Sciences may be classed as historical or physical according as they deal with the mind of man or with external nature. The forces and materials of nature remain always the same: oxygen and hydrogen, for instance, are in no way different to-day from what they were a million of ages ago, and, combined in the same proportions, would always have produced water. Man and his intellectual creations, on the other hand, have a history; that is, the same causes do not always act in the same way, nor do the causes themselves always remain the same. The sum of the forces set in motion by the human will goes on increasing in an accelerated ratio: each new generation is influenced and moulded by the one that preceded it, and that influence becomes itself a fresh factor in the sum of the forces and causes at work. In place of the simpler processes of nature, with their unvarying uniformity of action, we have an infinitely complicated development, each stage of which is the immediate growth of the previous one, and is in turn the origin and germ of all that are to follow. Unlike the forces and phænomena of nature, thought is infinitely progressive, for

Wherever we have to deal with the products of human thought, there we have a constant ever-varying evolution, conditioned, it is true, by the uniform laws of outward nature, but continually modifying and adapting them. It is through the conditions thus imposed on the development of thought that we can discover the direction it has taken, and our inquiry thus becomes in great measure a historical one. We have to see under what conditions, in what external shape, as it were, the development of thought has displayed itself at each particular stage of its progress.

Like sociology, or comparative law, the science of language is concerned with a product of the human intelligence, and must consequently be included among the historical sciences. Language, we have seen, is significant sound; sound without significance is not yet language. As it is the inward sense and meaning, therefore, which constitute the essence of language, the primary object of comparative philology ought to be to discover the nature, origin, and history of the signification we breathe into our words and sentences. This can only be done, however, by finding out the conditions under which this signification is put into them, and by questioning the external side of language, those articulate sounds, namely, whereby we communicate our meaning to another. Now the external side of language is purely physiological and governed accordingly by purely physical laws. Phonology, in short, is as much a physical science as sematology is a historical one; and if we claim for the science of language in general the rank of a historical science, it is only because the meaning, rather than the sound, is the essence of speech, and phonology the handmaid and instrument rather than the equivalent of glottology. The method pursued by the science of language is the method of physical science; and this, combined with the fact that the laws of sound are also physical—the same conditions producing the same sounds in all periods of human history,—has occasioned the belief that the science of language is a physical science. But such a view results in identifying phonology and glottology, in making a subordinate science equivalent to the higher one, and in ignoring all those questions as to the nature and origin of language which are of supreme importance to the philosophy of speech. If we treat glottology as a physical science we must content ourselves with an exposition of the laws of sound and a mere description of the languages of the world and their classification, so far as it is founded on phonology alone. It is evident that such a classification must be superficial and incomplete; the relationship of languages is primarily based on grammar and structure rather than on a community of roots, and even roots must agree in sense as well as in sound before they can be admitted in proof of linguistic kinship. The intimate and inseparable connection between the inward and the outward, between sense and sound, in articulate speech, is a symbol of the connection between the historical and the physical methods of investigating it; but inasmuch as the sense is more important than the sound, so, too, the historical side of linguistic science is more important than its physical side.

Language and languages are in a constant state of change: nowhere, indeed, can the maxim of Herakleitus, πάντα ῥεῖ, be better illustrated. This perpetual flux and change is necessitated by the very fact that language is a product and creation of the human mind. Thought is ever shifting, moving, developing, and so, too, is the language in which it seeks to embody itself. But language is not only changing on this its inner side, it changes also on its outward, its phonetic side. The physiological organs of speech may be affected by an alteration in climate, food, or other physical conditions: they are certain to be affected by the psychological desire to save trouble or to add emphasis in speaking.

The three great causes of change in language may be briefly described as (1) imitation or analogy, (2) a wish to be clear and emphatic, and (3) laziness. Indeed, if we choose to go deep enough we might reduce all three causes to the general one of laziness, since it is easier to imitate than to say something new, while clearness in expression not only saves our neighbour trouble, but also preserves us from unnecessary repetition. Nothing is gained, however, by too wide a generalization; and it is, therefore, better to keep the three causes of linguistic change distinct and separate.

Imitation has played a far more important part in the history of speech than is ordinarily admitted. Imitation is the primary instinct of the infant and the savage, and, under the name of fashion, is a ruling power among civilized men. The great imitative powers of barbarous tribes have often been remarked upon by travellers; and a marvellous facility in mimicry and imitation seems to exist in proportion to the scanty development of the reasoning faculties. In this respect, at all events, the savage has not much ground for boasting of his superiority to the ape. Among the less cultivated races, indeed, the passion for imitation frequently passes into a morbid mania, and strange stories are related concerning it. Thus Dr. R. Maak, in his “Journey to the Amur,” states that “it is not unusual for the Maniagri to suffer from a nervous malady of the most peculiar kind, with which we had already been made acquainted by the descriptions of several travellers.[84] This malady is met with, for the most part, amongst the wild people of Siberia, as well as amongst the Russians settled there. In the district of the Yakutes, where this affliction very frequently occurs, those affected by it, both Russians and Yakutes, are known by the name of Emiura; but here the same malady is called by the Maniagri Olon, and by the Argurian Cossacks Olgandschi. The attacks of the malady which I am now mentioning consist in this, that a man suffering from it will, if under the influence of terror or consternation, unconsciously, and often without the slightest sense of shame, imitate everything that passes before him.” So, too, Mr. Jagor, in his “Travels in the Philippines,”[85] tells us that the malady in question is well known in those islands under the name of Mali-mali, and in Java under that of Sakit-latar; and goes on to relate how his “companions availed themselves of the diseased condition of a poor old woman who met us in the highway, to practise some rough jokes upon her. The old woman imitated every motion as if impelled by an irresistible impulse, and expressed at the same time the most extreme indignation against those who abused her infirmity.” The description reminds us of the feats of our own “electro-biologists.”

It is to the desire of imitation that we owe our first knowledge of our mother-tongue. The child tries to imitate those about him, and as the faculties of imitation and memory are the only ones yet developed in him his efforts are usually successful. The distance at which we stand from the infantile state, and the development of our reasoning powers, are measured by the prominence given to individuality and our power of taking the initiative. The community in which each man acts like his neighbour is not yet a civilized community; Athens is typical of all that is highest in human culture, and Athens was emphatically the State in which individuality had the freest play. It is well for the child who has to learn the language of his parents that he is rather a member of an uncivilized community than of Periklean Athens.

