| 37. English has two commonly cultivated and imitated dialects: the Received Pronunciation (BrE) and the General American (AmE) variety. Whereas the two dialects have many common traits in their treatment of consonants, the two vary most distinctly in the nature of the vowels. Figures 4 and 5 list the principal IPA phones on the left along with the various phonemes of English on colored backgrounds. Keep in mind that phonemes exist auditorily, i.e., only in the ear, but linguists customarily use symbols of the closest articulatory and acoustic sounds to represent them. We will begin with a description of the principal allophones of the consonant phonemes and follow with those of the vowels. |
|
38. Stops.
Among the most commonly used consonants of English are the stops.
These correspond basically to the IPA series of plosives articulated as 1) bilabial; 2) dental, alveolar, postalveolar, and retroflex; 3) or velar. The articulation of the second group are called coronal and whether they are dental or retroflex depends on the nature of the articulation of the sound preceeding or following. If the sound following is a vowel, the coronal sound becomes dental, as in tea, if it is an alveolar, then it becomes retroflex, as in tree. The behavior of the coronal stops before the syllabic liquid was described in § A34. The unvoiced coronal stop becomes a simple glottal plosive before the syllabic nasal, so that cotton becomes  and mountain,  .
On the other hand wooden is  .Similarly with the velar stops: if the vowel following is a front vowel, the velar stop is palatal, as in “kick.” If the vowel following is a back vowel, the stop is velar, as in “cook.” Further the stop is often left unreleased at the end of a word or phrase. The release of the stop after “m” is required to distinguish hummed from humpt and dogged from docked since the “p,” viz., “k” is unreleased. Being unreleased is usual with the bilabials but rare with the velars. |
| 39. Affricates. The affricates are actually described by the IPA system as a combination of a coronal stop together with its corresponding similarly articulated fricative. Instead of the stop being released into a vowel or being left unreleased, it is released into the fricative sound. |
|
40. Fricatives.
In general the fricatives constitute the largest single class of phones defined by manner of articulation.
Linguists normally assign the four pairs listed into eight separate phonemes.
However, there appear to be five situations that tell us when the dental fricative is pronounced with its voiced alternative : 1) after a vowel when the syllabic “r” follows, as in brother, father, whether, bother, dither, etc. 2) after a long vowel in single syllable words of a single morpheme, as in seethe, smooth, lathe, loath, etc., but not teeth, beneath, uncouth; 3) preceding the plural ending in certain nouns in “th”, as in clothes (the “th” is usually dropped in this word, but the voicing would be parallel to that of other unvoiced fricatives here); 4) in the verb form of nouns ending in “th”, as in bathe, teethe, breathe, loathe; 5) the on-set of the definite article, the, demonstratives, this, etc., and adverbs, there, thence, thus, etc. Some of these environments are phonological, but most are morpho-phonemically defined, i.e., they depend on the word’s syntax and meaning. Another borderline phoneme is the voiced postalveolar fricative, which seems to appear only in words from French, as in rouge, garage, genre. Sibilants. The sibilants belong together as a subclass of fricatives mainly because of the way the plural marker on nouns {–ES1} and the present tense marker on verbs {–ES2} behaves. These morphemes are always pronounced as a separate syllable when the noun, viz. verb, ends in one of these four sounds, as fishes. |
| 41. Semi-vowels. The “h” is traditionally classed with the glides as sometimes acting like a consonant in modifying the pronunciation of a vowel, but at other times being silent. The morpho-phonemic rule for the pronunciation of the articles mentioned in § A29 must take this sound into account: do not drop the “n” of an and pronounce the vowel of the as a high front vowel when the article is followed by a word whose on-set is a vowel or unstressed semi-vowel. Hence we have a history but an historic event, and the history but the historic event /ðiyis'towrike'vent/. Its occasional palatal articulation is described in § A31. |
42. Glides.
Some linguists class the glides together with the semi-vowels.
Phonemically the back glide /w/ consists of a velar approximant with the simultaneous puckering of the lips  .
Such articulations as involve both ends of the vocal tract are called peripheral.
The front glide /y/ is phonetically a palatal approximant. The /w/ is not unlike the high back vowel , except that it clusters with other consonants in the on-set and may follow a vowel to form a coda. Similarly the /y/ sounds like a but also forms clusters in the on-set and may act as the coda in a syllable. If the nucleus vowel is a mid or high back vowel, the following back glide will form with it a diphthongized or long vowel: /uw/ and /ow/. The vowel and glide are homorganic. If the nucleus vowel is a mid or high front vowel, the following front glide, being homorganic, will likewise form a diphthongized or long vowel: /iy/ and /ey/. Some linguists of BrE prefer to compare these vowels with the temporally longer and uniformly articulated vowels of the many continental languages that have such vowels. However, it seems clear to analysts of AmE that such vowels are at least tense and usually if not always diphthongized. Their short counterparts are termed lax. |
|
43. Nasals.
Corresponding to the unreleased stops, , , , are the three consonants made with the soft velum lowered —the nasals.
When these three nasals cluster in the coda, it is with these homorganic stops.
The non-velar nasals (as well as the liquids) are capable of forming syllables without the articulation of a vowel.
Such syllabic consonants are attached to the previous syllable, but are acoustically (phonemically) taken as having a vowel.
