1911 Encyclopædia Britannica/Phonetics
PHONETICS (Gr. φωνή, voice), the science of speech-sounds and the art of pronunciation. In its widest sense it is the “science of voice,” dealing not only with articulate, but also with the inarticulate sounds of animals as well as men. The originally synonymous term, “phonology,” is now restricted to the history and theory of sound-changes. The most obvious of the practical applications of phonetics is to the acquisition of a correct pronunciation of foreign languages. But its applications to the study of the native language are not less important: it is only by the help of phonetics that it is possible to deal effectively with vulgarisms and provincialisms of pronunciation and secure uniformity of speech; and it is only on a phonetic basis that the deaf and dumb can be taught articulate speech. From a more theoretical point of view phonetics is, in the first place, the science of linguistic observation. Without phonetic training the dialectologist, and the missionary who is confronted with a hitherto unwritten language, can neither observe fully nor record accurately the phenomena with which they have to deal. These investigations have greatly widened the scope of the science of language. The modern philologist no longer despises colloquial and illiterate forms of speech. On the contrary, he considers that in them the life and growth of language is seen more clearly than in dead literary languages, on whose study the science of comparative philology was at first exclusively built up. It was not till philologists began to ask what were the real facts underlying the comparisons of the written words in Sanskrit, Greek, Latin, and the other Indo-European languages, embodied in such generalizations as Grimm's Law, that “letter-science” developed into “sound-science” (phonology). The rise and decay of inflexions, and the development of grammatical forms generally, are, from the formal point of view, mainly phonetic problems; and phonetics enters more or less into every department of historical and comparative grammar.
Methods of Study and Investigation.—Phonetics is the science of speech-sounds. But sounds may be considered from two opposite points of view—the organic and the acoustic. From the organic point of view a sound is the result of certain actions and positions of the organs of speech, as when we define f as a lip-teeth (dento-labial) consonant. This is the point of view of the speaker of a language. To the hearer, on the other hand, f is not a lip-teeth, but a hiss consonant similar to that denoted by th. This is the acoustic point of view. Theoretically, the organic study of phonetics is a branch of anatomy and physiology: that part of these sciences which deals with the organs of speech (see Mouth) and their functions (see Voice), while, from the opposite point of view, the study of phonetics is based on that branch of physical science known as acoustics (see Sound), together with the anatomy and physiology of the organs of Hearing (q.v).
Unfortunately, this basis is still imperfect. The principles of acoustics are well established, and we know much about the anatomy of the ear. But how the ear transmits to the brain the impression of sound is still a mystery. Again, although the mechanism of the vowel is clear enough, there is still no generally received acoustic theory of its formation. In fact, from the physical science point of view there is as yet no science of phonetics.
The real function of phonetics is philological and literary. The only sound basis of a theoretical knowledge of phonetics is the practical mastery of a limited number of sounds—that is to say, of the sounds which are already familiar to the learner in his own language. It is evident that the more familiar a sound is, the easier it is to gain insight into its mechanism and to recognize it when heard. It is indispensable to cultivate both the organic and the acoustic sense. These processes we are continually carrying out in ordinary conversation. All, therefore, that we have to do in dealing with native sounds is to develop this unconscious organic and acoustic sense into a conscious and analytic one. The first step is to learn to isolate each sound: to pronounce it, as far as possible, apart from its context; and to preserve it unchanged through every variation of length and force, and in every combination of sounds. The next step is to analyse its formation. Let the student, for instance, compare the two consonants in such a word as five by isolating and lengthening them till he can both hear and feel the voice-vibration in the second one. In the same way let him learn to feel the changes in the position of the tongue and lips in passing from one vowel to another. When the native sounds have been thoroughly studied in this way, the learner will proceed to foreign sounds, deducing each new sound from those which are already familiar to him.
But there are other methods besides those of direct observation by which the positions of the tongue may be objectively determined and measured with more or less accuracy. The interior of the mouth may be explored by the fingers. If the little finger is held against the gums during the articulation of the vowels in it, ate, at, the difference in the height of the tongue will at once become apparent in the formation of the first vowel the tongue is pressed strongly against the artificial palate, while in that of the second it only just touches it, and in that of the third it does not touch at all
Several forms of apparatus have been devised for a more accurate determination of the positions of the tongue and the other movable organs of speech. The best results hitherto as regards the vowel-positions have been obtained by Grandgent, who uses disks of cardboard of various sizes fixed to silver wires. A full description of this and other methods will be found in Scripture's Elements of Experimental Phonetics.
There are other methods whose results are obtained only indirectly. The simplest of these are the palatographic, by which are obtained “palatograms” recording the contact of the tongue with the palate. The apparatus most generally used consists of a thin, shell-like artificial palate, which is covered with chalk and placed in the mouth; when the sound is made, the articulation of the tongue is inferred from the contact-marks on the plate. This method is evidently limited in its application. It, too, has the drawback of not being applicable to the sounds formed in the back of the mouth. The outlines of palatograms are much vaguer than they appear in the published drawings of them; and it is a question whether the thickness even of the thinnest plate does not modify the record.
The methods hitherto considered are all comparatively simple. They require no special knowledge or training, and are accessible to all. But there are more elaborate methods—with which the name “experimental phonetics” is more specially connected—involving special training in practical and theoretical physics and mathematics, and requiring the help of often complicated and costly, and not easily accessible, apparatus. The investigation of the speech curves of phonograph and gramophone records is a typical example. Good examples of these methods are afforded by E. A. Mever's investigations of vowel-quantity in English (Englische Lautdauer, Uppsala, 1903). Their characteristic feature is their delicacy, and the minuteness of their distinctions, which often go beyond the range of the human ear. Although their results are often of value, they must always be received with caution: the sources of error are so numerous.The claims of instrumental phonetics have been so prominently brought forward of late years that they can no longer be ignored, even by the most conservative of the older generation of phoneticians. But it is possible to go too far the other way. Some of the younger generation seem to think that the instrumental methods have superseded the natural ones in the same way as the Arabic superseded the Roman numerals. This assumption has had disastrous results. It cannot be too often repeated that instrumental phonetics is, strictly speaking, not phonetics at all. It is only a help. it only supplies materials which are useless till they have been tested and accepted from the linguistic phonetician's point of view. The final arbiter in all phonetic questions is the trained ear of a practical phonetician: differences which cannot be perceived must—or at least may be—-ignored; what contradicts the trained ear cannot be accepted.
Sound-Notation; Spelling Reform.—Next to the analysis of the sounds themselves, the most important problem of phonetics is their representation by means of written and printed symbols. The traditional or “nomic” orthographies of most languages are only imperfectly phonetic. And, unfortunately, of the languages in most general use, two are exceptionally unphonetic in their orthographies, French showing the greatest divergence between sound and symbol, while English shows the maximum of irregularity and arbitrariness. The German orthography is comparatively phonetic: it has hardly any silent letters, and it generally has one symbol for each sound, each symbol having only one value, the exceptions falling under a few simple rules, which are easily remembered. There are other languages which have still more phonetic orthographies, such as Spanish, Welsh and Finnish. But even the best of them are not perfect: even when they are not actually misleading, they are always inadequate. On the other hand, no system of writing is wholly unphonetic. Even in French and English there are many words whose spelling not even the most radical reformer would think of altering. In fact, all writing which has once emerged from the hieroglyphic stage is at first purely phonetic, as far as its defective means will allow. The divergence between sound and symbol which makes spelling unphonetic is the result of the retention of phonetic spellings after they have become unphonetic through changes in the pronunciation of the words themselves. Thus, such English spelling as knight and wright were still phonetic in the time of Chaucer; for at that time the initial consonants of these words were still pronounced, and the gh still had the sound of ch in German ich. So also see and sea are written differently, not by way of arbitrary distinction, but because they were pronounced differently till within the last few centuries—as they still are in Irish-English.
