Académique Documents
Professionnel Documents
Culture Documents
On Pharynx-Larynx Interactions
Author(s): Loren Trigo
Source: Phonology, Vol. 8, No. 1 (1991), pp. 113-136
Published by: Cambridge University Press
Stable URL: http://www.jstor.org/stable/4420026
Accessed: 18-10-2015 16:54 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to Phonology.
http://www.jstor.org
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
On
pharynx-larynx
interactions*
Loren Trigo
Boston University
The role which the pharynx plays in the production of speech sounds has
long been the subject of debate, in particular as regards the independence
of the pharynx from other speech articulators and in particular the larynx.
Given that the linguistic and anatomical information about the larynx and
the pharynx is quite complex, the objective of this paper is to review the
literature relevant to the relationship between these two articulators and to
recast some of this information in terms of current articulator-based
theories of phonological features such as the ones reviewed in McCarthy
(1988). ? 1 presents two pharyngeal features [ATR/RTR]
'advanced
tongue root' and [LL/RL] 'lowered larynx' separately in two subsections,
with definitions, examples and arguments in favour of having two
pharyngeal features rather than one. ? 2 argues that pharyngeal features are
independent of the primary point of articulation features. ?3 discusses
how the pharynx may become involved in the articulation of consonants
that are not obviously pharyngealised.
1 Pharyngeal
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
by
by
by
by
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
[RL/LL] and
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
118
Loren Trigo
(2)
implosive
preglottalised
ejective
But the elevation of the larynx is not predictable from its internal state.
A prediction made by the system in (2) is that ejectives should all be
voiceless and implosives should all be voiced. This prediction is
contradicted by a few languages, including Igbo (Maddieson 1984), which
has two labial implosives that differ only in that one of them is voiced and
the other voiceless. A second prediction made by the system in (2) is that
ejectives should always elevate the pitch of a following vowel whereas
implosives should have no pitch effect. But Kingston (1985: 163-176) has
shown that the degree of vocal cord tension during the production of
ejectives and implosives is not universally predictable: in Tigrinya,
ejectives elevate the pitch of a following vowel (hence are pronounced with
stiff vocal cords) whereas in Quiche they lower it (hence are pronounced
with slack vocal cords); similarly, differences in periodicity and spectral
shape indicate that Niger-Congo implosives are pronounced with stiff
vocal cords whereas Hausa implosives are pronounced with slack vocal
cords. The existence of such contrasts suggests that the vertical displacement of the larynx is not a mechanical consequence of the intrinsic
laryngeal features.
If [RL/LL] is a feature phonologically independent of the laryngeal
features, we should find languages which exploit this independence in
making phonological distinctions. Evidence supporting this claim is weak.
In the Bushman language !X66 pharyngealised vowels can be breathy or
glottalised and similarly the non-pharyngealised vowels. The pharyngealised vowels are [RL], at least phonetically. According to Traill (1985),
vowels in !Xo6 can be (1) modal, (2) glottalised, (3) breathy, (4)
pharyngealised,
(5) breathy-pharyngealised
and (6) glottalisedpharyngealised:
(3)
RL
constricted glottis
spread glottis
2
h
v
?
v
+
-
4 5
S hs
v v
+ +
+
6
?s
v
+
+
-
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
On pharynx-larynx interactions 1 19
production of 'strident' vowels the larynx is elevated and pressed against
the epiglottis. The 'stridency' is due to an aryepiglottic constriction
through which the air flows causing vigorous vibration of the arytenoid
cartilages and epiglottis:
With one qualification, the articulatory posture ... is the same as that
involved in closing the laryngeal sphincter for non-linguistic purposes
such as coughing. The well-known purse-string effect of the contraction
of the aryepiglottic muscle and its extension, the transverse arytenoid
muscle, pulls the tips of the arytenoid together and upwards to meet the
cushion of the epiglottis, and adducts the false vocal folds (1985: 78).
The weakness of the evidence lies in the fact that Traill's X-ray tracings
of the 'strident' vowels also show a retraction of the root of the tongue;"
hence, it is not altogether clear that [RL] is the distinctive register feature
(as opposed to [RTR]). In order to make a stronger case for the
independence of the feature [RL/LL] from laryngeal features one would
have to find a language that distinguished modal and breathy vowels in
two registers based exclusively on [RL/LL]. However, as already
mentioned, there are relatively few X-ray tracings of register contrasts,
most descriptions being impressionistic. To establish the matter conclusively more data will be needed.
