Indo-European alphabet, and understand better its changes into Greek and Italian, and may lay a basis for any further investigation of the changes from Italian to the Romance languages'.

CH. IV.

difference

vowels and con

The old division of sounds into vowels and consonants Physiohas been much objected to, and with considerable reason, logical if it is taken to imply the absence of affinities between between vowels and consonants. But it will be retained, both for its convenience and because there is a real difference in sonants. their formation. This has been well shewn by Mr Melville Bell. When the voice in its passage through the mouth is not further modified by contact, partial or complete, of the lips or tongue, but flows through an open channel without any friction or hissing, then we have vowel sound. When on the other hand the sound is not complete until the action of some part of the organs of the mouth has ceased, then we have produced what we may call consonantal sound. Briefly, "a vowel is the result of an open position of the oral organs; an articulation (this is Mr Bell's term for consonant) is the result of an opening action of the organ." Thus the so-called "semivowel" w seems to differ infinitesimally from u, and linguists have often argued to what class, consonant or vowel, it should belong. But its consonantal character is quite clear by the above definition. The organs are in precisely the same position for sounding w as they are for sounding u,

1 The physiological side of phonetic change was only incidentally touched upon in the first edition of this work. This, it is believed, caused some indistinctness in the account there given.

2 Thus Prof. Whitney (On Lepsius' Standard Alphabet, p. 24) argues that while some consonants (k, g, &c.) are always consonants, and some vowels (a, e, o) are always vowels, yet some vowels (i, u) are consonantal, while I and r, and even the nasals to some degree, are vocalic: accordingly he declines to draw a hard line between, for example, u and w; but he holds that we ought to present the entire body of known sounds in lines, ranging from the most unmixed vowel-scund to the most absolutely obstructive consonant of each class. Thus what we may call the labial line would be a, au (the English sound), o, u, w, m, v, f, b, p. No doubt such a scheme is useful in bringing clearly forward the connection of sounds: but by the definition adopted in the text it is believed that a real difference between u and w is expressed.

3 Principles of Speech, p. 12.

CH. IV.

Difference

in the material of speech:

the mouth being slightly open. But no prolongation of u will make w. It is not heard till there has been the slightest closure of the inner edges of the lips: and when that closure has ceased, and the lips open, then the w is heard. Since however this action may be exceedingly slight, the difference between u and w may be infinitely reduced. We have thus distinguished sounds into vowels and consonants. But now comes another important distinction. The material of speech is breath, put into form by the position or action of the mouth. But this material may suffer an important change before it enters the mouth. If the glottis or aperture through which the breath passes from the windpipe be fully open', then breath pure and simple issues into the mouth, there to be moulded into sound. But the sides of this aperture are two ligaments called the chordae vocales, and these chords are capable of tension, and of being brought together so as to close the glottis altogether. When by their approximation the glottis is narrowed, they vibrate as the breath passes through and thus the breath is rendered sonorous, and becomes-not mere breath, but genuine voice. Now surd," or the same position of the organs of the mouth will give different results, according as breath or voice is emitted from the windpipe. Check the breath by closing the lips, and as the lips open the sound p is heard. But check voice in the same way, and the sound will be b.

"Hard,"

66

66 breath

sounds," distin

guished

from

66

soft," "sonant,"

sounds."

:

or "voice- Physiologically there is no other difference between p and b. Many terms are in common use to express this distinction. Thus p and the other similarly produced sounds are called hard, or surd, or voiceless, and b is called soft, or sonant, or voiced. None of these terms are quite unassailable: no consonant is properly "surd:" and "voice" in ordinary speech is applied to the current of air which rms pa, just as much as that which forms ba: "voicetoo is a mere negation; "breathed" would be better. the name matters little if we understand the idea.

1 See however note at page 64.

2 See M. Müller's diagram, Lectures 2, 113.

In general, when I am not speaking of the method in which each sound is produced, I shall retain the names hard and soft, as being well known, conveying an intelligible idea, and not likely to mislead if properly explained. All vowels are modifications of voice, that is, soft: consonants are either breathed or voiced, hard or soft.

CH. IV.

Double classification of consonants: ing to the

I. accord

nature

of the

contact.

