THE SYSTEM OF ENGLISH PHONEMES

The previous section has been concerned with the aims and methods of phonological analysis. In this section we are going to give a brief and readable description of the problems which scholars face trying to describe the English sounds from the func­tional point of view. We shall try to explain what is understood by quality of a sound, what articulatory characteristics may be considered constituents of quality and to determine which of them are important enough to arrest our attention as phonologically relevant.

By way of introduction we would like to remind you that

59

there are two major classes of sounds traditionally distinguished by phoneticians in any language. They are termed consonants and vowels. It would be fair to mention that the distinction is based mainly on auditory effect. Consonants are known to have voice and noise combined, while vowels are sounds consisting of voice only. From the articulatory point of view the difference is due to the work of speech organs. In case of vowels no obstruc­tion is made. In case of consonants various obstructions are made. So consonants are characterized by so-called close articulation, that is by a complete, partial or intermittent blockage of the air-passage by an organ or organs. The closure is formed in such a way that the air-stream is blocked or hindered or otherwise gives rise to audible friction. As a result consonants are sounds which have noise as their indispensable and most defining characteristic.

What we have said makes it logic-enough to consider each class of sounds independently. So we shall start with consonants.

Consonants

On the articulatory level each consonant may be identified by stating two general facts about it:

1) what sort of articulatory posture it is formed by;

2) whereabout in the mouth (or pharynx) it is produced.

Besides these major characteristics the particular quality of a consonant may depend on a lot of other factors, that is by what articulatory organ (or organs) an obstruction is made, how vocal cords work at the moment of production, what cavity is Used as a resonator, what is the force of articulatory effect and many others. So in our view the particular quality of a consonant would be best thought of as a complex bundle of features. Each sound is known to have three aspects: acoustic, articulatory and auditory and therefore can be studied on these three levels. For the sake of anal­ysis each aspect can be considered and described independently though we should take it reasonably obvious that there is no sharp dividing line between them. Trying to work out a classifica­tion of such complex units as speech sounds an analyst should specify those properties of sounds which are relevant to the sub­ject under discussion, so the attempts to classify sounds should have a theoretical foundation. Besides, each classification should

60

not only aim at linguistic description but it should be justified from "teaching of a language" point of view. Therefore it should try to include, if possible, both the principal relevant features and the ones that are redundant from phonological point of view but are considerably important for the articulation of the sound. Here we should note that phonological description of sounds will be made in terms of articulatory level.

As to the classification of English consonants there are few ways of seeing the situation.

According to V. A. Vassilyev (79) primary importance should be given to the type of obstruction and the manner of produc­tion of noise. On this ground he distinguishes two large classes of consonants:

a) occlusive, in the production of which a complete obstruc­tion is formed;

b) constrictive, in the production of which an incomplete ob­struction is formed.

The phonological relevance of this feature could be exempli­fied in the following oppositions:

[ti:] — [si:] tea — sea (occlusive — constrictive)

[si:d] — [si:z] seed — seas (occlusive — constrictive)

[pul] _ [ful] pull — full (occlusive — constrictive)

[bɜut] — [vɜut] boat — vote (occlusive — constrictive)

Each of the two classes is subdivided into noise consonants and sonorants. The division is based on the factor of prevailing either noise or tone component in the auditory characteristic of a sound. In their turn noise consonants are divided into plosive consonants (or stops) and affricates. The following scheme might be helpful to understand the system built in accordance with the above-mentioned order of articulatory characteristics:

Consonants

61

Another point of view is shared by a group of Soviet phone­ticians (authors of the book among them). They suggest that the first and basic principle of classification should be the degree of noise. Such consideration leads to dividing English consonants into two general kinds:

A — noise consonants

В — sonorants

It is easy to see that the term "degree of noise" belongs to auditory level of analysis. But it is hardly necessary to point out that there is an intrinsic connection between articulatory and au­ditory aspect of describing speech sounds, such that sometimes it is impossible to account for the former except in terms of the latter. In the above-mentioned case it is the term of auditory lev­el that defines the characteristic more adequately.