One might object that if [RL/LL] and the laryngeal features are
independent, as proposed here, then the fact that [RL/LL] register
distinctions tend to cooccur with changes in phonation becomes mysterious, but an explanation of this tendency is readily available. Despite
their functional independence, there is a mechanical link between pharynx
and larynx. It appears that the pharyngeal musculature can mechanically
affect the size of the glottis. To neutralise this effect the intrinsic laryngeal
musculature must be activated. Experiments conducted by Shin et al.
(1981)12 show that it is possible to open/close the glottis after removal of
the intrinsic laryngeal musculature by activating the part of the pharyngeal
musculature which also helps depress/elevate the larynx.'3 The laryngeal
constriction in these experiments is apparently not at the level of the vocal
folds but at the level of the aryepiglottic folds which are involved in
gestures of protective closure. The mechanical link between pharynx and
larynx explains why [LL] vowels tend to be breathy voiced while [RL]
vowels tend to be pressed voiced.
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
i-ce
i-kam
i-j k
'to carve'
'to catch'
'to be good'
i-syEn
i-wap
i-wok
'pity'
'to follow'
'to carry'
Why do glides have this effect? Glides are pronounced with a higher
tongue body position than are any of the vowels. Since, as already
mentioned above, the feature [ATR] seems to be intimately related to
tongue height, it appears that the feature which the glides are spreading is
the feature [ATR], which glides possess redundantly by virtue of their
articulatory configuration. Returning to the question of the 'tense'
vowels, it appears that the 'tense' vowels arise when a glide spreads its
[ATR] feature onto a following 'head' register (i.e. [RL]) vowel. Examples
of such spreading occur incidentally in the grammar and specific lexical
items. The target vowel is always a mid vowel:
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
On pharynx-larynx interactions
(5)
ATR
121
ATR
iry
iry
'tobe black'
RL
RL
a-kwap
-wa
-dyal
-yat)
a-kwop
wo
dyel
yet
'land'
'stand'
'persuade'
'skin'
Other 'mixed' register vowels arise in the build-up of words and I will not
consider them here. What is important is that 'tense' vowels are 'mixed'
register vowels, which are impossible to describe unless [ATR/RTR] and
[LL/RL] are distinct features."6
In conclusion,
I have distinguished
two pharyngeal features
[ATR/RTR] 'advanced tongue root' and [LL/RL] 'lowered larynx' in
addition to the laryngeal phonatory features [constricted glottis] and
[spread glottis]. The basis of the distinction between the feature [LL/RL]
and the laryngeal features is largely phonetic. The evidence for their
phonological independence is not conclusive. The cooccurrence of pharyngeal and laryngeal effects has been attributed to a mechanical link
between pharynx and larynx. The distinction between [ATR/RTR] and
[LL/RL] has been shown to be useful in the description of 'mixed'
register vowels in Turkana.
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
122
Loren Trigo
Dorso-pharyngeal uvulars
Dorsal uvulars
[-high]
[RTR]
[+back]
[RL]
Dorsum
Dorsum
Pharynx
1 place
1 place
2 place
[q]=[ki]
[q]
period) :17
(8)
rji-kayo
lu-ka-bob-o-k
e-kaalees
ti-ka-lbk-a-k
a-kamu
rji-kodyo
e-kod
lo-kori
a-bokok
OI-k3rI
e-k3rI
na-bakobok
a-kooki-aan-ut
a-kiru
a-makuk
ri-keno
i)a-kima-k
a-rukum
a-kEpu
e-risik
na-kituk
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
PL
'chest' register
xhx 'to leave vacant'
xh* 'classifier for doors'
Xh3lb
'trough'
ZhumE 'mouth'
'head' register
'to break'
xxS 'to dig with finger'
3i
baXoS
x?*
uvulars
P1H0 8
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
I
r
h
Standard
?asap
kilat
masa?
balas
negatef
?alem
sabon
dukoxj
batal
jujo:
yumah
Kelantan
?asa?
kila?
mast)?
balah
negatih
alir
saboxj
dukoij
bata:
jujo:
yumih
Terengganu
?asa?
kila?
mas)?
balah
Kedah
?asap
kilat
masa?
balaih
?alerj
saboij
dukoU
bata:
jujo
yum5h
?alem
sabon
dukot
betai
jujoS
HumTh
'smoke'
'lightning'
'cook'
'finish'
' negative'
' pious'
' soap '
' carry'
' cancel'
' sincere'
'house'
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
Proto-Nootkan
q?
*NqW?
