Gramley, Stephan; P?tzold, Kurt-Michael. A Survey of Modern English
Подождите немного. Документ загружается.
between the stops and the fricatives, a ‘mismatch’ which is caused by the lack of the
palatal fricatives /x/ and /
γ/ and the presence of the dental fricatives /θ/ and /ð/. Note that
the fricative /x/ (like the ach sound of German) with its allophonic variant [ç] (the ich
sound) has, as already mentioned, been retained in the regional Scots dialects, for example
in night /neçt/ or loch /lox/. Furthermore, many people use it in the pronunciation of
foreign words or names such as Bach or the interjection ugh /
x/.
Nasals The nasals do not occur in lenis-fortis (voiced-voiceless) pairs, for they are
sonorants and therefore are, phonologically speaking, always voiced. There are only three
nasals since the post-alveolar /
/ of Spanish (mañana), Italian (senor) or French
(compagnon) is not phonemic in English; instead, the analogous sound in English is seen
as a sequence of two phonemes /nj/ as in canyon /k
njən/. Furthermore, the historically
more recent addition to the nasals, the phoneme /
ŋ/, is not fully equivalent to /m/ and /n/
since it cannot occur initially in a word, nor does it occur after all the vowels of English
(in RP it follows /
, , , ɒ/; in GenAm /, , , ɑ, ɔ/, e.g. sing, sang, sung, song, the
latter with /
ɑ/ or /ɔ/ in GenAm depending on the region).
The lateral The lateral /l/ consists of two noticeably different allophones, clear [l] and
dark [
], which do not stand in phonemic opposition to each other. Indeed, there are
accents such as those of southern Ireland or the southwest of England in which clear [l]
appears exclusively, and accents from other areas, such as Scotland and some parts of
the United States, in which only dark [
] occurs. RP, as previously mentioned, is char-
acterized by the complementary distribution of the two allophones. This means that in
one set of circumstances only the one may occur and in another set of circumstances only
the other. Concretely, clear [l] is used before vowels (e.g. look, teller) while dark [
]
appears before consonants (e.g. help) or at the end of a word (e.g. goal); this includes
syllabic [
] as in bottle).
The semi-vowels The semi-vowels are difficult to adapt to the scheme of classifi-
cation used here because they are, phonetically speaking, not consonants at all, but vowels
which occur in the typical position of consonants, peripheral to the syllable (see 4.2). In
many classifications /w/ and /hw/ are classified as bilabial. The rounding of the lips which
is typical of /w/ and /hw/ is, however, of secondary importance and need not be present.
Note that pre-vocalic /r/ is often produced with lip-rounding as well. The criterion which
has been used in positioning the semi-vowels in the chart is the position of the tongue:
/j/ corresponds to the high front vowel /
/, because it has the same sound quality as //.
/j/ differs only inasmuch as it is extremely short (non-syllabic). Like /
/ it requires a
tongue position close to the alveolar ridge; hence it has been classified as alveolar. /r/
corresponds to the central vowel //, which is more or less post-alveolar; and /w/ corres-
ponds to /u/, a high back vowel, which takes a tongue position close to the velum.
The voiceless vowel /h/ occurs only before vowels and has a different resonance
depending on what vowel follows it, hence the term ‘voiceless vowel’. Preceding /i/, as
in heat, /h/ is [i]; preceding /
/, as in hat, it is [] etc. The small circle (‘under-ring’)
indicates devoicing; it is ‘whispered’.
The glottal stop [
ʔ] does not have the status of a phoneme, but is so obvious in some
accents of English that it will be treated below under 4.3.4.
