with the letters, with the sounds of which syllables are composed.

Some of this material is common to us and the brutes. It is hardly necessary to observe that the? vowels, pure or mixed, short or long, nasal or combined into diphthongs, may be recognised in the utterances of the dog, the cat, the horse, the ox, the sheep, the frog, the toad and the crow. Out of the sounds peculiar to each species it is easy to construct, without omitting a single note or quality of sound, the entire vowel scale: ǎ ā, an (nasal); ě, è, eŭ, en; ì, ì, in; ō, ŏ, on; oŭ, où; eũ; ŭ, ū, un; oa, oe, oi, ouă, oue, ouon (nasal), oui, ui, &c. Note how ē and ō are related to ā; o to ou; ou to u; i to é and u; while in diphthongs the final vowel only is continuous, the first ceasing to be heard as soon as it is uttered.

Another class of sounds give rise to similar observations: not only the vowels are susceptible of prolongation. Certain hissings and trills, which can give a continuous sound, and are very common among animals, have played so important a part in the formation of human speech that they cannot be too carefully studied. Nor can these be separated from other undefined utterances, midway between continuity and articulation, ingeniously called semivowels; to these we may add hard and soft breathings, which precede or follow vowels, semi-vowels, sibilants, liquids, and true consonants. All these appear to be variants or degenerate forms of the consonants to which they are really related; but the fact that they all, except the true consonants, may be found in the animal kingdom, may be urged in favour of their priority. They form the link between vocalism and consonantal language.

For while recognising the part played by the teeth, the throat, the palate, and the lips in the liquids, r, l, lh; in the palatals, j, ch, sh; in the sibilants, s, z; in the semi-nasals, m, n; and semi-labials, w, v, f; it is also impossible to separate them from certain vowels. y, j, ch, lh, derive from i, consequently also j, ch, sh, the liquids and the nasal n, which often changes with these last. Ou is the origin of v, w, m, f; s and, which are more independent, are not without vowel affinities through j, ch, sh and the liquids; r is reckoned a vowel by the grammarians of India; s has something of the character of an aspirate, which often takes the place of this letter, particularly in Greek and Zend. Now the aspirate, considered apart from the consonant, which it strengthens, is only a sort of toneless vowel; it may be compared to the prefatory murmur of an old clock before the hour strikes. It results from the effort of the breath made in giving the vowel distinctly or in articulating the true consonant.

In the present state of language the various semivowels often take the place of consonants. They have acquired this character by that which marks the decisive step towards articulation-i.e., the momentary arrest of the vowel-breathing by contact with the glottis, the tongue, the palate, teeth, and lips. From the moment this stoppage is produced, continuity is broken, and the issuing sound can only be heard together with a vowel or semi-vowel (however slightly audible), which precedes or follows it. (Such is the phenomenon of articulation; the word consonant, that which sounds. with something else, expresses its essential character. > The consonant is the substructure and the foundation of language. Man alone possesses it, and it is

This,

the greatest and most fruitful of his conquests. treasure is composed of but six letters: k, g; t, d; p, b; the gutturals, dentals, and labials. These cannot give a continuous sound, however we attempt to prolong them; they can only be the beginning, middle, or end of a syllable; it is impossible to separate them from a vowel, a sibilant, a liquid, or an aspirate, with which they form a sort of consonantal diphthong, ks, sk, kv, kch, and so forth. It seems probable indeed that these double sounds were the origin of the pure consonants.

Here I foresee an objection: gutturals, it will be said, are not unknown to animals; a number of birds and mammals pronounce k, t, p, b. But this is a vulgar error; it is we who attribute these articulations to the utterances of animals. The cock does not say cock-a-doodle-doo, nor the rook caw, nor the sheep baa. They utter the breathings akin to these consonants, which, so to speak, lead up to them; they come near to articulate utterance, but man alone has achieved it; not without effort, and with varying success, according to the vocal and hearing power of each human race or group.

This is not an assertion deduced from the logic of the theory of evolution. The most perfect languages, like the crudest, have retained the traces of a long hesitation, of a remarkable confusion, not only as in German, between the weak and the strong consonants, but between the three types of true consonants and the corresponding aspirate, and even between the true consonants and semi-vowels. We hear, as it were, across the ages the stammerings and hesitations of speech in its infancy.

Not only have the races unequal power in the use

of the gutturals, dentals, and labials, but, in certain dialects of Africa and Polynesia, the pronunciation is still so uncertain that the most delicate ear can hardly distinguish between k and t; the sound is doubtful, and approaches now the one, now the other. In like manner many children say tat for cat, many men fail to distinguish between cintième and cinquième. That which is obvious within the limits of the same language is seen on a wider scale in two dialects of the same origin, which have grown up at the same time and side by side.

I give a few examples taken from the Indo-European languages. In all these the names of numbers up to ten, except the number one, are identical; but it is not easy at first sight to recognise as sisters these casts from one mould. To be fully persuaded that eight and octo, zehn and deka, are the same words, we must have heard it stated more than once. The fact is certain, however, and I insist no further. Let us take

the words four and five, the only ones we need consider here. The Latin form quatuor, quadru, which has given quattro and quatre, corresponds to the Sanscrit tchatvaras; Zend tchathwar, tchatru; Pali tchattaro; Hindustani tchar; Lithuanian keturi; Slav tchetvero, tchetüri; Armenian tchorq, tchors; Greek TéTTapes and Téoσapes; also the Umbrian and Celtic patour, pewar; the Eolian Tίoupes; Anglo-Saxon and English fidvor, four thus we pass from guttural and dental diphthongs, kv, tv, tch, to various dentals and labials, t, p, ƒ, not to mention the double t alternating with the single or double s; or the transformation of the semi-vowel v into u and ou in quatuor, patour, and into o in fidvor. It may be noticed that there has been a struggle between the guttural and the labial, and now

the k survives, as in quatre, and now the v, strengthened into for p (in vier, fidvor, four, patour); it is less easy to understand the presence of the mixed dental tchatvaras and the pure dental; but it seems that at the time when Latin, Umbrian, Celtic, Greek, German, &c., were in process of formation from the parent language whence they all derive, there was still hesitation not only between k, t, or d, and p or ƒ, but even between these consonants and the forms tch, kv, tv. This conclusion becomes yet more obvious on a comparison of the various forms of the word five. Sanscrit pantchan; Lithuanian penki; Armenian hing; Umbrian pump; Gothic fimf; English five; German fünf; Greek TéμTе and Tévтe; Slav panti; Latin quinque; Italian cinque; French cinq; Irish coic. From these come derivatives as various as Quinctius, Pompeius, Pentecost, fifty. To these examples we may add a few well-known permutations: Latin coquere, coquus (cook, v. and subs.); Greek πÉπτ (whence рepsine); Sanscrit patch and pak; Low-Latin sequere, to follow; Sanscrit satch and sak; Greek Toμai; bœuf, bos, Boûs, gaus, kuh, cow; oil (eye), Latin oc-ulus, Greek oπтоμaι, oplaλuós, Sanscrit akeh (akchan). The Latin forms are here simpler than the corresponding Greek and Sanscrit forms, the Sanscrit kch, and the Greek κτ, πτ, σσ.

Since no one of the forms adopted respectively by one or other of the seven or eight families which constitute the entire group can claim the priority, and since none have been borrowed by any language from any foreign source, we are led to believe, as wel indicated above, that these are varying pronunciations of a primitive form which contained the germ of all of them.