CHAPTER III

CONSONANTS

Basic Competencies

After learning this chapter, the students are required to achieve the
following competencies. They are able to:
• distinguish between obstruent and sonorant sounds
• explain the classification of consonants based of manner of
articulation
• explain the classification of consonants based on place of
articulation
• explain the consonants not available in English
• explain the syllabic consonants

We have discussed in Chapter 2 that the class of consonants can be divided into a
number of sub-groupings on the basis of their place of articulation and their
manner of articulation. On the basis of their manner of articulation, they can also be
classified into several divisions; one of these divisions to be considered here is the
classifications of consonants into obstruent and sonorant. For obstruent, the airflow
is noticeably restricted, with the articulators either in complete closure or close
approximation. For sonorant, either there is no such restriction in the oral tract or
the nasal tract is open; either way, the air has free passage through the vocal tract.
The class obstruents can be further subdivided into, fricatives, stops, and
affricates. The class of sonorant consonants can be subdivided into nasals, liquids,
and glides.
This chapter discusses in details English consonants, together with their
varieties, and some other consonants available in the world languages. The basis of
the discussion is manner of articulation.

A. Fricatives
Fricative sounds are produced when the active articulator is close to, but not
actually in contact with, the passive articulator. This position means that as the air
exits, it is forced through a narrow passage between articulators, resulting in
considerable friction. Fricative sounds can be voiced or voiceless.
Place your lower lip loosely on the lower edge of your upper teeth and force
the air out of the mouth, after having filled your lungs aplenty. What will happen?
The obstacle you have created by the loose contact of the lower lip with the upper
teeth does not totally block off the air, and therefore the air will continue to flow
out. Because the teeth and the lip are touching each other, the air rushing past will
cause friction, and as a result, a noise will be produced.
The noise made by the air escaping between the lower lip and the upper
teeth will sound exactly like the f at the beginning of the word feel. The label
labiodental is given to this type of sound: labial is a Latin-based adjective meaning
‘of the lip’, and dental, related to dentist, means ‘of the teeth. Because the noise made
by the air escaping through the obstacle being described is caused by friction, we
refer to this type of sound as a fricative. The sound [v] like in the word vine is
produced like producing [f], but with vibrating the vocal folds (figure 3.1).
When we pronounce [T] as in the word thigh, we place the blade of the
tongue on either the inside or the edge of the upper teeth, allowing the appropriate
narrow gap for friction. The sound [D] as in the word this is produced in the same
way of producing [T]. They only differ in the quality of voice or voiceless (figure
3.2). These two sounds are known as dental fricatives.

Figure 3.1 Labiodental fricative Figure 3.2 Dental fricative

If you blow the air through the narrow gap between the blade of the tongue
and the upper tooth ridge, in the manner you did through the gap between the
lower lip and the upper teeth for [f], a hissing noise will be produced, which we
transcribe as [s], like in the word seem. The name given to this kind of sound is
alveolar fricative: alveolus is the Latin word for ‘socket’, and the teeth obviously
grow out of ‘socket, hence the expression alveolar ridge. The sound [z] is produced
in the same manner as the sound [s], but with vibrating the vocal folds (figure 3.3)
Another fricative we have not yet discussed is the first sound in the word
shy. The phonetic symbol for this sound is [S]. The articulation of this sound
involves drawing the blade of the tongue to the area where the tooth ridge joins the
hard palate—the part of the roof of the mouth which feels hard. Because its place of
articulation straddles the palate and the alveoli, the sound like [S] is known as

37
palato-alveolar fricative. If you add vocal cord vibration to this sound, you are
going to produce [Z], like in the word measure, pleasure, and rouge (figure 3.4).

Figure 3.3 Alveolar fricative Figure 3.4 Palato-alveolar fricative

The majority of variety of English has the fricative sounds as given in the
following table:

Place of Articulation Voice Symbol Example

labio-dental - [f] fox
+ [v] vixen
Dental - [T] moth
+ [D] this
alveolar - [s] snake
+ [z] zebra
palato-alveolar - [S] shrew
+ [Z] measure
Glottal - [h] haddock

The above table shows that the glottal fricative [h] does not have any
counterpart in many Englishes, though some speakers have a breathy voice [˙]
where the sound begins a stressed syllable which follows a vowel-final non-
stressed syllable, as in 'behave' or 'rehearsal'. The sound [h] does not occur at all, or
occurs sporadically, in many non-standard English Englishes, which thus make no
distinction between words such as 'hill' and 'ill'.
Davenport and Hannahs (1998) state that a number of varieties also have a
voiceless velar fricative [x]; this is particularly true of 'Celtic' Englishes (Irish,
Scottish and Welsh English) in words such as loch/lough, dreich (Scottish dreary)
and bach (Welsh dear) (figure 3.5).