The love of imitation is the instrument whereby one language is able to influence another. Sometimes we find a community giving up its own tongue altogether and adopting that of his neighbours. Such has been the case with the Kelts of Cornwall, with the Wends of Prussia, or with the Huns of Bulgaria. The Negroes of Haiti speak French, the Lapps Finnish, while according to Humboldt and Bonpland,[86] “a million of the aborigines of America have exchanged their native for a European language.” Social contact and not identity of race occasions a similarity of language, since language is the medium of communication between the members of the same community, not between the scattered branches of the same race. No doubt where the languages are essentially distinct, based on radically different conceptions of the sentence and its parts, even the desire of imitation will be often not strong enough to cause the one language to be borrowed by the speakers of the other. Here and there we come across children who have a difficulty in imitating the pronunciation or use of the words they hear, and such a difficulty is a main cause of the origination of dialects; but it is among the speakers of agglutinative or polysynthetic tongues when brought into contact with an inflectional language that the difficulty is best exemplified. The Negro of the United States still speaks a jargon which can be called English only by courtesy, and Humboldt states[87] that “nothing can exceed the difficulty experienced by the (South American) Indians in learning Spanish,” although they “manifest quickness of intellect” in other respects, and “the missionaries assert that their embarrassment is neither the effect of timidity nor of natural stupidity, but that it arises from the impediments they meet with in the structure of a language so different from their native tongue.” Potent as imitation is, it yet has a limit, and this limit is reached wherever the element of conscious intelligence intervenes. The savage, like the child, finds it hard to mimic the products of civilized man, in so far as these embody the application of the reasoning faculties, and the mode of thought elaborated through long ages by a cultivated race necessarily forms a stumbling-block to the Negro or the South American Chayma. The Ethics of Aristotle could not have been written in a Semitic language, and a Negro Goethe is a somewhat incongruous conception. Wherever the distance between the two languages or the two levels of culture is great enough, the attempt to imitate is either given up altogether or else becomes a failure. The modes of thought of the borrower are read into the language he borrows. The Chinaman endeavoured to assimilate English, and the result was the Pigeon-English of Canton, a jargon in which we have a framework of English reared upon Chinese grammar and Chinese pronunciation. The difficulty of reproducing a cultivated language of foreign origin, or a language based upon a wholly alien conception of things and their relations, may be illustrated by the difficulty of translating accurately books written in another tongue. However closely related two languages may be, the various shades of meaning they attach to corresponding words or idioms will necessarily differ, and the more cultivated the style of a writer, the more impossible will it be to represent it exactly in a translation.

Where a language is not borrowed bodily, or at any rate engrafted upon the old modes of thinking and expression, it may yet exercise a greater or less influence upon a neighbouring language. Words, sounds, idioms, suffixes, and even grammatical forms may be and constantly are borrowed from one dialect by another; and it is not too much to say that a thoroughly pure and unmixed language does not exist among the civilized races of mankind. Our own English is a superstructure of Norman-French and Latin upon a foundation of Anglo-Saxon, and nine-tenths of the Hindi language is Sanskrit. No people can have neighbours close to them without receiving something from them in the shape of inventions, products, or social institutions; and these almost inevitably are adopted under foreign names. Thus the French have taken meeting and comfortable from us, and we have received naïve and éclat in return from them. Such loan-words are of great use in tracing the history and distribution of civilization, as well as the geographical and social relationships of the past. Boomerang proves our intercourse with the natives of Australia, from whom we have derived both the idea and the name of the weapon; pew, the Dutch puyde, puye, “a pulpit” or “reading-desk,” from the Latin podium, reveals the close connection that existed between the Churches of England and Holland in the seventeenth century, while words like maize, hammock, canoe, and tobacco, derived as they are from Haytian through the medium of Spanish, show as plainly as ordinary history that the Spaniards must have been the discoverers of America and the introducers of its products into the West. By similar reasoning we infer that the Baltic provinces must have been inhabited by a Teutonic population at the time when the Romans received amber from them under the name of glæsum (our glass), and Professor Thomsen has proved that the Finns must have bordered on Scandinavians and Teutons some two thousand or more years ago from the number of words borrowed by Finnish from their languages.

Sounds, again, may be borrowed from one language by another, or native sounds modified through the influence of a foreign tongue. The easier of the Hottentot clicks have been borrowed by the Kafirs, and the Souletin dialect of Basque has admitted the French vowel u. Idioms, too, may pass readily from one tongue to another. Words like avenir and contrée in French, are the result of an attempt to express German idioms in the Romance of the conquered provincials, avenir or ad venire being a literal translation of the German zu-kunft, and contrée for contrata (terra), a curious representative of the German gegend, “country,” as derived from gegen, “against.” The great extension of the English plural in -s, confined as it was in Anglo-Saxon to a comparatively few words, seems due to Norman-French influence, and the use of the genitive and dative of the personal pronouns in English “of me,” “to me,” in the place of the Anglo-Saxon min and me, is modelled after a French pattern. Bulgarian and Roumanian seem to have caught the infection of Albanian usage in which the definite article is attached to the end of the word, as in the Roumanian domnu-l, “the lord,” and Persian has even adopted the Semitic order of words so repugnant to the general structure of the Aryan group, in saying dăst-ĭ-’Umăr, for “Omar’s hand.” For instances of borrowed suffixes, we have only to point to our English -ize and -ist from the Greek -ιζ-ω and -ιστ-ης, which tend to supersede the old corresponding suffixes of the language, and the French participial termination is imitated in the letter of Gawin Douglas to Richard II. (1385), where we find such phrases as “Zour honourable lettres contenand,” and “brekand the trewis.”

The borrowing of grammatical forms is of much rarer occurrence, inasmuch as grammar is the essence and life blood of language, and to borrow the forms of grammar, therefore, is to intermingle the psychological histories of two separate tongues. It is a metamorphosis of the whole inherited mode of thinking and of viewing the relations of things to ourselves and one another, and to mix two grammars together is like mixing two different and incompatible modes of thought. A supposed instance of a mixed grammar (that is, of a mixed language) generally turns out to have another explanation. Thus it has been believed that the modern Aryan languages of India have substituted agglutinated postfixes for flection, and so have adopted the grammatical machinery of their Dravidian neighbours. Thus in Gujerati, dêv-mā̃ means “in the god,” like the Hindustani ãdhe-mē̃, “in the blind,” and in Nepalese mânis-visê is “in man,” mā̃ or mē̃ being a contraction of the Sanskrit madhyê (= madhya-i), “in the middle,” and visê of visayê, “in the thing.” What has really happened in these cases, however, is this. The first noun instead of being provided with the locative suffix (-i) is compounded with another noun which still retains the suffix, and the locative signification accordingly resides not in the second member of the compound, but in its worn-away flection. Here, then, there is no example of grammatical confusion. There are other instances of “mixed grammar,” however, which cannot be so easily disposed of, and it would really seem that in rare cases there actually has been an interchange of grammatical forms between two unallied languages. Thus in Assamese, which appears to be at bottom an Aryan language, the plural affix (bilak) is inserted between the noun and the case-ending, so that from manuh-bilak, “men,” we get the genitive manuh-bilak-or, the dative manuh-bilak-oloi, the accusative manuh-bilak-ok, the locative manuh-bilak-ot, and the ablative manuh-bilak-e, where the postpositions are all of them said to be of non-Aryan origin. The language of Harar, in Northern Africa, again, though apparently belonging to the Semitic family of speech, makes use of postpositions, and reverses the Semitic order of words when employing the genitive; while, according to Schott, the Persian affix of the dative and accusative was originally a Turanian postposition. Cases like these must, of course, be carefully distinguished from those in which we are dealing with an artificial language and not with the spoken language of the people. A curious language of this kind, the Pehlevi, was formed in the courts of the Sassanian princes of Persia, in which the elements of Aryan and Semitic grammar were mixed together in a strange fashion, but such a language did not penetrate beyond the limits of the learned class. Of the same nature are such affected plurals as termini and fungi from terminus and fungus in English, or the genitive and dative Christi and Christo in theological German. They would not be understood beyond the boundaries of a narrow circle.[88]

