Timing
and sequencing in consonants and vowels.
Conventional alphabetic transcriptions such as

l
If

nt
tend to coerce us into imagining that the individual consonants and
vowels are put together in a simple sequence in time, like this:
 |
l
|
I |
f
|
 |
n
|
t
|
time --->
But this is probably not right. The phenomena of Consonant-Vowel
coarticulation and Vowel-Vowel coarticulation suggest that something
else is going on.
Vowel-to-Vowel coarticulation
For example: the central vowels in "to the" are often slightly
different in e.g. "go to the park" vs. "go to the pictures":
The vowels in "to the" are a bit higher before a close vowel than
before an open vowel. It is as if the stressed vowels exert an
influence over the preceding unstressed vowels. This influence has no
trouble going past or through any intervening consonants:
|
 |
<---
|
 |
<---
|
 : |
|
vs.
|
|
Ï |
<---
|
Ï |
<---
|
I |
|
t |
|
ð |
|
p |
|
k
|
|
t |
|
ð |
|
p |
|
k
|
The Swedish phonetician Sven Öhman demonstrated in the mid 1960's
that in nonsense Vowel1 + Consonant + Vowel2
sequences spoken by Swedish, English and Russian speakers, Vowel1varies
according to the identity of Vowel2. In some
cases where
Vowel1was unrounded and Vowel2
was rounded, he
observed that the lips sometimes began to move into the rounded
position ready for the rounded articulation of the Vowel2
during the articulation of Vowel1. In
consequence, the
intervening consonant was also lip-rounded (as you would expect for a
consonant before a rounded vowel, but here we begin to see an
explanation for that fact.)This suggests a timing pattern like this
(for e.g. [igu]):
i
|
rounding
<----------
|
u
|
|
g
|
|
Öhman (1966) wrote:
Since traces of the final vowel are
observable already in the transition from the initial vowel to the
consonant, it must be concluded that a motion toward the final vowel
starts not much later than, or perhaps even simultaneously with, the
onset of the stop-consonant gesture. A VCV utterance of the kind
studied here can, accordingly, not
be regarded as a linear sequence of three successive gestures. We have
clear evidence that the stop-consonant gestures are actually
superimposed on a context-dependent vowel substrate that is present
during all of the consonantal gesture.
This model is often known as the co-production
model of consonant-vowel timing, since it holds that consonants and
vowels are produced at the same time by two separate, parallel,
interacting processes of articulation.
Stages in the formation of stops.
Looking more closely at the fine temporal structure of (say) an
intervocalic voiceless aspirated stop (e.g. the medial [t] of city), various phoneticians have
identified a number of sequential stages in their production (see e.g.
Abercrombie 1967 chapter 9; Ashby and Maidment 2005 pp. 56, 93-5, 126).
Degree of aperture in the oral passage:
>-------< «
1 2
3 4
In (1), the closing stage, the active articulator moves away from its
position during the vowel towards its position for the stop. In (2),
the holding (closure) stage, the active articulator is held against the
passive articulator. In (3), the release stage, the active articulator
moves away from the passive articulator and the stricture of complete
closure is opened up. In (4), the plosion stage, the turbulent airflow
of an aspirated stop occurs.
Not all stops in all contexts have all these phases or stages (e.g.
unaspirated stops, unreleased stops etc.). In "incomplete" plosives,
such as the first stops in clusters such as "apt, apter, act, actor"
there are two overlapping closing and holding phases, but only the
second stop is released:
a >---k---<
1 2 3
>---t------< h« 
1
2 3 4
Aspiration (or its absence) is not the only type of stop release. We
also encounter, in various languages:
- nasal release, transcribed e.g. tn, pm, kŋ.
For example, Twi [etnẽtnẽŋ] 'many (things)
flowed'.
- lateral release, transcribed e.g. tl, c
. For
example: Navaho [nant
'ah] 'it is difficult'.
- gradual or delayed release, i.e. affricates, e.g. pf, ts, t
, kx. Considering affricates to be stops with a
gradual, "strident" release (Jakobson called them "strident stops"), we
can understand why the two parts of an affricate must be homorganic (=
having the same place of articulation), whereas other stop+fricative
clusters do not have to (e.g. [ps]).
References
Abercrombie, D. (1967) Elements of
General Phonetics. Edinburgh University Press.
Öhman, S. E. G. (1966) Coarticulation in VCV Utterances:
Spectrographic Measurements. Journal
of the Acoustical Society of America 39, 151-68.