Where there is no traditional orthography, as when Old English (Anglo-Saxon) was first written down in Latin letters, spelling was necessarily phonetic; but where there is a large literature and a class of professional scribes, the influence of the traditional orthography becomes stronger, till at last the invention of printing and the diffusion of one standard dialect over a large area occupied originally by a variety of other dialects make changes of spelling as inconvenient as they were once easy and natural. The ideal orthography for printers is one which is absolutely uniform over the whole territory of the language, and absolutely unchangeable. In such orthographies as those of the present English and French there is no longer any living correspondence between sound and symbol: they are, in intention at least, wholly unphonetic; they are preserved by graphic, not by oral, tradition.
But unphoneticness has its practical limits. A purely unphonetic degradation of an originally phonetic system of writing—one in which there is absolutely no correspondence between sounds and letters—could not be mastered even by the most retentive memory: it would be even more difficult than the Chinese writing. Hence a phonetic reaction is inevitable. In the middle ages the spelling was periodically readjusted in accordance with the changes of pronunciation-as far, of course, as the imperfections of the existing orthography would allow. This adjustment went on even after the introduction of printing. In fact, it is only within the last hundred years or so that the orthographies of English and French have become fixed. One result of this fixity is that any attempt to continue the process of adjustment assumes a revolutionary character. When, in 1849, the pioneers of the modern spelling-reform movement—A. J. Ellis and I. Pitman—brought out the Fonetic Nuz, few of those who joined in the chorus of ridicule excited by the new alphabet stopped to consider that this uncouthness was purely the result of habit, and that the Authorized Version of the Bible in the spelling of its first edition would seem to us not less strange and uncouth than in the new-fangled phonotypy of Messrs Ellis and Pitman. Nor did they stop to consider that phonetics and phonetic spelling, so far from being innovations, are as old as civilization itself. The Alexandrian grammarians were not only phoneticians—they were spelling-reformers; they invented the Greek accents for the purpose of making the pronunciation of Greek easier to foreigners. The Romans, too, were phoneticians: they learnt Greek by phonetic methods, and paid great attention to niceties of pronunciation. The Sanskrit grammarians were still better phoneticians.
As a matter of fact, English spelling was still phonetic as late as the time of Shakespeare—in intention, at least. But although people still tried to write as they spoke, the inherited imperfections of their orthography made it more and more difficult for them to do so. Hence already in the 16th century a number of spelling-reformers made their appearance, including classical scholars such as Sir John Cheke, and A. Gill, who was head-master of St Paul's School in London. Gill has left us extracts from Spenser's Faerie Queene in phonetic spelling; but, strange to say, nothing of Shakespeare's, although he and Shakespeare were exact contemporaries. But Gill's and the other alphabets proposed were too intricate and cumbrous for popular use.
Nevertheless, some important phonetic reforms were successfully carried through, such as getting rid of most of the superfluous final e's, utilizing the originally superfluous distinctions in form between i and j, u and v, by using i, u only as vowels, j, v only as consonants, instead of at random—a reform which seems to have begun in Italy. Another important reform was the introduction of ea and oa, as in sea and boat, which had hitherto been written with ee and oo, being thus confused with see and boot.
All these were as much phonetic reforms as it would be to utilize long s and tailed z (ʃ, ʒ) to denote the final consonants in fish and rouge respectively; a reform first suggested by A. J. Ellis, who was himself the first to call attention to the works of these early phoneticians and to utilize them in the investigations enshrined in his great work on Early English Pronunciation.
With all its defects, the present English spelling is still mainly phonetic; we can still approximately guess the pronunciation of the vast majority of words from their spelling. So when we say that English spelling is unphonetic we merely mean that it is a bad phonetic spelling; and all that spelling-reformers aim at is to make this bad into a good phonetic spelling, that is, an efficient and easy one. But the difficulties are great; and the more we know of phonetics, and the more we experiment with different systems of spelling, the more formidable do they appear. One of the difficulties, however, that is commonly supposed to stand in the way of spelling-reform is quite imaginary: namely, that it would destroy the historical and etymological value of the present system. Thus E. A. Freeman used to protest against it as “a reckless wiping out of the whole history of the language.” Such critics fail to see that historical spelling, if carried out consistently, would destroy the materials on which alone history can be based, that these materials are nothing else but a series of phonetic spellings of different periods of the language, and that if a consistent historical and etymological spelling could have been kept up from the beginning, there would have been no Grimm's Law, no etymology; in short, no comparative or historical philology possible.
The advantages of beginning a foreign language in a phonetic notation are many and obvious. In the first place, the learner who has once mastered the notation and learnt to pronounce the sounds the letters stand for, is able to read off at once any text that is presented to him without doubt or hesitation, and without having to burden his memory with rules of pronunciation and spelling. Another advantage of phonetic spelling is that when the learner sees the words written in a representation of their actual spoken form he is able to recognize them at once when he hears them. And if the learner begins with the phonetic notation, and uses it exclusively till he has thoroughly mastered the spoken language, he will then be able to learn the ordinary spelling without fear of confusion, and quicker than he would otherwise have done.
Spelling-reform may be carried out with various degrees of thoroughness. After the failure of many schemes of radical reform, an attempt was made to begin with those numerous spellings which are both unphonetic and unhistorical, or are against the analogy of other traditional spellings. Accordingly, in 1881 the Philological Society of London “aproovd (sic) of certain partial corections (sic) of English spellings,” which were also approved of by the American Spelling-reform Association; and a list of them was issued jointly by the two bodies, and recommended for general adoption. A similar movement has been started in France. But the general feeling appears to be that it is better to keep the ordinary spelling unchanged, and wait till it is possible to supersede it by one on a more or less independent basis.
If the existing Roman alphabet is made the basis of the new phonetic notation of any one language, the most obvious course is to select one of the various traditional representations of each sound, and use that one symbol exclusively, omitting, of course, at the same time all silent letters. A. J. Ellis's English Glossic is an example of such a phonetic spelling on a national basis. The following is a specimen:—
Ingglish Glosik iz veri eezi too reed. Widh proper training a cheild foar yeerz oald kan bee redili taut too reed, Glosik buoks.
But a system which, like this, writes short and long vowels with totally different symbols (i, ee) is only half phonetic: it is phonetic on an unphonetic basis.
A fully phonetic system, in which, for instance, long vowels and diphthongs are expressed by consistent modifications or combinations of the symbols of the short vowels, and in which simple sounds are, as far as is reasonable and convenient, expressed by single letters instead of digraphs such as sh, must necessarily discard any national basis., The best basis on the whole is obtained by giving the letters their original common European sounds, i.e. by returning to the Late Latin pronunciation, with such modifications and additions as may be advisable. As regards the vowels at least, this Latin basis is very well preserved in German and Italian. In French, on the other hand, the Latin tradition was greatly corrupted already in the earliest period through the rapid changes which the language underwent. Thus when the Latin u in luna assumed the sound it now has in French lune, the symbol u was still kept; and when the sound u afterwards developed again out of the diphthong ou, this digraph was used to denote the sound. So when the French system of spelling came into use in England after the Norman Conquest these unphonetic symbols were introduced into English spelling, so that such a word as Old English and Early Middle English hūs, “house,” was written hous in the Late Middle English of Chaucer, although the sound was still that of Scotch hoos, ou (ow) being also used to denote a true diphthong (ou) in such words as knou, know, from Old English cnāwan.