Talking about sonorants it is necessary to mention that they are consonants that phoneticians have traditionally a lot of argu­ment about. The point is that sonorants are sounds that differ greatly from all other consonants of the language. This is largely due to the fact that in their production the air passage between the two organs of speech is fairly wide, that is much wider than in the production of noise consonants. As a result, the auditory effect is tone, not noise. This peculiarity of articulation makes so­norants sound more like vowels than consonants. On this ground some of the British phoneticians refer some of these con­sonants to the class of semivowels, [r], [j], [w], for example. Acoustically sonorants are opposed to all other consonants be­cause they are characterized by sharply defined formant struc­ture and the total energy of most of them is very high. However, on functional grounds, according to their position in the syllable, [r], [j]. [w] are included in the consonantal category, but from the point of view of their phonetic description they are more perfect­ly treated as vowel glides.

According to the Soviet phoneticians sonorants are consid­ered to be consonants from articulatory, acoustic and phonologi­cal point of view and can be classified according to all the princi­ples of classification of consonants.

Looking back on the problem of providing an accurate criteri­on for estimating classificatory value of various properties dis­played by sounds we may say that the great articulatory and acoustic difference of noise consonants and sonorants could be very well relied upon in this respect. The phonological relevance

62

of this factor (the degree of noise) could be proved by the follow­ing oppositions:

[beik — meik] bake — make (noise consonant — sonorant)

[vi:l — wi:l] veal — wheel (noise consonant — sonorant)

The following scheme may illustrate the point of view shared by the authors of this book:

Summarizing briefly we may emphasize the fact that the two above-mentioned characteristics, that is the degree of noise and the manner of articulation are considered essential enough from the classificatory point of view, no matter which of them comes first in the description of the system. We could add that any par­ticular point of view will be judged by its effectiveness. Both above-mentioned classifications are valid, consistent and com­plete though we could assume that the latter is more preferable for teaching purpose because it is more concise and symmetrical.

The place of articulation is another characteristic of English consonants which should be considered from the phonological point of view. The place of articulation is determined by the ac­tive organ of speech against the point of articulation. According to this principle the English consonants are classed into:

1) labial,

2) lingual,

3) glottal.

The class of labial consonants is subdivided into: a) bilabial; b) labio-dental; and among the class of lingual consonants three subclasses are distinguished; they are: a) forelingual, b) medio-lingual and c) backlingual. The classification of consonants ac­cording to this principle is illustrated in the following scheme:

63

The importance of this characteristic as phonologically rele­vant could be proved by means of a simple example. In the sys­tem of English consonants there could be found oppositions based on the active organ of speech and the place of obstruction.

[pæn] — [tæn] pan — tan        (bilabial — forelingual)

[wai] — [lai] why — lie             (bilabial — forelingual)

[weil] — [jeil] weil — yale        (bilabial — mediolingual)

[pik] — [kik] pick — kick          (bilabial — backlingual)

[les] — [jes] less — yes            (forelingual — mediolingual)

[del] — [gei] day — gay           (forelingual — backlingual)

[sai] — [hai] sigh — high         (forelingual — glottal)

[fi:t] — [si:t] feet — seat          (labio-dental — forelingual)

Our next point should be made in connection, with another sound property, that is voiced — voiceless characteristic which depends on the work of the vocal cords. We should note that it has long been believed that from the articulatory point of view the distinction between such pairs of consonants as [p, b], [t, d], [k, g], [s, z], [f, v], [f, ʒ], [ʧ ʤ] is based on the absence or pres­ence of vibrations of the vocal cords, or on the absence or pres­ence of voice or tone component. However, a considerable body of experimental work on physiological and acoustic aspects of these sounds showed that this is not the only difference between them. It is obvious now that there is also energy difference. All voiced consonants are weak (lenis) and all voiceless consonants are strong (fortis). It is worth noticing that now there is a consid­erable controversy about what phonetic feature is involved in the above-mentioned oppositions. In the intervocalic position, for example, latter — ladder, the voicing difference is important, since it is the distinctive feature of the consonants. In word-initial and final positions the pronunciation of consonants traditionally considered to be voiced may well be voiceless. In these positions

64

it is the energy difference that serves as a differentiating feature. For example, "cap" — "cab", "not" — "nod", "pick" — "pig". In ini­tial position aspiration would be a more important feature for stops, for example "tick" — "Dick", "cap" — "cap", "pit" — "bit". In a word-final position it is the length of the proceeding vowel that would constitute the chief difference (the vowel of "bead" is longer than that of "beet").