The most usual way in which the grammar of one language is influenced by that of another is by the adaptation of existing words and forms to express new grammatical ideas and relations imported from abroad. Thus the Assyrians became familiarized with the distinction between present and past time through their acquaintance with the extinct Accadian of ancient Chaldea, and they accordingly set apart certain separate phonetic forms, which had previously existed side by side without any difference of meaning, to express the present and the past tense.[89] So Spiegel[90] believes that he has discovered the influence of Semitic grammar in the Zend use of the feminine to denote a neuter or abstract, and of the dual to denote a pair. The invariable rule of the ancient Maya of placing the adjective after its substantive, is sometimes violated in the modern language through the influence of Castilian,[91] and the Ragusan custom of using the Illyrian svoj, “his own,” in the place of njegòv, “his,” is referred by Brugman to the influence of Italian and German.[92]

But the principle of imitation comes chiefly into play in the sphere of language in changing the form and meaning of words so as to bring them into agreement with the form and meaning of other words. When the true history and significance of certain forms have been forgotten by those who use them, other words with a totally different history and significance are very likely to be assimilated to them. When language has once created a particular mould it is very liable to run all manner of words into it. This is what is meant by the action of false analogy in speech. Words, forms, and significations which ought to have been kept apart are erroneously made like one another; the instinct of imitation and the desire to save trouble combine to exclude the irregular from language, and to force all exceptions under a uniform rule. The modern Greek declines innumerable words which formerly belonged to different declensions after the type of ταμίας, turning βασιλέας, ἄνδρας, and the like, into nominatives singular, and in the English which is unchecked by a literary tradition I comed is already more common than I came. Analogy is constantly at work throughout the whole domain of language—in pronunciation, in formal grammar, in syntax, and in sematology—building up and reconstructing what phonetic decay and change of meaning have tended to pull down. English is rapidly forcing all exceptional cases under the rule that throws the accent back as much as possible; balcóny has become bálcony, and Milton’s line “O argument blasphémous, false and proud,” would no longer scan. There is good reason to believe that the vocabulary of the primitive Aryan was for the most part, if not entirely, accented on the last syllable; the course of centuries has been continually thrusting the accent back as much as possible, and Latin and the Æolic dialects of Greece which illustrate this tendency, only show their want of conservatism and relative decay. Though the old accent of pitch has become an accent of stress in most of the modern European tongues, the same process is still going on; and while Polish still accents its words on the penultima, the accentuation of Bohemian is upon the first syllable. The same fact reappears in the Semitic family of speech, where it can be shown that the penultima primarily received the accent, and that the accentuation of the modern Arabic which agrees with that of English is a later innovation.[93] Greek words like φῡ́ω, θῡ́ω, and τῑ́ω, where the length of the vowel compensated for the loss of an iota (*φυίω), were brought under the general rule of the language which made one vowel before another short,[94] and when Horace addresses the fountain of Bandusia as “splendidior vĭtro,” the quantity assigned to vĭtro, a contracted form of vistrum for vid-trum (from the root vid, “to see”), arises from the mistaken notion that because a naturally short vowel could be lengthened before a mute followed by a liquid every vowel in such a position might be treated as indifferently long or short. So, again, the termination of the Latin nominative plural in -es was properly short, as may be seen from a comparison with the Greek; but the long vowel resulting from the combination of this termination with the final vowel of stems in -i (such as nubi-es) was extended to other cases, and the nominative plural of consonantal stems like voc (vox) was accordingly regarded as ending in a long syllable.

Apart from accent or quantity, however, the pronunciation of words is largely affected by the influence of analogy. Our English preference for diphthongal sounds is changing either and neither into aither and naither, in spite of the fact that the only other word in the language by which such a pronunciation could be supported is the misspelt height from high. The Frenchman “gallicizes” the words he borrows or the proper names he uses just as the Englishman “anglicizes” his; it is easier for the one to say Londres and Biarri’ than London and Biarritz, and for the other Paris and Marsaels than Pari’ and Marseies. Up to the last Charles James Fox called Bordeaux wine “Bordox,” maintaining that it had been domesticated in England, and ought accordingly to follow English customs. The action of analogy throws much light on Grimm’s laws respecting the shifting of sounds in the various branches of the Aryan family, which will be specially treated in the next chapter. When once a particular variety of pronunciation has come into vogue it absorbs and kills all deviating modes of pronunciation as surely as the cardoon in Central America has killed the native plants in its neighbourhood. We are all creatures of fashion, and the instinct of imitation is at work from the moment we first cease to be infants,—“speechless” embryos of humanity.

In the matter of grammar, a familiar instance of the way in which analogy can change the current forms of speech is afforded by the extension of the English perfect in -ed, the last relic of the affixed dide, the reduplicated past tense of do. The Latin amamini is the plural masculine of the old middle participle which we find in the Sansk. bharamâṇas, the Greek τυπτόμενοι, and the Latin alumnus (alomenus from al-o) or Vertumnus, the “changing” year. But when it had firmly established itself as a substitute for the second person plural of the present of the middle-passive voice, with estis understood, its true origin and meaning came to be forgotten, and as amamini was conjugated with amamur and amantur, so the anomalous amemini was conjugated with amemur and amentur, and amabamini with amabamur and amabantur. The coexistence of the older and later forms of the third personal pronoun in Greek, σφέ (Sansk. swa, Lat. se), and ἕ caused the one to be employed as a plural and the other as a singular, although the pronoun was originally reflective and of all genders; and the new plural pronoun was then provided with cases as well as with a dual formed on the analogy of those of the first and second pronouns. In the case of the dative alone a difficulty occurred, since here ἡμῖν or ὑμῖν could not be distinguished in form from σφί(ν) still used as singular by Herodotus; but the difficulty was overcome by having recourse to the noun-declension and creating a σφίσι as a parallel to ναῦσι. The contracted plural accusative πόλεις could not be derived from the original πολιας (for πολιανς) by any known rule of Greek phonology; it owes its existence to the habit of making the accusative plural like the nominative. The whole of the so-called fifth declension in Latin has grown up from the unconscious blunders of speech. A before m tended to become e, as in siem for siam, and accordingly by the side of materiam was heard materiem. The accusative materiem was then confounded with accusatives like nubem, and so a new nominative came into being, materies by the side of materia. Meanwhile the vowel of the accusative case-ending had influenced the vowel of the other case-endings, and changed the old ablative materiâ and genitive materiai into materie and materiei. The same process was next extended to the plural, materiarum, materiabus, and materias became materierum, materiebus, and materies, and nothing remained but to assimilate nominative and accusative as in nouns of the third declension whose accusative plural also terminated in -es.

