flight of a bird, but we cannot per- ceive the quicker movement of a cannon-ball. This, better than anything else, shows how dependent we are on our senses; and how, if our senses are our weapons for the discovery of truth, they are likewise the chains that keep us from soaring too high. Up to this point everything, though wonderful enough, isclear and intelligible. As we hear a note, we can find out, with mathematical accuracy, to how many vibra- tions in one second it is due; and if we want to produce + It is customary to reckon by single vibrations in France and Ger- many, although some German writers adopt the English fashion of reckoning by double vibrations or complete excursions backwards and forwards. Helmholtz uses double vibrations, but Scheibler uses single vibrations. De Morgan calls a double oscillation a ‘ swing-swang.’ _ = ea Oe — THE ALPHABET, 97 the same note, an instrument, such as the siren, which gives a definite number of impulses to the air within a given time, will enable us to do it in the most mechanical manner. When two waves of one note enter the ear in the same time as one wave of another, the interval between the two is an octave. When three waves of one note enter the ear in the same time as two waves of another, the interval between the two notes is a Jifth. When four waves of one note enter the ear in the same time as three waves of another, the interval between the two notes is a fourth. When jive waves of one note enter the car in the same time as four waves of another, the interval between the two notes is a major third. When six waves of one note enter the ear in the same time as five waves of another, the interval between the two notes is a minor third. When five waves of one note enter the ear in the same time as three waves of another, the interval between the two notes is a major siath, All this is but the confirmation of what was known to Pythagoras. He took a vibrating cord, and, by placing a bridge so as to leave 2 of the cord on the right, $ on the left side, the left portion vibrating by itself gave him the octave of the lower note of the right portion. So, again, by leaving 3 on the right, 3 on the left side, the left portion vibrating gave him the fifth of the right portion. But it is clear that we may hear the same tone, 1.e. the result of exactly the same number of vibra- Op H 98 ' CHAPTER IIT. tions in one second, produced by different instruments, such as our vocal organs, a flute, a violoncello, a fife, or a double bass. They are tones of the same pitch, and yet they differ in character, and this differ- ence is called their quality. But what is the cause of these various qualities? By a kind of negative reasoning, it had long been supposed that, as quality could neither arise from the amplitude nor from the duration, it must be due to the form of the vibrations. It has now, however, been proved that this is so, by applying the microscope to the vibrations of different musical instruments, and thus catching the exact out- line of their respective vibrations—a result which before had been but imperfectly attained by an instru- ment called the Phonautograph. What is meant by the form of waves may be seen from the following outlines :-— THE ALPHABET. 99 The Qualities of Vowels. It has likewise been shown that the different forms of the vibrations which are the cause of what we call quality or colour, are likewise the cause of the presence or absence of certain harmonies, or by-notes; in fact, that varying quality and varying harmonics are but two expressions of the same thing. Harmonics are the secondary tones which can be perceived even by the unassisted ear, if, after lifting the pedal, we strike a key on a pianoforte. These harmonics arise from a string vibrating as if its motion were compounded of several distinct vibrations of strings of its full length, and one-half, one-third, one-fourth, &c., part of its length. Each of these shorter lengths would vibrate twice, three times, four times as fast as the original length, producing corre- sponding tones. Thus, if we strike c, we hear, if listen- ing attentively, 0’, a’, 0”, EB”, a’, B” flat, 0”, &e. : feu pagew “ineeseces 7s 7 2 £201 RE Sri To Te ET OE nad ES SS ee ie boa 2 = eres oe ee 1 i Y. 3 4 5 6 7 8 Cc Cc rene Ce E’ Gigs Geotlatee Ge That the secondary notes are not merely imagin- ary or subjective can be proved by a very simple and amusing experiment. If we place little soldiers —very light cavalry—on the strings of a pianoforte and then strike a note, all the riders that sit on strings representing the secondary tones will shake, and possibly be thrown off, because these strings He? 100 CHAPTER III. vibrate in sympathy with the secondary tones of the string struck, while the others remain firm in their saddles. Another test can be applied by means of resounding tubes, tuned to different notes. If we apply these to our ear, and then strike a note the se- condary tones of which are the same as the notes to which the resounding tubes are tuned, those notes will sound loudly and almost yell in our ears; while if the tubes do not correspond to the harmonics of the note played, the resounding tubes will not answer in the same manner. We thus see, again, that what seems to us a simple impression, the one note struck on the pianoforte, consists of many impressions which together make up what we hear and perceive. We are not conscious of the harmonics which follow each note and deter- mine its quality, but we know, nevertheless, that these by-notes strike our ear, and that our senses receive them and suffer from them. The same re- mark applies to the whole realm of our sensuous knowledge. There is a broad distinction. between sensation and perception. There are many things which we perceive at first and which we perceive again as soon as our attention is called to them, but which, in the ordinary run of life, are to us as if they did not exist at all. When I first came to Oxford, I was constantly distracted by the ringing of bells; after a time I ceased even to notice the dinner-bell. There are earrings much in fashion just now—little gold bells with coral clappers. Of course they pro- duce a constant jingling which everybody hears except the lady who wears them in her ears. In THE ALPHABET. 101 these cases, however, the difference between sensation and perception is simply due to want of attention. In other cases our senses are really incapable, with- out assistance, of distinguishing the various con- stituents of the objective impressions produced from without. We know, for instance, that white light is a vibration of ether, and that it is a compound of the | single colours of the solar spectrum. A prism will at once analyse that compound, and divide it into its component parts. To our apprehension, however, white light is something simple, and our senses are too coarse to cieceetieh its component elements by any effort whatsoever. We now shall be better able to understand what I consider a most important discovery of Professor Helmholtz." It had been proved by Professor G. S Ohm? that there is only one vibration without har- monics, viz. the simple pendulous vibration. It had likewise been proved by Fourier, Ohm, and other mathematicians,* that all compound vibrations or sounds can be divided into so many simple or pendu- lous vibrations. But it is due to Professor Helmholtz that we can now determine the exact configuration of many compound vibrations, and determine the presence and absence of the harmonics which, as we saw, caused the difference in the quality, or colour, or tumbre of sound. Thus he found that in the violin as compared with the guitar or pianoforte, the pri- mary note is strong, the secondary tones from two to six are weak, while those from seven to ten are much ' Helmholtz, l. ce. p. 82. RON Dade Stine pense 102 CHAPTER III. more distinct.!. In the clarionet® the odd harmonics only are perceptible, in the hautboy the even har- monics are of equal strength. Let us now see how all this tells on language. When we are speaking we are in reality playing on a musical instrument, and a more perfect instrument than was ever invented by man. It is a wind-in- strument, in which the vibrating apparatus 1s sup- plied by the chord vocales, while the outer tube, or bells, through which the waves of sound pass, are furnished by the different configurations of the mouth. The Vocal Organs. I shall try, as well as I can, to describe, with the help of some diagrams, the general structure of this instrument, though in doing so I ean only retail the scant information which I gathered myself from our excellent Professor of Physiology at Oxford, Dr. Rolleston. He kindly showed and explained to me by actual dissection, and with the aid of the newly- invented laryngoscope® (a small looking-glass, which enables the observer to see as far as the bifurcation of the windpipe and the bronchial tubes), the bones, the cartilages, the ligaments and muscles, which together form that extraordinary instrument on which we play our words and thoughts. Some parts of it are extremely complicated, and I would not venture to act even as interpreter of the dif- ferent and sometimes contradictory views held by 1 Helmholtz, 7. c. p. 143. Als: GC.) DaghOias 3 Czermak, Uber den Kehlkopfspiegel und seineVerwerthung. Leipzig, 1860; 2nd ed. 18638. THE ALPHABET. 103 Miller, Briicke, Czermak, Funke, and other dis- tinguished physiologists, on the mechanism of the various cartilages, the thyroid, cricoid, and arytenoid, which together constitute the levers of the larynx. It fortunately happens that the most important organs which are engaged in the formation of letters lie above the larynx, and are so simple in their structure, and so open to constant inspection and examination, that, with the diagrams here inserted, there will be little difficulty, I hope, in explain- ing their respective functions. There is, first of all, the thorax (1), which, by alter- nately compressing and dilating the lungs, performs the office of bellows. The next diagram (2), shows the trachea, a carti- laginous and elastic pipe, which terminates in the lungs by an infinity of roots or bronchial tubes, its upper extremity being formed into a species of head, called the laryna, situated in the throat, and com- posed of five cartilages. The uppermost of these cartilages, the epiglottis (3), is intended to open and shut, like a valve, the aperture of the glottis, i.e. the superior orifice of the larynx (fissura laryngea pharyngis). The epiglottisis a leaf- shaped elastic cartilage, attached by its narrower end to the thyroid cartilage, and possessing a midrib overhanging and corresponding to the fissure of the glottis. The broader end of the leaf points freely upwards towards the tongue, in which direction the entire cartilage presents a concave, as towards the larynx a convex, outline. In swallowing, the epi- glottis falls over the larynx, like a saddle on the back CHAPTER III. 104 In the formation of certain letters a of a horse. horizontal narrow fissure may be produced by de- Ee Se . ee ’ 5. External intercostals. 6. Rectus abdominis 7. Internal oblique. pressing the epiglottis over the vertical false and 46¢q . Ae Pao e ne 3a & et :3 § m (=) Mees ci go Bn 2 oS) eee hs HO FH ee SOS gO fa} Has o re S o = P= ata THE ALPHABET. 105 Superior Cornea Inferior Corn Bronchial Tubes Bronchial Tubes 106 CHAPTER III. Opening uf Nasal duct v7 : : f ’ Within the larynx (4), rather above its middle, between the thyroid and arytenoid cartilages, are two elastic ligaments, like the parchment of a drum split in the middle, and forming an aperture which is called the interior or true glottis, and corresponds in THE ALPHABET. 107 direction with the exterior glottis. This aperture 1s provided with muscles, which enlarge and contract it at pleasure, and otherwise modify the form of the larynx. ‘The three cartilages of the larynx supply the most perfect mechanism for stretching or relaxing the chords, and likewise, as it would seem, for dead- ening some portion of them by pressure of a protu- Fig. 4. berance on the under-side of the epiglottis (in Ger- man, Epiglottis-wulst). These chords are of different lengths in children and grown-up people, in man and in woman. Their average length in man is 18} mm. when relaxed, 23} mm. when stretched ; in woman, 122 mm. when relaxed, 153 mm. when stretched: thus giving a difference of about one- 108 CHAPTER III. third between the two sexes, which accounts for the different pitch of male and female voices.! The tongue, the cavity of the fauces, the lips, teeth, and palate, with its velum pendulum and uvula per- forming the office of a valve between the throat and nostrils, as well as the cavity of the nostrils themselves, are all concerned in modifying the impulse given to the breath as it issues from the larynx, and in pro- ducing the various vowels and consonants. Vowels. After thus taking to pieces the instrument, the tubes and reeds as it were of the human voice, let us now see how that instrument is played by us in speaking or in singing. Familiar and simple as singing or music in general seems to be, it is, if we analyse it, one of the most wonderful phenomena. What we hear when listening to a chorus or a sym- phony is a commotion of elastic air, of which, to quote from Helmholtz, the wildest sea would give a very inadequate image. The lowest tone which the ear perceives is due to about 30 vibrations in one second, the highest to about 4,000. Consider then what happens in a Presto, when thousands of voices and instruments are simultaneously producing waves of air, each wave crossing the other, not only like the surface waves of the water, but like spherical bodies, and, as it would seem, without any percep- tible disturbance ;? consider that each tone is accom- * Funke, Lehrbuch der Physiologie, p. 664, from observations made by J. Miller. * Weber, Wellenlehre, p. 495. ns THE ALPHABET. 109 panied by secondary tones, that each instrument has its peculiar tvmbre, due to secondary vibrations ; and, lastly, let us remember that all this cross-fire of waves, all this whirlpool of sound, is moderated by laws which determine what we call harmony, and by certain traditions or habits which determine what we call melody—both these elements being absent in the songs of birds—that all this must be reflected like a microscopic photograph on the two small organs of hearing, and there excite not only percep- tion, but perception followed by a new feeling even more mysterious, which we call either pleasure or pain; and it will be clear that we are surrounded on all sides by miracles transcending all we are accus- tomed to call miraculous, and yet disclosing to the genius of an Euler or a Newton laws which admit of the most minute mathematical determination. If we pronounce a vowel, what happens? Breath is emitted from the lungs, and some kind of tube is formed by the mouth through which, as through a clarionet, the breath has to pass before it reaches the outer air. If, while the breath passes the vocal chords, these elastic /aminw are made to vibrate periodically, we sing, and the number of the vibra- tions determines the pitch of our voice, but it has nothing to do with its timbre, i.e. its vowel. We may vary the pitch of our voice, without changing its vocal timbre. What we call vowels are neither more nor less than the qualities, or colours, or timbres of our voice, and these are determined by the form, not by the number, of the vibrations, this form being determined by the form of the buccal tubes. 110 CHAPTER III. This had, to a certain extent, been anticipated by Professor Wheatstone in his critique’ on Professor Willis’s ingenious experiments, but it has now been rendered quite evident by the researches of Professor Helmholtz. It is, of course, impossible to watch the form of these vibrations by means of a vibration microscope, but it is possible to analyse them by means of resounding tubes, like those before de- scribed; and thus to discover in them what, as we saw, is homologous with the form of vibration, viz. the presence and*absence of certain harmonics. Ifa man sings the same note on different vowels, the harmonics which answer to our resounding tubes vary as they would vary if the same note was played on different instruments, such as the violin, the flute, or the clarionet. In order to remove all uncertainty, Professor Helmholtz simply inverted the experiment. He took a number of tuning-forks, each furnished with a resonance box. By advancing or withdrawing this box he could impart to their primary tones various degrees of strength, and extinguish their secondary tones altogether. He tuned them so as to produce a series of tones answering to the harmonies of the deepest tuning-fork. He then made these tuning- forks vibrate simultaneously by means of a galvanic battery, and by combining the harmonics, which he had first discovered in each vowel by means of the sounding tubes, he sueceeded in reproducing _arti- ficially exactly the same vowels.? We know now what vowels are made of. They 1 London and Westminster Review, Oct. 1837, pp. 34, 37. 2 Le. p, 188, ee Se eee THE ALPHABET. ig a) are produced by various forms imparted to the voice, or to the air which is made to vibrate in its passage through the vocal chords. They vary like the timbre of different instruments, and we in reality change the instruments on which we speak when we modify the buccal tubes in order to pronounce 4@, @, 2, o, w (the vowels to be pronounced as in Italian or in Spanish), Is it possible, then, to produce a vowel, to evoke a certain timbre of our mouth, without giving at the same time to each vowel a certain musical pitch? This question has been frequently discussed. For a long time it was taken for granted that vowels could not be uttered without pitch. Yet, if a vowel was whispered, it was easy to see that the vocal chords were not vibrating, as they are when we sing, and that they began to vibrate only when the whispered vowel was changed into a voiced vowel. J. Miiller proposed a compromise. He admitted that the vowels might be uttered as mutes, and without any definite tone from the vocal chords, but he maintained that these mute vowels were formed in the glottis by the air passing the non-sonant chords. This view,! though in the main correct, has been somewhat modified by later observations, which have shown that in whispering the vocal chords are drawn together, while at the same time the back part of the glottis between the arytenoid cartilages remains open, assuming the form of a triangle. The breath passing ' Funke, Handbuch der Physiologie, p. 678. Different views of Willis and Briicke, p. 678. * Helmholtz, p. 171. Professor Czermak remarks, that the same effect Wi. CHAPTER III. through this aperture may produce imperfect vibra- tions, and these imperfect vibrations would produce the muffled tone that accompanies whispered vowels. In cases of aphonia, where the vocal chords cannot be made to vibrate freely, it is still possible to pro- nounce the different vowels, and the vox clandestina, though a mere whisper, is able to rise and to fall. Though it is true, therefore, that the vowels can be pronounced without the definite pitch of the perfect voice, it is still held by high authorities, though de- nied by others equally high, that, even in whispered vowels, some kind of pitch may be distinguished ; nay, that there is a pitch peculiar to each vowel, whether voiced or whispered. This was first pointed out by Professor Donders, and afterwards corrected and confirmed by Professor Helmholtz.2 We can best perceive this if we pronounce a whispered ii, and then allow it gradually to become a whistling, in which case we shall always get the same tone; a most useful discovery as a substitute for a tuning-fork.? It will be necessary, I think, to treat these indications of musical pitch in whispered vowels as imperfect tones, that is to say, as noises approaching to tones, or as irregular vibrations, nearly, yet not quite, changed into regular or isochronous vibrations; though the exact limit where a noise ends and tone begins has, may be and is produced by the larynx assuming different other conforma- tions. ‘ Uber den Spiritus asper,’ p. 7. See, however, the same author’s remarks in his Physiologische Vortriige, 1869, p. 101. * Sir William Thomson, for instance, denies this. 2 l.c. p. 172. That there is some connection between the quality and the pitch of vowels is also seen from the fact, that very high pitch is incompatible with the quality of the vowels wand o. * Czermak, Physiologische Vortrige, p. 118. THE ALPHABET. 1138 as far as I can see, not yet been determined by any philosopher,’ and the subject requires further careful consideration. Vowels in all their varieties are really infinite in number. Yet, for practical purposes, certain typical vowels, each with a large margin for dialectic variety, have been fixed upon in all languages, and these we shall now proceed to examine. We cannot take any account of the endless dialectic or local or even per- sonal variations that take place in the pronunciation of vowels, because, however interesting for special purposes, they are of no importance for the elucida- tion of the general principles of phonetics, with which alone we are here concerned. How far the subdi- vision of the sounds of the alphabet can be carried may be seen, for instance, in Mr. A. J. Ellis’s Palewotypic Alphabet, which contains about 270 signs for as many different sounds. When the sounds of a spoken lan- guage are submitted to so minute an analysis, it is not surprising that there should be so much variety of opinion between different authorities, and that the same letter should be described in the most divergent ways. Different elocutionists persuade themselves that there is a difference between the w in French dune and the % in German iiber, between the ew in French pew and the 6 in Goethe, and yet that the 6 in the German (otter, is the same as the w in gutter ! But though the Science of Language declines to re- cognise any but dynamic or functional distinctions of vowels and consonants, that is, any distinctions except * See Briicke, Grundziige, p. 16. 1 I 114 CHAPTER III. those which are connected with a real change of meaning, it will be useful to the scholar also to learn to what perfection the elocutionist has brought the minute analysis of spoken sounds. It is true that for his own purposes the student of Comparative Philology must always keep before his eyes the system of the typical sounds of any family of speech, however much they may be hidden behind the ever-changing play of dialect. But for this very purpose, for the study of dialects, and more particularly for the study of dialects that have not yet been reduced to writing, a knowledge of such systems as that of Mr. Melville Bell will prove extremely useful, and deserves more attention than it has hitherto received. Mr. Melville Bell complains that there is no representative of ‘Visible Speech’ in England. But surely both Mr. Ellis and Mr. Sweet have been most energetic apostles of that system, though, whether rightly or wrongly, they may occasionally have deviated from the opinions of its author. I cannot do more here than give a slight abstract of ‘ Visible Speech, and must refer for fuller information to Mr. Bell’s own publications. Mr. Melville Bell’s System of Phonetics. Mr. Melville Bell in his latest works! divides all speech-elements into three classes :— (1) those produced by vocalised breath or voice,— vowels and voiced consonants ; (2) those produced by unvocalised breath,—whis- pered vowels and breath consonants ; (3) those produced by the mouth alone,—percussions. 1 University Lectures on Phonetics, 1887, p. 59. Visible Speech and Vocal Physiology, 1889. THE ALPHABET. 115 These speech-elements require for their production, (1) the lungs, to supply breath ; (2) the glottis, to change breath into voice ; (3) the pharynx, to compress it ; (4) the tongue and lips, to parcel it; and (5) the cavities of pharynx, mouth, and nose, to mould it. The sounds produced by voice are the vowels and the voiced consonants. The sounds produced by breath are the vowels, if whispered, and the breath consonants. The sounds produced by the mouth alone, without either voice or breath, are the percussions, as heard in p, t, k, if not preceded or followed by breath. Tones, as described by Mr. Melville Bell, are turns of the voice as it rises and falls in speaking, commonly called cantilena. Glides (Ubergangslaute) are produced by the transi- tion from one organic position to another. Thus in ai-ry, there is a voice glide between the az and the +. If the top of the soft palate is slightly depressed and the nasal passage uncovered, all vowels become nasalised. Vowels. Mr. Melville Bell next gives a list of all possible vowel sounds, though he admits that several of them never occur in the languages known to us. Taking the top of the mouth as an arch, he shows that the tongue may take an equally high position close to the front, the top, or the back of the arch. This gives us the three high vowels, one in front, as in bee, jh! 116 CHAPTER IIT. an unused vowel at the back (00 delabialised),! and one between the two, called mixed, as in church, pronounced in American fashion. Each of these three vowels can be pronounced in three different ways. In pronouncing ee, we keep the tip of the tongue high, facing the front of the palate. If the front cavity is enlarged by gradually lowering the tongue, we get the vowel a as in‘ale, and lastly e as in ell. These three vowels are called High, Mid, and Low Front vowels. Taking the High Back vowel (which is not used) as our starting-point, we can modify it by enlarging the back cavity by lowering the tongue. We then get the three High, Mid, and Low Back vowels, described as delabialised oo, 6, and aw. Taking the High Mixed vowel as our starting-point, we can modify it by enlarging the mouth cavity by lowering the tongue. We then get the three High, Mid, and Low Mixed vowels, as heard in church (American), in alté? (German e), and pénny (Cockney huckster). Each of the nine vowels hitherto described can be modified if, in pronouncing them, we round our lips. Here the High vowels have a narrow, the Low a broad, the Mid, an intermediate labial aperture. If we pronounce the High Front ee of dee and round the lips, we get the German vw. If we pronounce the Mid Front a of ale, and round the lips, we get the French 1 To delabialise is meant for removing the action of the lips from such vowels as 00, 0, and aw. * 2 T put the accent on the vowel, if there is any doubt as to which vowel is meant. THE ALPHABET. GLE w. If we pronounce the Low Front e of eld, and round the lips, we get the French eu. Applying the same process to the High, Mid, and Low Back vowels, we get the vowels 00, 0, and aw. It was in fact by delabialising these vowels that the three primary Back vowels were discovered, though they are seldom used. Thirdly, by applying the same process of rounding to the Mixed vowels, we get a blending of 00 with i, of 6 with w (French), and of aw with ew (French). The first sound is heard in look, as pronounced in the North of [veland ; the second in the French homme, and the third in the initial element of the Irish diph- thongal sound of J, in J mind. We have now eighteen possible vowels. Every one of these, as Mr. Melville Bell informs us, admits of what he calls widening, ora loose and more indefinite pronunciation of the primary vowels, the organic positions remaining otherwise the-same. The follow- ing list will best show the difference between primary or narrow, and secondary or wide vowels, as under- stood by Mr. M. Bell. Secondary Primary (narrow). (wide). Primary (narrow). Secondary (wide). High-front: eel ill tiber (Germ.) une (Fr.) Mid-front: ale alr dai (Fr.) school (Scotch) Low-front: end and peu (Fr.) now (Cockney) High-mixed : church (Am.) the look (N. Ivel.) awful Mid-mixed : alté (Germ.) sofa homme (FYv.) Sorrow Low-mixed : zur (Somerset) sir I (Irel.) mirror (Chicago) 118 : CHAPTER III. Secondary Primary (narrow). (wide). Primary (narrow). Secondary (wide). High-back : laogh (Gaelic) —tidn pool pull Mid-back: up ask old ore Low-back : up (Scotch) ah yawn you Consonants. All consonants are the result of friction, compression, or interception of the breath in its passage from the lungs through the mouth. (1) If the breath-channel is contracted between the back of the tongue and the soft palate, we get the sound of ch in German nach. (2) If the breath-channel is contracted between the middle of the tongue and the soft palate, we get the sound of ch in German ich, or English hue. (3) If the breath-channel is contracted between the tip of the tongue and the gum or the front edge of the palatal arch, we get the sound of 7, as heard in three (Scotch). (4) If the breath-channel is contracted between the edges of the approximatal lips, we get the sound made in blowing to cool. (5) If the first of these consonantal sounds is modified by the lips, we get the sound ch as heard in leuch (laughed) in Scotch. (6) If the second sound is modified by the elevation of the forepart of the tongue, it is changed to sh. (7) If the third sound is modified by the elevation of the middle of the tongue, it 1s changed to s. (8) If the fourth sound is modified by the retraction of the tongue towards the back, the sharp blowing THE ALPHABET. 