In view of what has just been said it is perfectly obvious that the presence or absence of voice in the above-mentioned opposi­tions is not a constant distinctive feature. Thus it may be said that the oppositions [p — b], [t — d], [k — g], [f — v], [s — z], [ʃ — ʒ], [ʧ — ʤ] are primarily based on energy difference, that is on fortis — lenis articulation, which are their phonologically relevant features. It is for this reason that such characteristics as voice­less — voiced have given place to "fortis" — "lenis" terms.

There is one more articulatory characteristic which is usually included into the set of principles on the basis of which the Eng­lish consonants are classified, that is the position of the soft pal­ate. According to this principle consonants can be oral and na­sal. There are relatively few consonantal types in English which require the lowered position of the soft palate. They are the na­sal occlusive sonorants [m], [n] and [ŋ]. They differ from oral plo­sives in that the soft palate is lowered allowing the escape of air into the nasal cavity.

It is a well-known fact that no differences of meaning in Eng­lish can be attributed to the presence or absence of nasalization (there are no two consonants in English which differ in the posi­tion of the soft palate). It is for this reason that it cannot be a phonologically relevant feature of English consonants, so it is an indispensable concomitant feature of English nasal consonants. .

So far we have attempted to show the way the basic articula­tory characteristics can be interpreted from the phonological point of view and what classificatory value these characteristics may possess as items of a system.

There are, however, other problems of a phonological char­acter. In the English consonantal system it is the problem of af­fricates, that is their phonological status and their number. It would be true to say that though it is a long-standing problem it is obvious to anyone who is acquainted with the state of current general theory that its importance is not lessened, nor is the ana­lysts' concern for it diminished.

3—3483

65

The question is: what kind of facts a phonological theory has to explain?

1) Are the English [ʧ, ʤ] sounds monophonemic entities or biphonemic combinations (sequences, clusters)?

2) If they are monophonemic, how many phonemes of the same kind exist in the system of English consonants, or, in other words, can such clusters as [tr, dr], [ts, dz] and [tθ, dð] be considered affricates?

To define it is not an easy or obvious matter. If we tried to analyse all the theories concerning the problem, the task would prove daunting. So we shall try to direct our attention to essen­tials.

One thing is clear, that is the above-mentioned sounds are complexes because articulatorily and acoustically we can distin­guish two elements. The articulatory difference between [ʧ, ʤ] on the one hand, and [t, d] on the other is based on the speed of releasing the obstruction. When [ʧ, ʤ] are pronounced the re­leased stage is performed slower than in case of [t, d]. But this is not the only difference between an affricate and a plosive. Spe­cial instrumental analysis shows that there is no synchrony in releasing the obstruction by the central part of the tongue and its sides. It is the centre of the front part of the tongue that comes first in release stage while the sides of the tongue still form a clo­sure. At the next stage the obstruction is released by the sides of the tongue while the central part of the tip forms a narrowing against the alveolar ridge through which the air escapes. The ar­ticulatory movement of the parts of the tongue is smooth and continuous. It is difficult to say where the bordering line between the first and the second stage could pass.

Considering phonemic duality of affricates it is necessary to analyze the relation of affricates to other consonant phonemes to be able to define their status in the system of consonants. As you know from the course of practical phonetics, traditionally it is the type of obstruction that serves a basis bf comparison. The two main types of obstruction are complete and incomplete. In accordance with such classification affricates cannot be referred to either of the groups, since they are known to consist of both: the closure and the narrowing. That is why it seems justified to single out a group of affricates, or occlusive-constrictive conso­nants.

Theoretically in each language there might be as many

66

affricates as there are fricatives but in reality the number of them is limited and there are languages where there are none.

As was mentioned above, the problem of affricates is a point of considerable controversy among phoneticians. According to Soviet specialists in English phonetics, there are two affricates in English, they are: [ʧ, ʤ]. D. Jones points out there are six of them: [ʧ, ʤ], [ts, dz] and [tr, dr] (64). A. C. Gimson increases their number adding two more affricates: [tθ, dð] (57).