72 ENGLISH AS A LINGUISTIC SYSTEM
4.3.2 Force of articulation and voicing
All of the obstruents are members of pairs of phonemes which share the same features
with regard to place and manner of articulation. However, the members of each pair differ
from each other in that the one is pronounced with more force or energy (‘hard’, ‘fortis’)
and is always voiceless, i.e. the vocal cords do not vibrate while it is being pronounced;
the other member of each pair is pronounced with relatively less force or energy (‘soft’,
‘lenis’) and is often, though not always, voiced. The pair /
ʃ/–// differs from the other
obstruent pairs because /
/, which occurs relatively seldom, appears only at the begin-
ning (e.g. Zhivago) or end (rouge) of words which have been borrowed from other
languages and have been only incorporated into English imperfectly. In a medial position
(measure) it is perfectly normal.
The nasals, semi-vowels and the lateral are always regarded phonologically as voiced,
and indeed they usually are voiced phonetically as well. However, in a voiceless environ-
ment they may as a matter of actual phonetic realization become devoiced, e.g. flee [fli]
or twice [tw a
s]. /h/ is always voiceless. The voiced obstruents are fully voiced only in
a voiced environment, such as between two vowels (e.g. /v/ in giving or /d/ in sadder).
At the beginning of a word the sonority or voicing sets in gradually in the course of artic-
ulation, which means that voicing may be incomplete. In the final position of a word the
voicing may be missing totally. In both of these cases the lack of voicing is not gener-
ally noticed because the distinction between the obstruent pairs is maintained more by
force of articulation than by voicing. The /v/ of /la
v/ is, for example, seldom voiced. The
contrast to the /f/ of /la
f/ is maintained, rather, because /v/ is lenis while /f/ is fortis.
Furthermore, the length of the vowel differs since stressed vowels are longer before voiced
than before voiceless consonants (see 4.4.2). In addition, the opposition between the
‘voiced’ and the ‘voiceless’ stops in initial position is supported by the aspiration which
is associated with the fortis phoneme as opposed to the unaspirated lenis phoneme. In
English no stops are aspirated after an /s/, as mentioned above in 4.2.
In AmE and increasingly in AusE the distinction between voiceless /t/ and voiced /d/
is neutralized in intervocalic position (provided the following syllable is unstressed); both
are pronounced with voicing. Pairs of words such latter-ladder cannot, for this reason,
be distinguished on the grounds of sonority or force of articulation. For some speakers
the vowel preceding the stop may, however, be longer when a /d/ follows (see also 12.1.3).
4.3.3 Restrictions in distribution
The system of 24 consonants which has been assumed here does not mean that the various
consonants are fully equivalent. The infrequency of /
ŋ/ in comparison to /m/ and /n/ and
that of /
/ in comparison to /ʃ/ has already been pointed out. In much the same way initial
/
ð/ is restricted to the so-called grammatical or function words, which include pronouns
(they, thou etc.), the definite article (the), the demonstratives (this, that, these, those) and
basic adverbs such as there, then, thus etc. In addition, /
ð/ never occurs directly before
another consonant within the same syllable. The only exception is where an inflectional
ending follows, but here the /
ð/ is separated from the following consonant by a morpheme
boundary, e.g. the regular inflectional {-D} (= /d/ of the past tense and past participle) as
in smoothed or {-S} (= /z/ of plural nouns, possessive nouns or of the third person singular
simple present of verbs) as in paths, path’s or breathes.
1111
2
3
4
5
6
7
8
9
1011
1
2
3
4111
5
6
7
8
9
20111
1
2
3
4
5111
6
7
8
9
30111
1
2
3
4
5
6
7
8
9
40111
1
2
3
44
45
46
47111
THE PRONUNCIATION AND SPELLING OF ENGLISH 73
4.3.4 Consonant changes
/hw/ is the only consonant which seems to be disappearing completely. Many speakers
today use /w/ where once /hw/ was pronounced. In this way numerous <w-> and <wh->
words have become homophonous, so wear and where, wheel and we’ll, which and witch
etc. Nonetheless, many speakers still use /hw/ as an emphatic variant of /w/, as in Why?!
/hwa
/. As a result there are cases in which people have been known to produce an un-
historical /hw
ɑυ/ wow!. The /hw/–/w/ opposition is still maintained in various American
and British accents (e.g. the Northern dialect area in the United States and Scotland in
Great Britain).