Analogy will sometimes alter the whole structural complexion of a language. The Coptic, formerly an affix-language like Old Egyptian or the Semitic tongues, has become a prefix-language, denoting by prefixes the relations of grammar; and this metamorphosis seems due to the influence of the neighbouring Berber and cognate dialects. The tendency must have first shown itself in a few instances, and then by degrees have extended to the whole language. It has been held that the Aryan conjugation with a vowel between the root and the suffix, as in the Sanskrit bhav-â-mi or the Latin (e)s-u-m, has grown up in the same manner, verbs like the Sanskrit ad-mi, “I eat,” alone surviving as the remains of a past in which the personal pronoun was attached immediately to the verbal root. This, however, is very doubtful, the latter class of verbs being more probably the result of phonetic decay which has obliterated the connecting vowel, or more correctly the final syllable of the stem.

Syntax has not escaped the all-prevailing action of analogy and imitation. The relics of English flection are rapidly disappearing under its influence, and the use of the conjunctive were will soon be as obsolete as that of be. The relative pronoun was originally a demonstrative like our that, which drew attention to the idea contained in the principal clause, but with the extension of its use as a relative its demonstrative signification was lost, and it came to be used in instances where the demonstrative could not be employed.

Examples of the power of analogy in changing and extending the meaning of words are almost needless. The process is going on before our eyes every day. A new object or a new idea is named from its likeness to something with which we are familiar. The Kuriaks call the ox the “Russian elk” (Ruski olehn), just as the Romans spoke of the elephant as the Luca bos, and we are all familiar with the significant name of the Sugarloaf Mountain. There is a long distance from the primary signification of post as something “placed” or “fixed” to its signification as the arrival of correspondence, but every stage of the way can be traced and shown to be the work of analogy. The post fixed in the ground became a station, and when such stations were established for the conveyance of messages, news was said to travel “by post.” To transfer the name “post” from the machinery whereby the news was carried to the news itself was at once obvious and easy. The foot of a mountain is as much a metaphorical expression as the arm of the sea or the arm of law, and every metaphorical expression is an example of analogy. Three-fourths of our language, indeed, may be said to consist of worn-out metaphors. In no other way can terms be found for the spiritual and the abstract. Spirit is itself “the breath,” the abstract that which is “drawn apart.” Our knowledge grows by comparing the unknown with the known, and the record of that increase of knowledge grows in the same way. Things are named from their qualities, but those qualities have first been observed elsewhere. The table like the stable originally meant something that “stands,” but the idea of standing had been noted long before the first table was invented. The only abstract notion the Tasmanians had attained was that of resemblance. When they wanted to express the conception of roundness they had to say “like the moon” or some other round object, and similarly in the case of other abstract adjectives.

But as in pronunciation and grammar, so too in the matter of signification the analogy may sometimes be a wrong one. The men who coined the term “whale-fishery” were ignorant of the fact that the whale is a mammal, and that its only resemblance to a fish consists in its living in the sea. The name of guinea-pig, again, as applied to the small animal imported from Brazil, is singularly inappropriate. At other times the process whereby a new idea or object has been brought into relation with what was already familiar has been fair and legitimate. Thus the sense of the French canard as “idle gossip” can be traced back step by step to the primary meaning of the Low-Latin canardus. The feminine of canard is cane, and just as cane is the German kahn, “a skiff,” so canardus properly signified “a small boat.” Then by the force of analogy the words came to denote “a duck,” and as the duck was frequently used to decoy other birds by its cry, canard ended in signifying a mere decoy, a mere empty cry calculated to deceive.

Mythology, as we shall see hereafter, is in large measure based upon the metaphors of speech. The phænomena of nature were explained by likening them to those human actions with which primitive man was acquainted, and when in course of time a higher level of knowledge had been reached, and the original meaning of the traditional epithets had been forgotten, they came to be taken literally and interpreted as referring to beings of a super-human world. The dawn had been likened to a rosy-fingered maiden, the sun to a charioteer, and so the myths of Eôs, the ever-fleeing maiden, and of Phœbus Apollo, the heavenly charioteer, came into existence. Mythology is not so much a disease of language as a misunderstanding of its metaphors and a misconception of the analogical reasoning of our early forefathers.

Exactly the converse of this are those popular etymologies whereby words whose meaning is unknown or forgotten are assimilated to others with which the speakers are familiar. A gardener has been heard to call asphalt “ashes-spilt,” and thus render an explanation of the word to his own mind, and the modern spelling of the German sündfluth is due to the popular belief that the word, really a compound of sint, “great,” the Anglo-Saxon sin, “everlasting,” was invented to denote the deluge of Noah, which punished the “sins” of mankind. Luther still writes sindfluth (sindefluth), and in his translation of the Bible uses it in other passages besides those which relate to the Noachian flood (e.g., Ps. 29, 10, and Sirach 39, 22). Proper names have naturally suffered, especially from the attempt to give a meaning to them. Burgh de Walter has become Bridgewater and Widder Fjord, “the Creek of Wethers,” Waterford. The name of Madrid is explained by a popular legend which makes a boy, pursued by a bear, fly to a tree and cry to his mother “Madre id, Madre id” (“Mother, he comes”);[95] the Lepontii, we are told by Pliny,[96] received their title from having been the companions of Hercules who were “left behind” (λιπόντες!); and the Kirgises were so named from forty maidens, the mothers of the race, qyrg being “forty” in Turkic and qyz “a maiden.”[97] Similarly the modern Greeks have changed the meaningless Athens into Ἀνθῆναι, “the Flowery,” while Krisa has become Χρυσό, “the Golden.”[98] Where all other means failed the name was explained by the clumsy device of turning it into the name of an individual, and so there arose those eponymous heroes like Hellen and Asshur from whom tribes and nations were supposed to have been designated. The same process of etymologizing by the help of false analogy meets us in literature as well as in popular speech. The Homeric Poems are full of instances of the fact. In the Odyssey the old epic epithet ἐπηέτανος, “long lasting” (from ἐπὶ, ἄει, and τείνω), has come to be derived from ἔτος, which had lost its initial digamma (ϝετος, Sanskrit vatsas), and is accordingly employed in the sense of “lasting all the year,” while the Aorist infinitives χραισμεῖν and ἰδεῖν were taken to be presents and so provided with the futures χραισμήσω and ἰδήσω. Our own absurd mode of spelling presents us with parallel cases. Because should, the past tense of shall, has an l, could, the past tense of can, is given one; and further, the comparative of forth, has been written and pronounced farther as if derived from far.