By returning, then, to the original values of the letters we get the “Romic” or international (Continental) basis as opposed to the Glossic or national basis. Thus the passage quoted above appears as follows in Sweet's “Broad Romic” notation:—
iŋliʃ glosik iz veri iizi tu riid. wið prope treiniŋ a tʃaild fɔə jiəz ould kən bii redili tɔt to riid glosik buks.
Another important general distinction is that between “broad” and “narrow” systems of notation. A broad notation is one which makes only the practically necessary distinctions in each language, and makes them in the simplest manner possible, omitting all that is superfluous. From a practical point of view the necessary distinctions are those on which differences of meaning depend. A distinction of sound which is significant in one language may be insignificant in another. Thus the distinction between close é and open é, ê is significant in French, as in pécher, pêcher; so if in French phonetic writing the former is denoted by (e), it is necessary to find a new symbol (ϵ) for the open sound. But in languages such as English and German, where the short e is always open, there is no practical objection to using the unmodified (e) to denote the open sound, even if we regard (e) as the proper symbol of the close sound. And in those languages in which the short e is always open and the long e always close it' is enough to mark the distinction of quantity, and leave the distinction of quality to be inferred from it (e, ee). In such a case as this it is, of course, possible to apply the principle of ignoring superfluous distinctions in the opposite way: by writing the long and short vowels in such a language (e, ϵ), leaving the quantity to be inferred from the quality. But the former method is the more convenient, as it does not require any new letter. The “broad” principle is especially convenient in writing diphthongs. Thus in English Broad Romic we write the diphthongs in high and how with the same vowel as ask (hai, hau, ask), although all these (a)'s represent different sounds in ordinary southern English pronunciation. But the pronunciation of these diphthongs varies so much in different parts of the English-speaking territory, and the distinctions are so minute that it would be inconvenient to express them in writing; and as these distinctions are non-significant, it would be useless to do so. (ai) and (au) are symbols, not of special diphthongs, but of two classes of diphthongs: they can stand for any diphthongs which begin with a vowel resembling the Italian a, and end with approximations to i and u respectively. Theoretically it would be just as correct in English and German to write these diphthongs (ae, ao). But these notations are misleading, because they suggest simple sounds.
In comparing the sounds of a variety of languages, or of dialects of a language, and still more in dealing with sounds in general, we require a “narrow,” that is a minutely accurate, notation covering the whole field of possible sounds. It is evident from what has been said above that such a universal scientific alphabet is not suited for practical work in any one language. But the symbols of such a notation as Sweet's “Narrow Romic” are of the greatest use as keys to the exact pronunciation of the vaguer symbols of the Broad Romic notations of each language.
To prevent confusion between these two systems of notations Broad Romic symbols are enclosed in ( ), Narrow Romic in [ ], which at the same time serve to distinguish between phonetic and nomic spellings. This in English i (i) = [i] means that the English vowel in finny is the “wide” sound, not the “narrow” one in French fini, although in the Broad Romic notations of both languages (fini) is written for finny and fini alike.
Narrow Romic was originally based on A. J. Ellis's “Palaeotype,” in which, as the name implies, no new letters are employed. The symbols of Palaeotype are made up, as far as possible, of the letters generally accessible in printing-offices, the ordinary Roman lowercase letters being supplemented by italics and small capitals (i, i, i) and turned letters (ə,ɔ), many digraphs (th, sh) being also used. This notation was a reaction from Ellis's earlier phonotopy, in which a large number of new letters were used. Some of these, however, such as ʃ=(sh), ʒ=(zh), were afterwards adopted into Broad and Narrow Romic. In his Palaeotype Ellis also discarded diacritical letters, which, as he rightly says, are from a typographical point of view equivalent to new letters. In Narrow Romic a certain number of diacritical letters are used, such as (ñ, à), most of which are already accessible. Palaeotype is a Roman-value notation, the main difference as regards the values of the symbols between it and the later systems, being that it is more complex and arbitrary. Ellis afterwards had the unhappy idea of constructing a “Universal Glossic” on an English-values basis, which is even more cumbrous and difficult to remember than Palaeotype.
Sweet's Romic systems were made the basis of the “International” alphabet used in Le Maître Phonétique, which is the organ of the International phonetic Association, directed by P. Passy. Although this system is at the present time more widely known and used than any other, and although it is constructed on the international Romic principle, it is not really an international system. It is rather an attempt to make a special adaptation of the Romic basis to the needs of the French language into a general notation for all languages. But the phonetic structure of French is so abnormal, so different from that of other languages, that the attempt to force a Broad Romic French notation on such a language as English is even more hopeless than it would be to reverse the process. Although well suited for French, this alphabet must from a wider point of view be regarded as a failure: it is too minute and rigid for practical, and yet not precise enough for scientific purposes. In short, although it has done excellent service, and has helped to clear the way for a notation which shall command general acceptance, it cannot be regarded as a final solution of the problem.
Of the numerous other notations now in use, some still adhere to the diacritic principle of Lepsius's Standard Alphabet (1855), intended for missionary use, but found quite unfit for that purpose because of the enormous number of new types required. Most of them prefer to use new letters formed by more or less consistent modifications of the existing italic letters. A J. Lundell's Swedish dialect alphabet and O. Jespersen's Danish dialect alphabet are good specimens of this tendency. In the latter Roman letters are used for special distinctions, just as italic letters are used in the Romic systems.
But in spite of all diversity, there is much agreement. As regards the vowels, the following approximate values are now pretty generally accepted:—
Vowel-length is in some systems denoted by doubling (aa), in others by special marks (a: &c.), the diacritic in ā being used only in the nomic orthographies of dead and oriental languages.
The only consonant-symbols that require special notice are the following:—
All the systems of phonetic notation hitherto considered are based on the Roman alphabet. But although the Roman alphabet has many advantages from a practical point of view, it is evidently impossible to build up a consistent and systematic notation on such an inadequate foundation of arbitrary signs. What is wanted, for scientific purposes especially, is a notation independent of the Roman alphabet, built up systematically—an alphabet in which there is a definite relation between sound and symbol.
This relation may be regarded either from the organic or the acoustic point of view. The tendency of the earlier attempts at an a priori universal alphabet was to symbolize the consonants organically, the vowels acoustically, as in E. Brucke's Phonetische Transscription (1863). It is now generally acknowledged that the vowels as well as the consonants must be represented on a strictly organic basis. This was first done in A. M. Bell's Visible Speech (1867), which appeared again (1882) in a shorter form and with some modifications under the title of Sounds and their Relations. Bell's pupil, H. Sweet, gave a detailed criticism of Visible Speech in a paper on Sound-notation (Trans. of Philological Society, 1880–1881), in which he described a revised form of it called the Organic Alphabet, which he afterwards employed in his Primer of Phonetics and other works. Sweet's Narrow Romic notation already mentioned is practically a transcription of the Organic Alphabet into Roman letters.
Such notations are alphabetic: they go on the general principle of providing separate symbols for each simple sound. But as the number of possible shades of sounds is almost infinite, even the most minutely accurate of them can do so only within certain limits. The Organic Alphabet especially makes a large use of “modifiers”—characters which are added to the other symbols to indicate nasal, palatal, &c., modifications of the sounds represented by the latter, these modifiers being generally represented by italic letters in the Narrow Romic transcription; thus (ln) = nasalized (l).
In the Roman alphabet such symbols as f, v are arbitrary, showing no connection in form either with one another or with the organic actions by which they are formed; but in the Organic symbol of v, for instance, we can see the graphic representation of its components “lips, teeth, voice-murmur.” By omitting superfluous marks and utilizing various typographical devices the notation is so simplified that the symbols, in spite of their minute accuracy, are often simpler than in the corresponding Roman notation. The simplicity of the system is shown by the fact that it requires only about 110 types, as compared with the 280 of Lepsius's very imperfect Standard Alphabet.