Especially noticeable is the disappearance of non-prevocalic <r> in many accents of
English. While /r/ is pronounced wherever it is written in GenAm, in Irish English, in Scot-
tish English and in various parts of the southwest of England, it is missing in such accents
as RP, the English of New England and wide areas in the American South, Australia and
New Zealand. In these latter accents an /r/ can occur only before a vowel. In talking about
this split in the accents of English, it is convenient to speak of rhotic and non-rhotic
accents, i.e. those which have and those which have not retained /r/ in all positions.
In those accents which have retained non-prevocalic /r/ the quality of this phoneme
differs considerably. In America, Northern Ireland, parts of Scotland and the English
southwest this /r/ is realized chiefly in the quality of the preceding vowel, which is
r-coloured. In parts of southern Ireland and Scotland the /r/ may be rolled at the tip of
the tongue, trilled [r] or flapped [
ɾ].
Just as non-prevocalic r has become vocalic (‘r-coloured vowels’), so, too, is non-
prevocalic /l/ not only dark, as mentioned above, but completely vocalic in such widely
separated accents as Cockney and Southern American: the tongue no longer touches
the top of the mouth; instead, only the dark resonance of the back vowel /
υ/, which is
associated with it, remains (see the examples of homophonous Paul and paw given above
for Cockney).
One of the stereotypes of BrE for an American is H-dropping, and, indeed, this is
regularly the case not only for much of BrE, but for much of AmE as well as far as the
change from /hw/ to /w/ is concerned. Beyond this, although an <h> is written in such
words as her, him, he, all native speakers drop the /h/ when these words are unstressed
(in the so-called weak forms). The stereotype which the Americans mean is the loss of
/h/ in such stressed words as hat, house, horse. In a great many urban working class
accents of England (but not of Ireland and Scotland) these words are pronounced ’at,
’ouse, ’orse.
The simultaneous pronunciation (co-articulation) of /t/ and the glottal stop [
ʔ] is typical
of many urban accents of Great Britain and of GenAm as well. In BrE, however, it very
often happens that /t/ is completely replaced by [
ʔ]. It is this phenomenon which explains
the humour of this remark made by a Glaswegian: ‘My name’s Pa’erson, with two ts’
(McIntosh 1952: 53). In present day RP a /t/ is frequently realized as a glottal stop before
consonants (except /l/) as in hot day /h
ɐʔde/.
The two dental fricatives /
θ/ and /ð/ are often replaced by other fricatives. When this
happens voicing/force of articulation retains its original distribution. The Cockney accent
realizes the pair as /f/ and /v/ (muvver for mother, nuffink for nothing). New Yorkers often
use /t/ and /d/ (tanks for thanks, dis for this) or the affricates [t
θ] and [dð]; many Blacks
in America have /t/ and /d/ at the beginning of a word, but /f/ and /v/ at the end (dem for
74 ENGLISH AS A LINGUISTIC SYSTEM
them, wiv for with). In Ireland it is common to hear dental [t] or [tθ] and dental [d] or
[d
ð] for /θ/ and /ð/ (the<>, which indicates dental articulation, is called a ‘bridge’).
In addition and independent of this, almost all speakers pronounce words like clothes
or months without /
ð/ or /θ/, namely as /kləυz/ or /kloυz/ and /mns/ when they are
speaking casually.
Almost all accents of English have the pronunciation /ju/ for the spellings <u, ui, ew,
iew/ieu, eu, ue> unless the preceding consonant is dental or alveolar. (An exception is
the traditional accent of East Anglia, in which the /j/ does not occur for these spellings,
e.g. pew /pu/.) When there is a preceding dental or alveolar consonant (/s, z, n, t, d, l,
θ) as in suit, exuberant, new, tune, dew, revolution, thews, most accents of AmE have
/u/, the pronunciation varies between /u/ and /ju/ in RP (see 12.1.3).