The desire of clearness and emphasis, the second cause of change in language, is, like analogy, a creative and constructive power, and is often found at work in company with analogy. The object of speech is to communicate our thoughts to one another; where, therefore, our meaning is not clearly grasped, we begin to pronounce our words more distinctly than usual and to lay greater stress upon them. The result of this is a clear enunciation of all the syllables of a word, and sometimes a phonetic addition to the word itself. In this way we may explain the adventitious dental that has attached itself to the end of a word like sound, Latin sonus, French son, or the aspirate which is inserted in the wrong place by persons who are conscious of a difficulty in pronouncing it in the right place. So, again, in talking to a foreigner we instinctively raise the voice and repeat our remarks in a louder tone should he fail to comprehend them. The more readily our thoughts are understood, the less need there is of our dwelling upon the sounds which express them. Hence it is that with the progress of culture and education, and the consequent advance in quickness of perception, our words get worn away and slurred over, and a fragment only of the original word or the original sentence is often sufficient to convey our meaning. English and French are prominent examples of this fact, French cutting off its final consonants, and English softening its harder letters and avoiding the free play of the lips. Classical Italian, nurtured on the pedantic and metrical pronunciation of literary Latin and screened by the mountains of Tuscany, cannot, it has been well said, be spoken rapidly; but if we go to the Bolognese dialect, where these influences have not been at work, we shall find “A n’ vuoi t’ m’ in parl, S’nor,” doing duty for, “I won’t have you to speak to me about it, sir.”[99] While the educated Frenchman leaves the negative to be supplied by the mind when using pas, point, or jamais by themselves, the uneducated Englishman strengthens his negative by repeating it. Indeed, the repetition of the negative in order to emphasize the negation is a mark of most early languages, and runs parallel with the gesture and gesticulation which characterize the tongues of savages and barbarians. The muscular effort called forth by the latter necessarily extends also to the elocution, and a speaker generally finds that the clearness of his utterances is assisted by the exercise of the muscles of the arms and face.

Emphasis acts upon the outward sounds of a word as well as upon its inner meaning, and like analogy, though by the contrary process of differentiation, tends to build up new grammatical forms. The English thunder and jaundice go back to an Anglo-Saxon thunor and a French jaunisse, where the intrusive dental must be referred to the desire of clearness, since it can hardly be said to facilitate the pronunciation. So, too, in impregnable and groom, the French imprenable and Anglo-Saxon guman, we have other instances of the same striving after distinct and emphatic utterance, and the extension of the Greek πόλις (Sanskrit puris) into πτόλις, or of πόλεμος into πτόλεμος must be put down to a similar cause. People who wish to be very particular in the pronunciation of their words are apt to say kyind for kind, and the Italian luogho has arisen in no other way out of the Latin locus. The varying quality of a vowel, or an apparent exception to Grimm’s laws of letter-change may be explained by this principle of emphasis. Thus the Greek οἶδα, like the Sanskrit vêda or the Gothic vait, has a diphthong in the singular, whereas in the dual and plural the vowel is short (ĭ). This has resulted from the fact that the primitive Aryan laid the accent on the first syllable of the word in the singular; the less familiar flections of the dual and plural, however, were accented, and so preserved the short vowel of the root from being changed. In the same way the Old High German perfect laiþ in the singular observes the rule which makes an Old High German þ answer to an original d; in the plural, however, where the corresponding Sanskrit form accents the suffixes and not the root (as in the singular) the rule is violated and we have lidum, liduþ and lidun. So, too, by the side of the Old High German brôþar (bruder), answering to a primitive bhrâ´tar, we find môdar (muther) and fadar (vater) answering to a primitive mâtár and pitár (pâtár); while the accent of the Vedic saptán and the Greek ἑπτά, “seven,” shows why the Old High German seban and the Gothic sibun have b instead of the regular f.[100]

Emphasis enriches the vocabulary, first of all by introducing synonyms, and then by making a distinction of meaning between them. To set two synonyms side by side is the best way of giving clearness and intelligibility to our thoughts. Much of the charm of our authorized version of the Bible is due to the attempt of the translators to bring out the meaning of a Greek or Hebrew word by using two equivalents, one from a Romanic, the other from a Teutonic source. There comes a time, however, when we begin to contrast and differentiate the two synonyms; and so love comes to include much more than its New Testament synonym charity, and pastor, the synonym of shepherd, is confined to ecclesiastical language, while custom only allows us to say “much obliged,” and “very grateful.”[101]

Of a similar nature is the process whereby two varying forms of the same word become distinguished in use and signification. Thus the Latin tepor and tempus both go back to an earlier tapas, “heat,” but the strengthening of the first syllable of the one, and the change of s into r in the other, caused them to break apart and in course of time to be employed with a totally different meaning. The difference of sense brought with it a difference of gender, and thus introduced a grammatical change. The analogy of other nouns in final -or or -os preserved the masculine use of tepor, while tempus followed the gender of neuters like genus. The history of the termination of the nominative singular of Latin comparatives has been much the same. This was indifferently -ior or -ios (-ius), like the Greek -ίων and the Sanskrit -yan from an earlier -yans, and in Valerius Antias[102] we find prior still used for the neuter in the phrase “senatus-consultum prior,” while the title of the fourth book of Cassius Hemina’s Annals was, “Bellum Punicum posterior.” Arbor and robur were originally identical, and M. Bréal has shown that this was also the case with cruor and crus.[103] The two latter words both represent the Sanskrit kravis and the Greek κρέας in the sense of “bloody flesh” or “bloody limb,” and their differentiation was aided by the introduction of a new word, caro, in the sense of “flesh.” Caro originally meant simply “part” or “portion,” a sense in which the Umbrian karu is still employed in the Engubine Tables,[104] and the Oscan carneis in the Tabula Bantina. Roots, too, as well as derivatives, may be differentiated and gradually assume independent meanings. Thus in Greek, if we follow the usual theory, the old root ar or ara has been split up into three, ἀρ-, ἐρ-, and ὀρ-, in accordance with the threefold representation of the Sanskrit ă in European Aryan. Accordingly by the side of ἀρόω, the Latin arare, the Gothic arjan (Old English ear), which appropriated to itself the sense of “ploughing,” we have also ἐρέσσω (remus) in the sense of “rowing,” and ὄρ-νυμι (orior) in the sense of “rising” to one’s work. This differentiation of the three roots, however, seems to have come about after the separation of the several members of the Aryan group, as we find no trace of it in the Asiatic branch of the family, and it must, therefore, have really taken place in the fully-formed words of the European tongues.[105] Greek with its delicate sense of vocalic difference shows a special tendency towards utilizing vowel changes for grammatical purposes. Thus the reduplicated syllables in δίδωμι and δέδωκα were originally identical, but in course of time, while the sound of ĭ was appropriated to the present tense, the sound of e came to mark the perfect. In the same way Greek verbs in -αω, -εω, -οω all go back to the form which we have in the Sanskrit -ayâmi, but later usage tended to assign a transitive meaning to the form in -οω, and an intransitive one to that in -εω, while that in -αω floated between the two. It is probable that the three Semitic case-endings in u, i, a, which respectively denoted the nominative, genitive, and accusative, all went back to a primary indeterminate -a. In the Negro Dinka language certain plurals are formed by lengthening or sharpening the vowel of the singular, like rōr, the plural of ror, “wood,” nim, the plural of nom, “head,” līb, the plural of lyep, “tongue,” or tut, the plural of tuot, “goose;” and since we find that a verb becomes passive by simply lengthening the final i of the formative elements (as ran a-tšī tšōl, “the man has been called,” by the side of ran a-tši tšōl, “the man has called”), it is possible that the vowel change in all these cases is due to differentiation for the sake of clearness and emphasis. Such at least has been the origin of the tones which form so marked a feature in Chinese. Dr. Edkins has shown that the confusion between words of different signification occasioned by the loss of various initial and final letters in pronunciation was obviated by the substitution of tones, and the effects of phonetic decay have been thus neutralized by the action of the contrary principle of emphasis.

