Contents: [aspiration] [cluster simplification] [Palatalization] [tapping] [velarized /l/] [vowel length]

Aspiration

English voiceless stops /p, t, k/ have allophones such as aspirated stops [p, t, k] and unaspirated stops [p, t, k]. They are aspirated in the following positions (Kreidler, 1989:116):
(a) stressed syllable-initial position (for example, in 'apply' [pla] vs. 'application' [plken])
(b) unstressed word-initial position (for example, in 'parade' [pred])
(c) word-final position after /s/ (for example, in 'ask, test' [sk, test])
(d) (free variation) word-final position except after /s/ (for example, 'act, book' [kt or kt, bk or bk])
[Exmple of waveform, spectrogram or actual sound]

Often cluster simplification, through omission of a consonant in the middle position, may occur in words and phrases like these: 'as(k)s, ris(k)ed, six(th)s, Ar(c)tic, san(d)wich, frien(d)ly, Secon(d) Street, the firs(t) day, etc.' Especially, /t/ and /d/ are frequent in clusters and thus frequently subject to deletion. Kreidler (1989:256) states, "Consider these sets of phrases, all with a word-final /t/, which may be deleted.
(a) next time, the first day, last December
(b) last May, most people, the best night
(c) next winter, the best one, last week
(d) next afternoon, last October, the best idea
(e) the most, the best, the last
The five groups are meant to represent, roughly, a range of probability for the deletion of the consonant /t/. The /t/ is most likely to be dropped before a following /d/ or /t/, as in group (a); somewhat less likely before other consonantal segments - group (b); still less likely before the glide /w/ - group (c); rather unlikely before a folowing vowel - group (d); and quite unlikely in final position - group (e)."

[Exmple of waveform, spectrogram or actual sound]

When an alveolar obstruent, /t, d, s, z/, occurs at the end of a word and the next word has an initial /j/, the alveolar obstruents may change to the corresponding palatal obstruents, becoming, respectively, /, , , /. Often the /j/ is deleted when the change occurs. The examples are: won't you [wount ju] or [wounu], did you [dd ju] or [du], unless you [nls ju] or [nlu], as you [z ju] or [u]. Palatalization is an optional process across word-boundaries, as in the above examples.

[Exmple of waveform, spectrogram or actual sound]

In American English, 'fatty' and 'daddy' tend to be pronounced as [fi] and [di]. That is, a single alveolar stop becomes a voiced tap of the tongue-tip, [], between two vowels, of which the second vowel is unstressed. That is, this phenomenon does not occur before a stressed vowel (for example, in 'attack') or after another consonant (for example, in 'tasty, optic'). However, the tapped consonant may be preceded by /r/ (for example, in 'party, according' [pa, k]).
Ladefoged (1993:65) adds, "This does not apply to /t/ before syllabic [n] as in 'mutton' [m ] because the /t/ there has become a glottal stop. In addition, there is a great deal of variation among speakers. Some people have [] after lax vowels in words such as 'little, better,' but not after tense vowels as in 'writer, later.' Some have [] in 'motto' but not in 'veto.' "

[Exmple of waveform, spectrogram or actual sound]

English /l/ sound has two allophones such as so-called clear l [l], as in 'leave, healing' [liv, hil], and dark l or velarized l [], as in 'shield, heal' [id, hi].
Kreidler (1989:101) states, "A clear l is produced with the front of the tongue high in the mouth and the back of the tongue low. A dark l is made with the back of the tongue raised; the center is low; the front may be raised, so that the whole tongue has more or less the shape of a spoon, or the tongue-front may be down. However, the variation depends mostly on what position /l/ has in a syllable and only partly on what kind of phonemes follow."
Ladefoged (1993:65) states, "In most forms of American English, all examples of /l/ are comparatively velarized, except, perhaps, those that are syllable initial and between high front vowels, as in 'freely.' In British English /l/ is usually not velarized when it is before a vowel, as in 'lamb, swelling,' but it is velarized when word final or before a consonant, as in 'ball, filled.' . . . In my own speech, the whole body of the tongue is drawn up and back in the mouth so that the tip of the tongue no longer makes contact with the alveolar ridge. Strictly speaking, therefore, this sound is not an alveolar consonant but more like some kind of back vowel in the speech of some English speakers."

[Exmple of waveform, spectrogram or actual sound]

The same vowels tend to be produced longer in the following cases (Ladefoged, 1993:90. 95):
(a) in open syllables (for example, compare 'sea' vs. 'seat'),
(b) in stressed syllables (for example, compare 'inner' vs. 'innate'), and
(c) before voiced sounds in the same syllable (for example, compare 'pig' vs. 'pick')
The current English Pronunciation Simulator system deals with only the third case, and expresses a longer vowel using '' (for example, in 'pig' [pg] vs. 'pick' [pk].