All the systems hitherto considered are also alphabetic in a wider sense: they are intended for continuous writing, the more cumbrous “narrow” notations being, however, generally employed only in writing single words or short groups. An “analphabetic" basis was first definitely advocated by Jespersen, who represents each sound by a group of symbols resembling a chemical formula, each symbol representing not a sound, but an element of a sound: the part of the palate, tongue, &c., where the sound is formed, the degree of separation (openness) of the organs of speech, and so on. The two great advantages of such a system are that it allows perfect freedom in selecting and combining the elements and that it can be built up on the foundation of a small number of generally accessible signs.
As regards Jespersen's scheme, it is to be regretted that he has not worked it out in a more practical manner: that in his choice of the thirty odd symbols that he requires he should have gone out of his way to mix up Greek with Roman letters, together with other characters which would be avoided by any one constructing even a scientific alphabetic notation. And his use of these symbols is open to much criticism. In fact, it cannot be said that the analphabetic principle has yet had a fair trial.
The Organs of Speech.—Most speech-sounds are formed with air expelled from the lungs (voice-bellows), which passes through the two contractible bronchi or bronchial tubes into the also contractible wind-pipe or trachea, on the top of which is fixed the larynx (voice-box). Across the interior of the larynx are stretched two elastic ledges or cushions called “the vocal chords.” They are inserted in front of the larynx at one end, and at the other they are fixed to two movable cartilaginous bodies “the aretynoids,” so that the passage between them—the glottis—can be narrowed or closed at pleasure. The glottis is, as we see, twofold, consisting of the chord glottis and the cartilage glottis. The two can be narrowed or closed independently. The chords can also be tightened or relaxed, lengthened and shortened in various degrees.
When the whole glottis is wide open, no sound is produced by the outgoing breath except that caused by the friction of the air. Sounds in whose formation the glottis is in this passive state are called “breath” sounds. Thus (f) is the breath consonant corresponding to the “voice” or “voiced” consonant (v). In the production of voice, the chords are brought close enough together to be set in vibration by the air passing between them. In the “thick” register of the voice (chest voice) the chords vibrate in their whole length, in the “thin” register or falsetto only in part of their length. If the glottis is narrowed without vibration, “whisper” is the result. In the “weak whisper” there is narrowing the whole glottis; in the “strong whisper,” which is the ordinary form, the chord glottis is entirely closed, so that the breath passes only through the cartilage glottis. In what is popularly called “whisper”—that is, speaking without voice—the breath sounds remain unchanged, while voiced sounds substitute whisper (in the phonetic sense) for voice. Thus in whispering such a word as feel the (f) remains unchanged, while the following vowel and consonant are formed with the glottis only half closed. Whispered sounds—both vowels and consonants—occur in ordinary loud speech in many languages. Thus the final consonants in such English words as leaves, oblige are whispered, except when followed without a pause by a voiced sound, as in obliging, where the (ʒ) is fully voiced.
Above the glottis—still within the larynx—comes the “upper” or “false” glottis, by which the passage can be narrowed. On the top of the larynx is fixed a leaf-like body, the “epiglottis," which in swallowing, and sometimes in speech, is pressed down over the opening of the larynx. The contractible cavity between the larynx and the mouth is called the “pharynx.” The roof of the mouth consists of two parts, the “soft” and the “hard palate.” The lower pendulous extremity of the soft palate, the “uvula,” in its passive state leaves the passage into the nose open. In the formation of non-nasal sounds, such as (b), the uvula is pressed up so as to close the passage from the pharynx into the nose. If (b) is formed with the passage open, it becomes the corresponding nasal consonant (m). The other extremity of the (hard) palate is bounded by the teeth, behind which are the gums, extending from the teeth-rim to the arch-rim—the projection of the teeth-roots or alveolars.
There is great diversity among phoneticians as regards the mapping out—the divisions—of the palate and tongue, and their names. Foreign phoneticians generally adopt very minute distinctions, to which they give Latin names. Bell in his Visible Speech makes a few broad fundamental divisions. In the arrangement adopted here (mainly based on his) sounds formed on the soft palate are called “back,” and are subdivided into “inner” = nearer the throat, and “outer” = nearer the teeth, further subdivisions being made by the terms “innermost,” “outermost,” the position exactly half way between these two last being defined as “intermediate back.” Sounds formed on the hard palate or teeth may be included under the common term “forward,” more accurately distinguished as “teeth” (dental), “gum,” “front” (palatal, afterwards called “top” by Bell), which last is really equivalent to “mid-palatal,” including the whole of the hard palate behind the gums. All of these divisions are further subdivided into “inner,” &c., as with the back positions.
Of the tongue we distinguish the “back” (root), “front” or middle, “point” (tip), and “blade,” which includes the point and the surface of the tongue immediately behind it. The tongue can also articulate against the lips, which, again, can articulate against the teeth. The lip passage can be closed, or narrowed in various degrees. Sounds modified by lip-narrowing are called “lip-modified” (labialized) or “round” (rounded), the last being specially used in speaking of vowels.
Speech-sounds.—The most general test of a simple as opposed to a compound sound (sound-group) is that it can be lengthened without change. As regards place of articulation, no sound is really simple: every sound is the result of the shape of the whole configurative passage from the lungs to the lips; and the ultimate sound-elements, such as voice, are never heard isolated. The most indistinct voice-murmur is as much the result of the shape of the super glottal passages as the clearest and most distinct of the other vowels; and its organic formation is as definite as theirs is, the only difference being that while in what we regard as unmodified voice all the organs except the vocal chords are in their passive, neutral positions, the other vowels are formed by actively modifying the shape of the super-glottal passages—by raising the tongue towards the palate, narrowing the lips, &c.
The most important elements of speech-sounds are those which are dependent on the shape of the glottis and of the mouth passage respectively. It is on the relation between these two factors that one of the oldest distinctions between sounds is based: that of vowel and consonant. In vowels the element of voice is the predominant one: a vowel is voice modified by the different shapes of the super glottal passages. In consonants, on the other hand, the state of the glottis is only secondary. Consonants are generally the result of audible friction, as in (f), or of complete stoppage, as in (p). If the glottis is at the same time left open, as in (f, p), the consonant is “breath” or “voiceless”—if it is narrowed enough to make the chords vibrate, as in (v, b), the consonant is “voice” or “voiced”; intermediate positions producing the corresponding “whispered” consonants. Vowels are characterized negatively by the absence of audible friction or stoppage: if an (i) is formed with the tongue so close to the palate as to cause buzzing, it becomes a variety of the front consonant (j). There is, of course, no difficulty in forming a vowel with the glottis in the position for breath and whisper. Thus breath (i) may often be heard in French in such words as ainsi at the end of a sentence, the result being practically a weak form of the front-breath consonant (ç). The division between vowel and consonant is not an absolutely definite one. As we see, the closer a vowel is—that is, the narrower its configurative passage is—the more like it is to a consonant, and the more natural it is to devocalize it. Some voice consonants, on the other hand, have so little buzz that acoustically they constitute a class between consonants and vowels—a class of “vowel-like” or “liquid” consonants, such as n, m, l).
