Foreign.— F French, g German, it Italian. 10. b ê te f, l â che F, n ò it, d û F, f eu F, v eu f F, v in F, v en t f, v on t f, un f, s oi n f, s oi F, l ui f (bEEt, l aa sh, no, dyy, f ǝ , voef, veΛ, vaΛ, vo Λ , oeΛ, sŭeΛ, sŭa, lyí). 11. da ch tei ch , ta g e siege, w ahl, all g (dakh táikjh, taaghe, ziigjhe, bhaal). 12. pag li a it, beso gn e F (pà·ljà, b ǝ̯ zonj).

2. Nature of Speech-Sounds.— Speech-sounds result from shocks given to the air by the organs of speech, received by the drumskin of the ear, and transmitted to the auditory nerves in the cochlea. The apparatus is explained by Helmholtz, who deals with musical sounds. But speech is not musical, and its sounds are much obliterated when rendered musical. An original quality of tone generated by the vocal chords is modified by the cavities through which it passes, as explained by Helmholtz (Sensations of Tone) on the principle of resonance. There are three ways in which speech-sounds may be produced—(1) by the air in the mouth, without additional breathing, by smacks and clicks; (2) by drawing in air, as orally in chirps, whistles, sobs, gasps, and nasally in snuffles, snores; (3) by expelling air, as in the greater number of speech-sounds. The last are either fated, the vocal chords being wide apart and hence not vibrating, but allowing breath to pass freely, or voiced, the vocal chords being close together and vibrating fully, or else whispered, the vocal chords approximating but not touching, and their edges only vibrating. The last is only a variation of the second and needs no further consideration. Flated and voiced sounds are either fixed, the position of the vocal organs remaining unchanged throughout, or changing, the position constantly altering from one fixed position to another, forming “glides.”

3. Generic and Specific Speech-Sounds.— Fixed speech-sounds, intended to be the same, vary from speaker to speaker, and in the same speaker at different times. Those who attempt to write sounds from dictation rapidly find that they have to disregard these specific differences, and simply discriminate genera. And much difference of opinion has always existed as to the discrimination and number of genera.

4. Vowels, that is, vocals, are so called because their positions allow the voice-sounds to pass with least obstruction. The three genera (ii, aa, uu), which have always been distinguished, differ greatly in the positions of the tongue and lip, that is, in their mouth cavities, and hence resonance. The usual method of describing speech-sounds is by the shape of the cavity, which, however, could be shown to be insufficient for many reasons. As differently shaped cavities resound to a note of the same pitch, Helmholtz proposed the last for discrimination. The pitches of (ii, aa, uu) are widely different, (ii) having the highest and (uu) the lowest; but the extreme diversity of results in attempting to assign the actual pitch of vowel cavities shows that this will not suffice. Resonance cavities do not create but merely modify original vowel qualities of tone, and these last seem to depend upon the will of the speaker, guided by his powers of appreciation and imitation, both extremely variable, partly hereditary, partly depending on conformation of brain, and partly acquired during adolescence.

Melville Bell, Sweet, Storm, and Sievers, and all who have latterly examined the subject distinguish at least two series of vowel genera, that is, two forms of each genus, called “narrow” and “wide but they are far from being agreed as to what the difference consists in and how it is produced. Sweet differs from Bell, and Sievers does not wholly agree with Sweet. All, however, call (ii, uu) narrow, and (i, u) wide.

Besides these two series Bell introduced another distinction applying to both, termed “rounding,” consisting in a greater or less closure of the lips, slight for (aa), much for (uu), and intermediate for (oo ). But this character is not scientifically precise, because all the vowels can be produced with the mouth wide open (by means of a compression of the arches of the palate), and still more easily with the mouth at least as much closed as ordinarily for (uu). Other phonetists wish to introduce distinctions based upon the shape of the apertures between the lips.

There is also a feeling of intermediateness between vowelsounds. Thus (yy) is felt by many to lie “between” (ii, uu), and (oeoe) between (oo, ee). But we also have other intermediates which arise spontaneously when listening to new languages and dialects. Thus in west Somerset there is a vowel between (ə, i), one between (y, ə ), and another between (ə , oe), and the positions for these vowels have not been ascertained. These are only specimens of numerous cases. Hence the positional discrimination breaks down at present. Nevertheless it is very good so far as it goes, but must not be pressed to extremes.