The sequence /h/ + /j/ as in pew [p
h
ju], cue [k
h
ju], Hugh /hju/ etc. is realized as [ç]
(the sound of <ch> in German ich). (In pew and cue the [h] is the result of aspiration
following /p/ and /k/ in word initial position.) [ç] is basically a voiceless fricative
allophone of /j/; however, because of the meaningful opposition who–Hugh–you
[hu]–[çu]–[ju] the [ç] has marginal phonemic status.
4.3.5 Phonological processes
Palatalization Wherever historical /j/ has occurred before an unstressed syllable, but
especially in the suffixes -ion and -ure some degree of palatalization of preceding /s, z,
t, d/ has taken place everywhere in the English speaking world, though not always in a
fully predictable way. Such palatalization means that /s, z, t, d/ are pronounced slightly
further back (at the palate rather than the alveolar ridge), as in the following examples:
unpalatalized palatalized
-ion: /s/ (missile)/
ʃ/ (mission)
/z/ (fuse)/
/ (fusion)
/t/ (motive)/
ʃ/ (motion)
-ure: /s/ (fissile)/
ʃ/ (fissure)
/z/ (please)/
/ (pleasure)
/t/ (advent)/t
ʃ/ (adventure)
/d/ (verdant)/d
/ (verdure) (GenAm only)
In many accents the process of palatalization has been uneven, and there is change in
present day RP (see 12.1.3).
Simplification of final consonant clusters Whenever several consonants occur
together at the end of a word, one of them is frequently left out. In the case of the few
words which, according to the spelling and the phonotactics of English, can have a cluster
of four consonants, such simplification is very common, e.g. exempts or twelfths are
simplified from /egzempts/ and /twelf
θs/ to /egzemps/ and /twelfs/. In addition, it is rela-
tively normal, especially in casual speech, to drop final consonants in shorter clusters
when the following word begins with a consonant as well, e.g. west side becomes wes’
side [wesa
d] and left leg becomes lef’ leg /lefleg/.
1111
2
3
4
5
6
7
8
9
1011
1
2
3
4111
5
6
7
8
9
20111
1
2
3
4
5111
6
7
8
9
30111
1
2
3
4
5
6
7
8
9
40111
1
2
3
44
45
46
47111
THE PRONUNCIATION AND SPELLING OF ENGLISH 75
Assimilation A number of the cases of consonant loss or change so far described are
really cases of assimilation. A large part of the allophonic variation in English is due to
this. The loss of the aspiration of fortis stops in word-final position, often even unreleased,
or the voicing and flapping of intervocalic /t/ in AmE and AusE may, for example, be
explained in this way. In the former case assimilation is in the direction of a pause or
silence; in the latter, /t/ adapts to the sonority of the preceding and following vowels.
These allophones are not consciously noticed as is also the case with the following:
alveolar /s, z, t, d, n/ are dental [s
, z, t, d, n] next to /θ/ or /ð/, as in this thing, widths or
right there. In addition, the /t/ may have nasal release (the air is released through the
nose; = [t
n
]) before a nasal as in button and lateral release before /l/ (= [t
l
]) as in bottle.
Or /k, g/ are pronounced with closure further forward against the palate when a front
vowel follows (king, get) than when a back one follows (could, good).
Assimilation is also involved in palatalization. Indeed, whenever the pronunciation of
one sound becomes in some way similar to that of a neighbouring sound, it is possible
to speak of assimilation. Often only one single feature is changed. The following are well
known examples in modern English of assimilation which occurred long ago and have
remained frozen or irreversible:
• a change in voicing and force of articulation: in have to ‘must’ /h
ftə/, the /v/ of
have has become a voiceless, fortis /f/ due to the influence of the following /t/;
• a change in the manner of articulation: the original /d/ of soldier has become /d
/
under the influence of following /j/: (RP) /s
əυldə/ (GenAm) /soυldər/. (This is a
case of palatalization; the place and manner of articulation have changed.)