One of the modes in which this principle comes into play is what Professor Max Müller has called Dialectic Regeneration. The words and grammatical forms which have become effete in the literary dialect, are often replaced by others taken up from the fresh fountain of “provincial” speech. There is nothing any longer to attract attention in what has become so prosaic an expression as “the four cardinal points,” striking as the phrase once was; but when Carlyle goes to the Scotch and borrows from it the “four airts,” we are at once arrested by the unusual character of the word, a special emphasis is laid upon it, and we begin to realize its full meaning. It is in a period of social revolution, like that of the Norman Conquest in England, that Dialectic Regeneration is best seen at work on the literary language. As soon as the latter loses the support of the educated classes, it fails to withstand the attack of the less favoured but more deeply rooted dialects which have surrounded it, and, as in the case of literary Anglo-Saxon with its inflections and learned terms, it disappears for ever. The unwritten languages of savages and barbarians are in a continual state of flux and change. Old words and expressions which have ceased to possess the needed amount of clearness and emphasis have to make way for new ones. The slang of the schoolboy, or the cant of thieves and costermongers, exemplifies the same fact. It is not so much the desire of revolting against the proprieties of a civilized society, or of framing a secret jargon which shall be unintelligible to others, that produces these wild outgrowths of language; it is rather the feeling that the conventional terms have become mere symbols, or, as Hobbes said, the counters of wise men, and that the ideas which are perceived and felt clearly should be expressed with equal clearness and force. Man is not wholly ruled by the wish to save himself trouble and attain his object with the least effort; the healthy love of physical exertion for its own sake is also a powerful motive in human life. It is only with the growth of civilization and thought that the exertion is transferred from the muscles to the brain, that words become so many algebraic signs, and that syntax takes the place of elocution. It has been often noticed that the tendency of the modern languages of Europe is towards a monotonous level of both accent and tone; but it must be remembered that, as long as poetry exists, there will exist also a tendency in the opposite direction, as well as a protest against the reduction of all language into a mere reflection of the dry light of reason. Laziness will not explain everything in speech any more than it will in the ordinary dealings of mankind. As Sievers states:—“We even now often find it stated in works on the science of language, that all phonetic change results from a striving to facilitate the pronunciation and simplify the articulation; or, in other words, that change of sound always consists in a weakening of sound and not in a strengthening of it. We may allow that although many of the phænomena observable in the history of speech can be brought under this rule, the general application of the statement is absolutely false.... The idea of facilitating the pronunciation, if it is to be any longer maintained, must be regarded as an essentially relative one. Speaking generally, we must never forget that the different degrees of difficulty in uttering various sounds are in themselves extraordinarily slight, and that real difficulties in forming them are usually experienced only in the case of sounds belonging to a foreign language.... In short, real difficulties in pronunciation are never specially felt by the members of a community which speaks a given language, and with them only a further development of their language is possible.”

This brings us to the third and last cause of change of language, laziness, or, as it has also been termed, the principle of least effort. As the results of laziness show themselves principally in the alterations undergone by the sounds of speech, this cause of change is commonly known under the name of Phonetic Decay.[106] But the meanings of words as well as the expression of grammatical relations are as much subject to decay as the sounds of speech; the outward form of age which can be traced back to the Low Latin ætaticum and the classical ætas, has suffered no less from the wear and tear of time than its inward signification, which goes back to a root meaning “to go.” Like the present strata of the earth which are the débris of the earlier rocks, the present strata of language are the worn-out relics of older formations. The power of laziness, more especially in the shape of phonetic decay, is conspicuous in almost every word we utter; it is the first agent of linguistic change that strikes the student, and it has accordingly attracted more than its due share of attention. The influence of laziness has been insisted on to the exclusion of the two other equally important causes of change in speech, and the growth of grammatical consciousness, the discovery of new grammatical relations and the development of fresh mental points of view, have even been ascribed to its action. No doubt its influence is great and far-reaching, but we must be on our guard against regarding laziness as sufficient of itself to explain all the phænomena of language. Phonology is rather affected by it than either morphology or sematology. Owing, however, to the large place assigned to it in works on comparative philology, it will not be necessary to dwell upon it here in any great detail. We naturally seek to make ourselves understood by our neighbours with the least possible amount of trouble. Muscular and still more mental fatigue is distasteful to us, and the less we have to exert our vocal organs and powers of thinking when making our meaning clear to another, the better satisfied we are sure to be. Hence it happens that we constantly use words with a very dim appreciation indeed of their full and exact significance. We select that part of the meaning only which for some reason or other has made an impression upon our minds, and very often this part of the meaning is merely subsidiary and accidental to the proper signification of the word. But we are too lazy to realize that proper signification, and so pass words on to others the mere shadow and fragment of their former selves. It may often happen that a sense originally imported into a word by the context in which it accidentally found itself becomes appropriated to it to the gradual exclusion of its real signification. The word silly, for example, which once meant “blessed,” like its German cousin selig, from being applied euphemistically to half-witted persons, has entirely lost its true meaning. A word like impertinent is still in process of being changed. Its positive pertinent has hitherto preserved its proper sense, at all events in literature; but the popular mind has already forgotten the meaning of the negative, and only a short while ago a member of Parliament was called to order for describing a remark as “impertinent.” Here the accidental application of a word has caused its primary meaning to fall into neglect. Still more striking is the fate which has befallen words like transpire and eliminate. The newspapers speak of events “transpiring” in absolute disregard of the fact that events can hardly “breathe through,” while eliminate has been used not in the sense of removing out of the way but of bringing in.[107] It is so much easier to guess at the meaning of a word from the context in which it occurs than to trace it back to its real signification, and so long as our use of it is intelligible there is little care among ordinary speakers as to whether that use is correct or not.