119 sound is changed into a hollow whistling sound, the English wh. This gives us eight primitive consonants, all breath- consonants. We have only to substitute for breath vocalised breath or voice, and that number is doubled. This gives us :— Breath. Voice. Back ch, in nach (Germ.). g, in tage (Germ.). Top h, in ich (Germ.). y, Im yen. Point r, in three (Scotch). r, before vowel. Lip Blowing to cool. v, in wie (Germ.). Back mixed ch in leuch (Scotch). g (labialised, Germ.). Top mixed _ sh, in she. j (French), je. Point mixed s, in see. Zz, In zeal. Lip mixed — wh, in which. Ww, in we. There are still some consonants in which the breath issues, not by a central aperture, but laterally, whether on both sides or on one. These are in English :-— Breath. Voice. 1, in else. 1, in ells. th, in thin. th, in thine. f, in four. Vv, in voice. If instead of emitting breath, unvocalised or vocalised, through these channels, we shut them against the breath or against the voice, we get the consonants :— Breath. Voice. k, in key. @ (hard), gain. Sound between k and t. Sound between g and d. t, in town. d, in done. p, in poet. b, in bone. The English nasal consonants ng, a sound between ng and n, 2, and m, are formed by shutting the mouth 120 CHAPTER III. passage and emitting breath or voice through the nose. The nasal passage is closed when the soft palate is lifted, it is opened when the soft palate descends. Nasal consonants may be vocal and non- vocal. This gives us altogether forty-eight consonants. To these must, however, be added the h, as represent- ing a mere emission of breath, without any friction, and the whisper, produced by the narrowing of the throat-passage. This 1s, no doubt, a very imperfect sketch of Mr. Melville Bell’s system. It is particularly so, because I could not avail myself of the ingenious alphabet which he has framed in order to give a pictorial re- presentation to every one of his letters. Still it will give an idea both of the strong and the weak points of what he calls Visible Speech. The weak points Mr. Melville Bell is himself the first to admit. Both vowels and consonants admit in reality of so many minute variations that no system of notation can ever do Justice to them. The strength of the system consists in the classification of vowels and consonants, in their definition and their localisation. Critics eon- tend that his classification and subdivision of vowels and consonants has either been carried too far or not far enough. We saw that several of his letters were admitted by Mr. Melville Bell himself to be useless for spoken languages, as far as we know them, and it is certainly a fact that other elocutionists differ from Mr. Melville Bell in assigning to each of his categories the sounds known to us in English, French, and German. These critics may be, as Mr. Melville Bell suggests, in- —-. THE ALPHABET. AbD capable and prejudiced, still Mr. Sweet, Professor Sie- vers, Dr. Vietor, Dr. Paul Passy and others can hardly be classed as such. Indeed, on several points I feel inclined to agree with them. For practical purposes, more particularly for writ- ing down spoken dialects and languages not yet reduced to writing, any one of these systems will no doubt prove very useful. I have confined myself to that of Mr. Bell in its latest form (1887), as the most original and the most widely accepted system. Image of the Ear and Movement of the Tongue. We must not forget that in using any of these systems we have to learn not only how to pronounce, but likewise how to hear. The ear receives an im- pression, and the vocal organs have to make an effort to imitate that impression. Nothing is more difficult than to hear accurately what is spoken in a language which we do not understand. An American gentleman, long resident in Constantinople, writes :— ‘There is only one word in all my letters which I am certain (however they may be written) of not having spelt wrong, and that is the word bactshtasch, which signifies a present. I have heard it so often, and my ear is so accustomed to the sound, and my tongue to the pronunciation, that I am now certain [| am not wrong the hundredth part of a whisper or a lisp. There is no other word in the Turkish so well impressed on my mind, and so well remembered. Whatever else I have written, bachshtasch ! my earliest acquaintance in the Turkish language I shall never forget.’ ' The word intended is Bakhshish. 2 Constantinople and its Environs, by an American long resident, New York, 1835, vol. ii, p. 151; quoted in Marsh, Lectures, second series, p. 87. 1A) CHAPTER III. Yet even the best elocutionists are sometimes liable to strange illusions, and the sounds which they have correctly defined before uttering them, are by no means always the same, when uttered. The Chinese word which by French scholars is generally represented as eul, is rendered by different authorities 61, eulh, eull, vl, rll, urhl, rhl. It is curious that the same word is sounded at Canton 7, in Annam 72, in Japan 17.1 Well do I remember how long it took before I could hear that and was not ant, that of was ov, that tongue was twng. If one has once heard correctly, the effort of imita- tion is much less difficult. Nay, even in speaking our own language, our pro- nunciation is constantly varying, and if a man is asked to pronounce a word a second time, so that we may hear it better, he almost invariably pronounces it differently. If each letter is kept between the narrow limits assigned to it, much will have been gained, but we shall never get a really scientific classification of the sounds of the human voice till we can measure them, as we measure heat, light, and now electricity also. Helmholtz has shown how vowels may be analysed and reproduced according to their analysis. It is not impossible that the phonograph may in time supply students of acoustics with the means of measuring every shade of sound produced on the revolving cylin- der by the human voice. There are the impressions made by the point set to vibrate by the speaking voice. Why cannot these impressions be magnified so as to * Léon de Rosny, La Cochinehine, p. 294. THE ALPHABET. 123 become really Visible Speech, and to submit to actual measurement? Barlow’s experiments seem to me to point the way, but it is not for me to say more on such a subject. | The actual Alphabet. We now return to the humbler task of describing the vowels and consonants with which the student of the Science of Language has chiefly to deal. Their system is, no doubt, less perfect than the purely physiological system elaborated by Mr. Melville Bell. But we must not forget that they answer the purpose for which they were intended, inasmuch as the prin- cipal languages of the world have been able with that small array of vowels and consonants to express all they had tosay. They must be looked upon as typical sounds only, each admitting of a broad margin, i. e. of a considerable dialectic variety. The only question is with how many, or with how few of such typical sounds the work of language can be carried on. No one can fail to see, for instance, that the & has a different place of contact, as pronounced in king, care, car, coal, cool, and caw. In a physiological alphabet, therefore, we ought properly to have six /’s, nay even more, if we watch the k as followed by different consonants, asin ks, kl, kra. But for our own purposes one / is sufficient, and if we have to mark dynamic differences in the k, they do not concern its pronunciation, but rather its liability to labialisation in certain languages, a peculiarity unrecognised in any physiological al- phabet. The Sanskrit short a is pronounced very differently even by educated natives in different parts of India, 124, ; CHAPTER ITI. but we should always have to write it by a, even if pronounced like 6 or % Dynamically, however, Sanskrit @ represents three sounds, @, %, and 0, and though Sanskrit has dispensed with this threefold dialectic variety for the purpose of grammatical distinctions, as, for instance, in Greek réuve, érapor, and téwos, the scholar finds it useful to mark that latent distinction in the Sanskrit vowel a, though no human ear could ever detect it. I still think that for a right appreciation of the letters used in the Aryan languages nothing can exceed the usefulness of the old Indian Pratisikhyas, particularly that of the Rig-veda. Even the Semitic alphabet, though of a very different character, can to a great extent be accommodated within the broad categories established by the ancient phoneticians of India. All that I shall attempt here is to give diagrams of the position of the vocal organs required for the utterance of the principal vowels and consonants. These diagrams are very rough, and do not pretend to give more than an approximative picture. ‘For didac- tic purposes,’ as Professor Haeckel remarks,! ‘simple schematic figures are far more useful and instructive than pictures which preserve the greatest faithfulness to nature and are carried out with the greatest exacti- tude.’ Such minutely exact pictures may be seen, however, in Mr. Norman W. Kingsley’s article, [dlus- trations of the Articulation of the Tongue, in Techmer’s Zeitschrift, vol. ili. pp. 225-248. They are simply ' Haeckel, Ziele und Wege, p. 37. THE ALPHABET. 125 copies taken by a very ingenious process of the interior of the mouth while engaged in pronouncing certain vowels and consonants. But the author knows too well in how many different ways the same sound may be produced by different individuals, nay by the same individual, to wish us to accept these pictures as more than approximative. ‘It is not supposable, he writes, ‘that all persons in making the same sound place the active accessory organs—the tongue, palate, &e.—in the same identical position. Variations to a greater or less extent can be observed in every one. Exactly the same resonating cavity in shape is not likely to exist in any two mouths. With the fixed portion of any buccal cavity differing somewhat in form from every other, the changeable portions, such as the tongue and palate, adapt themselves to circumstances and produce a resonating cavity of the same clang-character. The variations in the position of the articulating organs as seen in different persons in producing the same sound are then understood. So long as the integrity of the accessory organs 1s preserved, a resonating cavity of like clang-character can be formed.’ Vowels. 1. In pronouncing w we round the lips and draw down the tongue so that the cavity of the mouth assumes the shape of a bottle without a neck. Such bottles give the deepest notes, and so does the vowel uw. According to Helmholtz its inherent tone 1S F. 126 CHAPTER III. EXAMPLES : ! Open syllable, long, who; Fr. ou; Germ. du. Open syllable, short fruttion ; Fr. owir; Germ. zuriick. Closed syllable, long, pool; Fr. poule ; Germ. Stuhl. Closed syllable, short, pull; Fr. pour; Germ. bunt. 2. If the lips be opened somewhat wider, and the tongue somewhat raised, we hear the sound of o. Its pitch, according to Helmholtz, is B’ flat. Fig. 6. EXAMPLES: Open syllable, long, ago; Fr. beau; Germ. Ofen. Open syllable, short, zoology; Fr. zoologie; Germ. Zoologie. Closed syllable, long, bone; Fr. cone; Germ. Mond. Closed syllable, short, soft; Fr. bol; Germ. fort. ' I give instances of short and long vowels, both in open and closed syllables (i.e, not followed or followed by consonants), because, in English THE ALPHABET. 1 3. If the lips are less rounded, and the tongue somewhat depressed, we hear the sound of a. Fico. 7 EXAMPLES: Open syllable, long, Arzgust (subs.); Fr. deest;' Germ. deest. Open sylable,short,augist (adj.); Fr. deest; Germ. deest. Closed syllable, long, nought; Fr. deest ; Germ. deest. Closed syllable, short, what ; Fr. deest ; Germ. ceest. 4. If the lips are wide open, and the tongue in its natural flat position, we hear the sound of a. In- herent pitch according to Helmholtz, B” flat. This seems the most natural position of the mouth in sing- ing; yet for the higher notes singers prefer the vowels e and 7, and they find it difficult to pronounce «@ and w on the highest.’ particularly, hardly any vowels pair when free or stopped. On the qualitative, and not only quantitative, difference between long and short vowels, see Briicke, J. c. p. 24, seq. and R. von Raumer. 1 A sound similar to this is said to exist in the dialect of Orleans and the centre of France. ? Briicke, p. 13. 128 CHAPTER III. Fig. 8. EXAMPLES: Open syllable, long, mamd@; Fr. bas; Germ. da. Open syllable, short, pépd ;! Fr. rabat; Germ. dabei. Closed syllable, long, pass; Fr. Basle; Germ. lahm. Closed syllable, short, deest; Fr. bal; Germ. Lamm. yt 5. If the lips are fairly open, and the back of the tongue raised towards the palate, the larynx being EXAMPLES : Open syllable, long, hay ; Fr, né; Germ. geh. Open syllable, short, derial; Fr. légal ; Germ. Gebet. Closed syllable, long, lake; Germ. geht. Closed syllable, short, debt; Fr. dette; Germ. Fett. * Ihave given papd as an instance of the short pure a in English, but even in this word children soon learn to pronounce pupaw instead of papd. The fact is that there is no short pure a in English, either in open or in closed syllables, and even in long syllables the pronunciation of the @ is seldom quite pure. According to the peculiarities of local THE ALPHABET. 129 raised at the same time, we hear the sound of e. The long @is seldom quite pure in English, and particu- larly in singing we clearly hear a furtive 7 at the end of this vowel, day sounding like dé%. The long o in the same manner is frequently followed by a short %, no sounding like né-%. The buccal tube resembles a bottle with a narrow neck. The natural pitch of ¢ is B’”’ flat or ¥’. 6. If we raise the tongue higher still, and narrow the lips, we hear the sound of 2. The buccal tube represents a bottle with a very narrow neck of no more than six centimétres from palate to lips. Such a bottle would answer to oc”. The natural pitch of 7 t/t) is said to be D’”’. EXAMPLES: Open syllable, long, he; Fr. vie; Germ. sve. Open syllable, short, pithy; Fr. vitesse ; Germ. Sibirien. Closed syllable, long, been; Fr. pire; Germ. mir. Closed syllable, short, been, pro- nounced bin; Fr. mirroir; Germ. mit. 7. There is, besides, the most troublesome of all vowels, the neutral vowel, sometimes called Uvrvocal, dialects we sometimes hear farm pronounced like fawirm, sometimes like fairm. The true pronunciation ofthe Italian @matd must be learnt in Italy. LY, K 130 CHAPTER ITI. better Unvocal. Professor Willis defines it as the natural vowel of the reed, Mr. Ellis as the voice in its least modified form. Some people hear it everywhere, others imagine they can distinguish various shades of it. If I could trust my own ear, I should say that this vowel was always pronounced with non-sonant or whispered breath; that it is in fact a whispered, not a voiced, vowel. We know it best in short closed syllables, such as but, dust, &e. _It is supposed to be long in absurd. Sir John Herschel hears but one and the same sound in spurt, assert, bird, virtue, dove, oven, double, blood. Sheridan and Smart imagine they can distinguish between the vowels heard in bird and work, in whirl’d and world. There is no doubt that in English unaccented syl- lables have a strong tendency towards it, e.g. against, ided, village, suppér, fully, muttén. Town sinks to tun or tw in Paddington, ford to furd or frd in Oxford; and though some of these pronun- ciations may still be considered as vulgar, they are nevertheless real. These are the principal vowels, and there are few languages in which they do not occur. But we have only to look to English, French, and German in order to perceive that there are many varieties of vocal sound besides these. There is the French wu, the Ger- man 7d, which lies between 7 and w;? asin French, lu, pur, sur; in German, friih, fiir, Stid, Stinde. Professor Helmholtz has fixed the natural pitch of % as a”. 1 Ellis, § 29, n, 7,1, and m are vocalic. * “While the tongue gets ready to pronounce 7, the lips assume the position required for w,’—Du Bois-Reymond, Kadmus, p. 150. THE ALPHABET. AN 3 j There is the French ew, the German 6, which lies be- tween.e and o, as in French pew, hewreua, peur, neuf ; German Konig, enport, or short in Bécke.! Professor Helmholtz has fixed the natural pitch of 6 as c’”” sharp. There is d@ as heard in bear, in German Védter, in French prétre, in Italian erba. Its natural pitch is Gat OreD: *. Several vowels as pronounced in English in un- accented syllables are what Briicke calls imperfect vowels. They have been ranged under their corre- sponding typical sounds, but they have a phonetic character of their own. Thus there is the peculiar short @ in closed syl- lables in English, such as hat, happy, man. It may be heard lengthened in the affected pronunciation of half. There is the peculiar short 2, as heard in the Eng- lish happy, reality, hit, knit. There is the short e in closed syllables, such as heard in English debt, bed, men, which if lengthened comes very near to the German @ in Védter, and the French ¢ in prétre, or é in pére, not quite the English there. | Lastly, there are the diphthongs, as heard in English by, boy, bow, which arise when, instead of pronouncing one vowel immediately after another with two efforts of the voice, we produce a sound during the change * The German 6, if shortened, seems to dwindle down to the neutral vowel, e.g. Ofen, ovens, but éffnen, to open. See Du Bois-Reymond, Kadmus, p. 173. With a little practice, however, we can perceive a difference between the vowel win English hut, and the vowel 6 in Ger- man Hérner ; and it is easy to distinguish between the German Gétter and the English gutter. Kee 132 CHAPTER III. from one position to the other that would be required for each vowel. If we change the « into the 7 position and pronounce a vowel, we hear ai as in aisle, A singer who has to sing I on a long note will often end by singing the Italian 7. If we change the a into the w position and pronounce a vowel, we hear aw, as in how. Here, too, we find many varieties, such as i, a1, ei, varying in different languages, nay in the dialects of one and the same language. This may seem a long and tedious list, though it is, in fact, butavery roughsketch,and I must refertothe works of Mr. Melville Bell, Ellis, and others, for many minute details in the chromatic scale of the vowels. Though the tube of the mouth, as modified by the tongue and the lips, is the principal determinant in the production of vowels, yet there are other agencies at work, the velum pendulum, the posterior wall of the pharynz, the greater or less elevation of the laryna, all contributing at times to modify the cavity of the throat. It is said that in pronouncing the high vowels, the bones of the skull participate in the vibration,! and it has been proved by irrefragable evidence that the velwm pen- dulwm is of very essential importance in the pronun- ciation of all vowels. Thus Professor Czermak,? by introducing a probe through the nose into the cavity of the pharynx, felt distinctly that the position of the velum was changed with each vowel; that it was lowest for a, and rose successively with e, 0, wu, 4, reaching its highest point with 7. ' Briicke, p. 16. * Sitzungsberichte der k. i. Akademie zu Wien (mathemat. natur- wissenschaftliche Classe), xxiv. p.5. Physiologische Vortrége, p. 114. THE ALPHABET, 1383 He likewise proved that the cavity of the nose was more or less firmly closed during the pronunciation of certain vowels. By introducing water into the nose he found that while he pronounced 7, wu, 0, the water would remain in the nose, but that it would pass into the fauces when he came to e, and still more when he uttered a.1 These two vowels, a and e, were the - only vowels which Leblane,? a young man whose larynx was completely closed, failed to pronounce. Nasal Vowels. If, instead of emitting the vowel sound freely through the mouth, we allow the velum pendulum to drop and the air to vibrate through the cavities which connect the nose with the pharynx, we hear the nasal vowels* so common in French, as wn, on, mm, an. It is not necessary that the air should actually pass through the nose; on the contrary, we may shut the nose, and thus increase the nasal twang. The only requisite is the removal of the velum, which, in ordinary vowels, covers the choanw more or less completely.‘ Consonants. There is no reason why languages should not have been entirely formed of vowels. There are words consisting of vowels only, such as Latin ¢o, I go; ea, 1 Funke, J. c. p. 676. * Bindseil, Abhandlungen zur allgemeinen vergleichenden Sprachlehre, 1838, p. 212. % Briicke, p. 27. * The different degrees of this closure were tested by the experiment of Prof. Czermak with a metal looking-glass applied to the nostrils during the pronunciation of pure and nasal vowels. Sitzwngsberichte der Wiener Akademie, xxviii. p. 575, xxix. p. 174. 134 CHAPTER III. she; eoa, eastern; the Greek Gideis (jiders, with high banks), but for its final s; the Hawaian hooiaioai, to testify, but for its initial breathing. Yet these very words show how unpleasant the effect of such a language would have been. Something else was wanted to supply the bones of language, namely, the consonants. Consonants are called in Sanskrit vyaiigana, which means ‘rendering distinct or mani- fest, while the vowels are called svara, sounds, from the same root which yielded suswrrus in Latin. As scholars are always fond of establishing general theories, however scanty the evidence at their dis- posal, we need not wonder that languages like the Hawaian, in which the vowels predominate to a very considerable extent, should on that very ground have been represented as primitive languages. It was readily supposed that the general progress of lan- guage was from the slightly articulated to the strongly articulated; and that the fewer the conso- nants, the older the language. Yet we have only to compare the Hawaian with other Polynesian lan- guages in order to see how erroneous this view would be. In these cognate languages the consonantal skeleton exists, and it is quite clear that these con- sonants were dropped in Hawaian. Consonants are much more apt to drop than to spring up. Dean Ramsay in his Reminiscences records a conversation between two Scotchmen, a shopman and a customer, relating to a plaid hanging at the shop- door. It consists entirely of vowels. Customer (inquiring the material) : Oo ? (wool). Shopman: Ay, 00 (yes, wool). THE ALPHABET. 135 Customer: A’ 00 ? (all wool). Shopman : Ay, a’ 00 (yes, all wool). Customer : A’ ae 00 ? (all same wool). Shopman : Ay, a’ ae 00 (yes, all same wool). Here we know that the consonants existed, but were dropt. Prof. Buschmann expresses the same opinion with regard to the Polynesian languages: ‘Mes recherches m’ont conduit 4 la conviction, que cet état de pauvreté phonique polynésienne n'est pas tant l'état naturel d'une langue prise & sa naissance, quune détérioration du type vigoureux des langues malaies occidentales, amenée par un peuple qui a peu de disposition pour varier les sons.’ The very name of Havai, or more correctly Hawaz’i, confirms this view. It is pronounced in the Samoan dialect, Saval’i Tahitian, Havai’1 Rarotongan, Avaiki Nukuhivan, Havaiki New Zealand, Hawaiki from which the original form may be inferred to have been Savaiki.? All consonants fall under the category of noises, and there are certain noises that could hardly be avoided even in a language which was meant to con- sist of vowels only. If we watch any musical instru- ments, we can easily perceive that their sounds are always preceded by certain noises, arising from the first impulses imparted to the air before it can pro- duce really musical sensations. We hear the puffing 1 Buschmann, Jles Marg. p. 36,59. Pott, Htymologische Forschungen, ss HAR ory fetal s 136 CHAPTER III. and panting of the siren, the scratching of the violin, the hammering of the pianoforte, the spitting of the flute. The same in speaking. If we send out our breath, intending it to be vocalised, we often hear the rushing out, the initial impulse produced by the inner air as it reaches the outer. Breathings. If we breathe freely, the glottis is wide open,’ and the breath emitted can be distinctly heard. Mere breathing, however, is not yet our h, or the spiritus usper, An intention is required to change mere breathing into 1; the velum pendulum has to assume its proper position, the larynx is stiffened, the glottis narrowed * in order to produce an accumulation and intensification of the breath; this breath is then jerked out by the action of the abdominal muscles. This is the / in its purest state, the Greek spiritus asper, free, as yet, from any degree of hoarseness that may be imparted to it by subsequent barriers. These barriers are formed by narrowing different portions of the larynx or the throat, and they have given rise, particularly in the Semitic languages and in some German dialects, to a great variety of guttural breath- ings which, even with the help of the laryngoscope, it ' Czermak, Physiologische Untersuchungen mit Garcia’s Kehlkopf- spregel, Sitzungsberichte der k.k. Akademie der Wissenschaften, vol. Xxxix. 