*yw
Makah
Nitinat
Nootka
q?
q?pa:k
Sapa:k
Sapa:k
qW?
qW?ica:k
Sica:k
Sica:k
'rotten'
zaci:
xaci:
haci:
'deep down'
Xw
Xw
ciXwa:
ciXwa:
ciha:
'willing'
'ghost'
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
place
supralaryngeal
laryngopharyngeal
[q]
oral
supralaryngeal
laryngopharyngeal
[SJ
depletion
secondary
place
[q]
secondary
place
oral
depletion
[i
There are two reasons why we want to say that uvulars in Proto-Nootka
and Kedah involve a secondary pharyngeal place component. First,
Clements (1989) has argued that each secondary (vocalic) place component
corresponds to an articulatorily analogous primary (consonantal) place
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
a i s
a i s
1 2 1
1
a
V
s
decomposition
i a k
1 2 1
i
1 2 1
fiV
a i s
a i
-
oral
depletion
i a k
-
decomposition
121
I\I
i a 1%
12 0
kv
a i h
ia
12
oral
depletion
I \
i a i
In this section I shall discuss how the pharynx may become involved in the
production of consonants that are not obviously pharyngealised: voiced
obstruents and laryngeals.
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
+ high
' chest'
+ back
back
i
u
i
- back
'head'
+ back
+low
The 'chest' register vowels occur after the voiced and the 'heavy'
aspirated obstruent stops /b d dc g pfid tfi
t
k7/. The 'head' register
vowels occur word initially and after the remaining consonants i.e.
/p t t 4 k m n p re s I r (w) y ?/. Suffixal vowel alternations conditioned by
the preceding root consonant determine the following vowel correspondences:
(19)
a.
kapfiuct-ct
sznar-ra
b. xj--kiuy-i
rjaijkRi?-i
1J-Yp3y-6
C. J`Y-lhu
la3
a-
'to use'
'to embroider'
'to provoke'
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
Harmony (multi-targeted)
[ATR]
[-nas] [-nas1
+sonJ
Evidence that a pharyngeal register feature is being spread can be
gathered from the typology of blockers of Harmony. These are either
voiceless obstruents (T, S) or nasals (N):
(21)
transparent (L, G)
opaque (T, S, N)
' wash'
tfRil
' tongue'
abysso
byrxs
buwx
diyy
'health'
'fruit'
'here'
'language'
' stone)
bxta
k%iman 'weapon'
p%RYsa
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
3.2 Laryngeals
Another type of consonant which may involve the pharynx (in particular
the feature [LL/RL]), even though pharyngealisation may not be evident,
is the group which includes laryngeal consonants. Lindau (1975: 134)
reports that the lower pharynx can be active in the production of some [h]s
and [?]s:
The width of the pharynx is smaller for [h] than for surrounding
vowels ... but the vertical position of the larynx seems to be the same as
that of the vowels ... in the production of at least one type of glottal stop
the pharynx is considerably constricted and the larynx higher (Lindqvist
1969).
According to Lindqvist (1969, 1972) there are two different kinds of
laryngeal constriction: one at the level of the vocal folds and one higher
up at the level of the aryepiglottic folds. The latter constriction can be seen
in Lindqvist's pictures of a phonetician's glottal stop taken through a
fibreoptic laryngoscope. Catford (1983) reports that in a number of
Caucasian languages, pharyngeal approximants [h Si] and laryngeals [h ?]
contrast phonemically with the aryepiglottal approximants [H i] and stop
[?]; the 'strident' vowels described by Traill (see above) involve a similar
aryepiglottal constriction. The aryepiglottic folds extend from the sides of
the epiglottis to the apexes of the arytenoid cartilages. The role of aryepiglottic closure is basically protective (as in swallowing) and is
effected by the aryepiglottic muscles within the folds, the oblique
arytenoid muscles and the thyroarytenoid muscles. I have associated
aryepiglottic closure with the feature [RL/LL] (see above).
Given the pharyngeal involvement in their production, aryepiglottal
laryngeals will be expected to pattern phonologically with pharyngeals as
do laryngeals in some languages, e.g. Semitic (cf. McCarthy 1989).
However, the claim that the laryngeals which pattern after pharyngeals are
aryepiglottal has not been substantiated.
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
132
Loren Trigo
4 Conclusion
I have argued for the existence of two pharyngeal features [ATR/RTR]
'advanced tongue root' and [LL/RL] 'lowered larynx' independent of
the laryngeal features. The distinction between the feature [LL/RL] and
the laryngeal features is drawn by establishing their phonetic independence. Evidence for their phonological independence is not conclusive.