There are several examples of these in § A38 above.
Even in interjections, for example the “um” of deliberation and the “uh-hum” of cautious assent, the written forms conform to the VC and VCC patterns of English.
The first example is normally a glottal stop followed by an unreleased /m/.
An emphatic version, sometimes spelled “humpf” might convey a strong disbelief or disagreement.
The second example is a long unreleased /m/ during which a glottal stop is co-articulated. The velar nasal is the only one that cannot form a syllable on-set. Since it appears as coda in a limited number of morphemes, it should be possible to posit a null-allophone of the velar stop to describe its occurrence there — a silent velar stop. It is interesting in this regard that the Greeks too seemed to be confused when writing this phone almost, but falling short of treating it as a separate phoneme. Before a /g/ they wrote it as another /g/, i.e., “γγ” even though before a /k/ it was written as an /n/. |
|
44. Liquids.
There are two liquids, the lateral /l/ and the non-lateral /r/.
Phonetically they are approximants, but the lateral approximant has its own series being distinguished from the others by the tongue allowing air to escape on both sides. As mentioned above in § A34, the non-rhotic dialects alter the pronunciation of the vowel when an “r” follows. In the phrase “so far away” the “r” returns,  , intruding between the two vowel sounds.
The intrusive “r” also appears when two vowels in separate words would otherwise come together, as in “to India alone,”  and “the idea of it,”  .
These examples suggest that the intrusion might be made to prevent the occurrence of a triphthong.
|
| 45. Whether we are describing BrE or AmE, we contrast the monophthongs or simple vowel sounds from the diphthongs or double vowel sounds. BrE has been described with as many as thirteen simple vowels, seven diphthongs, and two triphthongs. My own rhotic dialect of AmE reduces this to seven simple vowels (four of them diphtongized), four diphthongs, and one triphthong. It is possible that the phonologists describing these dialects are 1) unable to discern the articulatory differences of their ingrained phonemic systems and/or 2) fail to attribute differences in vowel sounds to their consonantal environment. Part of this may be due to the fact that there are both stressed and unstressed varieties of vowels. The phone called schwa (an inverted “e”), when unstressed may theoretically be the phonetic realization of any of several phonemes. |
46. High & mid vowels.
The peripheral vowels, i.e., those made in the front or back of the mouth, come in two varieties, the long ones and the short ones.
It was Trager and Smith in describing the mid-American dialect who suggested that the long variety was a diphthongized version of the short ones.
Linguists often want to make the BrE into a diphthong for the back diphthongized vowel of AmE /ow/ in boat, home, no, snow.
In a similar fashion they also make the BrE into a diphthong for the front diphthongized vowel of AmE /ey/ in bait, shave.
In addition to this, they take advantage of their non-rhotic accent to add a few more diphthongs and even a couple of triphthongs.
In such cases it seems most economical to analyze these furtive vowels as reflexes of the softened approximant: fear /fiyr/ BrE  , AmE  ; poor /puwr/ BrE  , AmE  .
|
|
47. Low & central vowels.
The BrE dialects tend to deepen or lower the AmE mid vowels.
The open back unrounded-a of father has a rounded phone that appears as the short-o in BrE pot.
Confusing the issue is the fact that the AmE pronunciations of pot may be rounded but that of saw usually is. The inverted vee of the IPA refers to the fully unmarked “neutral” vowel of but. We have chosen to represent the vowel phoneme in this word with /œ/. This phoneme may be stressed or unstressed, but when it is unstressed it is formed higher in the mouth. The higher neutral phone, the schwa, is the result of unstressing any vowel. This unstressed variety may be taken as a phoneme based on sandhi phenomena where it may be entirely lost. In cases where there is no stressed example the proper phoneme reflex of a schwa is indeterminate. For example, in the suffix that makes composition adjectives, as in oaken, or in the softening formative, as in trickle, the unstressed schwa appears with the /n/, viz. /l/, sometimes as its null allophone, i.e. with a syllabic /n/, viz. /l/, as in wooden and wheedle. The sandhi rules that treat the pronunciation of the regular past tense /-œd/, present tense singular /-œs/, and the noun plural /-œs/ also suggest a null allophone. In some cases, such as the schwa of the indefinite article, we sometimes see the full form /æn/. |
| 48. Diphthongs. Both BrE and AmE have the three diphthongs listed on figure 5, and pronounce them in largely the same way. In Australia and certain regions of England the /ay/ pronunciation has spread to words having /ey/. Many linguists maintain the existence of additional diphthongs in the non-rhotic dialects on the principle of primacy in articulatory and acoustic phonetic phenomena. These are held to rule over auditory considerations in the absence of demonstrable contrast in the definitions of distinctive features. |
| 49. Figure 6 summarizes the analysis of the phonemes of English on the basis of two dimensions of gross articulatory features. The feature of close vs. open is a relative measure of the vocal tract and is plotted on the horizontal axis as well as the vertical axis for vowels where it might better be termed “tense vs. lax.” The dotted line separates consonants based on the presence or absence of voice, but the vowels based on their relative length. The normal place of articulation for obstruents (and to a less obvious extent again for sonorants) varies from the front to the back of the vocal tract as one passes from “close” to “open.” In the following sections we will formalize the analysis in terms of the constituency of the various elements and their classification. |