1858, p. 563. * Czermak, Uber den Spiritus asper und lenis. Sonder-Abdruck aus dem LIT. Bande der Sitzungsberichte der kats. Akad. der Wissenschaften (December 7, 1865). Though Professor Ozermak is right in saying that the glottis is narrowed, if compared with its shape in men breathing, yet it is equally correct to say that the glottis for 4 is wide open as compared with its aperture in the pronunciation of other letters. THE ALPHABET. iba is difficult accurately to analyse or to describe. With regard to dead languages, as for instance the ancient Greek, it is a hopeless task to attempt to determine the exact formation of their true guttural breathings. But, without wishing to commit myself to any opinion as to the exact degree of harshness imparted by the ancient Greeks to their mveiua dacv, it will be con- venient to retain the name of spiritus asper for the _ least modified form of the guttural breathing. Now it is clearly possible, while the breath is thus passing through the more or less compressed throat, to bring the vocal chords near to each other, so that the breath in passing should produce a kind of friction or imperfect vibration. As the (‘), the spiritus asper, described before, is the type of all the modifications of non-sonant breath, this letter would be the type of all the modifications of sonant breath, or of ex- haled voice. The Sanskrit h must come very near to it, for it is described as a breath or wind, like s, but at the same time as sonant. As I wish to retain for the non-sonant breath, in its purest form, the name of spiritus usper, I should wish to assign to the typical form of sonant breath the name of spiritus lenis, without, however, committing myself to any opinion as to the exact pronunciation of mveipa Widdy in different parts of Greece, or at different periods in the history of the Greek languages.! * Professor Czermak, in trying to define the nature of the TVEULA yiAov in Greek, explains it as ‘the explosive sound at the beginning of a vowel where the tone breaks forth, having for its only, and often hardly perceptible, extraneous admixture, the peculiar acoustic pheno- menon of the first explosive opening of the glottis, appearing other- wise in its full strength and purity.’ Professor Czermak, in fact, seems 138 CHAPTER III. We often hear the spiritus lenis, like a slight bubble, if we listen to the pronunciation of any in1- to understand by mvedpua YAdv the coup de la glotte, the sound produced by the explosive contact of the two sides of the glottis. If that had been the Greek mvedpa yiAdv, the name would not have been chosen very happily, for the coup de la glotte is not the breath itself, the mvevya, but the sound produced by a check imposed upon the sonant breath, The adjective yAdv applied to mvedpa does by no means prove, as Pro- fessor Czermak imagines, that the mvedpua yAdv must have been formed, like the dpwva yAa, by an explosive opening of a complete contact. To a Greek such an idea had never occurred, and would certainly not have been conveyed by the adjective yiAdv. The adjective Ady is no doubt opposed to Sacv, but, aceording to the best authorities, the dpwva dacea were themselves pronounced originally by an explosive opening of a previous complete contact, @ being originally ph and not f. The fact is that the Greek classification of letters, and, in consequence, their terminology, were of the vaguest kind. They divided the dpwva or mutes into dacéa, i. e. rough letters, and into wad, i. e. letters that were without that roughness. The péoa, or medie, were supposed to stand between the two, but, if pressed on the subject, the Greeks would most likely have admitted that the péca, too, were free from the roughness of the 5acéa, and, in that sense,yAd. When they gave to mvevyua or breath, too, the name of dacv, all they meant to indicate by it was the roughness of the breathing, and this the Romans rendered very properly by spiritus asper. In mvedpa WAdy, therefore, we have really no more than a nega- tive definition of another breath which is free from roughness, and this the Romans understood so well that they did not translate mvedpa prov by spiritus tenuis, but by spirituslenis. The adjective yAor is likewise used in a merely negative sense in é YiAdy and d yiAdv. The natural meaning, therefore, of this term would seem to be a breath which is not rough, and in this sense I apply it to the sonant breath as just described. If the spiritus lenis in Greek had been what Professor Czermak asserts it was, it is strange that it should not have been ranged among the agwva wird. But these are questions which, at this distance of time, it is im- possible to answer positively. What is of importance to us is this, that it is possible to define the following four letters, the non-sonant glottal breath, the sonant glottal breath, the glottal non-sonant check, and the glottal sonant check. But though we can define these four letters, the three last are apt to run into each other in actual use. Nor is this to be wondered at, considering that in the glottal series the organs which check the breath are the same as those which impart to it its sonant nature. The change of simple breath (‘) into simple voice (’) implied a check of the forth-rushing breath, which, initially, might = , THE ALPHABET. 139 tial vowel, as in old, art, ache, ear, or if we pronounce ‘my hand, as it is pronounced by vulgar people, ‘my and.’ According to some physiologists,! and ac- cording to nearly all grammarians, this initial noise can be so far subdued as to become evanescent, and we all imagine that we can pronounce an initial vowel quite pure.*? Yet I believe the Greeks were right in admitting the spiritus lenis as inherent in all initial vowels that have not the spiritus asper ; and the laryngoscope clearly shows in all initial vowels a sustained narrowing of the vocal chords, quite distinct from the narrowing and sudden opening that takes place in the pronunciation of the h. There is another very important distinction between spiritus asper and lenis. It is impossible to sing the spiritus asper, that is to say, to make the breath which produces it, sonant. If we try to sing ha, the voice does not come out till the h is over, We might as well try to whistle and to sing at the same easily be mistaken for the check that constitutes the explosive tenuis ; nor would it be easy, in spite of the most hair-splitting definitions, to distinguish the sound of the glottal explosive media from that of the glottal sonant breath. Briicke doubts whether the glottal sonant breath can be ranged as a distinct letter. ‘Sonant consonants,’ he says (p. 85) ‘spring from non-sonant consonants simply by means of narrowing the glottis till it produces a sound ; and if this is done with the 4, the result must be the pure tone of the voice without any additional rustle.’ In strict logic this is true, but in actual language we neither get a perfectly pure (°),nor a perfectly pure (’), and the slightest trace of hoarseness would give to the (°) and to the (’) their peculiar consonantal body. 1 Briicke, p. 9. * Briicke, p, 85.‘ If in pronouncing the spiritus asper the glottis be narrowed, we hear the pure tone of the voice without any additional noise.’ The noise, however, is quite perceptible, particularly in the vow clandestina. | 140 CHAPTER III. time. The reason of thisis clear. If the breath that is to produce / is to become a tone, it must be checked by the vocal chords, but the very nature of h consists in the noise of the breath rushing forth wncehecked from the lungs to the outer air. The spiritus lenis, on the contrary, can be sounded, because, in pro- nouncing it, the breath is checked near the vocal chords, and changed to voice. The distinction which, with regard to the first breathing or spiritus, is commonly called asper and lenis, is the same which, in other letters, is known by the names of hard and soft, surd and sonant, tenuis and media.2 The peculiar character meant to be described by these terms, and the manner in which it is produced are the same throughout. The authors of the Pratisakhyas knew what has been confirmed by the laryngoscope, that, in pronouncing what are called ¢enues, hard or surd letters, the glottis is open, while, in pronouncing the medic, soft or sonant letters, the glottis is closed. In the first class of letters, the vocal chords are simply neutral; in the second, they are so close that, though not set to vibrate periodi- cally, they produce a hum, or what has been called a fricative noise (Reibungsgerausch). Anticipating the * See R. von Raumer, Gesammelte Schriften, p. 371, note. Johannes Miiller says, ‘The only continua which is quite mute and cannot be accompanied by the tone or the humming of the voice, is the h, the aspirate. If one attempts to pronounce the / loud, with the tone of the chordz vocales, the humming of the voice is not synchronous with the h, but follows it, and the aspiration vanishes as soon as the air is changed into tones by the chorde vocales.’ * Czermak, Physiologische Vortrdge, p. 120: ‘Die Reibungslaute zerfallen genau so wie die Verschlusslaute in weiche oder ténende, bei denen das Stimmritzengerausch oder der laute Stimmton mitlautet—und in harte oder tonlose, bei denen der Kehlkopf absolut still ist.’ THE ALPHABET. 141 distinction between &, t, p, and g, d, 6, I may quote here the description given by Professor Helmholtz of the general causes which produce their distinction. ‘The series of the media, b, d, g, he says, ‘ differs from that of the tenues, p, t, k, by this, that for the former the glottis is, at the time of consonantal open- ing, sufficiently narrowed to enable it to sound, or at Fig. 11. Fig. 12. \ \ ON YM wp AL AALK RC SS \ eae AF fit) COCO TA) Cha \) ' j ‘| ‘) j —(h); e.g. hand. *"—3 eg, least to produce the noise of the vow clandestina, or whisper, while it is wide open with the tenues,! and therefore unable to sound.’ ‘Medize are therefore accompanied by the tone of the voice, and this may even, when they begin a syllable, set in a moment before, and when they end a syllable, continue a moment after the opening of the mouth, because some air may be driven into the ' See Lepsius, Die Arabischen Sprachlaute, p. 108, line 1. 142° CHAPTER III. closed cavity of the mouth and support the sound of the vocal chords in the larynx.’ ‘Because of the narrowed glottis, the rush of the air is more moderate, the noise of the air less sharp than with the tenues, which are pronounced with the glottis wide open, so that a great mass of air may rush forth at once from the chest. We now return to an examination of the various modifications of the breath, in their double character of hard and soft, or surd and sonant. The simple breathing in its double character of surd and sonant, can be modified in eight different ways by interposing certain barriers or gates formed by the tongue, the soft and hard palate, the teeth, and the lips. If, instead of allowing the breath to escape freely from the lungs to the lips, we hem it in by a barrier formed by lifting the tongue against the uvula, we get the sound of ch, as heard in the German ach or the Scotch loch.? If, on the contrary, we slightly check the breath as it reaches that barrier, we get the sound which is heard when the g in the German word Tage 18 not pronounced as a media, but as a semi- vowel, Tage. ' This distinction is very lucidly described by R. von Raumer, Gesammelte Schriften, p. 444. He calls the hard letters flate, blown, the soft letters halate, exhaled. He observes that exhaled letters, though always sonant in English, are not so in other languages, and therefore divides the exhaled consonants, physiologically, into two classes, sonant and non-sonant. See also Investigations into the Laws of English Orthography and Pronunciation, by Prof. R. L. Tafel, New York, 1862. * The same sound occurs in some of the Dayak dialects of Borneo. See Surat Peminyuh Daya Sarawak, Reading Book for Land and Hill Dayaks, in the Sentah dialect. Singapore,1862. Printed at the Mission Press. . a THE ALPHABET, 143 A second barrier is formed by bringing the tongue in a more contracted state towards the point where the hard palate begins, a little beyond the point where the & is formed. Letting the breath pass this isthmus, we produce the sound ch as heard in the German China or ich, a sound very difficult to an Fig. 14. AAA = — a YS ese ay < \ \ y \\ ‘h (ch); e.g. Loch. y (ch); e.g. ich (German). *h (g) 3 e.g. Tage (German). y (y); &g. yea. Englishman, though approaching to the initial sound of words like hume, huge. If we soften or voice the breath as it reaches this barrier, we arrive at the familiar sound of y in year.? A third barrier, produced by advancing the tongue towards the teeth, modifies the breath into s, the voice ' Ellis, English Phonetics, § 47. * There is no evidence whatever that the Sanskrit palatal flatus J (s) was ever pronounced like ch in German China and ich. Most likely it was the assibilated sound which can be produced if, while keeping the organs in the position for German ch, we narrow the passage and strengthen the breath, This, however, is merely an hypothesis. 144, CHAPTER III. into z, the former completely surd, the latter sonant ; for instance, stm or rice; and seal or rise. A fourth barrier is formed by drawing the tongue back and giving it a more or less concave (retroussé) shape, so that we can distinctly see its lower surface brought in position towards the back of the upper teeth or the palate. By pressing the breath through \ A\\\y \" ,) \\\ \ s; e.g. the rise, rice, sin. z; e.g. to rise, zeal. 23; e 2. azure. this trough, we get the letter sh as heard in sharp, and s as heard in pleaswre, or j in the French jamais, the former mute, the latter sonant. The pronuncia- tion of the Sanskrit lingual sh requires a very elabor- ate position of the tongue, so that its lower surface should really strike the roof of the palate. But a much more simple and natural position, as described above, will produce nearly the same effect. A fifth barrier is produced by bringing the tip of the tongue almost point-blank against the back of THE ALPHABET. 145 the upper teeth, or, according to others, by placing it against the edge of the upper teeth, or even be- tween the edges of the upper and lower teeth. If, then, we emit the breath, we form the English th, if we emit the voice, the English dh; the former mute, as in breath, the latter intonable, as in to breathe, and both very difficult for a German to pronounce. NAVE i! . \ | UA \ th (p); e.g. breath. f; e.g. life. dh (8); e.g. to breathe. v; eg. to live. A sixth barrier is formed by bringing the lower lip against the upper teeth. This modifies the breath to f, the voice to v, as heard in life and to live, half and to halve. A seventh barrier is possible by bringing the two lips together. The sound there produced by the breath would be the sound which we make in blow- ing out a candle; it is not a favourite sound in civilised languages. If voiced, however, the sound LI: L 146 CHAPTER III. is very common; it is the w in German as heard in Quelle, i.e. Kwelle;+ also sometimes in the German Wind, &e. An eighth barrier is formed by slightly contracting and rounding the lips, instead of bringing them together flat against each other. Here the breath assumes the sound of wh (originally hw), in wheel, which; whereas the voice is the common English double w, as heard in weal.” We have thus examined eight modifications of the breath and voice, beginning with spiritus asper and lenis, and ending with Pee ihe the labial breathing of wh and w. They are all emitted either eruptively or prohibitively, and de- termined by certain nar- rowings of the mouth. Considering the great plability of the muscles of the tongue and the mouth, we can easily imagine other possible \ nalrowings; but with Ww (wh) ; e.g. which. the exception of some W; eo. we. A peculiar letters of the Semitic and African languages, we shall find these eight sufficient for our own immediate purposes. 1 Briicke, l.c. p. 34. * As my definition of the wh as a whispered counterpart of w, has been declared entirely false by an American critic, Mr. Whitney, and as I cannot pretend to speak with authority on the correct pronunciation .of English, to say nothing of American, I quote my authorities. Mr. THE ALPHABET, 147 The peculiar guttural sounds of the Arabs, which have given rise to so much discussion, have at last been scientifically defined by Professor Czermak. After hearing these letters pronounced by an Arab, he tried to imitate them, and by applying the laryn- goscope to himself, he was able to watch the exact. formation of the Hha and Ain, which constitute a separate class of guttural breathings in the Semitic languages. This is his account. If the glottis is narrowed and the vocal chords brought near toge- ther, not however in a straight parallel position, but distinctly notched in the middle, while, at the same time, the epiglottis is pressed down, then the stream of breath in passing assumes the character of the Arabic Hha, o> as different from h, the spiritus asper, the Arabic x. If this Hha is made sonant, it becomes Ain. Starting from the configuration as described for Hha, all that takes place in order to changeit into Ain is that the rims of the apertures left open for Hha are brought close together, so that the stream of air striking against them causes a vibration in the fisswra laryngea, and not, as for other sonant letters, in the real glottis. These ocular observations of Czermak,} Ellis, in his Universal Writing, p. 6, says: ‘ Also distinguish weal, wheel, veal, feel, where wh represents the whisper of w. Some ortho- epists and most foreigners confuse wh with hu.’ Mr. Bell, in his Prin- ciples of Speech, p. 52, says, ‘When the aperture of the lips is slightly enlarged by the separation of their anterior edges, and the breath passes between the inner edges of the lips, the effect is that of the English wh, w; the former being the voiceless, the latter the vocal form of the same articulation.’ * Sitzungsberichte der mathematisch-naturwissenschaftlichen Olasse der kaiserlichen Akademie der Wissenschaften, vol. xxix. p. 576, seq. Professor Lepsius, Die Arabischen Sprachlaute, has but partially L 2 148 CHAPTER III. coincide with the phonetic descriptions given by Arab grammarians, and particularly with Wallin’s account. If the vibration in the fissura laryngea takes place less regularly, the sound assumes the character of a trilled 7, the deep guttural 7 of the Low Saxons. The Arabic ¢ and ¢ I must continue to consider as near equivalents of the ch in loch and ’h in German tage, though the pronunciation of the approaches some- times to a trill, like the 7 grasseyé. Trills. We have to add to this class of letters two which are commonly called trills, the rand the/. They can be pronounced. both as sonant and surd, but they differ from the other modifications by a vibration of certain portions of the mouth. Many people are deficient in their pronunciation of the different 7's, which are well described by Mr. Ellis. ‘In the trills, he writes, the breath is emitted with sufficient force to cause a vibration, not merely of some membrane, but of some much more extensive soft part, as the uvula, tongue, or lips. In the Arabie grh (grhain), which is the same as the Northumberland burr (burgrh, Hégrhiut for Harriet), and the French Provencal 7 grasseyé (as, Paris c'est la France, Paghri ¢’est la Fgrhance), the uvula lies along the back part of the tongue, pointing to the teeth, and is very distinctly vibrated. If the tongue is more raised and the vibration indis- adopted the views of Briicke and Czermak on what they call the Gutturales Vere in Arabic. See also a curious controversy between Professor Briicke and Professor Lepsius, in the 12th volume of the Zeitschrift fiir vergleichende Sprachforschung. - 1 Universal Writing and Printing, by A. J. Ellis, B.A., 1856, p. 5. THE ALPHABET. . 149 tinct or very slight, the result is the English r in more, poor, while a still greater elevation of the tongue pro- duces the r as heard after palatal vowels, as hear, mere, fire. These trills are so vocal that they form distinct syllables, as swrf, serf, fur, fir, virtue, honour, and are with difficulty separable from the vowels. Hence, when a guttural vowel precedes, the effect of the 7 is scarcely audible. Thus laud and lord, Jather and farther, are scarcely distinguishable,’ Professor Helmholtz describes 7 and J as follows :— ‘In pronouncing r¢ the stream of-air ig periodically entirely interrupted by the trembling of the soft palate or of the tip of the tongue, and we then get an intermittent noise, the peculiar jarring quality of which is produced by these very intermissions. In pronouncing / the moving soft lateral edges of the tongue produce, not entire interruptions, but oscilla- tions in the force of air.’ ! If the lips are trilled the result is brh, a sound which children are fond of making, but which, like the corresponding spiritus asper, is of little import- ance in speaking. If the tongue is placed against the teeth, and its two lateral edges, or even one only, are made to vibrate, we hear the sound of J, which can easily be voiced, as well as the ¢. We have thus exhausted one class of letters which all agree in this, that they can be pronounced by themselves, and that their pronunciation can be con- tinued. In Greek, they are all included under the name of Hemiphona, or semi-vowels, while Sanskrit grammarians mention as their specific quality that, CeCe DL Lo, 150 ; CHAPTER III. in pronouncing them, the two organs, the active and passive, which are necessary for the production of all consonantal noises, are not allowed to touch each other, but only to approach. Checks or Mutes. We now come to the third and last class of letters, which are distinguished from all the rest by this, that for a time they stop the emission of breath altogether. They are called by the Greeks aphona, mutes, because they are without any voice. They are formed, as the Sanskrit grammarians say, by complete contact of the active and passive organs. They will require very little explanation. If we bring the root of the tongue against the soft palate, we hear the consonantal noise of k. If we bring the tongue against the teeth, we 1 In Panini, i. 1, 9, y, r, 1, v, are said to be pronounced with ishatsprishtam, slight touch; s,sh,s, h, with viveitam, opening, or ishadvivritam, slight opening, or asprishéam, no contact. THE ALPHABET. 15I hear the consonantal noise of ¢. If we bring the lower against the upper lip, we hear the consonantal noise of p. The real difference between those three articula- tions. consists in this, that in p, two flat sur- faces are struck against each other ; in ¢, a sharp against a flat surface ; in k a round against a hollow surface. These three principal contacts can be modified almost indefinitely, in some cases without percep- tibly altering the articu- lation. If we pronounce ku, ka, ki, the point of contact between tongue and palate advances con- siderably without much influence on the character of the initial consonant. The same applies to the ¢ contact.’ Here the essential point is that the tongue should strike against the wall formed by the teeth. But this contact may be effected— 1. By flattening the tongue and bringing its edge against the alveolar part of the palate. 2. By making the tongue convex, and bringing the lower surface against the dome of the palate (these are the lingual or cacumina] letters in Sanskrit).? 1 Briicke, p. 38. * Formerly called cerebral, a mistranslation of mfitrddhanya, thoughtlessly repeated by many Sanskrit scholars and retained by others, on the strange ground that the mistake is too absurd to mislead anybody. Briicke, p. 37. TD CHAPTER III. 3. By making the tongue convex, and bringing the upper surface against the palate, the tip against the lower teeth (dorsal ¢ in Bohemian). 4. By slightly opening the teeth and stopping the aperture by the rounded tongue, or by bringing the tongue against the teeth. Most languages have only one ¢, the first or the fourth, some have two; but we seldom find more than two sets of dentals distinguished phonetically in one and the same dialect. If we place the tongue in a position intermediate between the guttural and dental contact, we can pro- duce various consonantal sounds which go by the general name of palatal. The click that can be pro- duced by jerking the tongue, from the position in which ich and yea are formed, against the palate, shows the possibility of a definite and simple conso- nantal contact analogous to the two palatal breath- ings. Some physiologists, however, and among them Briicke,! maintain that ch in English and Italian consists of two letters, ¢ followed by sh, and should not be classed as a simple letter. In Sanskrit, how- ever, the palatal check, the k, must be treated as a single letter, for it does not lengthen a preceding short vowel, as all really double consonants would do. What the exact pronunciation of this palatal letter may have been at different periods of the history of Sanskrit, is impossible to say. It is curious, however, to observe that, while the simple & and g do not lengthen a preceding vowel, the aspirated kh does so, 1 Briicke, p. 638, seq. He would, however, distinguish these concrete consonants from groups of consonants, such as é, y. THE ALPHABET. 158 and is in consequence written kkh. The k, as some- times heard in English, in kind, card, cube, cow, sounding almost like kyind, cyard, cywbe, cyow, may give us an idea of the transition of k into ky, and finally into English ch. In the northern dialects of Jutland a distinct 7 is heard after k and g if followed by @, e, 0, 6; for instance, kjwv’, kjwr, qjekk, kjerk, skyell, instead of kee, ker, &e.! It is not always perceived that these three con- sonants k, t, p, and their modifications, represent in reality two quite different effects. If we say ka, the effect produced on the ear is very different from ak. In the first case the consonantal noise is produced by the sudden opening of the tongue and palate ; in the second by their shutting. This is still clearer in pa and ap. In pa we hear the noise of two doors opening, in ap of two doors shutting. In empire we hear only half ap; the-shutting takes place in the m, and the p is nothing but the opening of the lips. In topmost we hear likewise only halfa p; we hear the shutting, but the opening belongs to the m. The same in uppermost. It is on this ground that these letters have sometimes been called dividuc, or di- visible, as opposed to the first class, in which that difference does not exist ; for whether I say sa or as, the sound of s is exactly the same. Sonant Checks, or Mediz. We should now have finished our survey of the alphabet of nature, if it was not that the consonantal 1 See Kuhn’s Zeitschrift, xii. 147 ; M.M., On the Pronunciation of ¢ before e, 1, y, ae, eu, oe, in the Academy, 15 Febr. 1871. 154 CHAPTER III. stops k, t, p, are liable to certain modifications, which, as they are of great importance in the formation of language, deserve to be carefully considered. What is it that changes / into g and ng, t into d and n, p into band m? B is called a media, a soft letter, a sonant, in opposition to p, which is called a tenuis, a hard letter, or a surd. But what is meant by these terms? See Sir H. Rawlinson, Behistun, p. 146; Spiegel, Parst Gram- matik, p. 34. * Bindseil, p. 318; Pott, 2. c. xii. 453. PHONETIC CHANGE. 181 Persian, which admits Arabic words, has 31 con- sonants, of which 22 are really Persian, the rest Arabie. Arabic has 28 consonants. The Kafir (Zulu) has 26 consonants, besides the clicks. } | Hebrew has 23 consonants. English has 20 consonants. Greek has 17 consonants, of which 3 are compound. Latin has 17 consonants, of which 1 is compound. Mongolian has 17 or 18 consonants. Finnish has 11. Polynesian has 10 native consonantal sounds; no dialect has more—many have less.1 Some Australian languages have 8, with three variations. The Melanesian languages are richer in consonants. The poorest, the Duauru, has 12; others 18, 14 and more.