The changes in sounds which result from active narrowing of the passages admit of an important distinction as “sound modifying" and “sound-colouring,” although the distinction is not always definite. Nasality and rounding are examples of sound-modifying processes. Thus we hear a certain resemblance between (b) and (m), (i) and (y), but we regard all these four as distinct and practically independent sounds. Contraction of the pharynx, on the other hand, as also of the false glottis and windpipe, have only a sound-colouring effect: if a vowel is formed with such contractions its quality (timbre) is altered, but it still remains the same vowel. It follows from the definition of speech-sounds that they admit of a twofold classification: (1) organic and (2) acoustic. As already remarked, the older phoneticians used to classify the consonants organically, the vowels mainly from the acoustic point of view. The first to give an adequate organic classification of the vowels was the author of Visible Speech. Bell gave at the same time an independent acoustic classification of the consonants as well as the vowels. His acoustic classification consists simply in arranging the sounds in the order of their “pitches” (tone-heights). The pitches of the breath consonants are absolutely fixed in each individual pronunciation, while those of spoken vowels can be varied indefinitely within the compass of each voice by tightening the vocal chords in various ways and shortening their vibrating portions· the tighter and shorter the vibrating body, the quicker its vibrations, and the higher the tone. But when a vowel is whispered or breathed nothing is heard but the resonance of the con figurative passages, especially in the mouth, and the pitches of these resonant cavities are as fixed as those of the breath consonants; in other words, a whispered (or breathed) vowel cannot be sung. Although the absolute pitches of voiceless sounds may vary from individual to individual the relations of the pitches are constant: thus in all pronunciations (ç) and whispered (i) are the highest, breath (w) in what and whispered (u) nearly the lowest in pitch among consonants and vowels respectively.
If phonetics were an ideally perfect science there would be no occasion to discuss whether the acoustic or the organic study of the vowels and the other speech-sounds is the more important: a full description of each sound would necessarily imply (1) an exact determination of its organic formation, (2) an acoustic analysis of the sound itself, both from the objective physical point of view and from the subjective one of the impression received by the ear, and (3) an explanation of how (2) is the necessary result of (1). Even this last question has already been solved to some extent. In fact, the connection between the organic formation and the acoustic effect is often self-evident. It is evident, for instance, that (i) and (ç) owe their clear sound and high pitch to their being formed by short, narrow passages in the front of the mouth, while (u) owes its low pitch to being formed in exactly the opposite way, the sound being farther muffled and the pitch consequently still more lowered by the rounding.
One reason why it is impossible to classify the vowels exclusively on acoustic principles is that two vowels formed in quite different ways may have the same pitch. Thus the “high front-round” (y) and the "high-mixed” (i) have the same pitch, the tongue-retraction of the mixed position of the latter having the same effect as the rounding of the former. It is evident, therefore, that the fundamental classification of the vowels must, like that of the consonants, be purely organic. And although for practical purposes it is often convenient to classify sounds partly from the acoustic point of view, a full scientific treatment must keep the two points of view strictly apart, and make a special chapter of the relations between them.
Vowels.—The most obvious distinction between vowels is that which depends on the share of the lips in their articulation. In such non-round vowels as (i) and (a) the lips are passive, or even separated and spread out at their corners, by which the vowels assume a clearer resonance. If, on the other hand, the lips are actively approximated, they become the round vowels (y) and “open” (o) respectively.
owels are formed with different degrees of rounding. As a general rule, the narrowness of the lip-passage corresponds to the narrowness of the mouth-passage. Thus, in passing from the vowel of too to those of no and saw the back of the tongue is progressively lowered, and the rounding is diminished in the same proportion.
But there is also abnormal rounding. Thus, if we pronounce (o) with the lips in the position they have in forming (u), the resulting “over-rounded” vowel sounds half-way between (o) and (u); the second element of the diphthong (ou) in go is formed in this way. Conversely, the (u) in put is “under-rounded” in the North of England: the tongue position is kept, but the lips are only brought together a little at the corners, as in (ɔ).
The mouth positions of the vowels are the result of two factors: (1) the height of the tongue—its nearness to the palate—and (2) the degree of its retraction. Bell distinguishes three degrees of height: in his system (u) is “high,” the (o) of boy is “mid,” and the (ɔ) of saw is “low.” He also has three degrees of retraction: in “back” vowels, such as (u), the root of the tongue is drawn to the back of the mouth, and the whole tongue slopes down from back to front. In “front” vowels, such as (i), the front of the tongue is raised towards the hard palate, so that the tongue slopes down from front to back.
Most of these slope-positions yield vowels of a distinct and clear resonance. There is also a class of “flat” vowels. such as (a), in which the tongue is in a more or less neutral position. If the tongue is raised from the low-flat position of (əə) in bird to the high position, we get the (i) of North Welsh dyn “man,” which, as already observed, is acoustically similar to (y).
The flat vowels were called “mixed” by Bell, in accordance with his view that they are the result of combining back and front articulation. And although this view is now generally abandoned, the term “mixed” is still retained by the English school of phoneticians.
In this way Bell mapped out the whole mouth by the following cardinal points:—
In this arrangement “high-back,” &c., are fixed points like those of latitude and longitude. Thus normal “high” means that the tongue is raised as close to the palate as is possible without causing consonantal friction, and “back” implies retraction of the same kind. Intermediate positions are defined as “raised,” “lowered,” “inner,” “outer.”
The most original and at the same time the most disputed part of Bell's vowel-scheme is his distinction of “primary” and “wide.” All vowels fall under one of these categories. Thus, the primary French (i) and the corresponding English wide (i) are both high front-vowels, and yet they are distinct in sound: the English vowel is a semitone lower in pitch. Bell explained the greater openness of the wide vowels as the result of greater expansion of the pharynx; and he considered the other class to be most nearly allied to the consonants—whence their name “primary”—the voice-passages in the formation of primary vowels being expanded only so far as to remove all fricative quality. But alterations in the shape of the pharynx have only a sound-colouring, not a sound-modifying, effect; and Sweet showed that the distinction depends on the shape of the tongue, and accordingly substituted “narrow” for Bell's “primary.” He also showed that the distinction applies to consonants as well as vowels: thus the narrow French (w) in oui is a consonantization of the narrow French (u) in sou, while the English (w) preserves the wide quality of the (u) in put.
In forming narrow sounds there is a feeling of tension in that part of the tongue where the sound is formed, the tongue being clenched or bunched up lengthwise, so as to be more convex than in its relaxed or “wide” condition.
The distinction between narrow and wide can often be ignored in practical phonetic writing, for it generally depends on quantity; length and narrowness, shortness and wideness going together. When the distinction is marked, wide vowels may be expressed by italics, as in German (biinə, bin).
Bell's category of “mixed-round” vowels had from the beginning been a source of difficulty to students of Visible Speech. But it was not till 1901 that Sweet showed that they are only mixed as regards position: they are really the corresponding back-round vowels moved forward into the middle of the mouth while preserving the slope of back vowels, instead of having the tongue flat as in the (unround) mixed vowels. They are “out-back” vowels: there is an exaggeration of the outer back position of such a back-round vowel as the English (u) compared with the full back (u) in German muttre.
In the same way by moving the tongue backwards while forming a front vowel another series of “in-front” vowels is obtained.
The “in-mixed” vowels are obtained by shifting the neutral mixed positions into the full back position, keeping the tongue flat, so that these vowels might also be called “back-flat.”
The out-back, in-front and in-mixed vowels are included under the common designation of “shifted,” as opposed to “normal” vowels.
There is a large number of other vowel-schemes, of which a survey will be found in W. Vietor's Elemente der Phonetik. Many of the older ones are in the form of triangles, with the three chief vowels a, i, u at the three corners, the other vowels being inserted between these extremes according to their acoustic relations. Since the appearance of Visible Speech many attempts have been made to fit his new vowels into these older schemes.
Of all the vowel-schemes the one now most generally known is perhaps that of the International Phonetic Association already mentioned. In this scheme the distinction of narrow and wide, though admitted and occasionally marked, is not an integral part of the system, the vowels being classified first as “velar” (back) and “palatal” (front), and then according to openness as “close,” “half-close,” “medium,” “half-open” and “open.”