The cooccurrence of pharyngeal and laryngeal effects is attributed to a
mechanical link between pharynx and larynx. The distinction between
[ATR/RTR] and [LL/RL] proves useful in the description of 'mixed'
register vowels in Turkana. Pharyngeal features are shown to be independent of the so-called primary place features. It is argued that they
classify as secondary place features. Finally, it is shown that the pharynx
(in particular the feature [LL/RL]) may become involved in the articulation of voiced obstruents and laryngeals.
NOTES
*
[1]
[2]
[3]
[4]
[5]
[6]
[7]
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
On pharynx-larynx interactions
133
[14] 'In languages like Turkana tongue root advancing seems to produce a wide
pharynx and, possibly as an articulatory correlate, breathy vowels. [-ATR]
vowels in Turkana sound tense, or harsh' (Dimmendaal 1983: 27).
[15] 'Vowels with the feature [- ATR] have a hard voice phonetically. The [+ ATR]
vowels normally sound somewhat breathy, but in the environment of specific
[-ATR] vowels, the [+ATR] vowels with the feature [-high, -low] do not
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
134
Loren Trigo
have this concomitant feature. Instead they are realized as tense vowels' (1983:
18). 'The tense ([+ATR]) mid vowels in Turkana seem to be due to an anticipation of the state of the vocal apparatus for the production of [-ATR]
vowels, while a [+ATR] vowel is being produced' (1983: 29).
[16] Other types of 'mixed' register vowels have been noted to exist, but it is not
clear how these other types should be analysed. See Tucker (1975).
[17] Additional evidence that uvulars are [-high] in at least some languages can be
gathered from Greenlandic Eskimo (Kenstowicz & Kisseberth 1979: 250). The
uvularisation data presented in the text come from Dimmendaal (1983) and
Noske (1987, field notes).
[18] Since in Turkana the vowels which cause uvularisation must be [+back] vowels
[o z a], this suggests (but does not prove) that uvulars in Turkana are also
[ + back].
[19] All the oral depletion processes of Malay should be stated in terms of the end
of the syllable rather than the end of the word because they also apply wordinternally (see Teoh 1988). Compensatory lengthening applies in Malay whenever there is an empty timing slot in coda position. These empty timing slots
can be created in two ways. One is by the oral depletion of liquids. The other
is by the formation of prenasalised nasal and oral stops: /tombarj/ [to:mbarj]'to
fall', /gurindam/ [guri:'dam] 'a type of poetry', /barjga/ [ba:0ga] 'to be
proud', /bantuk/ [ba:nto?] 'form', /bapci/ [ba:fci] 'census', /m3lj-bororj/
[m3:m-bororj]'wholesaler', /maTj-daki/ [m5:0-daki] 'to climb' (ACTIVE), /paggali/ [pa:)-gali] 'digger', /mat-catu/ [m5:0-catu] 'to ration' (ACTIVE), /ma1jpapi/ [m3:-payii]'to sing' (ACTIVE) (prenasalised nasals simplify to simple nasals).
Glides do not prenasalise: /m3tr-warji/[m5ij-waiji] 'to cause to finance', /m5rjya-kan/ [m5rj-yak-an] 'to cause to agree for'.
[20] It should be noted that oral depletion in Malay is a family of rules, not a single
rule. For discussion see Teoh (1988).
[21] The hypothesis that pharyngealisation is incompatible with labialisation is based
on the assumption that the two modifications induce the same acoustic effect,
and hence cannot be distinguished from one another. Kingston (1987) shows
this assumption to be factually incorrect: labialisation flattens the spectrum
whereas pharyngealisation makes it compact.
[22] In fact the surface pronunciation has a nasalised [a]: [kapiaS], [pasiiS]. The
nasalisation is probably due to a well-known acoustic effect of pharyngeals.
[23] Vowels in the Mon-Khmer languages of Vietnam generally vary far more in
relation to final consonants than to initial. Rhyming lists have been developed
on the basis of which subsystems of vowels in relation to the various final
consonants have been determined. The influence of final consonants on vowels
has been widely noted in some of the other language groups of the Far East as
well. For bibliographic information on this subject see Phillips (1973).
REFERENCES
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
On pharynx-larynx interactions
135
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions
136
Loren Trigo
This content downloaded from 200.144.93.190 on Sun, 18 Oct 2015 16:54:47 UTC
All use subject to JSTOR Terms and Conditions