® Causes of Phonetic Change. One of the strangest facts with which the student of language is confronted, and for which, as far as possible, he has to account, is the change of letters, both vowels and consonants. In one sense the language of Tennyson is the same as that of Shake- speare, that of Shakespeare the same as that of Chaucer. that of Chaucer the same as that of Alfred; and yet, * Cf. Hale, p. 231; Von der Gabelentz, Abhandlungen der philo- logisch-historischen Classe der Kiniglich-Stichsischen Gesellschaft der Wissenschaften, vol. iii, p. 258. Leipzig, 1861. * Hale, p. 482. * See Von der Gabelentz, J. c. 182 CHAPTER IV. when we see it written, the language of Alfred is so different that Tennyson himself would find it impossible to understand it. The same applies to all languages. Whether they have been reduced to writing, or whether they live only as spoken by the people, they all change, nay, we may add, they cannot help changing. When touching on the growth of language, as distinguished from the history of language,’ I pointed out as the main causes of this change Phonetic Decay and Dialectic Growth. Some scholars have objected to the name of Phonetic Decay, and, to avoid useless controversy, I am quite willing to call it Phonetic Variety or Change. Others have assigned different names to these two motive powers, distinguishing them as Successive Change and Parallel Variety, or in German, as Laut-wandel and Laut-wechsel. So long as these names are clearly defined, there is no objection to any one of them. Benfey” admitted, in addition, what he called grammatical change. This, however, is of a different character altogether. It is quite true that the change of div into deva and daiva, of lip into lecpo and leloipa, or of mensa into mensae may be called a change quite as much as that of hafoc into hawk. Butin all these cases the change has a purpose. It produces a change of meaning, and must be treated as intentional or dynamic. The changes, on the con- trary, of which we are here treating are not intentional, they are not meant to produce a change of meaning, * Vol.i. chap. 2. * Die Spaltung einer Sprache in mehrere lautverschiedene Sprachen, in Nachrichten der K. Gesellschaft cu Gottingen, 1877, 24 Aug. Se PHONETIC CHANGE. 183 and they require in consequence a totally different explanation. Difference between Phonetic Change and Dialectic Growth. Phonetic Change, which is generally, if not always, Phonetic Decay, is necessarily successive. Thus hawk presupposes A.S. hafoc, to lie presupposes A.S. licgan, and /éogan. And whatever may be said of the in- herent rights of language to shape words according to its own pleasure, we are perfectly justified in saying that dilwviwm was corrupted to deluge, that pipio was reduced to pigeon, and that sapius decayed and became sage. It is surely corruption or decay, if words like salvia and sapius can no longer be dis- tinguished, or when sonus, subundare, A.S. sund, swimming, and A.S. geswnd dwindle all down to sound. But whether we call this process decay or change, or, as some would prefer, growth and development, we can and ought to distinguish it very carefully from Dialectic Change or Growth. If we compare, for instance, the different dialects of Aryan speech, we ought not to treat modern German drei as a corruption of Gothic threis, nor Gothic threis as a corruption of Latin tres, nor Latin tres as a corruption of Greek treis, nor Greek treis as a corruption of Sanskrit trayas. All these are parallel, not successive forms, and no one can say which was before or after the other. The th in Gothic threis is as little a phonetic corruption of t, as t in Gothic twazi is a phonetic corruption of d in duo, or d in door a phonetic corruption of 6 in Greek thyra, or of f in foris. 184, CHAPTER IV. No doubt, in many cases the Sanskrit form seems to us phonetically more primitive than corresponding forms in Greek, Latin, or Gothic. But the principle holds good nevertheless that they cannot be descended one from the other. It is quite true also that we often see the same change of letters produced by Phonetic Decay and by Dialectic Growth, but we shall see that nevertheless the principle of these two kinds of change is different. It is differentiation in Dialectic Growth, it is dissolution in Phonetic Decay. Dialectic Change. It was formerly the fashion to speak of a Proto- Aryan language from which Sanskrit, Greek, Latin, Teutonic, Slavonic, and Celtic were all derived, just as French was derived from Latin, or English from Anglo-Saxon. That theory, however, has hardly held its own for a longer time than the theory which it was meant to replace, namely that all Aryan languages were derived from Sanskrit. And yet there was some truth in that theory, if only rightly understood. To imagine that there was a settled Proto-Aryan language, as settled as Sanskrit, and that it became modified afterwards, according to strict phonetic rules, is, no doubt, impossible. That process can be studied to great advantage in the transition of Sanskrit into Prakrit dialects. But we have only to study languages, before they are reduced to writing, in order to see that the natural state of language is always dialectic, and dialectic, not in the sense in which Italian, Spanish, and French are dialects, derived from Latin, but as we often find in PHONETIC CHANGE. 185 the smallest Polynesian island two or three dialects existing side by side, not one of which has a right to claim precedence before the others. Indifferentiated Letters. A very common feature in these spoken dialects is _ the uncertain character of their consonants and vowels. We imagine that in every language, whatever the number of letters may be, each letter must at all events be definite, a k, or a p, or a t, ag, ora b,orad. But that is not so. There are races, for instance, who are quite unable to distinguish, either in hearing or in speaking, between some of the most normal letters of our alphabet. Dr. Washington Mathews, in his description of the Hidatsas, whose language belongs to the Dacota stock, informs us that it is difficult to say whether they pronounce mia, wia, or bia for mother, dopa, nopa, lopa, or ropa for two. In the language of the Mohawks the word for man is written rongwe, longwe, ronkwe, or lonkwe.1 No two con- sonants seem to us more distinct than k and t. Never- theless, in the language of the Sandwich Islands, these two sounds run into one, and it seems next to im- possible for a foreigner to say whether what he hears is a guttural or a dental. Chamisso (Werke, ii. 76) states that in these islands & and ¢ have the same value, likewise v, /,and m; and he confesses (ii. 95) that though his ear was well schooled, he was always doubtful between d, dh, and s, between ch, k, and g. Thus we find that the same word is written by + See Horatio Hale, ‘An Experiment in Phonetics, Journal of the Anthropological Institute, 1885, p. 236. 186 CHAPTER IV. English missionaries with k, by French missionaries with ¢, and they both agree that it takes months of patient labour to teach a Hawaian youth the difference between k and t, g and d,l, n, and 7.1 When a boy is told to pronounce fish, he will say pihi, when told to repeat knife, he utters netpa. No wonder that under these circumstances the English word steel should appear in the Hawaian dictionary as kila. Double letters are not tolerated, hence st became t. No word ever ends in a consonant, hence final a had to be added; and ¢ being pronounced like k, steel was necessarily changed to kila.? Such a confusion between two prominent conso- nants like & and ¢ would destroy the very life of English. The distinction between carry and tarry, car and tar, key and tea, would simply be lost. Yet the Hawaian language struggles successfully against these disadvant- ages, and has stood the test of being used for a transla- tion of the whole Bible, without being found wanting. If we consider that 7+ is in many languages a guttural, and / a dental, we may place in the same category of wavering pronunciation the confusion between these two letters, r and J, a confusion re- marked not only in the Polynesian, but likewise in many of the African languages. Speaking of the Setchuana dialects, Dr. Bleek remarks: ‘One is jus- tified to consider r in these dialects as a sort of floating letter, and rather intermediate between J and 7, than a decided 7 sound.’ 3 * Chamisso, Works, vol. ii. p. 76. * Buschmann, Iles Marg. p. 103; Pott, Etym. Forsch. ii. p. 188. * Mr, Powell, in his Introduction to the study of Indian Language a ll Ree ee PHONETIC CHANGE, 187 It is this absence of differentiation in certain con- sonants which seems to me to account for several so-called phonetic changes in dialects of the same language, which otherwise would defy all principles of phonetics. We are told that the missionaries in Hawaii were so perplexed as to whether they ought © to write & or t, that at last they had to appeal to the king. The king decided in favour of k, and after that his own name, which Ellis, in his Polynesian Researches, wrote Tamehameha, was changed into Kamehameha, and has remained so ever since. Is it not clear, therefore, that if during a period when the pronunciation still wavered between & and t, certain families had migrated from Hawaii to other islands, two dialects might have arisen in time, the one without any k’s, the other without any ?’s? And yet it would be quite wrong to say that & had become t, or t had become & And is it not equally wrong therefore to say that because we find in Greek tettares, and in Sanskrit katvar, in Latin quatuor, therefore Greek t was changed into Sanskrit k, and into Latin gu, or vice versa ? I feel convinced therefore that the key to much of the phonetic diversity which we observe between Sanskrit, Greek, Latin, and the other Aryan lan- guages must be looked for in a previous state of language, in which, as in the Polynesian dialects, the principal points of consonantal contact were not yet felt as definitely separated from each other. (second edition, Washington, 1880, p. 12) has fully treated of these sounds, which he calls synthetic sownds, and has pointed out their importance for phonological studies. 188 CHAPTER IV. There is nothing to show that in thermés, Greek ever had an initial guttural, and to say that Sanskrit gh became Greek th, is in reality saying what is im- possible. No Sanskrit letter can become a Greek letter; in fact, no letter ever becomes. People pro- nounce letters, and they either pronounce them pro- perly or improperly. If the Greeks pronounced th in thermés properly, without any intention of pronounc- ing gh, then the th, instead of gh, requires another explanation, and I*cannot find a better one than the one just suggested. When we find three dialects, like Sanskrit, Greek, and Latin, exhibiting the same word with guttural, dental, and labial initials, we gain but little if we say that Greek is a modification of San- skrit, or Latin of Greek. No Greek ever took the Sanskrit word and modified it; but all three received it from a common source, in which its articulation was as yet so vague as easily to lend itself to these various interpretations in different colonies. Though we do not find in any Greek dialect the same mixture of guttural and dental contact which exists in the Hawaian language, it is by no means uncommon to find one Greek dialect preferring the dental,t when another prefers the guttural; and I do not see how this fact can be explained, unless we assume that in an earlier or, as it is now called, a prehistoric state of the Greek language the pronunciation fluctuated or hesitated between k and ¢. I should prefer this explanation likewise in many cases when we see in cognate languages or dialects * Doric, méka, xa, GdAoxa, for dre, bre, GAAoTE; Doric, dydqos ; JMolic, yvédos ; Doric, 54 for 7. PHONETIC CHANGE. 189 an interchange between surd, sonant, and aspirated letters. To an educated ear these three varieties are not less marked than the three different points of con- tact in kh, t, and p. It is not only in such highly cultivated languages as Sanskrit and Greek that these three grades, tenuis, media, and aspirate, are used for - the differentiation of words. In the Dacota (Sioux) language, as the Rev. S. R. Riggs informs us, a clear distinction is made between b, p, and an emphatic p. The same applies to dental, guttural, and palatal let- ters. Thus be is to hatch, pe is sharp, and p’e is close. Da is to ask, ta moon, ?a to die. Simply to say that k becomes kh, and kh becomes g, seems again a de- fiance of all principles of phonetics; unless an ex- planation can be given how and why such successive changes should take place. The Rev. W. Ridley, in his grammatical outlines of the Kamilaroi, Dippil, and Turrinbad languages, spoken by Australian aborigines (‘New South Wales, 1866, p- 4), remarks: ‘They habitually soften the sound of their mutes, so that it is difficult to determine, in many instances, whether the consonant sound is 0 or p,dort,g ork. Mr. Curr, in his instructive work on the ‘Australian Race’ (Melbourne, 1886), tells us that the sounds represented by our letters f, s, @, and z do not exist in the languages of Australia; 7, g, and v are of rare occurrence, and probably absent in many. The sound of ch is absent in some, but abounds in others. The same applies to 7 as an initial, while as a final it is rolled out in some districts with great force and harshness. It is then so different from our own 7, that aboriginal names Yarr and Walgerr have been 190 CHAPTER IV. written down as Yass and Walgett. ‘ In taking down vocabularies from the Blacks, he continues, ‘it is often difficult to decide whether certain sounds should be expressed by 0 or p, d or ¢, k or g, nor is it possi- ble, as far as my experience goes, to make the Blacks aware of these distinctions of sound.’' ‘No Poly- nesian dialect, says Mr. Hale, ‘makes any distinction between the sounds of b and p, d and ¢, g and k, l and r, or v and w.2 This is not a case, therefore, of phonetic corruption, of allowing an established &, t, p to sink down to g, d, b, or of simply suppressing the voice that was originally heard in g,d,6. It is a case analogous to what the Rev. John Batchelor observed among the Ainos.2 ‘7’? he writes, ‘is pronounced neither like ¢ nor d in English, but as something be- tween the two. The same may be said of p and 0,’ If colonies started to-morrow from any of these centres of language, what took place thousands of years ago, when the Hindus, Greeks, and Romans left their common home, would take place again. One colony would elaborate the indistinct, half-cuttural, half-dental articulation of their ancestors into a pure guttural; another into a pure dental; a third into a labial. One settlement would fix on the sonant, another on the surd consonants. The Romans who settled in Dacia, where their language still lives in the modern Wallachian, are said to have changed every gu, if followed by a, into p. They pronounce 1 See also Australian Vocabulary, by G. F. Moore, 1843, p.x; Lawes, Grammar of Motu Language, p. vii. * Hale, Polynesian Grammar, p. 2338. 3 Ainu-English-Japanese Dictionary and Grammar, by the Rev. John Batchelor ; Tokyo, 1889. PHONETIC CHANGE. 191 aqua as apa; equa as epa.' Are we to suppose that the Italian colonists of Dacia said aqua as long as they stayed on Italian soil, and changed aqua into apa as soon as they reached the Danube? Or may we not rather appeal to the fragments of the ancient dialects of Italy, as preserved in the Oscan and Um- brian inscriptions, which show that in different parts of Italy certain words were from the beginning fixed differently, thus justifying the assumption that the legions which settled in Dacia came from localities in which these Latin qu’s had always been pro- nounced as p’s?? It will, no doubt, sound to many classical scholars almost like blasphemy to explain the phenomena in the language of Homer and Horace, by supposing for both a background like that of the Polynesian dialects of the present day. Some comparative philologists, too, will rather admit what is called a degeneracy of gutturals sinking down to dentals and labials, than look for analogies to the Sandwich Islands. Yet the most important point is, that we should have clear conceptions of the words we are using, and I confess that I cannot conceive how in the word for four a real & in Sanskrit could become ¢ in Greek, or ¢ in 1 The Macedonian (Kutzo-Wallachian) changes pectus into keptu, pectine into keptine, Cf. Pott, Htym. F.ii.49. Of the Tegeza dialects, the northern entirely drops the p; the southern, in all grammatical ter- minations, either elide it or change it into k. Cf. Sir G. Grey’s Library, i.p. 159. In Sicilian dialects fiore and fiume appear as ciore and ciwe. Academy, 1871, p. 147. Some of these changes have been rightly explained as mere acoustic illusions, and as cases of metathesis 3 see Paul, Principien der Sprachgeschichte, p. 59. * The Oscans said pomtis instead of quinque, &c. See Mommsen, Unteritalische Dialecte, p, 289. 192 CHAPTER IV. Greek degenerate into f in Gothic. I do not doubt the phonetic possibility,—for what is impossible in Phonetics? I doubt the historical reality of such changes. I can conceive different definite sounds arising out of one indefinite sound ; and those who have visited the Polynesian islands describe this fact as taking place at the present day. What then takes place to-day, can have taken place thousands of years ago; and if we see.the same word beginning in San- skrit, Greek, and Latin, with k, ¢, or p, it would be sheer timidity to shrink from the conclusion that there was a time in which that word was pronounced less distinctly ; in short, in the same manner as the k and ¢ in Hawaian. -I am glad to say that this distinction between Dialectic Change and Phonetic Corruption, and the account given by me of the nature of Dialectic Change many years ago, though strongly opposed at first, has been accepted by some of the most thoughtful students of language. Ineed only mention Mr. Horatio Hale, in his article ‘On some doubtful or intermediate Arti- culations’ in the Journal of the Anthropological Insti- tute, 1885, p. 283, and M. Maspero, in his essay on the ‘Personal Pronouns in Egyptian’ in the Mémovres de la Société de Linguistique, Paris, 1872. Referring to the occurrence of k and ¢ in these pronouns, he writes :— ‘La solution la plus raisonnable de ce probleme me parait étre celle que M. Max Miller propose, afin d’expliquer la pré- férence que certains dialectes indo-européens accordent a la dentale, dans la plupart des cas ot d’autres dialectes de la méme famille admettent la gutturale. Au lieu de supposer ee ee ee) PHONETIC CHANGE. 193 une dégénérescence organique de l’articulation primitive, qui aurait permis & la gutturale de s’affaiblir en dentale, il faudrait supposer que Varticulation du pronon de la 2° personne flottait primitivement entre K et T. La prononciation ne séparalt nettement la gutturale de la dentale que pour attribuer 4 chacune d’elles le rdle special que nous lui connaissons.’ Phonetic Idiosyncrasies. Tt must be conceded that single individuals or single families may sometimes influence the fates of a lan- guage. Personal defects in pronunciation, at first congenital, may spread by imitation, and in that case it would sometimes become very difficult to decide whether the effect should be treated as coming under the category of Phonetic Decay or of Dialectic Growth. We know that many people cannot pronounce /, and they say 7 or even m instead. They say grass or crowds instead of glass and clouds. I have heard ritten instead of little. Others change + to d, and say downd instead of vownd. Others change J to d, and say dong for long. The defects of infantine pronun- ciation also must not be forgotten, and we know how long some children will say tat for cat, tiss for kiss, &c. It cannot be denied that all this may tell and pro- duce phonetic changes, due, not so much to muscular laziness as to muscular inaptitude. | The Rev. W. G. Lawes tells us, in the second edition of his Grammar and Vocabulary of the Language spoken by the Motu Tribe (New Guinea), Sydney, 1888, that when he first went to Niue or Savage Island, the old men pronounced ¢ before 7 and e¢ as t, the children as ts, while, at a later visit, this infantine ts had become the general pronunciation. Il, 0 194 CHAPTER IV. It should, however, be remembered, that in all these cases we can tell what is primitive, and what is recent, while we have no right to say that ¢ in Greek tessares is recent, simply because we find an initial euttural or labial in other Aryan languages. Even the fact that in this case the guttural is found in a larger number of Aryan languages than the dental, proves nothing as to its being more primitive than the dental. If an individual, or a family, or a tribe cannot pro- nounce a certain letter, or imagines it cannot pronounce it naturally, nothing remains but to substitute some other letter, as nearly allied to it as possible. The Romans, for instance, were by nature destitute of aspirated consonants. They had neither kh, th, ph, nor gh, dh, bh. There is no excuse whatever for supposing that they originally possessed these letters, and that they exchanged them afterwards for others. If phonetic experts can prove that the letters Ligaen and b, which we find in Latin when in Sanskrit we find gh, dh, bh, in Greek ch, th, and ph, require less effort, well and good. Only it does not follow that the Romans, or their most distant ancestors, ever made that effort and failed. As little as we can prove that — the Greeks ever said yepuds for Oepuds, because the Sanskrit has gharmds, can we postulate that the Romans ever said thormus, because the Greeks said Oepuds. These changes are due to dialectic variety, not to phonetic decay. f These idiosyncrasies have to be carefully studied, for each language has its own, and it would by no means follow that because a Latin / or even b corresponds to PHONETIC CHANGE. 195 a Sanskrit dh, therefore every dh in every language may lapse into f or 6. | Greek has a strong objection to words ending in consonants; in fact, it allows but three consonants, and all of them hémtphona, to be heard as finals. We only find n, 7, and s, seldom k, at the end of Greek words... The Roman had no such scruples. His words end with a guttural tenuis, such as hic, nunc; with a dental tenuis, such as sunt, est; and he only avoids a final labial tenuis as not melodious. We can hardly imagine Virgil, in his hexameters, uttering such words as lwmp, trump, or stump. Such tendencies or dis- positions, peculiar to each nation, must exercise considerable influence on the phonetic structure of a language, particularly if we consider that in the Aryan family the grammatical lifeblood throbs chiefly in the final letters. Th and F. We know that th in English is a perfectly easy and legitimate sound, Its pronunciation comes quite natural toan Englishman. But it requires a consider- able effort on the part of most foreigners. It probably did so on the part of the Romans, when trying to speak Anglo-Saxon. Hence it happened that instead of th we sometimes find f, the dental instead of the labial aspirate. At first sight, such a change may seem very violent. J remember well, when Burnouf pointed out that the modern Persian name Feridun was a corruption of the Zend Thraétona, how several scholars doubted the possibility of such a change. But we have only to look at the diagrams of th and f O 2 196 CHAPTER IV. to convince ourselves that the slightest movement of the lower lip towards the upper teeth would change the sound of th into f.1 Children sometimes begin with pronouncing f instead of th, nay it is often difficult to distinguish their f’s and th’s. In vulgar English, ‘nothing’ sounds sometimes like ‘ nufjing, and ‘ had another’ is made to rhyme with ‘did not love her. ? In Russian we know that the Greek 6 appears as /, e.g. Feodor instead of Theodor. Now here we have clearly a case of pho- netic corruption. Th is right, f is wrong. Th came first, f came after- wards. But this cor- ruption is not due to economy of muscular Fig. 26. aN habits and poculianaee ( on the part of foreigners thane ae who were forced by ex- (the dotted outline is th.) ternal circumstances to adopt a foreign language. Not being able to pro- nounce a sound which was strange to their buccal 1 See M. M., On Veda and Zendavesta, p. 32. Arendt, Beitrdge zur vergleichenden Sprachforschung, i. p. 425. 2 “On what principle is it that the Yorkshireman travelling between Huddersfield and Saddleworth reads the name of Slaithwaite station as Slawit, or that the Wriothesley family dwindles in the public mouth into the insignificance of Rockley?’ London Quarterly, Oct. 1864, p. 209. Bunyan’s rhymes prove that he must have pronounced daughter like dafter ; see Earle, Philology of the English Tongue, p. 127. PHONETIC CHANGE, 197 organs, they took what lawyers call the ci-pres, the nearest approach. It is generally easy therefore 0 represent the process of this kind of phonetic corruption by anatomical diagrams, showing the natural transition from one position of the vocal organs to the other. - Thus it can be clearly perceived from the following diagram,’ how the Latin clamare requires complete contact between the root of the tongue and the soft palate, which contact is merged by sudden transition into the dental position of the tongue with a vibration of its lateral edges. In Italian this lateral vibra- tion of the tongue is dropt, or rather is replaced by the slightest possible approach of the tongue towards the palate, which follows al- most involuntarily on the opening of the guttural contact, producing chia- mare, instead of clamare. The Spaniard slurs over the initial cuttural contact altogether; hethinkshe has pronounced it, though his tongue has never risen, and he glides at once into the / vibration, the opening of which is followed by the same mouillé sound which we observed in Italian. Fisy27; Clamare, chiamare, llamar. * This diagram was drawn by Professor Richard Owen. 198 CHAPTER IV. K and T. In some cases it is, no doubt, difficult to say why one letter should seem easier to pronounce than another. For instance, when a language possesses both the k and the ¢, it is difficult to see why in some words t should be changed into k. This case, however, is quite different from that of the indifferentiated letters of the Polynesian languages which we considered before. All we can say in this case is that to a certain class of people, the & contact must have appeared more natural, and that others imitated their peculiarity. The fact itself, however, cannot be doubted. In Canada the lower classes habitually pronounce ¢ as k, saying mékier and motkié for métier and mozteé! This cannot be due to the fact that in Canada French was a foreign language. For at home also the French language underwent the same corruption, chiefly among the lower classes. Thus Moliére in Le Médecion malgré lui, makes Jaqueline say hériquié instead of héritier. In the same play quarquié occurs for quar- tier, amiquié for anitié. M. Agnel, in his Observa- tions sur la prononciation et le langage rustique des environs de Paris, pp. 11, 28, testifies to the existence of the same corruption among the peasants near Paris and Havre, where charkier may be heard for charretier, abricokier for abricotier, crapu for trapu. In one case this corruption has affected even the classical French, for there seems to be a unanimous opinion that craindre stands for Latin tremere.* 1 Student's Manual of the English Language (Marsh and Smith), p. 349. 2 See also Metiviers, Dictionnaire Franco-normand, 1870, p. 5. The ee ee PHONETIC CHANGE. 