All the vowels may be also flated and whispered; that is, the position and dictating vowel-intention remaining, the totally or partially open vocal chords forbid voice and produce sound more or less recognized as substitutes for the true vowels. Write (ii) voiced, ('ii) whispered, (“ii) flated. This distinction becomes of more importance for consonants.

5. Glottids and Physems.— A glottid is the action of the vocal chords in altering the form of the glottis or tongue-shaped space between them. (1) The glottid is clear when there is no attempt to utter the vowel until the chords are brought together, yet the utterance takes place at that instant. This may be written (,ii) initial. Similarly, a vowel may end with a clear glottid (,ii,), no flatus escaping after the vowel ceases. This clear glottid is usually inferred and not written. (2) The glottid is gradual, written (ן), when flatus passes through the vowel position before the chords are sufficiently approximated for voice, or after they are separated, thus (ןiiן) is really (“ii+ 'ii + ii +'ii +“ii). This is an exceedingly common habit with some speakers. (3) The check glottid (;), Arabic hamza *, arises from keeping the chords tightly closed so that they cannot vibrate, and then releasing them with an explosion. It may be final in reverted order in Arabic, and it is common as an initial in German, as ;eine;er;innerung, and is used as the catch accent in Danish, as ma;nd, a man, distinct from man = F on. (4) An exaggeration of (;) gives Arabic (g⅛in) £, the bleat, with a rattle in the cartilaginous glottis.

Physems are the bellows-actions of the lungs. (1) The jerk (h) or sudden puff of either vocalized or flated breath, accompanying either clear or gradual glottid. The first, with voice only, is the singer’s and Bengali aspirate; the second, with flatus, is the Scotch or German aspirate. (2) The wheeze (h ), Arabic stated by Czermak to arise from suddenly forcing breath through the cartilaginous glottis.

6. Vowel Glides and Vanishes.— So far the positions of the vowel above the larynx have been supposed to remain unchanged. In this case many degrees of length may be distinguished, as (ă) very short, (a) short, (à) medium, (aa) long, (aà) drawled, (aaa) extravagantly prolonged. If the vowel sign consists of two parts, as (ah), only the first is marked doubled or tripled for these lengths, as (ăh, aah), &c. In English it is felt very difficult to preserve the positions for long (ee, aa, oo), and these vowels gravitate to, without by any means reaching, (i, ə, u). The first and last may be written (edj, odw), implying what are termed vanishes or gliding alterations of sound, accompanied by alterations of position as the vowel ceases. This change is generally unintended and mostly used unconsciously.

7. Diphthongs.— But there are conscious changes to quite different positions. The first and last vowels are then taken as fixed, one of them having the chief stress, and there is a vowel glide between them. These form diphthongs; the stress and glide being the chief characteristics are marked by ('), and the two elements are juxtaposed. The glide is generally short and close in English, longer in German, still longer and looser, or “slurred,” in French and Italian. There are many typical classes, i. With weak final (i), unanalysed (a'i), analysed (ái, á i, ə' i, Ǝ' i , éi, é i, E 'i, ae'i), &c., all common. ii. With weak final (u), unanalysed (a'u), analysed (áu, á u, ǝ 'u, Ǝ' u , éu, é u, E' u, ae' u ), &c., all very common, iii. Weak final (y), theoretic German eu (óy, áy), Devonshire ow (oe'y 1 5 ). iv. Weak initial (i) or (ĭ), used for (J) in Italy, France, Wales, &c. v. Weak initial (y) in Fr. vi. Weak initial (u) or (ŭ), used for (w) in Italy, Spain, France, Wales, &c. vii. Murmur diphthongs ending in weak (e ), common in English, but generally with the option of trilling an (r) after it, and hence written (ɹ), as in ear, air, oar, lord, poor, pure, pyre, power (ii ɹ , eeɹ, oo ɹ, lAAɹd, puuɹ, pĭ uu ɹ , pa' i ɹ, pa'uɹ); the r is always trilled in Scotland, viii. The vanish diphthongs (ee ’j, oo ’ w ), just considered. ix. Inchoant diphthongs, first grave, where the speaker begins too low and corrects himself, as (í i, ú u), and secondly acute, where he begins with the mouth too open and corrects himself as he proceeds, as (Ǝ'o); both are common in English dialects.