Other instances of assimilation are dependent on the style of speech. What in careful,
formal style is (RP) /
ðs hə/ or /wɒt du ju wɒnt/ and (GenAm) /ðs hr/ or /wt du ju
w
ɑnt/ become (RP) /ðʃə/ or /wətʃə wɒnt/ and (GenAm) /ðʃr/ or /wədəjə wɑnt/ in
the casual style of colloquial language.
Morphophonemic alternations If assimilation were seen as a purely sound condi-
tioned phenomenon, many cases of alternation in form could not be explained because
they are limited to certain grammatical and lexical classes of words. The form /be
/ bay,
for example, is given the ending /z/ in order to form the plural bays. Since it is not possible
to have a fortis /s/ here, this appears to be a case of assimilation to the preceding vowel.
However, since there is also the word /be
s/ base, in which there has not been a similar
instance of assimilation, it becomes clear that the /z/ of bays is a case of assimilation
restricted to particular circumstances. It involves only the inflectional ending and is there-
fore termed morphophonemic. It can be illustrated by the following examples.
(a) The ending {-S} (for the regular plural and the possessive of nouns and for the regular
third person singular of the present simple form of the verb) as well as the {-D} (for
the past tense, the past participle forms of regular verbs and the derivational morpheme
{-ed} as in blue-eyed or heavy-footed) are realized in differing ways depending on what
phonological environment they occur in. The morpheme {-S} is realized as:
/-
z/ when the word to be inflected ends in a homorganic fricative (/s, z, ʃ, , tʃ, d/),
e.g. mixes /m
ksz/, bushes /bυʃz/;
76 ENGLISH AS A LINGUISTIC SYSTEM
/-z/ when the word to be inflected ends in any other phonologically lenis or voiced
phoneme including a vowel, e.g. boys /b
ɔz/, lugs /lgz/, child’s /tʃaldz/;
/-s/ when it ends in any other phonologically fortis or voiceless phoneme, e.g. bikes
/ba
ks/, raps /rps/, life’s /lafs/.
The morpheme {-D} is realized as:
/-
d/ when the word that is to be inflected ends in a homorganic (alveolar) stop (/t, d/),
e.g. headed /hed
d/, heated /hitd/;
/-d/ when the word to be inflected ends in any other phonologically lenis or voiced
phoneme including a vowel, e.g. allowed /
əlaυd/, rammed /rmd/, saved /sevd/;
/-t/ when it ends in any other phonologically fortis or voiceless phoneme, e.g. licked
/l
kt/, brushed /brʃt/.
Exceptions include such common, but irregular forms as wife-wives or burn-burnt.
(b) Lexical words of Latin origin are the second example of morphophonemic alterna-
tion. This has to do with the numerous words which end in the syllable /-
k/. Such words
as public, historic etc. show a change from /k/ to /s/ when a suffix beginning with <i> or
<e> is added. This alternation is due to assimilational processes in Latin and does not
apply to words of Germanic origin, e.g. RP:
public /p
blk/ publicity /pblst/
publisher /p
blʃə/ (here palatalized)
historic /h
stɒrk/ historicity /hstərst/
but no such alternation in the following words:
stick /st
k/ sticker /stkə/
picnic /p
knk/ picnicker /pknkə/
English spelling, though inconsistent, makes the semantic–etymological relationships
involved in words derived from the Latin clear, inasmuch as the letter <c> can represent
both /k/ and /s/.
4.3.6 Phonotactics
Phonotactics is concerned with how sounds are distributed, that is, where they can occur
in a word (beginning, middle, end), which sounds can occur together, and in what rela-
tive order they can occur. Several points of distribution (involving /
ŋ, h, , ð/) have already
been mentioned. To round out the picture in this area a few examples of regular combi-
nations of sounds will be presented.