Consonants.—These are the result of audible fiiction or stoppage, which may be accompanied either with breath, voice or whisper. Consonants admit of a two-fold division (1) by form, and (2) by place. Thus (p, b) are by place lip-consonants, while by form they are stopped consonants or “stops.”
If the mouth-stoppage is kept, and the nose-passage is opened, the stop becomes the corresponding “nasal”; thus (b) with the soft palate lowered becomes the nasal (m).
In “open” consonants the sound is formed by simply narrowing the passage, as in the back-open-breath (x) in Scotch and German loch. In some open consonants, such as the lip-teeth (f), there is slight contact of the organs, but without impeding the flow of breath.
In “divided” consonants there is central stoppage with openings at the sides, as in the familiar point-divided (l). These consonants are sometimes “unilateral”—with the opening on the side only—the character of the sound not being sensibly modified thereby.
When open and divided consonants are formed with the nose passage open they are said to be “nasalized.” Thus (m) with complete lip-closure becomes the nasalized lip-open-voice consonant.
“Trills” (or rolled) consonants are a special variety of un-stopped consonants resulting from the vibration of flexible parts against one another, as when the lips are trilled, or against some firm surface, as when the point of the tongue trills against the gums in the Scotch (r), or the uvula against time back of the tongue, as in the Northumbrian burred (r), and the French and German (r), where—especially in German—the trill is often reduced to a minimum or suppressed altogether.
As regards the place of consonants, there is, as already remarked, great diversity among phoneticians, both in mapping out the palate and tongue and in the names given to these divisions. The classification and nomenclature given here is, in the main, that of Bell.
By place then, we distinguish seven main classes of consonants: back, front, point, blade, fan, lip, and lip-teeth.
“Back” (guttural) consonants are formed between the root of the tongue and the soft palate. In most languages the positions of these consonants vary according to those of the accompany in vowels. thus the back-stop and back-nasal in king are more forward than in conquer.
“Front” (palatal) consonants are formed between the middle of the tongue and the hard palate, the point of the tongue lying passively behind the lower teeth. It is easy to make the front open-voice (j) in you into the corresponding stop (ɟ) by narrowing the passage till there is complete closure, as in Hungarian nagy (nɔɟ) “world.” In the same way the open breath (ç) in German ich may be made into the stop (c) = Hungarian ty. (ɟ) nasalized becomes (ñ)—Italian gn, Spanish ñ, French gn in vigne. The front-divided-voice consonant is the Italian gl and Spanish ll. These are all simple sounds, distinct from the (lj), (nj) in French and English million and English onion.
“Point” consonants when formed against the teeth are called “point-teeth” (dental). English (þ) in thin is the point-teeth-open-breath consonant, (ð) in then the corresponding voice consonant. If (ð) is modified by turning the tip of the tongue back into the inner position—about on the arch-rim—it becomes the untrilled (r) in English rearing, in which position the tongue is easily trilled, the trilling becoming more and more difficult the more the tongue is approximated to the point-teeth position. In French and many other languages all the point consonants (t, d, n, l), &c., are formed on the teeth, except (r), which is always more retracted than the other point consonants. If the tip of the tongue is turned so far back as to articulate with its lower edge against the arch of the palate—that is, farther back than for the “inner” position—it is said to be “inverted.” Inverted (r) is frequent in the dialects of the south-west of England. The opposite of inversion is “protrusion,” in which the tip of the tongue articulates against the upper lip.
“Blade” consonants are formed by the blade or flattened tip of the tongue against the gums, as in English (s, z), or against the teeth, as in the corresponding French sounds. If these consonants are modified by turning the tongue a little back, so as to bring the point more into play, they become the “blade-point” consonants (ʃ, ʒ), as in fish, measure. (ʃ) is acoustically a dull (s). In some languages, such as German, sounds similar to (ʃ) and (z) are formed partly by rounding, which lowers the pitch of the hiss in the same way as retraction does, so that the tongue-articulation is only imperfectly carried out. When the rounding is very marked there is only a slight raising of the front of the tongue, as in some Swedish dialects; and if the tongue-articulation is progressively shifted back, and the rounding diminished in the same proportion, (ʃ) can at last develop into the pure back-open consonant (x), as in the present pronunciation of Spanish x and j.
The English point consonants (t, d, n, l) are formed on the gums just behind the teeth, the point of the tongue being fattened, so that they are almost blade consonants.
“Fan” (spread) consonants—the “emphatic” consonants of Arabic—are modifications of point and blad)e consonants, in which the sides of the tongue are spread out, so that the hiss of such a consonant as (s) is formed partly between the sides of the tongue and the back teeth, which gives a peculiar deep, dull quality to these sounds.
“Lip” consonants, such as (p, m), and “lip-teeth” consonants, such as (f, v), offer no difficulty. The simple li-open-breath consonant does not occur in English; it is the sound produced in blowing out a candle. The corresponding voice sound is frequent in German—especially in Middle Germany—in such words as quelle.
If the lip-open consonants are modified by raising the back of the tongue, they become the “lip-back” consonants (wh, w) in English what, we, which may also be regarded as consonantized (u). In them the lip articulation predominates. In the “back-lip” consonants, as in German auch, the reverse is the case.
This last is one of a large number of “lip-modified” consonants, of which the already-mentioned German sch is a further example.
In a similar way consonants may be “front-modified.” (i) is peculiarly susceptible to such modifications. In French and other languages it is formed with the tongue more convex than in English, and consequently with a tendency to front-modification. Front-modified (s) and point (r) may be heard in Russian in such words as gusǐ “goose,” tsarǐ “emperor,” where the final vowels are silent.
Some consonants are formed below the mouth.
When the glottis is sharply opened or closed on a passage of breath or voice an effect is produced similar to that of a stop in the mouth, such as (k). This “glottal stop” is the sound produced in hiccuping; and is an independent sound in some languages, such as Arabic, where it is called “hamza.” In German all words beginning with a stressed (accented) vowel have a more or less distinct glottal stop before the vowel.
Of the passages below the glottis, the bronchial and the windpipe are both susceptible of contraction.
Spasmodic contraction of the bronchial passages is the main factor in producing what is known as “the asthmatic wheeze.” If this contraction is regulated and made voluntary it results in the deep hiss of the Arabic ḥā. If this sound is voiced, it causes a peculiar intermittent vibration of voice, which is habitual with some speakers, especially in Germany. If this effect is softened by slightly expanding the bronchial passages, an (r)-like sound is produced, which is that of the Arabic ’ain.
Contraction of the windpipe produces a sound similar to the Arabic ḥā, but weaker, which when followed by a vowel has the effect of a strong aspirate. When voiced it becomes a mere colourer of the accompanying voice-murmur, or vowel, to which it imparts a deep timbre.
Non-expiratory Sounds.—All the sounds hitherto described imply out-breathing or expiration. Many of them can also be formed with in-breathing or inspiration. In English it is a not uncommon trick of speech to pronounce no in this manner, to express emphatic denial.Some consonants are formed without either in- or out-breathing, but solely with the air in the throat or mouth. In forming “suction stops” or “clicks” the tongue or lips are put in the position for a stop, and the air is sucked out from between the organs in contact, so that when the stop is loosened, a smacking sound is produced by the air rushing in to fill the vacuum. Thus the point-click is the interjection of impatience commonly written tut! In many savage languages clicks are a part of ordinary speech.