We could ask, what might account for such a difference in their opinions?

The fact is that Soviet phoneticians look at English affricates through the eyes of a phoneme theory, according to which a phoneme has three aspects: articulatory, acoustic and functional, the latter being the most significant one. As to British phoneti­cians, their primary concern is the articulatory-acoustic unity of these complexes, because their aim is limited by practical rea­sons of teaching English.

Before looking at these complexes from a functional point of view it is necessary to define their articulatory indivisibility. This procedure is generally carried out according to the rules worked out by N. S. Trubetskoy (34). According to his point of view a sound complex may be considered nonphonemic if:

a) its elements belong to the same syllable;

b) it is produced by one articulatory effort;

c) its duration should not exceed normal duration of either of its elements.

Now let us try to apply these criteria to the above-mentioned sound complexes.

Rule I. Syllabic indivisibility. If we compare the following words:

butcher ['buʧ-ə] —lightship ['lait-ʃip]

mattress ['mætr-is] — footrest ['fut-rest]

curtsey ['kɜ:-tsi] — out-set ['aut-set] eighth [eitθ] — whitethorn ['wait-θɔ:n]

we could see that in the words given in the left column the sounds [ʧ], [tr], [ts], [tθ] belong to one syllable and cannot be di­vided into two elements by a syllable-dividing line. We could compare these complexes to the Russian [ц] phoneme which also cannot belong to different syllables. Cf. [иай-'цо], но [съ-'вет-ский]. We could assume that the articulation of the voiced counterparts does not differ from the voiceless ones.

67

Rule II. Articulatory indivisibility. We might say that special instrumental analysis shows that all the sound complexes in question are homogeneous and have the maximum of articulato­ry features in common; that is at the beginning of the articula­tion the organs of speech are in the position of the second frica­tive element [ʃ], [r], [s], [θ] or [ʒ], [z], but there is a complete ob­struction (a closure) formed by the tip and the sides of the tongue against the alveolar ridge and the side teeth. Then the closure is released and the air escapes from the mouth cavity, producing audible friction. In other words the above-mentioned complexes are produced by one articulatory effort.

Rule III. Duration. We should note here that the available data of that kind is not reliable enough. Moreover [ʧ, ʤ] com­plexes which are considered phonemes by all phoneticians, are not defined properly as to their length or quantity. With G. P. Torsuev (32), we could state that length of sounds depends on the position in the phonetic context, therefore it cannot serve a reliable basis in phonological analysis. He writes that the length of English [tʃ] in the words [tʃɛə] chair and [mæʧ] match is different, [ʧ] in march is considerably longer than [t] in mat and may be even longer than [ʃ] in mash. This does not prove, how­ever, that [ʧ] is biphonemic.

N. S. Trubetskoy himself admits that this condition is less im­portant than the two previous ones (34).

From what we have said it follows that this rule has no deci­sive value. That is why we could be certain that the analysed sounds are articulatively indivisible. So potentially they can be considered monophonemic. But in fact they could be considered monophonemic on condition they could enter the "phonological model of the language" (58).

The rules suggested by N.S.Trubetskoy (34) are based on ar­ticulatory and phonological indicators. They may well be called the grounds of phonology, because in great many instances they permit us to define the phonemic status of sound complexes. However, doing credit to articulatory and phonological criteria applied in the interpretation of such entities, scholars seem to at­tach decisive importance to morphonological criterion: Accord­ing to this criterion a sound complex is considered to be mono­phonemic if a morpheme boundary cannot pass within it, be­cause it is generally assumed that a phoneme is morphologically indivisible. If we consider [tf], [ʤ] from this point of view we