199 In all these cases, however, it should be remembered that the ¢ was there at first, and that its change to k was not due to differentiation, but to the phonetic idio- synerasies of certain individuals or certain classes. Lastly, there are some cases where it seems very doubtful whether the ear of some of our phonetic - authorities may not be as much at fault as the pro- nunciation of certain speakers. While in the cases before mentioned a real ¢ dwindled down to k, we are told by Webster, in the Introduction to his English Dictionary, that in English the letters cl are often pronounced like tl, e.g. tlear and tlean for clear and clean, and gl like dl, dlory for glory. Webster is, no doubt, a great authority, still I doubt the accuracy of this observation, at least among educated people. Cause of Phonetic Decay. We now come to the question, What is the cause of Phonetic Decay? For many years it was the custom among comparative philologists, when treating of phonetic changes, to say that s has become 7, or that m has been dropt, s has been elided, a and 7 have been contracted, ¢ softened, d hardened, &c. The question why letters should thus ‘change or become’ was never asked. Curtius comprehended all these pro- cesses under the name of Verwitterung, a meta- phorical expression taken from the decay which is produced by storm and weather, as if letters were things by themselves, exposed to external influences, King of Siam when speaking of maitri, the Buddhist word for love, mentioned that some Sanskrit scholars pronounced it maikree; see Mrs. A. H. Leonowens, The Governess at the Siamese Court, 1870, p. 197. 200 CHAPTER IV. and liable to the ravages of time. I was the first, I believe, who ventured to ascribe phonetic change to its vera causa, namely, to a natural desire of econo- mising muscular exertion, to a vis inertiae, or, in simpler language, to human laziness. Every letter requires more or less of muscular exertion. There is a manly, sharp, and definite articulation, and there is an effeminate, vague, and indistinct utterance. The one requires a will, the other is a mere laisser-aller. The chief cause of phonetic degeneracy in language is when people shrink from the effort of: articulating each consonant and vowel; when they attempt to economise their breath and their muscular energy, when they lay con- siderable stress on one syllable, and in consequence slur over the rest. It is perfectly true that, for practical purposes, the shorter and easier a word, the better, as long as it conveys its meaning distinctly. Most Greek and Latin words are twice as long as they need be, and I do not mean to find fault with the Romanic nations, for having simplified the labour of speaking. If the provincial of Gaul came to say pere instead of pater, it was simply because he shrank from the trouble of lifting his tongue, and pushing it against his teeth. Pere required less strain on the will, and less expenditure of breath : hence it took the place of pdtrem. So in English, night requires less expenditure of muscular energy than ndcht or Nacht, as pronounced in Scotland and in Germany; and hence, as people always buy in the cheapest market, night found more customers than the more expensive terms. Nearly all the Te PHONETIC CHANGE. 201 changes that have taken place in the transition from Anglo-Saxon to modern English belong to this class. Thus :— A.S. hafoe became hawk A.S. niwiht became nought » deg 3 day poehilatord 2s lord Beles Ol bn as fair ,» hléfdige ,, lady SOCIO aI! os, Say » séelig ns silly smrsprecan —,, speak » biton rf but Bertolpian * follow » héafod ee shead » morgen ,, morrow 5 nose-pyrel ,, nostril PE eacynine. 5; king » wWwif-man _,, woman saeworuld. =,, world! 5, Hofor-wic ,, York The same took place in Latin or French words naturalised in English. Thus :— Scutarius esculer = squire Historia histoire = story Egyptianus Egyptian = gipsy Extraneus estrangier = stranger Hydropsis —— = dropsy Capitulum chapitre = chapter Dominicella demoiselle = damsel Paralysis paralysie = palsy Sacristanus sacristain = sexton The best illustrations of the progress of phonetic decay are no doubt to be found in modern languages, such as the Romanic dialects in Europe, and the Pra- krit dialects in India. But the same process was going on in ancient languages also. Thus the Latin quintus stands for quinctus, just as Ital. santo stands ' Old High-German wér-alt=seculum, i.e. Menschenalter. Shake- speare in The Merry Wives of Windsor, iv. 4, 36, speaks of ‘ the super- stitious idle-headed eld.’ Cf. wérwulf, lycanthropus, werewolf, wihr- wolf, lowp-garou(l); were-gild, mann-geld, ransom, Cf. Grimm, Dentsche Grammatik, ii. 480. * See Lectures on the Science of Language, vol. i. p. 186. 202 CHAPTER IV. for sanctus. Umbrian mestrw shows phonetic cor- ruption more advanced than Italian maestro for magister. Umbrian deztw and fetw for dicito and facito represent but the first step which in the end led to Italian dite and fatte. There are, no doubt, some words in English which, if compared with their originals in Anglo-Saxon,seem to have added to their bulk, and thus to violate the general principle of simplification. Thus A.S. thwnor is in English thunder. Yet here, too, the change is due to laziness. It requires more exertion to with- draw the tongue from the teeth without allowing the opening of the dental contact to be heard than to slur from 7 on to d, and then only to the following vowel. The same expedient was found out by other languages. Thus, the Greek preferred to say dndres, instead of aneres; ambrosia, instead of amrosia.1 The French genre is more difficult to pronounce than gendre; hence the English gender, with its anomalous d. Similar instances in English are, to slumber =A.S. slumerian; embers=A.S. émyrian; humble = hu- mus. Euphony. It was formerly the custom of grammarians to ascribe these and similar changes to ewphony, or a desire to make words agreeable to the ear, the real object being to make them agreeable to the mouth— 1 In Greek p cannot stand before A and p, nor A before p, nor v before any liquid. Hence peonu(e)pla = peonuBpia ; yaupos =yauBpés ; Auaptov =7uBpotov ; woptés=Bpotrds. See Mehlhorn, Giriechische Grammatik, p- 54. In Tamil nr is pronounced ndr. Caldwell, Dravidian Gram- mar, p. 188, PHONETIC CHANGE. 208 that is to say, to save a certain amount of muscular effort. Greek, for instance, it was said, abhors two aspirates at the beginning of two successive syllables, because the repeated aspiration would offend delicate ears. Ifa verb in Greek, beginning with an aspirate, has to be reduplicated, the first syllable takes the tenuis instead of the aspirate. Thus thé in Greek forms tithémi, as dha in Sanskrit dadhami. If this were done simply for the sake of euphony, it would be difficult to account for many words in Greek far more inharmonious than thtthémz. Such words as xbév, chthon, earth, pOdyyos, phthdgqgos, vowel, begin- ning with two aspirates, were surely more objection- able than thithémi would have been. There is nothing to offend our ears in the Latin fefelli,1 from fallo, or in the Gothic reduplicated perfect hachald, from hal- dan, which in English is contracted into held,the A.S. being hedld, instead of hehold ; or even in the Gothic faifahum, we caught, from fahan, to catch.* There is nothing fearful in the sound of fearful, though both syllables begin with an f’ But if it be objected that 1 Tt should be remarked that the Latin f, though not an aspirated tenuis like #, but a labial flatus, seems to have had a very harsh sound. Quintilian, when regretting the absence in Latin of Greek ¢ and v, says, ‘Que si nostris literis (f et w) scribantur, surdum quiddam et bar- barum efficient, et velut in locum earum succedent tristes et horrid quibus Grecia caret. Nam et illa que est sexta nostratium (/) pene non humana voce, vel omnino non voce potius, inter discrimina dentium efflanda est ; que etiam cum vocalem proxima accipit, quassa quodam- modo utique quoties aliquam consonantem frangit, ut in hoc ipso frangit, multo fit horridior’ (xii. 10).—Cf. Bindseil, p. 287. r Pres. Perf. Sing. Perf. Plur. Part. Perf. Pass. Goth. haita haihait haihaitum haitan A.S. hatan héht (hét) héton haten O.E. hate hight highten hoten, hoot, hight. 204 CHAPTER IV. all these letters in Latin and Gothic are mere breath- ings, while the Greek x, 0, @ are real aspirates, we have in German such words as Pfropfenzieher, which to German ears is anything but an unpleasant sound. I believe the real cause of this so-called abhorrence in Greek is nothing but laziness. An aspirate requires great effort, though we are hardly aware of it, begin- ning from the abdominal muscles and ending in the muscles that open the glottis to its widest extent. It was in order to economise this muscular energy that the tenuis was substituted for the aspirate, though, of course, in cases only where it could be done without destroying the significancy of language. Euphony is a very vague and unscientific term. Each nation considers its own language, each tribe its own dialect, euphonic; and there are but few languages which please our ear, when heard for the first time. To my ear knight does not sound better than Knecht, though it may do so to an English ear; but there can be no doubt that it requires less effort to pronounce the English knight than the German Knecht. A desire for euphony seems to me in most cases but a disguised desire for a saving of muscular exertion, what is disagreeable to the ear being disagreeable to the voice also. There is no objection, however, to admit euphony as one of the less direct causes of pho- netic change. Thus the recurrence of the same letter in two successive syllables is often avoided, possibly for the sake of euphony, possibly for the sake of ease. There can be no doubt, for instance, that the two Latin derivatives aris and alis are one and the same. If we derive Saturnalis from Saturnus, and secularis PHONETIC CHANGE. 205 from seculum, normalis from norma, regularis from regula, astralis from astrwm, stellaris from stella, it is clear that the suffix in all is the same. Yet there is some kind of rule which determines whether alis or aris is to be preferred. If the body of the words contains an /, the Roman preferred the termination aris; hence secularis, regularis, stellaris, the only exceptions being that / is preserved (1) when there is also an 7 in the body of the word, and this 7 closer to the termination than the 7; hence pluralis, lateralis ; (2) when the / forms part of a compound consonant, as fluvialis, glaccalis.! The same explanation must probably be given for coeruleus from coelwm, for kephalargia and lethargia by the side of otalgia. All these are changes dependent on a dislike of the repetition of the same letter. But there are other changes of ¢ into r which it would be difficult to assign to euphony only, e.g. colonel, pronounced cuwr- mel (Old French, coronel; Spanish coronel); ros- signole = lusciniola.2 The Wallachian dor, desire, is supposed to be the same word as the Italian duolo, pain. In apétre, chapitre, esclandre, the same change of / into 7 has taken place.? On the other hand, 7 appears as / in Italian albero = arbor; celebro = cerebrum; mercoledi, Mercurii dies; pellegrino, pilgrim = peregrinus.* If certain scholars prefer to ascribe the change between two vowels of s into 7 in Latin, and the ' Cf. Pott, Etymologische Forschungen, 1st edit. ii. 97, where some exceptions, such as legalis, letalis, are explained. * See Corssen, Kritische Nachtrége, p. 86. * Diez, Vergleichende Grammatix, i. p.189. * Diez, lc. p. 209. 206 CHAPTER IV. dropping of s in Greek under the same circumstances, to a desire for euphony rather than to an economy of muscular energy, I see no objection, if only it is clearly understood that such changes are never in- tentional, but simply mechanical. To us it may seem as easy to say genesis as generis, yeveros as yeéveos and yevouvs. But we must remember that the nerves and muscles employed in speaking may assume certain habits and tendencies in each individual by imitation, and by inheritance in whole families and nations, and that what is easy and natural for pronunciation must be determined, in each case, by such habits and tendencies. Phonetic Habits. Though I have lived much longer in England than in Germany, and spoken more English than German, yet even now, after lecturing for one hour in English, the muscles of my throat feel tired, my throat becomes heated and dry, while in Germany I could lecture for two and three hours without any such feeling. What does this show? It shows that with me the combination of sounds peculiar to English requires a greater muscular effort, a greater exertion of will, than the usual run of sounds in German; but it does not prove that in themselves English sounds are more difficult to pronounce than German. Habit, whether self-formed or inherited, forms here as elsewhere ‘lines of least resistance,’ and these lines of least resistance determine what seems easy or difficult to pronounce in every lan- guage. PHONETIC CHANGE, 207 Double Consonants. We have still to treat of one other cause of Phonetic Decay, namely Double Consonants. Certain con- sonants, if they come together without intervening vowels, are troublesome to pronounce, particularly at . the beginning of words. Hence they are very hable to phonetic decay, either by being assimilated, or by one of these being dropt. But if it is the tendency of most languages to avoid or soften these troublesome combinations, we must not shirk the question, how it ever came to pass that such troublesome groups were framed and sanctioned. Strange as it may seem, I believe that these troublesome combinations of con- sonants were likewise the result of phonetic corruption, i.e. of muscular relaxation. Most of them owe their origin to contraction, that is to say, to an attempt to pronounce two syllables as one, and thus to save time and breath, though not without paying for it by an increased consonantal effort. It has been argued, with some plausibility, that language in its original state, of which, unfortunately, we know next to nothing, eschewed the contact of two or more consonants. There are languages still in existence in which each syllable consists either of a vowel, or of a vowel preceded by one consonant only, and in which no syllable ever ends ina consonant. This is the case, for instance, in the Polynesian languages. A Hawaian finds it almost impossible to pronounce two consonants together, and in learning English he has likewise the greatest difficulty in pronouncing cad, or any other word ending in a consonant. Cab, as pro- 208 CHAPTER ITV. nounced by a Hawaian, becomes caba. Mr. Hale, in his excellent ‘ Polynesian Grammar, ! says :— In all the Polynesian dialects every syllable must terminate in a vowel; and two consonants are never heard without a vowel between them. This rule admits of no exception what- ever, and it is chiefly to this peculiarity that the softness of these languages is to be attributed. The longest syllables have only three letters, a consonant and a diphthong, and many syllables consist of a single vowel. There are other languages besides the Polynesian, which never admit closed syllables, i.e. syllables ending in consonants. All syllables in Chinese are open or nasal,’ yet it is by no means certain whether the final consonants which have been pointed out in the vulgar dialects of China are to be considered as later additions, or whether they represent a more primitive state of the Chinese language. In South Africa all the members of the great family of speech, called by Dr. Bleek the Ba-ntu family, agree in general with regard to the simplicity of their syllables. Their syllables can begin with only one consonant, including, however, consonantal diphthongs, nasalised consonants, and combinations of clicks with other consonants reckoned for this pur- pose as substantially simple. The semi-vowel w, too, may intervene between a consonant and a following vowel. No syllable, as a general rule, in these South African languages, which extend north beyond the equator, can end in a consonant, but only in vowels, 1 Hale, J. ¢. p. 234. * Endlicher, Chinesische Grammatik, p. 112. PHONETIC CHANGE. 209 whether pure or nasal.! The exceptions serve but to prove the rule, for they are confined to cases where by the falling off of the generally extremely short and almost indistinct terminal vowel, an approach has been made to consonantal endings.” In the other family of South African speech, the Hottentot, compound consonants are equally eschewed at the beginning of words. It is clear, too, that all radical words ended there originally in vowels, and that the final consonants are entirely due to gram- matical terminations, such as », s, ts, and 7. By the frequent use of these suffixes the final vowel dis- appeared, but that it was there originally has been proved with sufficient evidence.? The permanent and by no means accidental or in- dividual character of these phonetic peculiarities is best seen in the treatment of foreign words. Practice willno doubt overcome the difficulty which a Hawaian feels in pronouncing two consonants together, or in ending his words by consonantal checks, and I have myself heard a Mohawk articulating his labial letters with perfect accuracy. Yet if we examine the foreign words adopted by the people into their own vocabu- lary, we shall easily see how they have all been placed on a bed of Procrustes. In the Ewe, a West-African language, school is pronounced suku, the German Fen- ster (window) fesre.4 * Bleek, Comparative Grammar, § 252; Appleyard, Kafir Language, 3 Sea Comparative Grammar, § 257; Hahn, Herero Grammar, § 3. > Bleek, Comparative Grammar, § 257-60. * Pott, Etymologische Forschungen, ii. 56. 1g & le 210 CHAPTER IV. In the Kafir language we find bapitizesha = to baptize ” 9 igolide = gold »” ” inkamela = camel ” ” ibere = bear ” ” umperisite = priest ” 5 ikerike == kirk ” » umposile = apostle » isugile = sugar 9 % ama-Ngezi = English’ If we look to the Finnish and the whole Uralic class of the Northern Turanian languages, we meet with the same disinclination to admit double conso- nants at the beginning, or any consonants whatever at the end of words. The German Glas is written last in Finnish. The Swedish smak is changed into maku, stor into swuri, strand into ranta. No genuine Finnish word begins with a double consonant, for the assibilated and softened consonants, which are spelt as double letters, were originally simple sounds. This applies equally to the languages of the Esths, Ostiakes, Hungarians, and Syrjanes, though, through their in- tercourse with Aryan nations, these tribes, and even the Fins, succeeded in mastering such difficult groups as pr, sp, st, str, &e. The Lap, the Mordvinian, and Tcheremissian dialects show, even in words which are of native growth, though absent in the cognate dia- lects, initial consonantal groups such as kr, ps, st, &e. ; but such groups are always the result of secondary formation, as has been fully proved by Professor Boller.2 The same careful scholar has shown that 1 Appleyard, Kafir Language, p. 89. ? Boller, Die Finnischen Sprachen, p.19. Pott, l. ec. pp. 40 and 56. See also Boehtlingk, Ueber die Sprache der Jakuten, § 152. ‘The Turko- PHONETIC CHANGE. AM the Finnish, though preferring syllables ending in vowels, has admitted n, s, 2, 7, and even ¢, as final consonants. The Esthonian, Lapponian, Mordvinian, Ostiakian, and Hungarian, by dropping or weakening their final and unaccented vowels, have acquired a large number of words ending in simple and double consonants ; but throughout the Uralic class, wherever we can trace the radical elements of language, we always find simple consonants and final vowels. We arrive at the same result, if we examine the syllabie structure of the Dravidian class of the South Turanian languages, the Tamil, Telugu, Canarese, Malayalam, &c. The Rev. R. Caldwell, in his excel- lent work, the ‘ Dravidian Comparative Grammar,’ has treated this subject with the same care as Professor Boller in his Essay on the Finnish languages, and we have only to place these accounts by the side of each other, in order to perceive the most extraordinary coincidences. The chief peculiarity of Dravidian syllabation is its extreme simplicity and dislike of compound or concurrent consonants ; and this peculiarity characterises the Tamil, the most early cultivated member of the family, in a more marked degree than any other Dravidian language. In Telugu, Canarese, and Malayalam, the great majority of Dravidian words, 1.e. words which have not been derived from Sanskrit, or altered through Sanskrit influences, and in Tamil all words without exception, including even Sanskrit deriva- tives, are divided into syllables on the following plan. Double or treble consonants at the beginning of syllables, like ‘stv,’ in ‘strength,’ are altogether inadmissible. At the beginning Tataric languages, the Mongolian, and Finnish show a strong aversion to double consonants at the beginning of words.’ P 2 2b. CHAPTER IV. not only of the first syllable of every word, but also of every succeeding syllable, only one consonant is allowed. If in the middle of a word of several syllables, one syllable ends with a consonant and the succeeding one commences with another consonant, the concurrent consonants must be euphonically assimilated, or else a vowel must be inserted between them. At the conclusion of a word, double and treble consonants, like ‘ gth,’ in ‘strength,’ are as inadmissible as at the beginning; and every word must terminate in Telugu and Canarese in a vowel; in Tamil, either in a vowel or in a single semivowel, as ‘1,’ or ‘r,’ or ina single nasal, as ‘n,’ or ‘m.’ It is obvious that this plan of syllabation is extremely unlike that of the Sanskrit. Generally, ‘1’ 1s the vowel which is used for the purpose of separating inadmissible consonants, as appears from the manner in which Sanskrit derivatives are Tamilised. Sometimes ‘u’ is employed instead of ‘i... Thus the Sanskrit preposition ‘pra’ is changed into ‘pira’ in the compound derivatives, which have been borrowed by the Tamil; whilst ‘Krishna’ becomes ‘Kirutfina-n’ (‘tt’ instead of ‘sh’), or even ‘Kiftina-n.’ Even such soft conjunctions of consonants as the Sanskrit ‘dya,’ ‘dva,’ ‘gya,’ &c., are separated in Tamil into ‘diya,’ ‘diva,’ and ‘ giya.’ 1 The Semitic languages are quite free from words beginning with two consonants without an inter- mediate vowel or shewa. ‘This is, in fact, considered by Ewald as one of the prominent characters of the Semitic family ;? and if foreign words like Plato have to be naturalised in Arabic, the » has to be changed to /, for Arabic, as we saw, has no p, and an initial vowel must be added, thus changing Platon into [flatin. It is hardly to be wondered at that evidence of this kind, which might be considerably increased, should ' Caldwell, Dravidian Comparative Grammar, p. 188. * Ewald, Gramm. Arabica, i. p. 23; Pott, Htym. Forsch. ii. 66. PHONETIC CHANGE. 213 have induced speculative scholars to look upon the original elements of language as necessarily consist- ing of open syllables, of one consonant followed by one vowel, or of a single vowel. The fact that lan- guages exist, in which this simple structure has been preserved, is certainly important, nor can it be denied, © that out of such simple elements languages have been formed, gradually advancing, by a suppression of vowels, to a state of strong consonantal harshness. The Tcheremissian sma, mouth, if derived from a root su, to speak, must originally have been swma. In the Aryan languages, the same process can easily be observed as producing the same effect, viz. double consonants, either at the beginning or at the end of words. It was in order to expedite the pronunciation of words that vowels were dropt, and consonants brought together: it was to facilitate the pronunciation of such words that one of the conso- nants was afterwards left out, and new vowels were added to render the pronunciation easier once more. Thus, to know points back to Sk. gfia, but this gia, the Lat. gndé in gndvi, or gnd in Gr. égnon, again points back to gan, contracted to gid. Many roots are formed by the same process, and they generally express a derivative idea. Thus gan, which means to create, to produce, and which we find in Sk. ganaz, Gr. génos, genus, kin, is raised to gan& and g fia, in order to express the idea of being able to produce. If I am able to produce music, I know music; if I am able to produce ploughing, I know how to plough, I can plough; and hence the frequent running together of the two conceptions, I can and I know, Ich kann 214, CHAPTER IV. and Ich kenne, Je sais and Je peux! As from gan we have gfia, so from man, to think (Sk. manas, Gr. ménos, mens, mind), we have mna, to learn by heart, Greek mémnémai, Lremember, mimnéesko. In modern pronunciation the m is dropt, and we pronounce m-nemonics. Again, we have in Sanskrit a root mlai, which means to fade; from it mldna, faded, m]ani, fading. Now, whence this initial double consonant ml? The Sanskrit root mlai or mla& is formed like gi& and mnA, from a simpler root mal or mar, which means to wear out, to decay. As gan became gia, so mar, mra. This mar is a very prolific root, of which more hereafter, and was chiefly used in the sense of decaying or dying, morior, dp(B)pdova, Old Slav. mréti, to die, Lith. mirtz, to die. These instances will suffice in order to show that in Sanskrit, too, and in the Aryan languages in general, the initial double consonants owe their existence to the same tendency which afterwards leads to their extinction. It was phonetic economy that reduced mara to mra; it was phonetic economy that reduced mra to ra and 1a. The double consonants being once there, the simplest process would seem to be to drop one of the two. This happens frequently, but by no means always. We see this process in English words such as knight, A.S. eniht; knife, A. 8. cnif; knee, A. S. enéo; to leap, A. 8. hléapan; ring, A. 8S. hring. We likewise observe it in Latin natus instead of gnatus, nodus instead of gnodus, English knot. We know ' Pott (#. F. ii. 291) compares queo and scio, tracing them to San- skrit ki. See Benfey, Kurze Sanskrit Grammatik, § 62; note. ; PHONETIC CHANGE. 215 that the old Latin form of locus was stlocus,) thus pointing to root std, whence the German Stelle; we know that instead of zs, itis, quarrel, litigation, the ancient Romans pronounced stlis, which has been compared with German streit. In all these cases the first consonant or consonants were simply dropt. Sometimes, however, a vowel is added again to facilitate the pronunciation. Many words in Latin begin with sc, st, sp. Some of these are found in Latin inscriptions of the fourth century after Christ spelt with an initial7: e.g. im istatwam (Orelli, 1,120, A.D. 875); Ispiritus (Mai, Coll. Vat. t. v. p. 446, 8).? It seems that the Celtic nations were unable to pro- nounce the initial s before a consonant, or at least that they disliked it.2 The Spaniards, even when reading Latin, pronounce estudiwm for studium, eschola for schola.* Hence the constant addition of the initial vowel in the Western or chiefly Celtic a CQuintil. i. 4, 16. 2 See Crecelius, in Hoefer’s Zeitschrift, iv. 166 ; Corssen, Aussprache, p-i. p. 289, 3 Richards, Antique Lingue Britannice Thesaurus (Bristol, 1753), as quoted by Pott, #. /. ii. 67, says (after letter S); ‘No British word begins with s, when a consonant or w follows, without setting y before it; for we do not say Sgubor, snoden, &c., but Ysgubor, ysnoden, And when we borrow any words from another language which begin with an s and a consonant immediately following it, we prefix a y before such words, as from the Latin schola, ysgol; spiritus, yspryd ; scutum, ysgwyd.’ 4 Tschudi, Peru, i.176. Caldwell, Dravidian Comparative Gram- mar, p. 170: ‘How perfectly in accordance with Tamil this is, is known to every European resident in Southern India, who has heard the natives speak of establishing an English iskool.’ This iskool is as good as establishing for stabilire ; or the Italian expressions, con istudio, per istrada, &c. ‘Tl en est de méme des mots germaniques devenus fran- gais, ainsi: stock, estoc; skarp, escarpé; skiff, esquif, &c.’—Terrien Poncel, Du Langage, p. 64. 216 CHAPTER IV. branch of the Romanic family; French escabeau, instead of Latin scabellum ; estame (étaim), Latin sta- men; espérer, instead of Latin sperare. Then again, as it were to revenge itself for the additional trouble caused by the initial double consonant, the French language throws away the s which had occasioned the addition of the initial e, but keeps the vowel which, after the loss of the s, would no longer be wanted. Thus spada became espée, lastly épée ; scala became eschelle, lastly échelle. Stabilive became establir, lastly établir, to stablish.! Different causes for Phonetic and Dialectic Change. Now it must be clear that all these changes which we have examined, whether due to economy of mus- cular exertion, or to what is called euphony, or to phonetic idiosyncrasies, rest on principles totally dis- tinct from those which made the Romans pronounce the same word as guatuor which we pronounce fowr. The transition from Gothie fidwér to English four, of Latin quatuor to French quatre, may properly be ascribed to phonetic corruption, but guatuor and jid- wor together can only be explained as the result of dialectic variation. If we compare quatuor, téssares, pésyres, and fidwér, we find a change of guttural, dental, and labial contact in one and the same word. There is nothing to show that the Greeks, or even their most distant Aryan ancestors, ever changed the guttural into the dental contact, or that the Teutonic nations ever considered the labial contact less difficult than the guttural and dental. We cannot show that * Diez, Grammatik, i. p. 224. PHONETIC CHANGE, a in Greece the guttural dwindles down to a dental, or that in German the labial is later, in chronological order, than the guttural. We must look upon gut- tural, dental, and labial as ‘three different phonetic expressions of the same general conception, not as corruptions of one definite original type. That which ~ is not yet differentiated may grow and break forth in many different forms; that which has become differ- entiated and definite, loses its capability of unbounded development, and its changes assume a downward tendency and must be considered as decay. Laws of Phonetic Change. What distinguishes phonetic from dialectic changes is that the former can be reduced to very strict rules, while the latter can not, at least not with the same unerring certainty. Phonetic decay, being due to a relaxation of muscular energy, admits of a simple physiological explanation, and depends on causes which are always the same. It is wrong, no doubt, to speak of phonetic laws in the same sense in which we speak of the law of gravitation. Phonetic laws can be no more than rules which are obeyed uni- formly, unless there is a cause sufficient to disturb them. It would be more correct therefore to speak of phonetic rules or of similarities in phonetic change. But the habit of speaking of phonetic laws has be- come so general that it would be very difficult now to change it. It stands to reason that the phonetic changes which are due to one and the same cause, namely muscular relaxation, must, unless there is a 218 CHAPTER IV. complete change of circumstances, be uniform and free from all exceptions. And this is so, not only in what may be called classical or well-regulated lan- guages, but likewise in spoken dialects, which have as yet no literary standards. In the growth of the modern Romanic languages out of Latin, we can perceive not only a general tendency to simplification, not only a natural disposition to avoid the exertion which the pronunciation of certain con- sonants, and still more, of groups of consonants, en- tails on the speaker: but we can discover tendencies peculiar to each of the Romanie dialects, and laws so strict as to enable us to say, that in French, and in French only, the Latin patrem would by necessity dwindle down to the modern pére. The final m is always dropped in the Romanic dialects, and it was dropped even in Latin. Thus we get patre instead of patrem. Now, a Latin ¢ between two vowels in such words as pater is invariably suppressed in French. Whether we call this a law, or a rule, or a tendency, certain it is that it admits of no exception. By means of it we can say a priort that Latin catena must in French become chaine; fata, a later feminine repre- sentation of the old neuter fatwm, fée; pratwm, a meadow, pré. From pratum we derive prataria, which in French becomes prairie; from fatum, fa- taria, the English fairy. Thus every Latin participle in atus, like amatus, loved, must end in French in @. The same law then changed patre (pronounced patere) into paere, or péere; it changed matrem into mere, fratrem into frére. These changes take place gradually, but irresistibly; and, what is most im- PHONETIC CHANGE, 219 portant, they are completely beyond the reach or control of the free will of man. Dialectic growth is equally beyond the control of individuals, but it does not submit to quite so strict and general rules. The acceptance of peculiar pro- nunciation, or of a dialectic word, or of a newly- - invented term, or of a peculiar grammatical form, depends on the pleasure of the majority far more than on the zeal of a single poet, or the exertions of a few srammarians. Phonetic changes of this kind are often the cause of grammatical changes. They can be accounted for after they have taken place, but they cannot be predicted with the same unvarying certainty as the phonetic changes due to muscular relaxation. Granted, for instance, that the loss of the Latin terminations was the natural result of a more careless pronunciation; granted that the modern sion of the French genitive dw is a natural corruption of the Latin de illo—yet the choice of de, instead of any other word, to express the genitive, the choice of illo, instead of any other pronoun, to express the article, could never have been predicted. No single individual could deliberately have set to work in order to abolish the old Latin genitive, and to replace it by the periphrastic compound de «lo. It was necessary that the inconvenience of having no distinct or distinguishable sign of the genitive should have been felt by the people at large who spoke a vulgar Latin dialect. It was necessary that the same people should have used the preposition de in such a manner as to lose sight of its original local meaning altogether (for instance, wna de multis, in Horace, i.e. one out of 220 CHAPTER IV. many). It was necessary,again, that the same people should have felt the want of an article, and should have used ] II. II IV. GOuniCu, 9s di 0 ke ay 48) ee thet IIL. Il. ; I. Greeks Ait ye 0rp ke? tip oa Weal III. II IV. High-German k t p (2) hide) fer ech 7 ties Let us now examine one or two more of these clusters of treble roots, like dhavr, dar, tar, and see how they burst forth under different climates from the soil of the Aryan languages. There are three roots, all beginning with a gut- tural and ending with the vocalised 7, In the ab- stract they may be represented as KAR, GAR, KHAR (or GHAR). In Sanskrit we meet first of all with GHAR, which soon sinks down to HAR, a root of which we shall have to say a great deal when we come to examine the growth of mythological ideas, but which for the present we may define as meaning to glitter, to be bright, to be happy, to burn, to be eager. In Greek this root appears in chatrein, to rejoice, &c. Gothic, following Sanskrit as far as it could, fixed 238 CHAPTER V. the same root as GAR, and formed from it geiro, desire; gairan and gairnjan, to desire, to yearn— derivatives which, though they seem to have taken a sense almost the contrary of that of the Greek chatrein, find. valuable analogies in the Sanskrit haryati, to desire, &e.! The High-German, foliow- ing Greek as far as possible, formed hiri, desire ; kerni, desiring, &. So much for the history of one root in the four representative languages, in San- skrit, Greek, Gothic, and High-German. We now come to a second root, represented in Sanskrit by GAR, to shout, to praise. There is no difficulty in Greek. Greek had not spent its media, and therefore exhibits the same root with the same consonants as Sanskrit, in gérys, voice; geryo, I proclaim. But what was Gothic to do, and the lan- guages which follow Gothic, Low-German, Anglo- Saxon, Old Norse? Having spent their medic on ghar, they must fall back on their tenues, and hence the Old Norse kalla, to call,? but not the A.S. galan, to yell. The name for crane is derived in Greek from the same root, géranos, meaning literally the shouter. In Anglo-Saxon eran and Old E. crane we find the corresponding tenuis. Lastly, the High-Ger- man, having spent its tenuis, has to fall back on its guttural breath; hence O.H.G. challén, to eall, and chranoh, erane. The third root, KAR, appears in Sanskrit as well as in Greek with its guttural tenuis. There is in 1 See Curtius, Griechische Etymologie, i. 166, and Objections, ibid. ii. 313. 2 Lottner in Kuhn’s Zeitschrift, xi. p. 165. GRIMM’S LAW. 239 Sanskrit kar, to make, to achieve; kratu, power, &e.; in Greek kratné, I achieve; and kratys, strong ; kdrtos, strength. Gothic having disposed both of its media and tenuis, has to employ its guttural spirant to represent the third series; hence hardus, hard, i.e. strong. The High-German, which naturally would have recourse to its unemployed media, pre- fers in the guttural series the Gothic spirant, giving us harti instead of garti, and thereby causing, in a limited sphere, that very disturbance the avoidance of which seems to be the secret spring of the whole process of the so-called Dislocation of Consonants, or Lautverschiebung. Again, there are in Sanskrit three roots ending in u, and differing from each other merely by the three dental initials, dh, d, and t. There is dht (dhu), to shake; du, to burn; and tu, to grow.! The first root, dhi, produces in Sanskrit dhfi-no- mi, I shake; dhti-ma, smoke (what is shaken or whirled about); dht-1li, dust. In Greek the same root yields thyo, to rush, as applied to rivers, storms, and the passions of the mind; thyella, storm ; thymés, wrath, spirit; in Latin, fwmus, smoke. In Gothic the Sanskrit aspirate dh is represented by d; hence dawns, vapour, smell. In Old High- German the Greek aspirate th is represented by ¢; hence twnst, storm. The second root, du, meaning to burn, both in a material and moral sense, yields in Sanskrit dava, conflagration; davath&é, inflammation, pain; in Greek dato, dédawmai, to burn; dé, misery. * See Curtius, Griechische Htymologie, i. 224, 196, 192. 240 CHAPTER V. Another Sanskrit root, du, to move about, to be busy, has as yet been met with in Sanskrit gram- marians only. But, besides the participle dina, mentioned by them, there is the participle dita, a messenger, one who is moved or sent about on busi- ness, and in this sense the root du may throw light on the origin of Gothic tawjan, German zauen, to do quickly, to speed an act.} The third root, tu, appears in Sanskrit as taviti, he grows, he is strong; in tavés, strong; tavisha, strong; tuvi (in comp.), strong; in Greek, as tajjs, great. The Latin tétus has been derived from the same root, though not without difficulty. The Um- brian and Oscan words for city, on the contrary, certainly came from that root, tuta, tota, from which tuticus in meddia tuticus,? town magistrate. In Lettish, tauta is people; in Old Irish, tuath. In Gothic we have thiuda,* people ; thiudisk-s, belonging to the people, theodiscus; thiudiskd, ethnikods; in Anglo-Saxon, thedn, to grow; thedd and theddisc, people; gethedd, language (il volgare). The High- German, which looks upon Sanskrit t and Gothic th as d, possesses the same word, as diot, people, diutisc, popularis ; hence Deutsch, German, and deuten, to ex- plain, lit. to Germanize. 1M. M., Rig-veda-Sanhitd, translated, vol. i. p. 63. * Aufrecht und Kirchhoff, Die Umbrischen Sprachdenkmiler, i. p. 155 ; Kuhn, Zeitschrift, vii. 166. See, for a new interpretation of meddix, Corssen, in Kuhn’s Zeitschrift, xi. 332. * Lottner, Kuhn’s Zettschrift, vii. 166. * Grimm, Deutsche Grammatik, first part, 8rd edition, 1840, Einleit- ung, p. x. ‘ Hxcurs tiber Germanisch und Deutsch.’ GRIMM’S LAW. | 241 Examples of Lautverschiebung. Let us now examine a few words which form the common property of the Aryan nations, and which existed in some form or other before Sanskrit was Sanskrit, Greek Greek, and Gothic Gothic. Some of them have not only the same radical, but likewise the same formative or derivative elements in ail the Aryan languages. These are, no doubt, the most in- _ teresting, because they belong to the earliest stages of Aryan speech, not only by their material, but likewise by their workmanship. Such a word as mother, for instance, has not only the same root in Sanskrit, Greek, Latin, German, Slavonic, and Celtic, namely, the root md, but likewise the same derivative tar throughout,! so that there can be no doubt that in the English mother we are handling the same word which in ages commonly called prehistoric, but in reality as historical as the days of Homer, or the more distant times of the Vedic Rishis, was framed to ex- press the original conception of genitriv. But there are other words which, though they differ in their derivative elements, are identical in their roots and in their meanings, so as to leave little doubt that, though they did not exist previous to the dispersion of the Aryans in exactly that form in which they are found in Greek or Sanskrit, they are nevertheless mere dialectic varieties, or modern modifications of earlier words. Thus star is not exactly the same word as stella; yet these two words show that, previous to 1 Sk. mata; Greek pyrnp; Lat. mater; O.H.G. muotar; O. Sh. mati; Lith. moti; Gaelic, mathair. II, R 242 CHAPTER V. the confusion of the Aryan tongues, the root star, to strew, was applied to the stars, as strewing about or sprinkling forth their sparkling light. In that sense we find the stars called stvz, plural staras, in the Veda. The Latin stella stands for sterula, and means a little star; the Gothic stair-nd is a new feminine derivative. As to the Greek astér, it is supposed to be derived from a different root, as, to shoot, and to mean the shooters of rays, the darters of light; but it can, with greater plausibility, be claimed for the same family as the Sanskrit star. ) It might be objected that this very word star violates the law which we are going to examine, though all philologists agree that it is a law that cannot be violated with impunity. But, as in other sciences, so in the science of language, a law is not violated, on the contrary, it is confirmed, by excep- tions, if a rational explanation can be given of them. Now the fact is that Grimm’s law is most strictly enforced on all initial consonants, but much less so on medial and final consonants. But whenever the tenuis is preceded at the beginning of words or syl- lables by an s, h, or f, these letters protect the f, f, p, and guard it against the execution of the law. Thus the root sté does not become sthd in Gothic; nor does the t at the end of noct-is become th, night being naht in Gothic. On the same ground, sé in stir and stella could not appear in Gothie as sth, but remain st as in stairnd. In selecting a few words to illustrate each of the nine cases in which the dislocation of consonants has taken place, I shall confine myself, as much as pos- GRIMM’S LAW. 943 sible, to words occurring in English; and I have to observe that, as a general rule, Anglo-Saxon stands throughout on the same step as Gothic. Consonants in the middle and at the end of words are liable to various disturbing influences, and I shall therefore dwell chiefly on the changes of initial consonants. Our first class consists of words which in English and Anglo-Saxon begin with the sonant g, d,and 0. If the same words exist in Sanskrit, we expect the aspirates gh, dh, bh, but never g, d, b, ork, t, p. In Greek we expect x, 6,@. In the other languages there can be no change, because they ignore the distinction between aspirates and sonant checks, except the Latin, which fluctuates between sonant checks and guttural and labial spirants. I. KH, Greek y; Sanskrit kh, gh, h; Latin h, f (g). G, Gothic g; Latin gv, g,v; Celtic g; Slavonic g, z. K, Old High-German k. The English yesterday is the Gothic gistra, the Anglo-Saxon geostra or geostrandeg, German gestern. The radical portion is gis, the derivative tra; just as in Latin hes-ternus, hes is the base, ternus the deriva- tive. In heri the s is changed to 7, because it stands between two vowels, like genus, generis. Now in Sanskrit we look for initial gh, or h, and so we find hyas, yesterday. In Greek we look for x, and so we find chthés. Old High-German, késtre. In Persian, di-ruz. - Corresponding to gall, bile, we find Greek cholz, Latin fel instead of hel. 1 Lotiner, Zeitschrift, vii. 167, R 2 94.4, CHAPTER Y. Similarly Gothic giu-ta, to pour out, is connected with Greek yéw, yu7és, and Sanskrit hu, to pour out libations, the Latin fundo, and futilis. The English goose, the A.S. gés, is the O.H.G. kans, the Modern German Gans! (It is a general rule in A.S. that before f, s, and th is dropped ; thus Goth. munth-s=A.S. mith, mouth; Latin dens, A.S. téth, tooth ; German ander, Sk. antara, A.S. éther, other.) In Greek we find chén, in Latin anser, instead of hanser, in Sanskrit hamsa, in Russian gus’, in Bohe- mian hus, well known as the name of the great reformer and martyr. Il. TH, Greek 6, ¢; Sanskrit th, dh; Latin f (b, d). D, Gothic d; Latin d, b; Celtic d; Slavonic d. T, Old High-German t. The English to dare is the Gothic gadawrsan, the Greek tharsein or tharretn, the Sanskrit dhvrzsh, the O.Sl. drizatt, O.H.G. tarran. The Homeric Ther- sites? may come from the same root, meaning the daring fellow. Greek, thrasys, bold, is Lithuanian drasus. The English doom means originally judgment ; hence, ‘final doom,’ the last judgment; Doomsday, the day of judgment. So in Gothic, dom-s is Judg- ment, sentence. If this word exists in Greek, it would be there derived from a root dhd or thé (tithémi), which means to place, to settle, and from which we have at least one derivative in a strictly legal- sense, namely, thémis, law, what is settled, then the ° oe of justice. 1 Curtius, G. H.:1. 222, : ? Curtius, G. H. i, 222. GRIMM’S LAW. 245 Professor Bréal has traced Latin fas to the same root. There is less reason why law, A.S. lagu, should not be connected with lex, and both be derived from a root “lah, to lay down (Aéxos, Lat. lectus), just as the German Gesetz was meant for what is settled, a statute. | sR et | Ill. PH, Greek #; Sanskrit ph, bh; Latin f (b). B, Gothic b; Latin b; Celtic and Slavonic b. P, Old High-German p. The A.S. béom, ‘I am,’ is the O.H.G. pi-m, the modern German bin, the Sanskrit bhavami, from a root which appears in the Greek phiio, and in Latin fui. The Gothic bédka! is the Latin fdgus, the O H.G. puocha. The Greek phégés, which is identically the same word, does not mean beech, but oak. It is diffi- cult to say whether this change of meaning was acci- dental, or whether there were circumstances by which it can be explained? Was phéqds originally the name of the oak, meaning the food-tree, from phagein, to eat? And was the name which originally belonged to the oak (the Quercus Esculus) transferred to the beech, after the age of stone with its fir-trees, and the age of bronze with its oak-trees, had passed away,? and the age of iron and of beech-trees had dawned on the shores of Europe? I hardly venture to say Yes; yet we shall meet with other words and other changes’ of meaning suggesting similar ideas, and encouraging 1 The A.S. b&ce, English beech, presupposes a Teutonic békd, fem. In buck-mast we have evidence of a former ddc. ? Sir Charles Lyell, Antiquity of Man, p. 9. 246 CHAPTER V. the student of language in looking upon these words as witnesses attesting more strikingly than flints and tags the presence of human life and Aryan language in Europe, previous to the beginning of history or tradition. | What is the English brim?! We say a glass is brim full, or we fill our glasses to the brim, which means simply ‘to the edge. We also speak of the brim of a hat, the German Brame. Now originally brim did not mean every kind of edge or verge, but only the line which separates the land from the sea. It is derived from the root bhram, which, as it ought, exhibits bh in Sanskrit, and means to whirl about, applied to fire, such as bhrama, the leaping flame, or to water, such as bhrama, a whirlpool, or to air, such as bhrémi, a whirlwind. Now what was called estus by the Romans, namely, the swell or surge of the sea, where the waves seemed to foam, to flame, and to smoke (hence estuary), the same point was called by the Teutonic nations the whirl, or the brim. After meaning the border-line between land and sea, it came to mean any border, though in the expression, ‘fill your glasses to the brim,’ we still imagine to see the original conception of the sea rushing or pouring in toward the dry land. In Greek we have a de- rivative verb phrimdsseivn,’ to toss about ; in Latin fremo, chiefly in the sense of raging or roaring, and ' 1 Kuhn, Zeitschrift, vi. 152. 2 Boéuw and Bpépos, which are compared by Kuhn, would violate the law; they express principally the sound, for instance in Bpovrn, ivi Bpeuérns, Curtius, G. H. ii.109. Grassmann, in Kuhn’s Zeitschrift, xii, 98. ee GRIMM’S LAW, 24:7 perhaps frendo, to gnash, are akin to this root. In the Teutonic languages other words of a totally dif- ferent character must be traced back to the same original conception of bhram, to whirl, to be confused, to be rolled up together, namely, bramble, broom, &e." We now proceed to the second class, namely, words which in Gothic and Anglo-Saxon are pronounced with k, t, p, and which, therefore, in all the other Indo-European languages, with the exception of Old High-German, ought to be pronounced with g, d, b. IV. G, Sanskrit g; Greek, Latin, and Celtic g; Slavonic g, z. K, Gothic k. KH, Old High-German ch. The English corn is the Gothie kawrn, Slavonic zrno, Lith. Zirnis. In Latin we find granum, in Sanskrit we may compare girna, ground down, though chiefly applied metaphorically to what is ground down or destroyed by old age. O.H.G. chorn. The English kin is Gothic kuni, AS. cynn, O.H.G. chunni. In Greek génos, Latin genus, Sk. ganas, we have the same word. The English child, A.S. cild, is in Old Saxon kind, the Greek génos, off- spring. The English queen is the Gothic géns, the A.S. cwén. It meant originally, like the Sanskrit g&ni, woman, because mother, just as king, the Ger- man kénig, the O.H.G. chuninc, the A.S. cyn-vng, meant originally, like Sk. ganaka, father.* Besides the forms with long vowel, the same word exists with 1 Brande, sorte de broussaille dans le Berry, bruyére & balai. 2 Brugmann, Vergleichende Grammatik, § 306. 3 See infra, p. 284. ‘948 - CHAPTER V. a short vowel, as Gothic qind, Old Saxon quéna, A.S. cwéne, Slav. Zena, Boet. Bava, Sanskrit ona.! The English knot is the Old Norse knitr, the Latin nodus, which stands for gnodus. V. OD, Sanskrit d; Greek, Latin, Celtic, Slavonic d. T, Gothic t. TH, Old High-German z. English: two is Gothie twat, O.H.G. zuet. In all other languages we get the initial soft d; Greek dwo, Latin duo, Lith. du, Slav. dva, Irish do. Dubius, doubtful, is derived from duo, two; and the same idea is expressed by the German Zweifel, Old High- German zwifal, Gothic tweifls. English tree is Gothic triw; in Sanskrit dru, wood and tree (daru, a log). In Greek drgs is tree, but especially the tree, namely, the oak.2 In Irish darach and in Welsh derw the meaning of oak is said to preponderate, though originally they meant tree in general. In Slavonic dijevo we have again the same word in the sense of tree. The Greek déry meant originally a wooden shaft, then a spear. English timber is Gothic timr or timbr, from which tumrjan, to build. We must compare it, therefore, with Greek démein to build, démos, house, Lat. domus, Sanskrit dama, the German Zimmer, room. VI. B, Sanskrit b or v; Greek, Latin, Celtic, and Slavonic b. P, Gothic p (scarce). PH, Old High-German ph or f. 1 See Brugmann, § 70. ; ? Schol. ad Hom. JI. xi. 86 Spurdpos, ae Spuv yap éxadcuv of madaol dd Tov dpxaiorépov nav 5évSpor. GRIMM’S LAW. 249 - There are few really Saxon words beginning with p, and there are no words in Gothic beginning with that letter, except foreign words, In Sanskrit, too, the consonant that ought to correspond to Gothic p, namely 4, is very seldom, if ever, an initial sound, its place being occupied by the labial spiritus v. | We now proceed to the third class, i.e. words begin- ning in English and Gothic with aspirates, or more properly with breathings, which necessitate in all other Aryan languages, except Old High-German, corresponding consonants such as k, t, p. In Old High-German the law breaks down. We find h and f instead of g and 6, and only in the dental series the media d@ has been preserved, corresponding to Sanskrit t and Gothic th. VII. K, Sanskrit k; Greek k; Latin c, qu; Old Irish ec, ch; Slavonic k. KH, Gothic h, g (f). Sanskrit h. G, Old High-German h (g, k). The English heart is the Gothic hazrté. Accord- ingly we find in Latin cor, cordis, in Greek kardia. In Sanskrit we should expect srizd, instead of which we find the irregular form hrzd. O.H.G. herza. The English hart, cervus, is the Anglo-Saxon heorot, the Old High-German hzruz. This points to Greek keraés, horned, from kéras, horn, and to cervus in Latin. The same root produced in Latin cornu, Gothic haurn, Old High-German horn. In Sk., siras is head, sringa, horn. The English who and what, though written with wh, are in Anglo-Saxon hwd and hwet, in Gothic hwas, hwé, hwa. Transliterating this into Sanskrit, 250 CHAPTER V. we get kas, ka, kad; Latin quis, que, quid; Greek kés and pds. VIII. T, Sanskrit t; Greek, Latin, Celtic, Slavonic t. TH, Gothic th and d. D, Old High-German d. The English that is the Gothic thata, the neuter of sa, sd, thata; A.S. se, sed, that; German der, die, das. In Sanskrit sa, s4, tad; in Greek, ho, hé, td. In the same manner three, Gothic threis, is Sanskrit trayas, High-German drev. Thou, Sanskrit tvam, Greek ty and sy, Latin tu, High-German du. Thin in Old Norse is thunnr, Sanskrit tanu-s, Latin tenwis, High-German diinn. IX. P, Sanskrit p; Greek, Latin, Celtic, Slavonic p. PH, Gothic f and b. B, Old High-German f and v. The last case is that of the labial spiritus in English or Gothic, which requires a hard labial as its sub- stitute in Sanskrit and the other Aryan dialects, except in Old High-German, where it mostly re- _appears as /f. The English to fare in ‘fare thee well’ corresponds to Greek pdros, a passage. Welfare, wohlfahrt, would be in Greek ewporia, opposed to aporta, helplessness. The English feather would correspond to a Sanskrit pattra, and this means a wing of a bird, i.e. the in- strument of flying, from pat, to fly,andtra. As to penna, it comes from the same root but is formed with another suffix. It would be in Sanskrit patana, pesna and penna in Latin, » GRIMM’S LAW. Q51 - The English friend is a participle present. The verb frijén in Gothic means to love; hence, fryénd, a lover. It is the Sanskrit pri, to love. The English few is the same word as the French pew. Few, however, is not borrowed from Norman- French, but the two are distant cousins. Pew goes. back to paucus; few to A.S. féawe, Gothic fawar; and this is the true Gothic representative of the Latin paucus. O.H.G. foh. GENERAL TABLE OF GRimmw’s LAw, 1 2 3 4/5] 6 7 8 9 { SektiGe ee | Sh) an (ny Lon (a) | een an. by k t p ERM 5 a ob 2 x ) ob ay pte fh Ve K T T Lauer eee cept (eV) f(a tet (bj) | eeaid- 1. b equ t p Old Irish . ; g d b g | da] b? | ¢(ch) | t (th) | (p)? Old Slavonic. . EZ d b gz|dj]{ b k t p Lithuanian . . 