Synthesis.—Besides analysing each sound separately, phonetics has to deal with the phenomena which accompany synthesis or the combination of sounds. Although a sentence may consist of a single word, and that word of a single vowel, sounds mostly occur only in combination with one another. The ordinary division into sentences and words is logical, not phonetic: we cannot mark off sentences and cut them up into words until we know what they mean and are able to analyse them grammatically. But the logical division into sentences corresponds to some extent with the phonetic division into “breath-groups,” marked off by our inability to utter more than a certain number of syllables in succession without pausing to take breath. Within each of these breath-groups there is no necessary pause between the words, except when we pause for emphasis. The only necessary phonetic divisions within the breath-group are those into syllables, sounds and intervening “glides.” But before considering these last it will be necessary to say something about the general factors of synthesis: quantity, stress and intonation.
Stress is, organically the result of the force with which the breath is expelled from the lungs; while acoustically it produces the effect of loudness, which is dependent on the size of the sound vibrations: the bigger the waves, the louder the sound, and the greater the stress, of which we may distinguish infinite degrees. If we distinguish only three, they are called weak, medium and strong. The use of stress in different languages shows the same variety as quantity. Some languages, such as French, make comparatively little use of its distinctions, uttering all the syllables of words and sentences with a more or less even degree of force. English, on the other hand, makes great use of minute distinctions of stress both to distinguish the meanings of words and to mark their relations in sentences.
With stress is closely connected the question of syllable-division. A syllable is a group of sounds containing a “syllabic” or syllable-former, which is, of course, able to constitute a syllable by itself. The distinction between syllabic and non-syllabic depends on sonority, the more sonorous sounds being the voiced ones, while of these again, the most open are the most sonorous, the most sonorous of all being the vowels, among which, again, the openest are the most sonorous. But these differences are only relative. When a vowel and a consonant come together the sonorousness of the vowel always overpowers that of the consonant, so that the two together only constitute one syllable. But in such a word as little the second (l) is so much more sonorous than the accompanying voiceless stop that it assumes syllabic function, and the whole group becomes dissyllabic to the ear. The beginning of a syllable corresponds with the beginning of the stress-impulse with which it is uttered. Thus in atone the strong stress and the second syllable be in on the (t), and in bookcase on the second (k), the first (k) belonging to the first syllable, so that the (kk) is here double, not merely long, as in book (bukk) by itself.
Intonation or variation of tone (pitch) depends on the rapidity of the sound-vibrations the more rapid the vibrations, the higher the pitch. Intonation is heard only in voiced sounds, as being the only ones capable of variations of pitch.
In singing the voice generally dwells on each note without change of pitch, and then leaps up or down to the next note as quickly as possible, so that the intervening “glide” is not noticed—except in what is called portamento. In speaking, on the other hand, the voice hardly ever dwells on any one note, but is constantly gliding upwards or downwards, so that an absolutely level tone hardly ever occurs in speech. But in the rising and falling inflections of speech we can distinguish between “voice-glides” (portamentos or slurs) and “voice-leaps,” although the distinction is not so definite as in singing.
Of the three primary forms of intonation the level tone ( ) can be approximately heard in well as an expression of musing—although it really ends with a slight rise, the rising (´) in the question well?; the falling (`) in the answer yes. There are besides compound tones formed by uniting the two last in one syllable. The compound rising tone (ˇ) may be heard in take care! the compound falling tone (ˆ) in the sarcastic oh! All these tones may be varied according to the intervals through which they pass. The greater the interval, the more emphatic the tone. Thus a high rise, which begins high, and consequently can only rise a little higher, expresses simple question, while the same word, if uttered with a low rise extending over an interval of between a fifth and an octave—or even more—expresses various degrees of surprise or indignation, as in the emphatic what! compared with the simply interrogative what?In English and most European languages, intonation serves to modify the general meaning and character of sentences. This is sentence-intonation. But some languages, such as Swedish and Norwegian, and Chinese, have word-intonation, by which words which would otherwise be identical in sound are distinguished. The distinction between Gr. oíkoi and oîkoi was no doubt one of intonation.
Glides.—Such a word as cat consists not only of the vowel and the two consonants of which it is made up, but also of “glides” or transitions between these sounds. The glide from the initial consonant to the vowel consists of all the intermediate positions through which the tongue passes on its way from the (k)-position to the (ae)-position. The number of these positions is infinite, but they are all implied by the mere juxtaposition of the symbols, for it is assumed that in all transitions from one position to another the shortest way is taken. Although the direction of a glide is dependent on the positions of the two fixed points between which it lies, its character may be varied both by the shape of the configurative passages—especially the glottis—and by stress and quantity.
In the word given above the “off-glides” from the consonants are both breath-glides, the glottis being kept open during the transition from the voiceless consonant to the following vowel, or, as in the case of the final consonant, to silence. The “on-glide” from the vowel to the (t) is, on the other hand, a voice glide, the closure of the glottis being maintained till the stop is made.
In French and most of the languages of the south of Europe voiceless consonants are followed by voice-glides. Thus in French qui there is no escape of breath after the (k), as there is in English Key. Other languages again have breath on-glides before voiceless stops.
If an independent strong stress is put on the breath-glide of English key, it is heard almost as a full independent consonant, and becomes an “aspirate.” Aspirated steps may be heard in the Irish-English pronunciation of such words as tell, and also in Danish, and in Sanskrit as pronounced in India. If the voice-glide after a voice stop is emphasized in a similar way the “sonant aspirates” of Sanskrit and its modern descendants are produced, as in Sanskrit dhanu.
Glides are especially important from an acoustic point of view. Acoustically speaking, indeed, voiceless stops are pure glide sounds, the stop itself being inaudible. In voice-stops, on the other hand, the stop itself can be made audible as well as the intervening glides. In English these latter are fully voiced when they come between voice sounds, as in ago; but when preceded by voiceless sounds or by a pause, as in go! they are formed with imperfect vocality, full voice being heard only just before the stop is loosened. So also initial English (z) as in zeal is formed with imperfect vocality under the same conditions, so that it sounds like (sz). In French and other languages which have voice-glides after voiceless consonants initial (g, z) &c. are fully voiced.
Consonant-glides may be further modified in various ways. In the formation of “implosive” stops, such as occur in Saxon German, Armenian and other languages, voiceless stops followed by voice-glides are modified by simultaneous closure of the glottis, the larynx being raised by means of its muscles, so that it acts like a plug, compressing the air between the closed glottis and the mouth-stop, so that when the latter is released a peculiar choky effect is given to the off-glide.
Rounded glides may be heard in Russian in such words as komnata, where the rounding of the (o) is anticipated in the preceding consonant, being heard, of course, only in the off-glide of the consonant. The acoustic effect is between that of (kwo) and ordinary (ko).
Glideless consonant-combinations remain to be considered. The general articulative principle of taking the shortest way between sounds in juxtaposition necessarily results in certain transitions being effected without any glide at all. This is regularly the case when the consonants have the same place, and differ only in form, as in (nd, dlt), where the point of the tongue remains unmoved through the whole sound-group. In such combinations as (mf) the very slight glide is often got rid of entirely by assimilating the place of the first consonant to that of the second, so that the (m) becomes a lip teeth consonant, as in English nymph.
Even when consonants are formed in different parts of the mouth it is often possible to join them without any glide. In English such combinations as (kt, pt) are glide less, the point of the tongue being brought into position before the preceding stop is loosened. In French and most other languages such consonants are separated by a breath-glide.
Combinations of stops and vowel-like consonants (tr, gl, kw) are glideless in English and most other languages. In English the breath-glide after a voiceless stop unvoices the beginning of the following vowel-like consonant; thus try is almost (trhrai).
Vowel-glides.—Vowels are begun and ended in various ways. In the “gradual beginning,” which is the usual one in English and French, the glottis is gradually narrowed while breath is being emitted. In the “clear” beginning the breath is kept back till the glottis is closed for voice, which begins without any “breathiness.” German favours the clear beginning, generally exaggerating it into a glottal stop.