8. Glides from and to Mutes, Post-Aspirates, Sonants.— The essence of the diphthongal character was the glide, which was independent of the sounds of the first and last elements. These might be absolutely mute, as in (piip, tAAt, kook) peep, taught, coke, in which (p, t, k) are mere positions without sounds. But the results are quite different from (ii, aa, oo ), because while the consonant positions are opened out the vowel is at the same time sounded. Similarly in the reverse order, when final. But here the enclosure of the breath is felt to be uncomfortable, and, if there is no vowel to fall upon, the mouth is opened and a puff of flatus (‘), called the “recoil,” is heard in England, as (piip') peep! Using then (+) for the gliding sounds, we have (p + ii + p + '); but there is no recoil in (p + ii + p + i + q) or (p + ii + p b + oo ’ w !) peeping, peep-bo! Various nations'have very different habits in this respect. In Indian languages (p') would be felt as a final post-aspirated mute. So initially in Germany, the (p) position is usually released, not on a vowel with a clear glottid, as in England and Italy, but on a vowel with a gradual glottid, as (pןii), and hence flatus is heard before the vowel. When this is exaggerated, as (pןhii) or (pHןhii), we have the true Indian post-aspirated mute.

But an attempt to utter the vowel through a mute position may be made before the position is quite opened out, or the vowel may be continued into it after it has been assumed. This gives the English, Italian, and Indian “sonant,” as in (b ee b) babe. The German is not quite the same. Here the glides are (b + ee + b), with possibly a voiced recoil (b + ee + b + ’), where (’) represents the most amorphous voice. This voiced recoil is strong in French, but seldom heard in English, except in declamation, is regular in modern Indian, and impossible to a German, who says at most (b ee bp') or (b ee p'); also Indians and Irish sometimes jerk out their vowel after sonants, as (bH ee bH'), producing the sonant post-aspirates. The ancient Indian never ended words in the pause with sonants, post-aspirated mutes, or post-aspirated sonants, but only with simple mutes, and avoided the recoil.

9. Glides to and from Hisses, Buzzes.— In the case of a hiss, flatus passes through the consonant position and is continued part of the time during which the vowel position is assumed, but towards the end of that time voice is put on. Hence in (s + ii) see, the glide (+) is partly flated and partly voiced, so that (s) acts in much the same way as a gradual glottid; similarly when final, as (s + ii + s) cease, where the hiss replaces the recoil. But the proportion of voice and flatus in the glide may vary. The voice may be put on during the hiss, and then the change takes place in the hiss position. The result, far less clear than a vowel, is a hiss (s), followed without a positional glide by the buzz (z), then an entirely vocal glide, the vowel, and a vocal glide, a buzz, and a hiss, as (sziizs) seize, sees. The initial (sz) is regular in Germany, where no vowel precedes, as sie sehen (szii z ee ’n), they or you see; and the reverse (zs) is regularly in English seize (siizs) in the pause, and similarly (haavf, briidhth, ruuzhsh, dɟǝdɟshj) halve, breathe, rouge, judge. In the south-west of England Saxon words beginning with s, f are pronounced with (z, v) initial, which passes through (sz-, fv-) to (s, f).

10. Glides to and from Flaps.— Flaps are consonants where there is a slack organ which flaps with the breath as it passes. The r is very varied, but properly voiced, though the flated form occurs. The flap may be made (1) with the lips, as (brh), used in Germany to stop horses; (2)

with the tip of the direct tongue, (r, r), used in Italy; with the tip of the reverted tongue, (r), used in the south of England and in modern (not ancient) Indian, where it is called “cerebral”; (3) with the uvula, (r ), common in France and north Germany, labialized (rw ) in Northumberland, and harsher in Greek and Arabic; (4) with the glottis, (ן), usual in Denmark; and so on. In the educated south of England the tongue is often raised to the (r) position, but not allowed to flap, and is treated as a buzz (r o ).