It is not possible, for example, to begin an English word with a combination of nasal
and stop (e.g. *mbit /mb
t/ or *dnime /dnam/). <Pn->, <gn-> and <kn-> are, as is well
known, only written: the stops are not pronounced. In the middle or at the end of a word
combinations of nasal plus stop are completely unproblematic in the case of the fortis
1111
2
3
4
5
6
7
8
9
1011
1
2
3
4111
5
6
7
8
9
20111
1
2
3
4
5111
6
7
8
9
30111
1
2
3
4
5
6
7
8
9
40111
1
2
3
44
45
46
47111
THE PRONUNCIATION AND SPELLING OF ENGLISH 77
stops and of /d/, while nasal plus the lenis stops /b/ or /g/ can only appear in the middle
of a word (see Table 4.2).
The phonotactics of English permits consonant clusters in which the semi-vowels
/j, r, w/ and the lateral /l/ can occur after almost all the stops and some of the fricatives
at the beginning of a word, e.g. /pr-, br-, fr-, tr-, dr-,
θr-, ʃr-, gr-, kr-/. Only /s/ can occur
before /p, t, k, m, n, f/ at the beginning of a word. At the end of a word considerably
more combinations are possible. A large number of these consonant clusters are due to
morphological endings like {-D} or {-S} discussed above or the {-th} (= /
θ/) of many
derived nominals (e.g. twelfth, width, depth).
4.4 THE VOWELS
Each of the vowels of English can be distinguished by three features: height of the tongue;
horizontal position of the tongue; and the complexity of the vowel. In this brief presen-
tation of the vowels, 20 in RP and 16 in GenAm, first the distinctive features (4.4.1 and
4.4.2) and then the status of non-prevocalic <r> (4.4.3) will be examined.
The nasality of a vowel (as in French, for example) is a further phenomenon; however,
it is peripheral since its presence varies individually and regionally. Although it is typical
of many varieties of AmE, it does not appear to be phonemic in any of them. Further-
more, lip-rounding, though not distinctive, is characteristic inasmuch as the front vowels
are spoken with spread lips; the central ones with neutral lips and the back ones with
rounded lips.
4.4.1 Position of the tongue
The sound quality of each vowel of English is determined by the horizontal position of
the highest point of tongue, which can be in the front, centre or back of the mouth (the
oral cavity). Three vertical levels are recognized: high, mid and low. Theoretically, a
combination of these dimensions should provide for nine possibilities. However, among
78 ENGLISH AS A LINGUISTIC SYSTEM
Table 4.2 Final and medial nasal-stop clusters
labial camp /-mp/ bomb /-m/
camping /-mp-/ bombing: /-m-/, but bombard /-mb-/
dental Lent /-nt/ land /-nd/
Lenten /-nt-/ landing /-nd-/
palatal think /-
ŋk/ long /-ŋ/
thinker /-ŋk-/ longing /-ŋ-/; singer: /-ŋ-/ (but longer /-ŋg-/)
Notes:
Before the inflectional ending {-ing} the letters <mb> and <ng> are pronounced without the stops /b/ and /g/.
The <g> of <ng> is, however, pronounced when followed by the comparative and superlative endings {-er} and
{-est} as in stronger, younger, or longest. The <g> is not pronounced when the agent ending {-er} is added to
a verb as in singer, wringer, banger etc.
In bombard /m/ and /b/ occur in two different syllables; cf. also iambic; /-mb-/ does not occur within a single
syllable.
the short vowels of English only six of these are realized. A further short vowel is /ə/
(often called schwa); however, it is essentially different since it is always unstressed. (See
below Figure 4.3.)
The system of the short vowels of English is of significance because it is the short
vowels which have remained relatively stable over several centuries. With notable and
clearly defined exceptions the short vowels are phonotactically limited to occurrence in
checked syllables, that is in syllables which end in a consonant. This means that words
cannot end in short vowels without a following consonant. Unstressed schwa /
ə/ is not
restricted in this manner. As a result there are no words of the form /be/ or /s
/.