38
 Figure 3.5 Velar fricative
Other languages have fricative sounds in other places of articulation. We
can find a language has bilabial fricatives and uvular fricatives, while another
language has palatal fricatives and pharyngeal fricatives. On the other hand, one
language has fewer fricatives compared to English. For example, Indonesian has
only labio-dental fricatives, alveolar fricatives, and glottal fricative. The following
table shows the complete fricative sounds occurred in the world languages:

Place of Articulation Voice Symbol Example

Bilabial - [Φ] Ewe [∏o] 'to beat'
+ [β] [E$BE$] 'the Ewe language'
labio-dental - [f] Fine [faIn]
+ [v] Vine [vaIn]
Dental - [θ] Think [TINk]
+ [
] With [wID]
alveolar - [s] Sip [sIp]
+ [z] Zoo [zu˘]
palato-alveolar - [S] Ship [SIp
+ [Z] Measure [mQ˘Z´]
retroflex - [ß] Toda [poß] 'place name'
+ [Ω]
Palatal - [C] Margi [Caò] 'moon'
+ [∆] Margi [∆aò∆aòd´ò] 'picked up
Velar - [x] Margi [xa@] 'big water pot'
+ [ƒ] Margi [ƒa$f´@] 'arrow'
Uvular - [X]
+ [“] French [“atç] 'rat'
pharyngeal - [(] Agul [mu(] 'barn'
+ [÷] Agul [mu÷] 'bridge'
Glottal - [h] High [haI]
+ [˙]

39
1. Distribution
The labio-dental sounds [f] and [v], the dental sounds [θ] and [D], the
alveolar sounds [s] and [z] and the voiceless palato-alveolar sound [S] occur in all
positions in English (word-initially, word-medially and word-finally). However,
the voiceless dental fricative sound [D] for word-initial position is restricted to a
small set of 'function words' such as an article and demonstrative pronouns (the,
this, that, these and those) and adverbs (then, there, thus, etc.).
The distribution for each of the voiced palato-alveolar [Z], the glottal [h] and
the velar [x] (in those varieties that have it) is in some way restricted in English. The
sound [] occurs only in few words, such as 'treasure' and 'beige', and never word-
initially (apart from loan words such as 'genre' and 'gigolo'). The glottal fricative
[h], on the other hand, occurs only word-initially (such as 'happy'), or word-
medially at a beginning of a stressed syllable (such as 'behead'), but never word-
finally. The velar fricative [x] never occurs word-initially in those varieties that
have this sound. Thus Scottish English has word-medial 'lo[x]an' (a small loch) or
word-final 'drei[x]' (dreary), but no words beginning in [x].

2. Voicing
English fricatives -- with the exception of [h] and [x] -- may be voiceless or
voiced, giving opposition such as 'safe' vs. 'save', 'wreath' vs. 'wreathe', 'sue' vs.
'zoo', and (somewhat marginally) 'ruche' vs. rouge'. Moreover, the voiced fricatives
undergo devoicing word-initially and word-finally, typically only being fully
voiced between other voiced sounds. Compare the 'v' in 'vague' or 'save' with that
in 'saving'; the initial and the final 'v's will be (partially) devoiced [v•] whereas the
'v' in 'saving' is voiced all through its production [v].
The voicing of a fricative also affects the length of the preceding sonorant
(nasal, liquid or vowel). Voiceless fricatives shorten the duration of any sonorant
they follow; compare the underlined sonorant sounds in 'fence' and 'fens', 'shelf'
and shelve', 'face' and 'phase'.