4 d b raat fol 1G k t Pp COUDICN, hus 2, g d b k | t | (p)?|be(f)| thd | fb Old High-German k t Pp ch |zz|fph}| hgk d fb The Theory of Grimm’s Law. So much for the facts comprehended under the name of Grimm’s Law. What is even more im- portant however than the facts, is the question, . whether they can be accounted for. Various theories have been started to account for this far-reaching change, and as they touch some of the fundamental principles of our science, we shall have to examine some of them more carefully. In spite of repeated protests, many scholars, chiefly encouraged by the example of Schleicher, will continue to treat consonants and vowels as things existing by themselves. They speak of a letter as 252 “CHAPTER V. produced at a certain time, then changing gradually by growing stronger or weaker, being assimilated or elided, and all this without any reference to the speaker, without whom after all no letter has any existence whatever. If scholars would always think clearly, and remain conscious of the metaphorical character of the language they are using, there would be little harm in their speaking of a Sanskrit dh being changed into a Greek 0, or of a Greek 6 being changed into a Gothic d. I am not so pedantic as to cavil at such statements, so long as they are used for the sake of brevity only. But when such phrases are taken literally, and when the change of Greek treis into Gothic threis, and Old High-German drei is represented as an historical process, it seems high time indeed to protest. Why have all accurate scholars so strongly protested against looking upon Sanskrit as the mother of Greek and Latin, if Greek, Latin, or Sanskrit may be represented as the mother of Gothic? Is Gothic to be treated as a more modern language than Sanskrit or Greek or Latin, because we happen to know it only in the fourth century of our era? And again, is Old High-German to be treated as a more modern dialect than Gothic, be- cause its literature dates from the eighth century only ? Are all the lessons of Greek dialectology to be thrown away, when we approach the dialects of Germany ? No Greek scholar would now venture to derive Attic from Doric, or Doric from Attic, nor would he allow the existence of a uniform Greek language, a kind of pre-Homeric Kow, from which the principal dialects of Greece were derived. Why then should we mete GRIMM’S LAW. 253 a different measure to German dialects, such as Low- German, High-German, and Scandinavian? Are Greek onviw, Lat. spuo, to spit, to be treated as phonetic corruption of Sk. shthyu (shthiw)? Is Sanskrit satam more modern than Latin centwm? There are rules of Dialectic Growth, though they are not so strict as the rules of Phonetic Decay. We may say, for instance, with perfect certainty that Sk. s never varies dialectically with Latin p, but we have no right to say that in the course of time kw dwindled down to p, or p to kw, however plausible the imperceptible degrees of phonetic transition between kw and p may be. If it is contrary to the principles of the Science of Language to derive Attic téssares from Doric tétores, or Doric tétures from Aeolic pessyres, why then should Old High-German drei be treated as the degenerate descendant of Gothic threis? No Sanskrit dh did ever become th, no Greek th did ever become Gothic d. Nay, we must go further and say that no Gothic d ever became a High-German ¢, as little as High-German ¢ ever became a Gothic d. Webeneinander and Nacheinander. The fact is that what Grimm called Lautverschie- bung has nothing to do with Phonetic Change, but is simply and solely a case of Dialectic Growth. Grimm looked upon Lautverschiebung as the result of a phonetic change, which took place very gradually. He actually fixed the beginning of the first change, the Gothic, about the second half of the first century A.D., and supposed that it was carried 254 CHAPTER V. through in the second and third centuries. More towards the West of Europe, he says, it may have commenced even at an earlier time, and have been succeeded by the second change, the Old High- German, the beginning of which is difficult to find, though we see it developed in the seventh century.’ There is one very plausible argument in support of this theory that the changes from d to ¢ and from ¢ to z were historical changes, following each other in regular succession, and that the first change from the classical to the Gothic stage took place about the second half of the first century after Christ, and the second change from the Gothic to the Old High- German stage about the sixth or seventh century. It is said that the name of Strassburg occurs in Gregory of Tours? (died 594) as Strataburgum ; in the Geo- grapher of Ravenna,’ in the middle of the seventh century, as Stratisburgo; whereas, in the eighth century, it has been changed into Strazpuruc. It is supposed, therefore, that, from the middle of the seventh to the middle of the eighth century, the third change took place, all mediw becoming tenues, all tenues becoming aspirate, and all aspirate medve. Now does anybody really believe that, some day or other, the people of Strassburg became aware that they called their town no longer Strataburgum but Strazpuruc, and that accordingly they changed the name in all official documents? Is there not a much more simple explanation, viz. that about the eighth century the High-German races became gradually 1 Grimm, Geschichte der Deutschen Sprache, i. p. 487. 2 Hist. Franc. ix. 36; x.16. _ 3 231, 7; 282, 2. GRIMM’S LAW. 255 more preponderant in Germany, whereas the Low- German tribes, the Goths and Saxons, in particular, disappeared more and more from the political and literary stage? In the famous Oath of Strassburg (A.D. 842) we still meet with such Low-German forms as dag, godes, thing. These High-German races, during their intercourse with their Low-German neighbours and enemies, had naturally become aware of the fact that, whenever they pronounced ¢, d, z, their neighbours pronounced di, th, t, and the same in the guttural and labial series. Under such circumstances a kind of habit became established, which led the speakers of High-German to replace without any conscious effort the sounds of Low-German by the corresponding sounds of High-German, and vice versd. We can watch the same curious process even now, when we try to speak a foreign language, and particularly when, while speaking High-German, we try to ex- press ourselves in Low-German.? Certain phonetic rules become established in our mind, which we obey without being aware of it. Thus, if the High- German tribes of the Frankish empire had once become impressed with the general idea, that where their Low-German predecessors or neighbours said k, t, p, g, d, b, h, th, f, they always said ch, 2, f, k, t, p, g, d, b, nothing was more natural than that they should apply the same rule to foreign words which they heard either from their Low-German compatriots or 1 A child which pronounced all 7’s as /’s was taught after some time how to pronounce the r. The result was, that it pronounced new words which really began with J with r, saying rong instead of long, &c, In Gaelic Pascha, Easter, is Caisg, in Welsh Pasg. 256 CHAPTER V. from the Roman provincials. Over and over again they had observed that, where in Low-German there was at, there was in their own language a z; there- fore, when they received a foreign word like Strata- burgum, they at once received it on the same terms, and changed Strata to Strdz. The second word was really German, and it would therefore at once be replaced by the High-German puruc. The same process 1s repeated in many foreign words which Old High-German borrowed either directly from Latin or indirectly from Low-German.! Thus pondus is in Gothic pund, in O.H.G. phunt; sinapi, G. sinap, O.H.G. senaf; persicum, O.H.G. phersich; cuprum, O.H.G. chuphar; strata, O.H.G. stréza; Turicum, O.H.G. Zurich; tegula, O.H..G. ziegal, &e. It is curious that O.H.G. zins, the Latin census, should in Old Saxon appear as tins. It is by no means neces- sary to suppose that these foreign words should all have passed through a Gothic channel before they reached Old High-German. Such a view would be necessary only if we looked upon Old High-German as the offspring of Gothic. All that is really re- quired for the explanation of the change of Latin words in Old High-German is to admit that the High-Germans possessed a phonetic sentiment which would lead them at once to translate any foreign t by z, d by t, th by d, and which therefore would make them adopt Strataburgum as Strazpwruc with- out a moment's thought as to whether it was origi- nally a Latin or a Low-German word, being satisfied 1 See W. ib sare fae, Die Umdeutschung fremder Worter. Basel, 1862. j GRIMM’S LAW. 957 that, before it should enter into High-German, it would have to submit to the same rules to which all other words seemed to have submitted. And if on these grounds I feel convinced that the consonantal system in High-German had become settled long before the seventh century, I feel equally certain that the consonantal system of Gothic does not date from the first century of our era only. We have no reason to suppose that what is called the classical system, or the first stage in Grimm’s Law, prevailed at any time in Gothic. The interesting researches of Dr. W. Thomson! have at all events established this fact, that at a much earlier period, when we see Low-German dialects, in some respects more primi- tive than Gothic, reflected on the surface of the Finnish language, their consonantal system was the same as at the time of Ulfilas. When we compare, for instance, ten, the A.S. tén, with Sanskrit dasan, Greek déka, Latin decem, we have no right to look upon ten as the result of phonetic corruption or decay. Yen may be called a phonetic corruption of a Teutonic typical form tehun (Gothic tahun), but tehun has as much right as Sanskrit dasan, so far as its consonantal structure is concerned. The loss of the medial h in tehwn, which represents an original k, is no doubt due to laziness of pronunciation. But not so the ¢ in place of d, or the / in place of s. These can be treated as dialectic only, i.e. as one out of many possible ways of permanently fixing the Aryan numeral ten, the pro- 1 Uber den Einfluss der Germanischen Sprachen auf die Finnisch- Lappischen, Halle, 1870, p. 124. iT; S) 258 CHAPTER V. nunciation of which must have varied from the first in various families, tribes, and nations, as we see it at the present day among tribes not united as yet by a common literature, whether in Africa, America, or Australia. Grimm’s Law in Africa and Polynesia. In Africa, for instance, we have what is meant by Grimm’s Law quite as much as in Europe. The various members of the Bantu family stand to each other very much in the same relation as Greek and Gothic. They share a large capital of words and forms in common, but they have at the same time diverged so much that even the members of the South-Eastern Branch of the Bantu family of speech, the Setshudna, Tekeza and Zulu Kafir, are now mutually unintelligible. As to deriving one from the other, it is impossible. They must therefore be treated as three independent varieties. And what do we see? Just what we see in Greek and Gothic. When Kafir has nasalised tenuis, Setshuana has the aspirate, Tekeza nasal only or spiritus asper or lenis. Kafir nk nt mp Setshuana, kh th ph Tekeza forwa n m There are exceptions, but Bleek, like Grassmann and Verner, has been able to account for most of them. Secondly, a nasalised media in Kafir and Tekeza appears as unnasalised tenuis in Setshuana. Kafir and Tekeza ng nd mb Setshuana k t p GRIMM’S LAW. 959 Thirdly, the nasalised v of Kafir and f of Tekeza (sometimes z) is represented in Setshuana by p. Kafir mv nz Tekeza nf mf Setshuana p p Fourthly, k, ¢, p in Kafir are represented in Set- shudna by x, 7, (the r being probably akin to 2), while the other cognate languages follow this rule: Kafir k t p Sesuto h f West Setshuana x r h Tekeza k OG. For further information on this subject I must refer to Bleek’s Comparative Grammar and to his article On Grinm’s Law on South Africa, It is curious that he too labours under the impression that some of these consonants must be looked upon as more primitive than others, and that therefore k is derived from ng, 7 from t, and not wice versa. But though this may be so in phonetic theory, it is not always so in historical truth, and Dr. Bleek has to confess, as we have, ‘that there are instances in which we are not quite certain of the direction which the current of transmutation has taken, and some in which it is quite possible that the different sounds occurring in the South-Eastern Branch languages are to be de- duced, not from each other, but from a primary form which is now only met with in other Bantu lan- guages. Thus, when a Kafir 2 corresponds to a Tekeza t, and to a Setshuana ¢s or yA—to which are we to give the palm of priority?’ Is this not exactly the same as when we have to say, ‘When an Old 8 2 260 CHAPTER V. High-German z corresponds to a Gothic ¢, and to a Sanskrit d—to which are we to give the palm of priority?’ Phonetically it may be to ¢, but historic- ally to none, because each represents an independent phase in the settlement of the language, such settle- ment taking place in different localities, and at different times, and, at all events in the beginning, not nacheinander, but nebeneinander. And not in Africa only, but wherever language can still be watched in its dialectic growth, phonetic phenomena which can be called by the name of Grimm’s Law have been discovered. Dr. Pope has an article in the Indian Antiquary (1876, p. 157) on Grimm’s Law as between Tamil and Kanarese, and changes analogous to the same Law and exhibiting the unsettled phonetic state of language previous to its being reduced to writing have been carefully described in Codrington’s Melanesian Languages (1885), pp. 198-219.1 Of course, phonetic rationalists will say: Surely, there must have been one primitive form for each word, and in this primitive form each consonant must have been fixed. If therefore there was an Aryan word for ten, its consonantal skeleton must have been D-K-N, which afterwards sank down in Gothic to T-H-N, in Old High-German to Z-H-N. But where is the must? First of all, the change of D to T, and of D to Z is in no sense of the word a sinking down, a weakening, or a corruption. Not ' See also Hale’s Polynesian Grammar, p. 232. The New Zealand poe is represented by foe in Tonga, just as Sk. pati is represented in Gothic by fath-s. GRIMM’S LAW. 261 even if we thought that the Old High-German form passed through an intermediate Gothic stage, would the change of T to Z be a corruption in the strict phonetic sense of the word. It involves no lightening of muscular effort, which is at the root of nearly all that is called phonetic corruption. But why should D-K-N be considered as the primitive form? Because it occurs in a majority of the Aryan languages? Fortunately majorities do not yet rule supreme in the Science of Language, which has often succeeded in discovering in one lonely so-called anomalous form the legitimate heir of a long line of ancestors. But let us take another word. Was the Sanskrit root BHAR more primitive than Greek PHER? Were both Greek pépw and Latin fero really derived from Sanskrit bhar&émi? And if not, why should Gothic bairan be an offshoot either of Sanskrit bhar, or of Greek and Latin fer, or possibly, like advimatar, of both? Again, when Gothic bairan stands to San- skrit bhar, exactly as O.H.G. peran does to Greek pher, why should O.H.G. peran be derived from Gothic bairan and not from Greek pher? Perhaps most scholars would be inclined, after a little reflection, to yield with regard to Gothic, and place it on a level with so-called classical languages, whether Sanskrit, Greek, or Latin. They would admit that the tenues are as good as the media, the mediz as good as the aspirates, whether surd or sonant, and that the aspirates or breathings are as good as the tenues. Was High-German derived from Gothic? But no such privilege is to be granted to High- 262 CHAPTER V. German. It is to be treated as a secondary language, as a corruption of Gothic, or at all events of some form of Low-German. Why isthat? Attic is more modern, and in many respects more corrupt than Doric. Does any scholar derive Attic from Doric? Is Welsh derived from Irish, or Spanish from Italian? Not even amado can be treated as a corruption of amato, though both presuppose a Latin amatus. What has the date of a literature to do with the age of a lan- guage? If High-German had come to our knowledge for the first time in Hebel’s Allemannische Gedichte, that would not make it modern as a language. The gradual spreading of High-German goes hand in hand with the spreading of High-German influence, whether _ political, religious, or literary. Whether it began in the fifth, or sixth, or seventh! century, itis still going on in the nineteenth. Braune (Bevtrdge,i. 1-56) tells us that the High-German change started from Oberdeutsch- land and spread northward, the first and most vigorous stage going furthest, the others getting weaker succes- sively. Under the first stage he comprehends the change of ¢ into z, of p and k after vowels into f and ch; under the second the-change of p’s, which had still been preserved (when initial, medial after con- sonants, and if strengthened) into ph, rarely into /; under the third the affrication of k and the change of the two remaining medic into tenues. He maintains that in Oberdeutschland the change in all its three stages is anterior to any of our literary documents, in Fran- conia the first stage completely so, while the second 1 Scherer, Geschichte der Deutschen Sprache, 1868, p. 63. See Die Sprache Deutschlands, von P. Piper, 1880, i. 223. GRIMM’S LAW. 263 can still be watched, and the third has never reached so far. The transition of th to d, he thinks, can be followed historically over the whole of Germany. In Oberdeutschland th vanishes in the second half of the eighth century, in East Franconia, saec. ix. init., in South Franconia, saec. ix. med., in Middle Franconia still later, and so likewise in Low Franconia. All this may be perfectly true, though the evidence is naturally very uncertain and fragmentary. But, if it is true,it proves no more than that certain phonetic changes rise to the surface at certain times, and reach certain literary and political centres at certain periods. It proves in no way that they spring into existence at the very moment when for the first time they become visible to us. In order to give an idea of the artificial contriv- ances which have to be resorted to if the changes comprised under Grimm’s Law are to be accounted for by the phonetic character of each letter, I shall give a few specimens of the more important theories. Grimm thought that the change began with the medize. Bopp thought that it began with the tenuis, which became an aspirate and an aspirate a media, When more minute physiological reasons were looked for to account for these changes, the great difficulty was, of course, to find out what exact sound was meant to be expressed by each letter in different MSS. of different writers in different parts of Germany and at different times. Always starting from the con- viction thata t became a th (z) and a th (z) ad, Raumer held that the aspirates contained a check and an aspiration, and that therefore when the pure spirant 264 CHAPTER V. had been reached (F and H) no further advance was possible. Hence he thought it was that Gothic f remained O.H.G. /, Gothic h O.H.G. h, while Gothic th varied between th, dh, and d. Between kh, th, ph, and g, d, 6, he admitted an intermediate stage, gh, dh, bh, and he looked upon the reinforcing of simple tenues and the vanishing of the aspiration in aspirates as the motive power of the whole process. Curtius ascribed the initiation of Lautverschiebung to the aspirates, which were changed into either mediz (Gothic) or tenues (O.H.G.). But when he ascribes these consonantal changes to ‘ vigour, boldness, and youthful energy,’ he is simply dealing in’ phonetic mythology, like many of his successors. If the change of d into ¢, and possibly even of ¢ into th, is youthful and vigorous, what is the change of th into d ? Scherer introduced still greater refinements, all based on the supposition that phonetic changes take place by slow degrees, and become more intelligible if we can account for every one of the minute degrees of change through which they passed. From a purely physiological point of view, such analytical researches are very useful, but as explaining an historical process they seem to be of very little help. I shall give one instance only. In order to explain the transition from Gothic th into O.H.G. d, Scherer writes: ‘It is more important for us to define as accurately as possible the pronunciation of the dh, which lies beyond the O.H.G. d, and the nature of this transition. We have here no other guide but English analogy. English s‘ (surd th) is a pure spirant, English 24 (sonant th) is often sounded with a slight initial check, as d+ 24,’ GRIMM’S LAW. 265 I doubt the fact, if Scherer means that there is more of d audible in thow than of ¢ in thin. But granting it, what should we gain ? Scherer continues: ‘This occasional, allowable, but not necessary check will be admissible likewise in the character of our O.H.G. dh. Nay, we may see in it, with Raumer, the very germ of the change, so that theoretically the sound to be changed would have to be represented by d* z+ (ddh). Hence it is not the spirant itself which is changed immediately into a media, but because the sonant spirant likes to take the support of a slight check, it might happen that this check was again deprived of the accompanying fricative sound.’ All this is very ingenious physiologically, but for our own historical purposes we gain nothing from it. Are we to suppose that one person, when he was a boy, said th, when a man, dh, and ddh, and when an old man, @; or that one generation said th, the next dh, the next ddh, the next d? Scherer himself shrinks from that conclusion, for he writes: ‘We must not look upon s*, 2*, d* 2‘, and d as four stages in a race- course, which had needs to be traversed before the poor hunted sound could find rest. D* z* may have been heard occasionally from the very first, after z* existed, and z* may have been heard occasionally to the very end, so long as there was d‘z*. Nay, from the begin- ning of the softening (becoming sonant) of th (s*) till the accomplishment of the change into d, the relation of the pronunciation dz‘ to the pronunciation 24 was probably unchangingly the same, and the former need — not have preponderated. If images could clear up — 266 ‘CHAPTER V. anything, I should say: the media hovers unseen over — the sonant spirant, and may appear at any moment ; and for that very reason it belongs to the nature of that sound.’ } I do not think that all this, not even the imagery, carries us further than the fact that instead of Gothic th, some Old High-German writers at different times and in different localities tried to indicate the sound which they heard, and which we ourselves shall never hear, by th, dh, and d, and that we may gather from their way of writing, that initially they heard some kind of aspiration besides the t or d, while medially that aspiration was not perceived, and therefore not written by them. As these attempts at writing what they imagined they heard, were the work of indi- viduals, we shall be much more justified in looking upon the changes which they tried to express in writing as scattered links of a lost chain than as representing what are called the slow and imper- ceptible degrees of transition in the same effort of pronunciation. Nothing is so fatal to all sound reasoning as this idea of minute and imperceptible degrees of transition. Everything can be explained by minute and imperceptible degrees of change, only we find that these imperceptible or almost impercep- tible degrees of change produce in the end no percep- tible result whatsoever. It does great credit to Mr. Sweet’s acumen as a phonetician that he formerly perceived this fallacy of imperceptible transition. In his History of English Sounds (p. 18) he says: ‘From this we can easily deduce another law, namely that the changes in early GRIMM’S LAW. 267 languages are not gradual, but per saltum. A clear appreciation of this principle is of considerable impor- tance, as many philologists have assumed that in such changes as that of a back into a front consonant (San- skrit k into k) the tongue was shifted forwards by imperceptible gradations,’ Exceptions to Grimm’s Law. Grimm’s Law is not without exceptions, but fortu- nately they are exceptions which prove the rule, that is to say, which can be accounted for from the very nature of the rule. Lottner. It was Lottner who in Kuhn's Zeitschrift, xi. 161, brought the first powerful indictment against Grimm’s Law, showing numerous cases and whole classes of cases in which it failed to act. Some of them had been pointed out by Grimm himself, more particu- larly with regard to Old High-German. Here, in fact, the exceptions were almost as numerous as the regular changes. Taking the texts of Isidorus, Otfried, and Tatianus as the principal representatives of Old High-German, Grimm constructed a table showing the different ways in which the Lautverschiebung was carried out by them. EPO ROE San Bh rad Egg f 6 ed < Cae 5 | eater EEE pericULttia.: t Fu fi ek CH Ht ie eer) Isidor, init.: B F G CH H D Z DH Tied sve El eV: G HH H D ZS -DH fine CAP He Cote te 75D. * The sign ! shows that the Lautverschiebung stops in Gothic. 268 CHAPTER V. Otfried, init.: B PH F GAs kK Fei D247 ante med.: _B oResh G CH H A Ar 8 in ee ae Eee sh ak dee Ar YE, Tatian, init.: B PH F Spo oa teil Lee med? sO Bs2H RV. G HH H TS Zee fin Curtius, U. ¢c, i. 141. 294, CHAPTER V. of the principal branches of the Aryan family, the coincidences between the Celtic and Teutonic names being of a doubtful character. If, then, we consider that the Sanskrit ayas, which meant, originally, the same as Latin ws and Gothic aiz, came to mean iron ; that the German word for iron is derived from Gothic aiz, and that Greek chalkés, after meaning copper, was used as a general name for metal, and conveyed occa- sionally the meaning of iron, we may conclude, I believe, that Sanskrit, Greek, Latin, and German were spoken before the discovery of iron, that each nation became acquainted with that most useful of all metals after the Aryan family was broken up, and that each of the Aryan languages coined its name for iron from its own resources, and marked it by its own national stamp, while it brought the names for gold, silver, and copper from the common treasury of their ancestral home. | Let us now apply the same line of reasoning to the names of fir, oak, and beech, and their varying significations. The Aryan tribes, all speaking dialects of one and the same language, who came to settle in Europe during the fir period, or the stone age, would naturally have known the fir-tree only. They called it by the same name which still exists in English as fir, in German as fohre. How was it, then, that the same word, as used in the Lombard dialect, means oak, and that a second dialectic form exists in modern German, meaning oak, and not fir? We can well imagine that the name of the fir-tree should, during _ the fir period, have become the appellative for tree in general, just as chalkds, copper, became the appella- APPENDIX. 