68

could be secure to grant them a monophonemic status, since they are indispensable. As to [ts], [dz] and [tθ], [dθ] complexes their last elements are separate morphemes [s], [z], [θ], so these elements are easily singled out by the native speaker in any kind of phonetic context. So these complexes do not correspond to the phonological models of the English language and cannot ex­ist in the system of phonemes. The case with [tr], [dr] complexes is still more difficult. According to morphonological criterion they have more grounds than the above-mentioned [ts], [dz], [tθ] and [dθ] to be considered monophonemic because they very often belong to the same morpheme. In such situations analysts apply the native speaker's knowledge, which might serve an ad­ditional criterion, for any linguistic analysis can be largely based on intuition about the rules to be recognized, combinations to be noted and results to be obtained. So talking about [ʧ], [ʤ], if we assume that in the word chair [t] is dispensable leaving share [ʃɛə] and [ʃ] is dispensable leaving tear [tɛə] and therefore it is a sequence of [t] + [ʃ], the native speaker's feeling cannot accept it as anything but a unit. Perhaps the reason is partly to do with [ʤ] which cannot be treated so easily as [ʧ]. If we dispense [d] in [ʤei], for example, we could get [ʒei], but [ʒ] is not a permitted initial phoneme in English because it occurs only in a few bor­rowed words. So it is not satisfactory, because it would be odd to treat one of the correlated pairs as a sequence, and the other as a unit. Another reason: if we treat [ʧ], [ʤ] as sequences what other sequences of this type would we find in the system of En­glish consonants? Parallel to the [dr] complex there are [kr], [pr], (θr]. But there is nothing parallel to [ʧ] and similarly to [ʤ]. So it may be said that the native speaker does not regard [ʧ], [ʤ] as composite sounds, that is composed of distinctive elements. On the other hand, [tr], [dr] are not normally regarded as anything but sequences. A. C. Gimson himself admits that he grants them monophonemic status on the basis of the articulatory criterion (57).

By way of conclusion we could say that the two approaches that have been adopted towards this phenomenon are as fol­lows: the finding that there are eight affricates in English [ʧ], [ʤ], [tr] [dr], [ts], [dz], [tθ], [dθ] is consistent with an articulatory and acoustic view, because in this respect the entities are indivi­sible. This is the way the British phoneticians see the situation. This point of view underestimates the phonological aspect and is in a way an extremity.

69

On the other hand, Soviet phoneticians are consistent in looking at the phenomenon from the morphological and the phonological point of view which allows them to categorize [ʧ, [ʤ] as monophonemic units and [tr], [dr], [ts], [dz], [tθ], [dθ] as biphonemic complexes. However, this point of view reveals the possibility of ignoring the articulatory and acoustic indivisibility of the complexes. In this case the pronunciation peculiarities of these complexes are not analysed properly. It must be distinctly understood that that is a genuine articulatory difference between phonemes [t], [d] pronounced in combination with other sounds and the [t], [d] as parts of clusters [tr], [dr]. It requires special at­tention and training. On this account textbooks in practical pho­netics should include effective instructions on teaching the pro­nunciation of these sound complexes.

So far we have attempted to show how a fairly general prob­lem of interpreting the system of English consonants is solved, what essential complication exists and what kind of criteria can be used in solving these problems.

Summarizing what have been described we could state that with the majority of Soviet specialists in English phonetics we consider relevant the following articulatory features:

1) type of obstruction,

2) place of obstruction and the active organ of speech,

3) force of articulation.

As was mentioned in the previous section, the phonetic sys­tem of a language is patterned. So we have tried to show what articulatory features could serve as a criterion for grouping con­sonants into functionally similar classes. The above-mentioned articulatory characteristics are undoubtedly the prime ones as they specify the essential quality of a consonant which is enough to describe it as an item of a system. On this level of analysis it is the point where the distinction between consonants becomes phonemic that matters. However, if we approach the matter from "teaching pronunciation" point of view it is natural we should want to gain some additional information about the articulation of a consonant, about such delicate distinctions that make the description complete from the articulatory point of view, for example, if the consonant is apical or dorsal; if it is dental, alveolar, post-alveolar, or palato-alveolar; if it is oral or nasal; if the narrowing is flat or round and a lot of others. These characteristics are of no importance from phonological point of

70

view but they provide necessary and instructive information for comparison between the English consonants and those of the mother tongue and so are considerably important for teaching purposes. It is for this reason that these characteristics are nor­mally included into the classification.

 

⇐ Предыдущая12345678910Следующая ⇒

Поделиться: