1

CH. IV.

R passes into w.

Classifica

tion of vowels.

the tongue may indeed touch the teeth, but this is not. essential to the sound (as for the proper th). Prof. Huxley tells1 of a man whose tongue had been amputated, who could yet sound th: the contact must have been given by the stump of the tongue against the edge of the palate. S may however pass into the true th, if in sounding it the tongue be allowed to touch the teeth, because then central contact takes place, and the sound must be emitted laterally. With respect to continuous labials I know no general rule that can be given. I think no language has more than one pair of them. R passes into w in lisping, which agrees with the natural order of the sounds; which order holds, as a rule, for continuous as well as for momentary sounds. But as they are less firm than momentary sounds there is more scope for assimilation, which may neutralise the common order.

It is more difficult to determine with exactness the true sound of vowels than of consonants. For many consonants there can be little or no variation: in whatever lands Ρ and b are sounded they must be sounded at the same place; there can be no great range for k and g; and so on. The points too of contact, which give distinct difference of sound, are limited. But we can never be sure that we have obtained all the varieties of vowel sound: they are unlimited. For our purpose however it is sufficient to point out the best marked points in the scale, in the assurance that if some of the Greek and Latin vowels did not exactly coincide with any of these, the difference could not have been great. As I said at the beginning, all vowels must be pronounced as in Italian, except English sounds which are given as examples in brackets; e. g. (ee) denotes the i sound. It is to be remembered that vowels are various open positions of the vocal tube,

I greatly doubt. The instance "says for sayeth" (Prof. Mayor, 1. c. p. 335) is, I think, an error; the two forms were distinct: " sayis" or sayes is old North English, both sing. and plur.; to which in old South English "saith" corresponds; though later "sayen" was the plural in use.

1 Elem. of Phys., p. 203.

modifying in various ways the breath which has received tone from the vibration of the stretched chordae vocales.

The two most important vowel classifications, with which I am acquainted, are those of Prof. Lepsius, and Mr Melville Bell. They differ considerably: that of Prof. Lepsius is best suited to render plain the historical development of vowel sound in Greece and Italy, and will in the main be followed here. But Mr Bell's system is important in many ways, and as I shall take a good deal from it, it will be best briefly to describe it first.

CH. IV.

Mr Bell recognises in all thirty-six vowel sounds, and Vowels and twelve "glides;" the latter being "transitional" sounds, glides. which differ from consonants, in that the vocal tube is so open that there is no friction in the mouth: they would therefore be vowels, if they had any "fixed configuration," any permanent position; but they are sounded too short for this, and are essentially incapable of being prolonged. They principally occur as one of the two elements of a diphthong, and in this character they will come under our notice later on. Glides in English are frequent: thus the y at the end of "day," is certainly not a full consonant; nor is it a vowel, for it cannot be prolonged: it is a transitional" sound between the two. Other examples are the w in "now," the r in "are," and "our:" the slight prolongation of vowel-sound heard between a and rin "va'ry," &c."

Of the thirty-six vowels, nine are primary; that is, possessing the smallest opening of the vocal tube necessary to distinguish them from consonants. Three of these are formed with the back of the tongue, and so the position of the mouth does not differ much from that for k and these are called "back" vowels. Three are "front" vowels: for these the back of the tongue is raised to the arch of the palate, much as for y. The other three are called "mixed," because they combine the properties of each of these classes, being formed with the back of the tongue, 1 Visible Speech, p. 69 and 94.

g:

P. E.

6

I. Nine
“primary
or "close'
vowels.

(1) Three

back,"

(ii) three "front,"

(iii) three

mixed,"

CH. IV.

and crosswise, three "high," three

“mid” and

three
"low."

II. Nine

"wide" or open' vowels, modifica

tions of the nine close vowels.

but with the edges also raised to the sides of the teeth. The three vowels in each of these three classes are called “high,” “mid," and "low," respectively, according to the degrees of elevation of the tongue. Thus, for example, the English (ee) in “eel,” is technically called a "high front” vowel; the (u) in "up" is a "low mid," while the vowel of the French "que" is a "mixed mid1." Next, we have nine modifications of these nine primary vowels, called "wide" vowels. In these, “the resonance cavity is enlarged behind the configurative aperture"," by drawing back the soft palate and expanding the pharynx. By doing this, for example, when the tongue is in the same position as for sounding "eel," we get "ill"-a "high-back wide," instead of a "high-back" vowel. Instead however of the terms "primary," and "wide," I shall use the terms "close," and "open," for these two classes respectively, as being already in use in some languages to express, I believe, the same distinction. Lastly, III. Eigh- each one of these eighteen vowels may be modified in a further way—thus giving us the whole thirty-six. The external sign of this modification is a contraction or rounding" of the aperture of the lips, representing, as Mr Bell believes, a corresponding contraction of the guttural passage (the real cause of the change of sound) and probably also of the upper part of the larynx. If we "round" the vowel in "eel," for example, we get the German ü, "high-front round," according to Mr Bell's arrangement, "in which the labial orifice is reduced to little more than a chink :" while the broad aperture of (ah), which is a "low-back wide," is "rounded" to (o) in "on," "odd," &c. ("low-back wide round"), "by contracting only the corners of the lips"."

teen

"round"

vowels,

modifica

tions of the

first eighteen.

Number of English vowels.

(

Out of these thirty-six sounds, twenty-two occur in English. But six of these, the "mixed-wides," differ from each other very slightly. They are the sounds heard in places" (High), “fatal" (Mid), "err" (Low), "pleasure"

This Table is taken from Visible Speech, p. 94, with two differences. I have employed the terms "open" and "close," instead of Mr Bell's "wide" and " "wide" and "primary;' and I have exhibited the " Low," Middle," and "High" letters together. The examples above the line in such divisions are Mr Bell's own; in a few cases however I have substituted examples, which he has elsewhere given, instead of some American and Cockney variations which he gives in this Table. I am responsible for everything below the line in each division. Where the space is blank, there is no English sound to give and none in dialectical English or foreign languages which seemed to me important for my present object. If indeed I had been consistent, I should have omitted all the English final syllables which are so acutely distinguished by Mr Bell; but I thought it worth while to include all English variations which could be discriminated by any ear.

CH. IV.

The

"neutral" vowel.

Reasons for commencing with a.

(High Round), "oratory" (Mid R.), "orator" (Low R.). It will be observed that five of these sounds occur in unaccented syllables: and that they all represent variations of what is commonly called the "neutral vowel," that sound to which the vowels of all unaccented syllables have a tendency to return, and which is also supposed to be heard in "but," "son," "fir," and to be sufficiently denoted, by a single symbol, as by the e of Prof. Lepsius. At all events, Mr Bell's distinctions will not be required by us, but the existence of the neutral vowel itself must be borne in mind, as it undoubtedly occurs in every language. It is sometimes called the Urvocal-unfortu nately, as that may be interpreted to mean that it is the vowel (instead of ah) from which other vowels can be traced. Its most typical formation is no doubt that given by Mr Bell for the a in fatal, where the tongue is in the position of perfect rest, without action of either the back or front part.

We start with the vowel a (ah), for two reasons. First, on physiological grounds-it is the vowel par excellence, the most open vowel position: the back of the tongue is depressed beneath the soft palate, which however does not seem to be raised1: the lips are in the most open position, not "rounded." The second reason is found in the history of the vowels: we find, as a matter of experi ence, that when there has been change between ɑ and i, or a and u, the change has always been from a to i, or from a to u, and not the reverse. Without, then, asserting that there ever was a time when a existed, and i and u did not, we may yet fairly call a an older vowel, because in many cases i and u are demonstrably younger. It has been pointed out by Schleicher that a is more distinctly vocal than either i or u, because while they are closely cognate to y and w, it always preserves its full vowel character. These three vowels a, i, and u belonged to the Indo-Euro

1 Max Müller (11. 124), gives the result of Prof. Czermak's experiments, that the soft palate is lowest for a, then for e, o, u, i in order.