In this way general terms come to be restricted to individuals, while words which denote the particular are extended to denote the universal. Deer, which, like the cognate German thier and Latin fera, originally signified wild animals of all kinds, is now confined to a particular species; while, on the other hand, the Latin emere, which properly signified “to take” in general, came to be restricted to the special meaning of taking when we “buy.” The older significations of words are continually decaying and being supplanted by new ones. Those who use them are too lazy to find out their exact significance.

The principle of laziness is equally active in the province of grammar. Here, too, the relations formerly conceived to exist between the several parts of the sentence may be forgotten altogether or replaced by other relations. The inflections of the Anglo-Saxon noun have been almost all lost, and the datives him and whom have become objective cases. Prepositions have taken the place of the case-endings, the adjective no longer “agrees” with its noun, but is now conceived of as a simple attribute, while all remembrance of the dative relation has faded out of the expressions “give me a book,” “send it away.” The subjunctive is fast ceasing to exist, and the modern Englishman troubles himself but little about the difference between be and is or between if I was and if I were.

It is in phonology, however, that the principle of laziness is most active. As far back as we can follow the history of language we see the stronger and harder sounds perpetually changing into weaker and easier ones; and so uniform and constant is this tendency that in the absence of counter-indications we are justified in referring most cases of phonetic change with which we may meet to the operation of decay. Mr. Douse[108] has lately made an ingenious but unsuccessful attempt to assign the phænomena of Grimm’s law to what he terms the principle of least effort, by supposing that the different phonetic systems of the several branches of the Indo-European family were evolved out of the tenues or hard consonants, at a time when these branches were still co-existing dialects of a single language, through the influence of “Reflex Dissimilation.” Reflex dissimilation is explained to be a more complicated and somewhat varying instance of that simple cross compensation which we see exemplified in the Cockney interchange of v and w, or the perverse persistency with which the same persons, who leave out the aspirate where it ought to exist, insert it where it ought to be omitted. In both cross compensation and reflex dissimilation, however, we have a compound action of the two antagonistic principles of laziness and emphasis.

The age of a language is marked by the extent to which it has been affected by phonetic decay, and when we find how large its influence has been upon the Old Egyptian and the Accadian of Chaldæa, as they appear in the earliest monuments we possess, we may form some idea of the length of time that must have elapsed since those languages were first being moulded and fixed. At the same time we must not forget that phonetic decay will act more readily upon some classes of languages than upon others. Wherever there is no clear consciousness of the distinction between root and grammatical suffix, as in our own inflectional family of speech, there we may expect a greater and more rapid amount of change than in agglutinative dialects where the relations of grammar are expressed by independent or semi-independent words. But even the latter cannot escape the law of gradual decay. To pass over the incorporating Basque in which words like dakarkiotezute, “ye eat it for them,” or detzadan, “that I should have them,” have to be decomposed into da, “it” or “him,” ekarri, “to eat,” ki, sign of the dative, o, “for him,” te, sign of the plural, zute, “ye,” and d, “him,” ez (izan), “to be” or “have,” za, sign of the plural, ta, “I,” and n, conjunctive affix, we find Yakute Turkish changing bin + śän (“I + thou”) into biś, “we,”[109] while the written Japanese taka-si and taka-ki, “high,” are pronounced takai. Chinese itself is not exempt from the universal rule. As Dr. Edkins[110] and M. de Rosny have shown, the modern Mandarin dialect has lost numerous initial and final consonants, and words like yi, “one,” and ta, “great,” were once tit and dap. Along the southern bank of the Yang-tsi-kiang and through Chekiang to Fuh-kien the old initials are still preserved, while in the northern provinces no less than three finals have been lost and the tones by which Chinese words of similar form are distinguished from one another are so many compensations for the loss of letters. Here again we have the principle of emphasis endeavouring to repair the damage wrought by the principle of decay.

A literary dialect is naturally less subject to the inroads of decay than an unwritten one. The spelling of words reacts upon their pronunciation and preserves it from extensive alteration. There is a wide chasm between that Tuscan Italian which has been preserved from corruption by the genius of Dante and the modern dialect of Bologna or Naples. In the age of Cicero the cave ne eas of polite society had become cauneas in the language of the people,[111] and how artificial was the attempt of pedants and purists to maintain the older pronunciation, even to the restoration of the final s which had already been dropped by Ennius, appeared pretty plainly as soon as the decline of the Roman empire and the extinction of the literary class deprived it of support. Latin at once fell away into the Romance dialects of modern Europe, just as literary Anglo-Saxon with its inflections and its learned vocabulary disappeared before the Norman Conquest. The language of the Assyrian inscriptions remains almost unaltered throughout the long period of nearly 2,000 years, during which we can watch its fortunes; but this language was the stereotyped one of literature and education, and differed very considerably from the spoken language of the people. The late linguistic character of Hebrew, the extent, that is, to which it has been influenced by phonetic decay as compared with its sister tongues, is an incontrovertible proof of the backward literary condition of its speakers. But even literature and cultivation are unable to preserve a language altogether from decay and change. The pronunciation of the educated slowly changes; words become clipped and shortened in spite of their spelling, and notwithstanding printers and schoolmasters the spelling in the end has to follow the pronunciation. Mr. Alexander Ellis has shown in his “Early English Pronunciation” how widely our modern pronunciation of English has departed from that of Shakspeare’s time, and the spelling of though, through, and enough bears witness to a period when they ended in a guttural aspirate. Our pronunciation is still undergoing change; the vowels are becoming more and more indistinct and merged in a common obscure ĕ; while such contractions as I’ll, I’d, won’t, and can’t can hardly be distinguished from Basque forms like those mentioned above. The educated Englishman speaks, as the French say, with his lips closed; he finds that he can be understood without the trouble of opening and rounding them, and his vowels are accordingly formed in the front rather than in the back part of the mouth. No wonder that he has a difficulty with the French eu; the effort to pronounce it is too great a strain upon the unexercised muscles of the lips, and so the English gentleman who told the waiter not to let the feu go out in his absence found on his return that his friend had been strictly watched and guarded as a dangerous fou.