In the gradual as well as the clear beginning the stress begins on the vowel. If in the former it is thrown back on the breath glide, the latter is felt as an independent element and becomes the “aspirate” or (h), which in English and most other languages is a glide not only in the throat but in the mouth as well, the tongue and lips gradually moving up into the position for the following vowel while the glottis is being closed.
There is also a “strong” aspirate, which occurs in Finnish and other languages, in the formation of which the full vowel position is assumed from the beginning of the aspiration, which is therefore a voiceless vowel.
In most languages, when an aspirate comes between voiced sounds it is formed with imperfect vocality, the contrast of which with the full vocality of the other sounds is enough to produce the effect of breath Thus in Enlglish behold the voice runs on without any actual break, the glottal closure being simply relaxed, not fully opened for breath, as in the emphatic aha! In some languages, such as Bohemian, this “voice-aspirate” is used everywhere, initially as well as medially.
Vowels are finished analogously, either by a gradual opening of the glottis, or by a cessation of aspiration while the glottis is still closed for voice. If stress is put on the gradual ending it becomes a distinct aspirate, as in the Sanskrit “visarga” in such a word as manah.
Organic Basis.—Every language has certain general tendencies which control the formation of its sounds, constituting its “organic basis” or basis of articulation. The tendency of the present English is to flatten and lower the tongue and draw it back from the teeth, while the lips are kept as much as possible in a neutral position. The flattening of the tongue makes our vowels wide and favours the development of mixed vowels, and gives the dull quality which is especially noticeable in our (l); and its retraction is unfavourable to the development of teeth sounds; while the neutrality of the lips eliminates front-round vowels. In such a language as French everything is reversed. The tongue is arched, and raised, and advanced, and the lips articulate with energy. Hence French sounds tend to narrowness, mentality and distinct rounding.
National Sound-systems.—Each language uses only a part of the general phonetic material. Each one has only a limited number of sounds; and each one makes only a limited use of the synthetic distinctions of quantity, stress and intonation. As we have seen, many of these differences between individual languages are the result of, or may be referred to, differences in their organic basis.
Just as cognate languages differ from each other in phonetic structure, so also dialects of the same languages differ from each other more or less. Thus the sound-system of Lowland Scotch—which is, historically, a dialect of Northern English—differs considerably from that of standard English. Standard English itself was originally that mixture of the Midland and the Southern dialect which was spoken in London in the middle ages, just as standard French is, historically, the dialect of that district of which Paris is the centre. Standard English, like standard French, is now more a class-dialect than a local dialect: it is the language of the educated all over Great Britain. But it is not yet perfectly uniform. It is still liable to be influenced by the local dialects in grammar and vocabulary, and still more in pronunciation.
Again, English, like all other living languages, changes from generation to generation. Pronunciations which are vulgar in one century may become fashionable in the next. Sounds which are distinct in one generation may be confounded in another, and new distinctions may be made, new sounds may arise. A spoken language is, therefore, necessarily a vague and floating entity, and English is no exception to the rule. The very fixity of its written form gives all the freer play to the influences which cause change.
A standard spoken language is, strictly speaking, an abstraction. No two speakers of standard English pronounce exactly alike. And yet they all have something in common in every sound they utter. There are some divergences, some peculiarities of pronunciation, which pass unnoticed, while others, less considerable perhaps in themselves, are at once felt as archaisms, vulgarisms or provincial isms, as the case may be, by the majority of educated speakers.
Sounds of English.—The following is a convenient classification of the vowels of standard English:—
Here the vowels are in four rows: (1) normally short, or, more correctly, monophthongic, (2) long, or half-diphthongic, (3) full diphthongs, (4) murmur-diphthongs.
Those under (1) are often lengthened in monosyllables such as ten, good, but they always remain absolutely monophthongic. The only one in the next row that is always strictly monophthongic is (əə): all the others, as we shall see, tend to become more or less diphthongic, especially in the south of England, being often exaggerated into full diphthongs of the (ai) and (au)-type in vulgar speech.
(a), as in come up, is the short vowel corresponding to the (aa) in calm. (aa) is the mid-back-wide vowel, and (a) differs from it only in being narrow. Acoustically (a) is a muffled or obscure (aa): an the same effect may be produced by advancing the tongue from the mid-back to the corresponding out-back position, preserving the wide articulation: this ronunciation of u is common in the south of England. Historically, these sounds are the result of unrounding and older (u).
(ə), as in sofa, is a mixed vowel, tending to wideness and mid position, which occurs only unstressed. (əə) in turn, earth, is low-mixed-narrow. It is the result of absorption of an older (r), weakened into (ə).
(æ), as in man, is low-front-wide, from older mid-back-wide.
(i) in it is high-front-wide. The long (ii) in eat is narrow in the north of England, while in the south it is wide (') followed by (j).
(e) in men is generally mid-front-wide. (ei) in mane is the same vowel either narrow or wide, raised in its latter half towards (i).
(u) in good is high-back-wide-round. Narrow (uu) in too becomes (uw) in southern English.
(o) in not is low-back-wide-round. In (ou), as in no, the midback-round vowel, either narrow or wide, is over-rounded in its latter half. (ɔ), as in all, is low-back-narrow-round.
The full diphthongs (ai, au, oi), as in eye, now, oil, all end in lowered high vowels. Their first elements are only roughly indicated by the transcription, and vary in the mouths of different speakers. That of (ai) is generally the out-mid-back-wide, that of (au) the broader low-mixed-wide, that of (oi) the mid-back-wide-round.
The murmur-diphthongs (iə) as in here, (eə) as in air, (uə) as in poor, all tend to broaden their first elements. That of (eə) is the low-front-narrow vowel. The other two begin with lowered forms of the wide (i) and (u) respectively. In (uə) the lowering is often carried so far as to make poor almost, or completely, into pore (pɔə).
The following arrangement of the English consonants will show their organic relations to one another.
|j||r;||þ,ð||s, z;||ʃ, ʒ||wh, w;||f, v|
|k, g||t, d||p, b|
The “aspirate” (h) may be regarded either as a throat-consonant or as a breath-glide.
Characteristic features of the English consonant-system are the large number of hisses and buzzes, the sharp distinction of breath and voice, and, negatively, the absence of the open-back consonants, and of the voiceless forms of the vowel-like consonants (l, r) and the nasals, most of which still existed in Old English.Bibliography.—The most important general works are: H. Sweet, A Primer of Phonetics (3rd ed, Ox ord, 1906); E. Sievers, Grundzuge der Lautphysiologie (5th ed., Leipzig, 1901); W. Vietor, Elemente der Phonetik des Deutschen, Englischen und Franzosischen (5th ed, Leipzig, 1904); O. Jesperson, Lehrbuch der Phonetik (Leipzig, 1904); M. Trautmann, Die Sprachlaute (Leipzig, 1884–1886); Le Maître Phonétique, organe de l'association phonétique internationale (apply to Dr P. Passy, Bourg-la-Reine, France). For the laws of sound-change, see the above-mentioned work of Sievers; H. Sweet, A History of English Sounds (Oxford, 1888); P. Passy, Les Changements phonetiques (Paris, 1890). For phonetics in language-teaching see H. Sweet, The Practical Study of Languages (London, 1899); O. Jesperson, How to Learn a Foreign Language (London, 1904). For phonetic shorthand, H. Sweet, A Manual of Current Shorthand (Oxford, 1892). For the application of phonetics and phonetic notation to the practical study of special languages, H. Sweet, A Primer of Spoken English (2nd ed., Ox ord, 1895); F. Beyer and P. Passy, Elementarbuch des gesprochenen Franzosisch (2nd ed, Cothen, 1905); W. Vietor, Deutsches Lesebuch tn Lautschrift (Leipzig, 1899). (H. Sw.)