Consonants' may be classified either by the nature and degree of completeness of the contact of their respective mouth-organs, or by the names of the organs concerned. Thus where the contact is complete, where the current of the breath is entirely stopped, we have shut or explosive sounds: these are the breath-sounds K, T, P, with the corresponding voice-sounds, G, D, B. To this class must be (i.) Moadded the aspirates GH, DH, BH, of the original Indo-Euro-mentary or explosive pean speech, which have been preserved unchanged by no consonants. nation except perhaps the Hindus. The distinguishing mark of these nine sounds is their incapability of prolongation if we attempt to prolong them we only prolong the time during which the breath is pent up against the tongue or the lips, during which no sound at all is heard: the sound as before explained is merely the outbreak of the breath or voice when the obstruction is removed; and this can last but for a moment: hence these sounds are well called Momentary. This title marks them off from all the others to follow, which are capable of prolongation until the air in the lungs is exhausted, these are therefore called Continuous, or (as distinguished from explosive) fricative sounds. In them sound is heard while the organs are still in a position more or less open according to the degree of contact: and if there were no further closing and opening of the organs, they would by definition be vowels: but, because the sound is not fully heard until the organs are relaxed from the position in which they were held, a certain action (sometimes exceedingly slight) is required, and therefore they are consonants.

1 The description of the English sounds in the following sketch is taken almost entirely from Mr Bell's Principles of Speech.

CH. IV.

(ii.) Continuous or frica

tive consonants:

(1) nasals,

I have said that the aspirates GH, DH, BH were momentary sounds: but they have some claim to an intermediate position between the two classes. They have not been retained in any European speech, and therefore their sound is strange to us, and must remain a matter of some doubt. The opinion commonly held about them is that they were not originally compound sounds (as might seem to be implied by these symbols): that is, gh would not be truly represented by such a compound as that in "loghouse," where there are two elements of about equal importance. The sound represented by the h in each case was much less than this, and also was not separated from the preceding momentary sound. It was probably the slight escape of breath, which is possible at the very instant of the sundering if the contact be exceedingly slight. This appears to be the sound given to the Sanskrit aspirates in India at the present day. This slightness of contact explains the passage of unaspirated into aspirated letters which is clearly traceable in Sanskrit: this takes place almost always when the letter is combined with another consonant, generally s: e.g. Indo-European STA becomes Sanskrit √stha. Now in all compounds one part is liable to suffer here the exertion in pronouncing the s diminishes the strength of contact for the t; and the result is a slighter t followed by a breath in the same action1. It is because of this slighter contact, and because of the possibility of slightly prolonging the breath, that these aspirates approximate to the continuous consonants: but they are much nearer to the momentary class, and will be included in it.

First among the continuous letters we take the nasals, NG, N, and M, because of their close relation to the shut consonants. To explain their nature we must enter a little more into physiology. There is a cavity called the

1 Sometimes the aspirate seems to be produced by the assimilating influence of a preceding r, e.g. in pra-thama "first," where the analogy of other languages leaves no doubt that the suffix was originally tama.

pharynx immediately behind and rather higher than the mouth it is the termination of the throat canal, prolonged behind the soft palate, which divides it from the mouth'; and it has a passage to the air by tubes through the nostrils. It is thus obvious that air issuing from the windpipe may finish its course in two ways: either passing above the tongue, and below first the soft palate, then the hard palate, and so to the mouth; or behind and above the soft palate and by the nasal tubes to the nose. Now this soft palate is moveable: it can be raised so as to cover entirely the opening of the nasal tube from the pharynx, and this is its position whilst we sound all the consonants hitherto described: during them the breath has access both to the pharynx and the mouth, but having no egress from the pharynx it is discharged from the mouth alone : the pharynx however plays an important part, for in every proper articulation the pharynx is distended with the breath, and it is the natural contraction of the pharynx alone, when the check in the mouth is removed and the air escapes, that causes the audibility of each of these letters. But if the soft palate be low enough to allow air to pass to the nostrils, as well as to the mouth-cavity, then the nasals are produced, whilst all the organs of the mouth are in precisely the same position as they were for the explosive consonants, and the sound is not complete till that position ceases. It is possible to produce them either with breath or voice; but it does not seem that any European nation has ever used breath-nasals. Therefore NG, N, and M, are nothing but nasal modifications of the voiced-sounds G, D, B. Similar modifications of vowels are clearly possible, and are actually used by

The soft palate or velum pendulum, with the tongue-like little uvula attached to it below, may be seen by any one who will stand before a glass and put his mouth into the position for sounding Ah.

This is Mr Bell's account: see Principles, p. 44.

The common phrase that a person "speaks through his nose" when he has a cold, is therefore clearly erroneous. He ought to speak through his nose, but cannot; the ends of the apertures being closed: therefore instead of m he sounds nearly a b: and instead of n either a d or an l.

CH. IV.