But though a literature and more especially a widely extended literary education form the chief obstacle to the action of phonetic decay, there are other social influences which operate to the same end. Wherever there is a fixed and stable society, cut off from close intercourse with its neighbours and handing down unchanged its customs and institutions, we are likely to find a more or less fixed and stable language. For language is the mirror of the community that uses it, and where the community alters but little the language will alter but little too. It is in this way that we must explain the fact that Lithuanian, though unprotected by a literature and spoken by the least progressive of the European members of the Aryan family, is yet the most conservative of all the Western languages of our group, or that the Bedouin of Central Arabia is said to speak at the present day a more archaic language than those of Nineveh or Jerusalem 3,000 years ago. Since the institution of an annual fair among the Rocky Mountains the idioms of the eastern and western Eskimaux, who at first were hardly understood by one another, became more and more assimilated;[112] and the stationary character of Icelandic may be ascribed as much to the isolation of the settled Norse community in the island as to the existence of a literature. Of course, the community must be one which has reached a certain level of culture, and its customs and institutions must imply organization and recognition of fixed principles. Where the customs and institutions are founded on mere unreasoning habit and precedent, we are dealing with a community of barbarians, and consequently with languages or dialects in a perpetual state of flux.

The changes wrought by phonetic decay are sometimes sufficient to alter the whole aspect of a language, and are at once the foundation and the riddle of etymology. Who would recognize in the French même, for instance, any derivative from the Latin pronoun se? And yet même goes back to the Low Latin semetipsissimum through the Old Provençal smetessme, the later Provençal medesme and the Old French meïsme. Words of different origin, like scale from the Latin scala and the Anglo-Saxon scalu and scealu, may come to assume the same form; while words of the same origin, like the French captif and chétif, from captivus, or noel and natal from natalis, may appear under different forms. The processes of assimilation and swarabhakti, of metathesis and epenthesis, to be described in the next chapter, are so many forms under which phonetic decay displays itself. The history of language is the history of the continual weakening of uttered sounds and the gradual lessening of the demands made upon the organs of speech, and attempts like that to reduce the triliteral roots of the Semitic tongues to biliteral ones are contrary to the whole tendency of language. Accent alone is able to hold out against the assaults of phonetic decay; it is only the accented syllable that remains unchanged when all around it is perishing, and, as in the case of age from ætaticum or dine from desinere, is often all that is left of the primitive word. It is again the struggle between the principle of emphasis and the principle of laziness, between conservatism and revolution. Only when the accent is shifted to another syllable can phonetic decay gain the victory, and the shifting of the accent is itself the work of the principle of decay.

The principle of laziness has much to do with the creation of dialects. Slight variations of pronunciation and of the usage of words are as inevitable in language as variations of species in zoology, and where there is no correcting standard these variations are perpetuated and intensified. Helped by the two other causes of linguistic change, the dialect of a household becomes in time the dialect of a clan or tribe, and as soon as its characteristics are sufficiently numerous and distinct, the dialect is transformed into a language. An isolated community will by slow degrees form a new language for itself. Just as the history and character of one society differ from those of another, so too must the dialect or language differ in which the society finds expression. Even where the rapid and intimate intercourse of modern civilization and the safeguard of a common and widely-studied literature stand in the way, as in the case of England and America, dialectical differences and peculiarities will yet spring up. In savage and barbarous communities the growth of innumerable dialects is a matter of necessity. The manifold languages of the Malayan and Polynesian Archipelago can be traced back to a common source, but the natives of two neighbouring islands are often unintelligible to one another; while von der Gabelentz says of the Melanesians, that “every small island has its own language or even several languages.”[113] Before the utter extinction of the Tasmanians, with a population of no more than fifty persons there were four dialects, each with a different word for “ear,” “eye,” “head,” and other equally common objects. The language of a shifting unorganized community will reflect the condition of those who speak it, and we are not surprised, therefore, at Captain Gordon’s assertion that “some” of the Manipuran dialects “are spoken by no more than thirty or forty families, yet (are) so different from the rest as to be unintelligible to the nearest neighbourhood.” Humboldt tells us[114] that in South America, together with a great analogy of physical constitution, “a surprising variety of languages is observed among nations of the same origin, and which European travellers scarcely distinguish by their features.” Greece, with its small extent of country and still smaller amount of population, was said a few years back to possess no fewer than seventy dialects,[115] and no less than eight principal dialects besides several subordinate ones exist among the modern Basques, whose whole population is under 800,000.[116] Indeed, considering the isolation of the Basques, socially, politically, and linguistically, as well as the narrow tract of country into which they have been compressed, it is remarkable that natives of places not forty miles distant from one another are yet mutually unintelligible.[117] But the natural condition of language is diversity and change, and it is only under the artificial influences of civilization and culture that a language becomes uniform and stationary. As soon as the coercive hand of civilization is removed it breaks out again into a plentiful crop of dialects. Of course, the vicissitudes through which semi-civilized peoples are continually passing greatly assist the process of change. Conquest and the mixture consequent upon it, famine, disease, and migration, are all powerful aids to dialect-making. The women of a tribe who stay at home, or who have been married out of another tribe, sometimes possess a language different from that of the men; thus, the Carib women in the Antille Isles used a different tongue from that of their husbands, while the Eskimaux women in Greenland turn k into ng and t into n.[118] Even religion and superstition play their part in the work; the sacred language of the “medicine-men” in Greenland, for instance, is for the most part an arbitrary perversion of the significations of known words; thus tak, “darkness,” is used in the sense of “the north,” and so gives rise to two new words of this secret speech, tarsoak, “earth,” and tarsoarmis, “roots.” The custom of tapu among the Pacific Islanders, according to which every word which contains a syllable identical with some part of the name of the reigning chief has to be dropped or changed, is due to the belief that all things belonging to a chief are consecrated and inviolable. Since the reign of Queen Pomare mi has been substituted for po, “night,” in Tahitian, and Hale tells us of this language[119] that its “manner of forming new words 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 hepai, ... au for tau, &c. In other cases the word substituted is one which had before a meaning nearly related to that of the term disused.... 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,’ papai for vai, ‘water,’ pohe for mate, ‘dead.’” Similar to the Polynesian tapu is the Chinese custom of tabooing the elements of the reigning emperor’s name, and the ukuhlonipa, which forbids the Kafir women to pronounce a word containing a sound like one in the names of their nearest relations. Thus, “Mr. Leslie states that the wives of Panda’s sons would never call him (Mr. Leslie) by his Kafir name of u’ Lpondo, on account of its partial identity with that of the chief, their father-in-law. In the name of the river Amanzimtoti, ‘Sweet Waters,’ in like manner, mtoti has been substituted for mnandi, hlonipaed or tabooed on account of its occurring in the name of Tsaka’s mother Unandi.”[120]