295 tive for metal in general. But how could that name have been again individualised and attached to oak, unless the dialect to which it belonged had been living at a time when the fir vegetation was gradually re- placed by an oak vegetation? Although there is as little evidence of the Latin quercus having ever meant fir and not oak, as there is of the Gothic aiz having ever meant copper and not bronze, yet, if quercus 1s the same word as fir, I do not hesitate to postulate for it the pre-historic meaning of fir. That in some dia- lects the old name of fir should have retained its meaning, while in others it assumed that of oak, 1s in perfect harmony with what we observed before, viz. that ws retained its meaning in Latin, while ayas in Sanskrit assumed the sense of iron. The fact that phégés in Greek means oak,’ and oak only, while fagus in Latin, bék in Anglo-Saxon, mean beech, requires surely an explanation; and, until a better one can be given, I venture to suggest that Teutonic and Italic Aryans witnessed the transition of the oak period into the beech period, of the bronze age into the iron age, and that while the Greeks retained phéegés in its original sense, the Teutonic and Italian colonists transferred the name, as an appellative, to the new forests that were springing up in their wild homes. I am fully aware that many objections may be 1 In Persian, too, b#k is said to mean oak. No authority, however, has ever been given for that meaning, and it is left out in the last edition of Johnson’s Dictionary and in Vullers’ Lexicon Persico-Latinum. Though the Persian b¢@k, in the sense of oak, would considerably strengthen our argument, it is necessary to wait until the word has been properly authenticated. 296 CHAPTER V. urged against such an hypothesis. Migration from a fir-country into an oak-country, and from an oak- country into a beech-country, might be supposed to have caused these changes of meaning in the ancient Aryan words for fir and oak. I must leave it to the geologist and botanist to determine whether this is a more plausible explanation, and whether the changes of vegetation, as described above, took place in the same rotation over the whole of Europe, or in the North only. Again, the skulls found in the peat deposits are of the lowest type, and have been con- fidently ascribed to races of non-Aryan descent. In answer to this, I can only repeat my old protest, that the Science of Language has nothing to do with skulls.2 Lastly, the date thus assigned to the Aryan arrival in Europe will seem too far remote, particu- larly if it be considered that long before the first waves of the Aryan emigrants touched the shores of Europe, Turanian tribes, Fins, Laps, and Basks, must have roved through the forests of our continent. My answer is, that I feel the same difficulty myself, but * See M. M.’s Lectures on the Turanian Languages, p. 89 : ‘ Ethnology v. Phonology,’ * The same opinion has lately found a powerful supporter in Professor Huxley. I refer particularly to his paper ‘ On the Methods and Results of Ethnology,’ published in the Fortnightly Review, No. 3, June 15, 1865; and his lecture on the ‘ Forefathers of the English People,’ pub- lished in Nature, March 17, 1870. ‘If we confine our attention,’ he says, ‘to the British Islands, we have absolutely no means of ascribing any special physical characters to the Celtic-speaking people. A British or Irish “Celt” might be tall or short, dark or fair, round-headed or long-headed ; and the remark of Professor Max Miiller, that it is as rational to speak of a dolichocepha- lic language as of a Celtic skull, is, for the Celts of Britain, perfectly justified.’ APPENDIX. 297 that I have always considered a full statement of a difficulty a necessary step towards its solution. I shall be as much pleased to see my hypothesis refuted as to see it confirmed. All that I request for it is an impartial examination.1 * Some notes on the causes of the change of the vegetation in ancient Denmark, in G. P. Marsh, Man and Nature, p. 8, seq. 298 CHAPTER VI. CHAPTER VI. ON THE PRINCIPLES OF ETYMOLOGY. Guessing Etymology. OLTAIRE, as is well known, defined etymology ag a science in which vowels signify nothing at all, and consonants very little. ‘ L’étymologie,’ he said, ‘est une science ow les voyelles ne font rien, et les con- sonnes fort pew de chose. Nor was this sarcasm quite undeserved by those who wrote on etymology in Vol- taire’s time, and we need not wonder that a man so reluctant to believe in any miracles, should have declined to believe in the miracles of etymology. Of course, not even Voltaire was so great a sceptic as to maintain that the words of our modern languages had no etymology, i.e. no origin, at all. Words do not spring into life by an act of spontaneous genera- tion, and the words of modern languages in particular are in many cases so much like the words of ancient languages that no doubt is possible as to their real origin and derivation. Wherever there was a certain similarity in sound and meaning between French words and words belonging to Latin, German, He- brew, or any other tongue, even Voltaire would have acquiesced. No one, for instance, could ever have doubted that the French word for God, Dieu, was the same as the Latin Deus; that the French ON THE PRINCIPLES OF ETYMOLOGY. 299 homme and on came from Latin hominem and homo ; the French femme from the Latin femena. In these instances there had been no change of meaning, and the change of form, though the process by which. it took place remained unexplained, was not such as to startle even the sensitive conscience of Voltaire. There was indeed one department of etymology which had been cultivated with great success in Voltaire’s time, and even long before him, namely, the history of the Neo-Latin or Romanic dialects. We find in the dictionary of Du Cange a most valuable collection of extracts from medizval Latin writers, which enables us to trace, step by step, the gradual changes of form and meaning from ancient to modern Latin; and we have in the much ridiculed dictionary of Menage many an ingenious contribution towards tracing those mediseval Latin words in the earliest docu- ments of French literature, from the times of the Crusades to the Siécle of Louis XIV. Thus a mere reference to Montaigne, who wrote in the sixteenth century, is sufficient to prove that the modern French géner was originally gehenner. Montaigne writes: ‘Je me suis contraint et gehenné, meaning, ‘I have forced and tortured myself’ This verb gehenner is easily traced back to the Latin gehenna,' used in the Greek of the New Testament and in the ecclesiastical writings of the middle ages not only in the sense of hell, but in the more general sense of suffering and pain. It is well known that Gehenna was originally the name of the valley of Hinnom, near Jerusalem (2573), the Tophet, where the Jews burnt their sons 1 Molitre says, ‘Je sens de son courroux des génes trop cruelles.’ 300 CHAPTER VT. and their daughters in the fire, and of which Jere- miah prophesied that it should be called the valley of slaughter : for ‘They shall bury in Tophet till there be no place. How few persons think now of the sacrifices offered to Moloch in the valley of Hinnom when they ask their friends to make themselves com- fortable, and say, ‘Ne vous génez pas. It was well known not only to Voltaire, but even to Henri Estienne,? who wrote in the sixteenth century, that it is in Latin we may expect to find the original form and meaning of most of the words which fill the dictionaries of the French, Italian, 1 Jeremiah vii. 31, 32. ? Henri Estienne, Traicte de la Conformité du Langage Francois avec le Grec, 1566. What Estienne means by the conformité of French and Greek refers chiefly to syntactical peculiarities, common to both languages. ‘En une epistre Latine que je mi l’an passé audevant de quelques miens dialogues Grecs, ce propos m’eschappa, Quia multo majorem Gallica lingua cum Greca habet affinitatem quam Latina; et quidam tantum (absit invidia dicto) ut Gallos eo ipso quod nati sint Galli, maximum ad lingue Greece cognitionem mporépnua seu mAEovEK- Tha afferre putem.’ Estienne’s etymologies are mostly sensible and sober; those which are of a more doubtful character are marked as such by himself. It is not right, therefore, as is so often done, to class so great a scholar as H. Estienne together with Perion, and to charge him with having ignored the Latin origin of French. (See August Fuchs, Die Romanischen Sprachen, 1849, p.9.) What Estienne thought of Perion may be seen from the following extract (Zraicte de la Conformité, p- 139): ‘Il trouvera assez bo nombre de telles en un livre de nostre maistre Perion: je ne di pas seulemét de phantastiques, mais de sottes et ineptes, et si lourdes et asnieres que n’estoyent les autres temoignages que ce poure moine nous a laissez de sa lourderie et asnerie, on pour- roit penser son ceuvre estre supposée.’ LEstienne is wrongly charged with having derived admiral, French amiral, from dApvpds. He says it is Arabic, and so it is. It is the Arab Emir, prince, leader, possibly with the Arabic article. French amiral; Span. almirante; It. almi- raglio, as if from admirabilis. Hammer’s derivation from amir al bahr, commander of the sea, is untenable. ON THE PRINCIPLES OF ETYMOLOGY. 301 and Spanish languages. But these early etymolo- gists never knew of any test by which a true deriva- tion might be distinguished from a false one, except similarity of sound and meaning; and how far this similarity might be extended may be seen in such works as Perion’s Dialogi de Lingue Gallice Ori- gine (1557), or Guichard’s Harmonie Etymologique des Langues Hebraique, Chaldaique, Syriaque, Greque, Latine, Italienne, E'spagnole, Allemande, Flamende, Angloise (Paris, 1606). Perion derives brebis, sheep (the Italian berbice), from prdbaton, not from the Latin verve, like berger from berbicarius. Envoyer he de- rives from the Greek pémpein, not from the Latin enviare. Heureux he derives from the Greek owrios. Now, if we take the last instance, it is impossible to deny that there is a certain similarity of form and meaning between the Greek and French ; and as there can be no doubt that certain French words, such as parler, prétre, awmdne, were derived from Greek, it would have been very difficult to convince M. Perion that his derivation of hewreux was not quite as good as any other. ‘There is another etymology of the same word, according to which it is derived from the Latin hora. Bonheur is supposed to be bona hora; malheur, mala hora; and therefore hewreux is referred to a supposed Latin form, horosus, in the sense of fortunatus. This etymology, however, is no better than that of Perion. It is a guess, and no more, and it falls to the ground as soon as any of the more rigid tests of etymological science are applied to it. In this instance the test is very simple. There is, first of all, the gender of malheur and bonheur, 302 CHAPTER VI. masculine instead of feminine.t Secondly, we find that malhewr was spelt in Old French mal aiir, which is malum augurium. Thirdly, we find in Provengal agur, augur, and from it the Spanish aguéro, an omen. SS a a“ ¢ , bY f 5 x fa Oe eN \ a iL patra; éx 5€ THs THY dpoiwy édhAEiWEews’ Oia TL ATO pev THS Ppovngews / A Lea. \ la UA b) BA Fy / Ba A€éyouey ppovely, awd Se THs Suxacoovyys ove ETL Tapovopacomev 5 TYXN apa Ul ny e2 / kal ov pvoet TA OvopaTa, 396 CHAPTER VII. too, we have the pithy expression of ancient philo- sophy. Words are not natural images, images thrown by nature on the mirror of the soul; they are statues, works of art, only not in stone or brass, but in sound, Such is the opinion of Democritus, though we must take care not to stretch his words beyond their proper intent. If we translate thésec by artificial, we must not take artificial in the sense of arbitrary. If we translate némé by conventional, we must not take it to mean accidental. The same philosopher would, for instance, have maintained that what we call sweet or sour, warm or cold, is likewise so théset or conven- tionally, but by no means arbitrarily. The war-cries of physet or théset, which are heard through the whole history of these distant battles of thought, involved not only philosophical, but political, moral, religious interests. We shall best understand their meaning, if we watch their application to moral ideas. Philolaos, the famous Pythagorean philosopher, held that virtue existed by nature, not by institution. What did he mean? He meant what we mean when we say that virtue was not an invention of men who agreed to call some things good and others bad, but that there is a voice of conscience within us, the utterance of a divine law, independent of human statutes and traditions, self- evident, irrefragable. Yet even those who maintained that morality was but another name for legality, and that good and bad were simply conventional terms, insisted strongly on the broad distinction between law and the caprice of individuals. The same in lan- guage. When Democritus said that words were not ELEMENTS OF LANGUAGE. 397 natural images, natural echoes, but works of art in sound, he did not mean to degrade language to a mere conglomerate of sound. On the contrary, had he, with his terminology, ascribed language to nature, nature being with him the mere concurrence of atoms, he would have shown less insight into the origin, less regard for the law and order which pervade language. Language, he said, exists by institution ; but how he must have guarded his words against any possible misapprehension, how he must have protested against the confusion of the two ideas, conventional and arbi- trary, we may gather from the expression ascribed to him by a later scholiast, that words were statues in sound, but statues not made by the hands of men, but by the gods themselves.1_ The boldness and preg- nancy of such expressions are the best guarantee of their genuineness, and to throw them aside as inven- tions of later writers would betray an utter disregard of the criteria by which we distinguish ancient and modern thought. Our present object, however, is not to find out what these early philosophers thought of language—I am afraid we shall never be able to do that—but only to guard against their memory being insulted, and their names abused for sanctioning the shallow wisdom of later ages. It is sufficient if we only see clearly that, with the ancient Greeks, language was not con- sidered as mere onomatopeia, although that name 1 Olympiodorus ad Plat. Philebum, p. 242, Ore dyaApara pwovnevTa nad Tatra éo7l Trav Ocav, &s Anudxpiros. It is curious that Lersch, who quotes this passage (iii. 19), should, nevertheless, have ascribed to Democritus the opinion of the purely human origin of language. (i. 13.) 398 CHAPTER VII. means, literally, making of names. I should not ven- ture to explain what Pythagoras meant by saying, ‘the wisest of all things is Number, and, next to Number, that which gives name.’! But of this I feel certain, that by the Second in Wisdom in the universe, even though he may have represented him exoterically as a human being, as the oldest and wisest of men,? ~Pythagoras did not mean the man who, when he heard a cock crow, succeeded in re- peating that sound and fixed it as the name of the animal. As to Plato and Aristotle, it is hardly ne- cessary to defend them against the imputation of tracing language back to onomatopwia. Even Epi- curus, who is reported to have said that in the first formation of language men acted unconsciously, moved by nature, as in coughing, sneezing, lowing, barking, or sighing, admitted that this would account only for one-half of language, and that some agreement must have taken place before language really began, before people could know what each person meant by these uncouth utterances.? In this Hpicwrus shows a more correct appreciation of the nature of language than many who profess to hold his theories at present. 1. ‘Lersch, /..c. 1.25, 2 Tbidi l, ccd 24 % Diogenes Laértius, Hptcurus, § 75: “Odev kal ra évopara ef apxns Bi) Oéoe yevéoOa, GAN’ adtas Tas pices THY avOpwrwv Ka Exacra evn) iia macxXovaas T4On, Kal ida AapBavovoas POV TECH ARES idiws Tov dépar exTE[TELY, re ouey oy by’ ExdoTav Tov Ta0wv Kab T&Y pavTacpHaTeY, ws av TOTE Kal % Tapa Tovs TémOUS THY COVaV Staspopa ein. Latepey dé Kowwas Kad’ Exacta €6vn Ta tdia AEP mpos TO Tas onhwoas H Rae dugiBodous yeveobau adAAnAos, Kal Tuy TopOTE pols dnAovpevas’ Tiva bé Kal ov ouvopupeva a pay Hare elopépovras, Tovs cuveddTas mapeyyunoar Tivds pOdyyous ay TOUS peyv MOLD OL GSS dvaparicat, Tovs 6& TO AOYLO UD EAopevous KaTa THY TrELoTHY aitiav otTwSs Epunvedoat,—Lersch, i. 39. ELEMENTS OF LANGUAGE. 399 He met the objection that words, if suggested by nature, ought to be the same in all countries, by a remark in which he anticipated Humboldt, viz. that human nature is affected differently in different. countries, that different views are formed of things, and that these different affections and views influence the formation of words peculiar to each nation. He saw that the sounds of nature would never have grown into articulate language without passing through a second stage, which he, from his peculiar point of view, represents as an agreement or an understanding to use a certain sound for a certain conception. Natural Selection or Rational Elimination. Let us substitute for this Epicurean idea of a con- ventional agreement an idea which did not exist in his time, and the full elaboration of which in our own time we owe to the genius of Darwin ;—let us place instead of agreement, Natural Selection, or, as I still prefer to call it, Natwral Elimination, and we shall then arrive, I believe, at an understanding with Epicurus, and even with some of his modern fol- lowers. Natural selection, whenever we can watch it, is invariably rational selection. It is not any acci- dental variety that survives and perpetuates itself; it is the individual that is best calculated to accom- plish the ends for which the type or species to which it belongs was called into being, that conquers in the great struggle for life. So it is in thought and lan- guage. Not every random perception is raised to the dignity of a general notion, but only the con- 400 CHAPTER VII. stantly recurring, the strongest, the most useful; and out of the endless number of general notions that suggest themselves to the observing and gathering mind, those only survive and receive definite pho- netic expression which are absolutely requisite for carrying on the work of life. Many perceptions which naturally present themselves to our minds have never been gathered up into general. notions, and accordingly they have not received a name. There is no general notion to comprehend all blue flowers or all red stones; no name that éncludes horses and dogs, but excludes oxen and sheep. The Greek language has never produced a word to ex- press animal as opposed to man, and the word zéon, which, like animal, comprises all living creatures, is post-Homeric.! Locke has called attention to the fact that in English there is a special word for killing a man, namely, murder, while there is none for killing a sheep; that there is a special designation for the murder of a father, namely, parricide, but none for the murder of a son or a neighbour. ‘Thus the mind, he writes,? ‘in mixed modes, arbitrarily unites into complex ideas such as it finds convenient; whilst others that have altogether as much union in nature are left loose, and never combined into one idea because they have no need of one name. And again, ‘Colshire, driiling, filtration, cohobation, are words standing for certain complex ideas, which, being seldom in-the minds of any but the few whose particular employments do at every turn suggest * Curtius, Graundziige,i.78. L. Geiger, Ursprung der Sprache, p. 14. * Locke, On the Understanding, iii. 5, 6. ELEMENTS OF LANGUAGE. 401 them to their thoughts, those names of them are not generally understood but by smiths and chymists, who having framed the complex ideas which these words stand for, and having given names to them or. received them from others upon hearing of these names in communication, readily conceive those ideas in their minds; as by cohobation, all the simple ideas of distilling and the pouring the liquor distilled from anything back upon the remaining matter, and dis- tilling it again. Thus we see that there are great varieties of simple ideas, as of tastes and smells, which have no names, and of modes many more, which either not having been generally enough ob- served, or else not being of any great use to be taken notice of in the affairs and concerns of men, they have not had names given to them, and so pass not for species.’! Of course, when new combinations arise, and again and again assert their independence, they at last receive admittance into the commonwealth of ideas and the republic of words. This applies to ancient even more than to modern times—to the early ages of language more than to its present state. It was an event in the history of man when the ideas of father, mother, brother, sister, husband, wife were first conceived and first uttered. It was a new era when the numerals from one to ten had been framed, and when words like law, right, duty, virtue, gene- rosity, love, had been added to the dictionary of man. It was a revelation—the greatest of all revelations— 1 Locke, l. c. ii, 18, 7 Il, pd 402 CHAPTER VII. ~ when the conception of a Creator, a Ruler, a Father of man, when the name of God was for the first time uttered in this world. Such were the general notions that were wanted and that were coined into intellec- tual currency. Other notions started up, lived for a time, and disappeared again when no longer required. Others will still rise up, unless our intellectual life becomes stagnant, and will receive the baptism of language. Who has thought about the changes which are brought about apparently by the exertions of indi- viduals, but for the accomplishment of which, never- theless, individual exertions would seem to be totally unavailing, without feeling the want of a word, that is to say, in reality, of an idea, to comprehend the influence of individuals on the world at large and of the world at large on individuals—an idea that should explain the failure of a Huss in reforming the Church, and the success of a Luther, the defeat of a Pitt in carrying parliamentary reform, and the success of a Russell? How are we to express that historical process in which the individual seems to be a free agent and yet is the slave of the masses whom he wants to influence, in which the masses seem irre- sistible, and are yet swayed by the pen of an unknown writer? Or, to descend to smaller matters, how does a poet become popular? How does a new style of art or architecture prevail? How, again, does fashion change ?—how does what seemed absurd last year become recognised in this, and what is admired in this become ridiculous in the next season? Or take language itself. How is it that a new word, such as ELEMENTS OF LANGUAGE. 408 to shunt,’ or a new pronunciation, such as gold instead of goold, is sometimes accepted, while at other times the best words newly coined or newly revived by our best writers are completely ignored and fall dead? We want an idea that is to exclude caprice as well as necessity—that is to include individual exertion as well as general co-operation—an idea applicable neither to the unconscious building of bees nor to the conscious architecture of human beings, yet com- bining within itself both these operations, and raising them toa new and higher conception. You will guess ‘both the idea and the word, if I add that it is like- wise to explain the extinction of fossil kingdoms and the origin of new species—it is the idea of Natural Selection or Rational Elimination that was wanted, and being wanted it was found, and being found it was named. It is a new category—a new engine of thought; and if naturalists are proud to affix their names to a new species which they discover, Mr. Dar- win may be prouder, for his name will remain affixed to a new idea, a new genus of thought. All Names are General Terms. There are languages, we are told, without numerals beyond four. All beyond four is lumped together in the general idea of many. There are dialects, such as the Hawaian, in which? black and blue and dark- green are not distinguished, nor bright yellow and white, nor brown and red. This arises from no ob- 1 See vol. i. p.:37. 2 The Polynesian, September 27, 1862; Hibbert Lectures, p. 41. pd2 404, CHAPTER VII. tuseness of sense, for the slightest variation of tint is immediately detected by the people, but from slug- gishness of mind. In the same way the Hawaians are said to have but one term for love, friendship, gratitude, benevolence, esteem, &c., which they call indiscriminately aloha, though the same people dis- tinguish in their dictionary between aneane, a gentle breeze, matani, wind, puhi, blowing or puffing with the mouth, and hano, blowing through the nose, asthma.' It is the same in the lower classes of our own country. People who would never use such words as quadruped, or mineral, or beverage, have different names for the tail of a fox, the tail of a dog, the tail of a hare.? Castrén, the highest authority on the languages, literature, and civilisation of the Northern Turanian races, such as the Mins, Laps, Tatars, and Mongo- lians, speaks of tribes which have no word for river, though they have names for the smallest rivulet; no word for finger, but names for the thumb, the ring- finger, &e.; no word for berry, but many names for cranberry, strawberry, blueberry ; no word for tree, but names for birch, fir, ash, and other trees. He states in another place (p. 18) that in Finnish the word for thumb gradually assumed the meaning of finger, the word for waterberry (empetrum nigrum) the meaning of berry. But even these, the most special names, are really general terms, and express originally a general quality ; * Hale, Polynesian Lexicon, s. v. * Pott, Htymologische Forschungen, ii. 439. * Vorlesungen iiber Finnische Mythologie, p. 11. ELEMENTS OF LANGUAGE. 405 nor is there any other way in which they could have been formed. It is difficult to place ourselves in the position of people with whom the framing of new ideas and new words was the chief occupation of their. life." But suppose we had no word for dog; what could we do? If we, with a full-grown language at our command, became for the first time acquainted with a dog, we should probably discover some simi- larity between it and some other animal, and call it accordingly. We might call it a tame wolf, just as the inhabitants of Mallicolo,? when they saw the first dogs that had been sent to them from the Society Islands, called them broods, their name for pig. Exactly the same happened in the island of Tanna. Here, too, the inhabitants called the dogs that were sent to them pigs (buga). It would, however, very soon be felt as an inconvenience not to be able to distinguish between a dog and a pig, and some dis- tinguishing mark of the dog would have to be chosen by which to name it. How could that be effected 2 It might be effected by imitating the barking of the animal, and calling it bow-wow; yet, strange to say, we hardly ever find a civilised language in which the dog was so called. What really took place was this. The mind received numerous impressions from every- thing that came within its ken.