The above form the central flaps; if the point of the tongue is fixed and the voice escapes by the side it causes minute lateral flaps of the tongue. The place of the point of the tongue discriminates the various sounds which differ but slightly—(l), advanced tongue at gums, Continental; (1), coronal, tongue near the crown of the palate, English; (l), reverted, in connexion with (r) in south-west England.

Both flaps, especially the latter, are extremely vocal, and the glides from and to them are like those from and to vowels, while they glide readily to and from mutes, sonants, hisses, and buzzes.

11. Glides to and from Hums, Orinasals.— For (p, t, k) both nasal or oral passages are cut off, the former by pressing the uvula against the back of the pharynx. Let this pressure be relaxed so that the nasal passage is opened, the oral passage remaining closed. The voice passes through the nose, forming the three hums (m, n, q). The glide from these to ordinary vowels is the same as from (b, d, g), and the peculiarity consists in the preceding hum and the closing of the nasal passage as the vowel position is assumed. If the nasal passage is left open at all the vowel is “nasalized,” and as it resounds partly in the nose and partly in the mouth it becomes an “orinasal.” Four principal orinasals exist in French, as an, on, un, vin (a λ , o λ, oe λ, ve λ) ; there are more in Portuguese, and many others in the modern Indian languages. The oral vowel is altered in character by nasalization, and it is not possible to assign the oral to the orinasal form precisely. If the oral passage is only slightly open, a “nasalized tone” is produced, as in Gaelic, some south German, and American dialects, written as (a i ). The hum also may be prolonged, and (’mpaa, ’mbaa, ’ntaa, ’ndaa, ’qkaa, ’qgaa) result. These forms exist in South African languages.

The final hum may be continued like a vowel. If the nose entrance is closed and the voice continued (l i m, laem, k oo m) become (l i mb, laemb, k oo mb), which, as the ordinary spelling shows, were probably once pronounced. But not only the nasality, the voice itself may be cut off, and then we have the mere stops (p, t, k), thus (l i mp, l i nt, l i qk), which in the pause have the recoil. Some phonetists consider (m, n, q) to become flated in this case, as (mh, nh, qh). This is no more necessary than to suppose a vowel to be flated before a mute, so that (laep, maet, haek) lap, mat, hack should be (l“aep, m“aet, h“aek), a usage unknown.

12. Palatalization and Labialization.— When a consonant precedes a diphthong of classes iv, v, vi, in art. 7 beginning with weak (i, y, u), there is a tendency to take these vowels as nearly as possible simultaneously with the consonant, expressed by writing (j, w j, w) after the consonant. To say (tj) at least two-thirds the length of the tongue from the tip backwards must lie against the palate, for (kj) two-thirds from the root forwards. The first occurs in Hungarian; the second was very common among older speakers of English before (ae), as (kjaend’l). Both (tj, kj) are apt to develop into (tɟ); compare nature, kirk, say (neetɐɹ, kjǝǝɹkj), with colloquial modern (n ee tɟɐ, tɟǝǝtɟ). Similarly the voiced sounds (dj, gj) become (dj), compare rig, ridge. These (tɟ, dɟ) are consonantal diphthongs = (tjshj, djzhj), as in ch est, j est, and are distinct from the Indian sounds (kɟ, gɟ) W, ∣Γ, which are true mutes, produced by bringing the tongue from the position for (j) tight up against the extreme back of the hard palate so as to produce a complete stop. The most important of the palatalized letters are (lj, nj), the Italian gli, gn in miglior, ognor (mìljò·r, ònjò·r), where the palatalization brings the Italian advanced (l, n) to the position of the English (l, n). The (lj) has degenerated to (ĭ) or (j) in France during the 19th century. It exists in Spanish ll, Portuguese lh. The (nj) exists as gn in French, ñ in Spanish, and nh in Portuguese.

Parallel to the palatal are the labial forms, of which English queen, guano (kwiin, gwaa·no) are examples. They seem to exist in abundance in French, as in toi, doigt (t w a, d w a). The palato-labial form (w j), as in juin (zh w jeΛ), is much disputed, and a diphthong (zhyéΛ) is usually assumed.

13. Syllables.— A group of speech-sounds increasing in volume from a mute, sonant, hiss, buzz, or flap to a full vowel and...