The notable exception to this phonotactic regularity is the use in RP (and the accents
of northern England) of the vowel /
/ as the realization of final <-y> or <-ie> in words
such as lazy or Suzie. Most other accents (for example, in southern England, North
America, Australia etc.) have long /i/ here, and increasingly often RP does as well.
The vowel /
ə/ is the form which many vowels may be regarded as ‘taking’ when they
occur in unstressed syllables in the natural flow of speech. Schwa is never stressed.
Because of the high incidence of unstressed syllables in English it is easily the most
frequent vowel in the language.
4.4.2 Complexity: length and diphthongization
Every deviation from the short nature of a vowel will be regarded here as a case of
complexity. Length is one such deviation. Length is indicated by <>. Since, in addition,
the long vowels have a distinct tendency in many varieties of English to be at least some-
what diphthongized, there is some justification for grouping length and diphthongization
together. In addition, the short vowels are produced without special muscle tensing of the
tongue (i.e. they are lax) while both the long vowels and the diphthongs are tense. The
degree of diphthongization varies considerably, but it is usual to speak of three long
closing diphthongs and two slightly diphthongized closing ones. Closing refers to the
closing movement of the mouth during the articulation of these diphthongs. The arrows
1111
2
3
4
5
6
7
8
9
1011
1
2
3
4111
5
6
7
8
9
20111
1
2
3
4
5111
6
7
8
9
30111
1
2
3
4
5
6
7
8
9
40111
1
2
3
44
45
46
47111
THE PRONUNCIATION AND SPELLING OF ENGLISH 79
υ
e
ə
1
ɒ
2
1
// is called ‘ash’
2
/ɒ/ is not present in GenAm
Figure 4.3 The short vowels
i u
ɔ
ɑ
Figure 4.4 The long vowels
in Figure 4.5 show the direction of the movement from the first to the second element of
the diphthongs.
The two slightly closing diphthongs /e
/ and /əυ/ (RP) or /oυ/ (GenAm) are realized
in many varieties (e.g. Standard Scottish English) or in certain phonetic environments
(especially before fortis stops) in GenAm as long monophthongs, e.g. gate [get] or goat
[got]. Three further diphthongs, /a
, aυ, ɔ/ are pronounced as diphthongs in almost all
accents. The centring diphthongs (see Figure 4.6) are present in RP, but not in GenAm
(see below 4.4.3).
In addition to the phonological differences in length, which are of central importance
here, there are also non-phonological length differences. All vowels, whether short or
complex, are relatively shorter when followed by a fortis consonant and relatively longer
when followed by a lenis one or, for those where this is possible, when no consonant
follows (in free or unchecked syllables). For this reason the /e
/ of late is shorter than the
/e
/ of laid or the even longer /e/ of lay; the // of back is shorter than that of bag. This
length difference may sometimes be used to distinguish writer from rider in GenAm,
where both words have a voiced and flapped /d/ (= [
ɾ]) in the middle. Since the /d/ of
writer was historically a /t/, this means that for some (but not all) speakers the /a
/ of
rider may be longer than the /a
/ of writer. Otherwise, the two words are indistinguish-
able. The same distinction may differentiate other pairs such as latter and ladder. It is
not, however, clear whether such differences in length are perceived in normal speech.
4.4.3 Non-prevocalic /r/
Up to this point a system of 17 (RP) or 16 (GenAm) vowels has been assumed. This
system and the comparability of these two standard accents becomes considerably more
complicated when non-prevocalic r (traditionally termed ‘postvocalic’ r) is included. Non-
prevocalic r refers to an orthographic <r> which is not followed immediately by a vowel
(either within the same word or linked to an initial vowel in the following word); instead,
it is followed by a consonant (hard) or comes at word end (here).
In the rhotic accents orthographic <r> is regularly pronounced in all environments. In
many such accents the vowel system is noticeably simplified in the sequence vowel + /r/.
(Note: this does not apply to all rhotic accents, for example, it does not apply to Scottish
80 ENGLISH AS A LINGUISTIC SYSTEM
e əυ
1
oυ
2
ɔ
aaυ
1
RP
2
GenAm
Figure 4.5 The closing diphthongs
ə υə
εə
Figure 4.6 The centring diphthongs
English.) In GenAm the oppositions that otherwise exist between /i/ and //, between /e/,
/e/ and /
/, between /ɔ/ and /oυ/ and between /u/ and /υ/ when /r/ follows are all neutral-
ized. This leads to a system of just ten vowels (see Table 4.3).
Furthermore, in GenAm /r/ usually has the features of a vowel (it is a semi-vowel).
(An actual consonant, such as rolled [r], occurs in few accents of English.) Combinations
of vowel + /r/ in GenAm are often really phonetic diphthongs whose final element is
an r-coloured schwa. This means the schwa has the sound quality of a retroflex [
! ],
produced with the tip of the tongue curled back towards the rear of the mouth, or a
constricted [
ɹ], articulated with lateral tension of the tongue. It is sometimes written with
the symbol [
]. This is the case, for example, with fear /fr/ = [fi] or cure /kjυr/ =
[kj
υ]. The central vowel // is r-coloured without the need for a schwa [], e.g. purr
/p
r/ = [p]. This can be the case with /ɑ/ + /r/ as well, which may show up as in car
/k
ɑr/ = [kɑ
ɹ
].
Table 4.3 shows the full GenAm and RP systems of vowels on the left in comparison
to the restricted system of vowels that occurs before non-prevocalic r. Note that the
neutralized vowels may be transcribed with large capital letters, to indicate their status
as archiphonemes. In the left-hand column there are only fifteen phonemes because
/
/ appears exclusively before an /r/ in GenAm. In the non-rhotic accents, in which
1111
2
3
4
5
6
7
8
9
1011
1
2
3
4111
5
6
7
8
9
20111
1
2
3
4
5111
6
7
8
9
30111
1
2
3
4
5
6
7
8
9
40111
1
2
3
44
45
46
47111
THE PRONUNCIATION AND SPELLING OF ENGLISH 81
Table 4.3 The vowels of General American and RP
Full system Before <r> [in RP + consonant or zero]
GenAm RP key word GenAm RP key word
/i//i/ bead /
r/ (or /Ir/) /ə/ beard
/
/// bid
/e
//e/ bade
/e/ /e/ bed /
εr/ (or /Er/) /eə/ (or /εə/) bared
/
/// bad
/
ɑ//ɑ/ GenAm: bod(y) /ɑr/ /ɑ/ barred
RP: bar
–/
ɒ/ bod(y)
/
ɔ//ɔ/ bawd /ɔr/ /ɔ/ bored
/o
υ//əυ/ bode
/
υ//υ/ Budd(ha), book
/u//u/ booed /
υr/ (or /Ur/) /υə/ (also: /ɔ/ boor
sometimes: /
ɔ/ your
/
/// bud //// bird
/
ə//ə/ baba /ər/ /ə/ barbered
/a
//a/ bide /ar/, /ajər/
1
/aə/, /aə/, /a/
2
buyer
/a
υ//aυ/ bowed /aυr/, /aυwər/
1
/aυə/, /aə/, /a/
2
bower
/ɔ//ɔ/ Boyd /ɔr/, /ɔjər/
1
/ɔə/, /ɔə/
2
Boyer
Notes:
1 The triphthongs /a
r, aυr, ɔr/ (= [a, aυ, ɔ]) are not stable and are therefore often pronounced as two
syllables. This may result in an epenthetic [j] or [w] (= conditioned by movements necessary in articulation).
Bronstein also counts [e
] as in mayor and [oυ] as in blower as triphthongs (1960: 201).
2 There is a strong tendency towards simplification of the triphthongs in RP, called smoothing. Instead of the
possible further triphthongs /e
ə/ or /əυə/ one finds smoothed diphthongs as in the words player /eə/ or mower
/
ə/ or the monophthongs /ε/ or // respectively.