3. Variation in Fricatives
The labio-dental fricatives [f] and [v] do not show a great deal of
assimilation, though [v] may often become voiceless word-finally preceding a
voiceless obstruent, as in 'ha[f] to' (have to), 'mo[f]e slowly' (move slowly), 'o[f]
course'. Indeed in faster speech, the sound may be lost altogether in unstressed
function words such as 'of' and 'have' as in 'piece of cake' or 'could have been',
where 'of' and 'have' are pronounced similar to the unstressed indefinite article 'a'
(the symbol for this is [´], known as schwa). This loss of a segment is known as
elision.

B. Stops
It is known that the common dominator for fricatives is the gesture with
which these sounds are produced: the two articulators are held close together, but
40
no so close that the air is prevented from getting through and causing friction.
Obviously, if the contact is tightened up, the air flow will be interrupted.
Surprisingly though it may seem, momentarily stopping the air flow is another
common method of producing speech sounds. The sounds produced in this manner
of articulation are known as stops
Stop consonants are characterized by involving complete closure in the oral
tract, preventing the airflow from exiting through the mouth. They may be oral
(velum raised) or nasal (velum lowered, allowing air to pass freely out through the
nose). Pulmonic egressive oral stops are often know as plosive and, as expected for
obstruents, are either voiced or voiceless. Nasal stops, being sonorants, are in most
languages voiced only.
If you close both lips tightly and then open them suddenly, you will hear
the sound [p], corresponding to the spelling of p in the word spy. However, you
need to have built up sufficient air pressure in the lungs for the ‘vocal instrument’
to sound at all. If you vibrate the vocal folds during the closure phase of the lips,
you will hear the sound [b] of the word obey. The sounds [p] and [b] share both
manner of articulation and place of articulation (figure 3.6). They are stops and
bilabial. Therefore, [p] and [b] only differ with regard to voice.

Figure 3.6 Bilabial stop

If you the blade of the tongue roughly in the same position as for [s], but
this time interrupt the air flow by tightening up the contact to the alveolar ridge,
you will get the sound [t] as in the word steam. If you increment [t] with voicing,
the result will be sound [d], like in the word adorn (Figure 3.7). What will happen if
you stop the air by placing the back of the tongue against the soft palate, in
approximately the same position as for the fricative [x] and [ƒ]? The respective stop
sound will now be [k], without voicing, as in skin, and [g], with voicing, as in again
(figure 3.8)

41
 Figure 3.7 Alveolar stop Figure 3.8 Velar stop

English has three pairs of voiceless/voiced stops. The following table shows
English stop consonants:

Place of Articulation Voice Symbol Example

bilabial - [p] Pig
+ [b] bear
alveolar - [t] Tiger
+ [d] dog
velar - [k] cat
+ [g] gorilla

There is also the glottal stop [/] heard, for example, in many British English
varieties (e.g. London, Manchester, Glasgow, Edinburgh as well as newer varieties
of RP) and some varieties of North American English (e.g. New Jersey,
metropolitan and upstate New York) as the final sound in 'rat', or for most speakers
in the negative 'uh-uh', or at the beginning of a voluntary cough. The glottal stop is
voiceless; it has no voiced counterpart, since the vocal cord cannot vibrate when
they are in contact.
As stated above, most languages have bilabial, alveolar and velar stops.
Moreover, a number of them may well have stops at other places of articulation,
such as palatal [c ] and [Ô] (Malayalam in India), or uvular [q] and [G] (Quechue in
Bolivia and Peru). The following table shows more complete description of plosive
consonants in the world languages:

Place of Articulation Voiced Symbol Example

bilabial - p pie /paI/
+ b buy /baI/
alveolar - t tie /taI/
+ d die /daI/
retroflex - ˇ Malayalam [kaˇ˘i] 'thickness
+ Í

42
 palatal - c
+ Ô
velar - k could /kUd/
+ g good /gud/
uvular - q
+ G
glottal - /
+

A small number of languages have some stops produced with an airstream

other than pulmonic egressive. These types of stops are not plosives. If the glottis is
closed then raised, the air above it (in the vocal tract) will be pushed upwards,
becoming compress behind the blockage in the oral tract. This air exits on release of
the closure in the oral tract. This airstream mechanism is known as glottalic
egressive, and the stops produced are known as ejectives. Ejectives are indicated by
an apostrophe following the stop symbol, as in [p'], [t'], [k']. Because they are
produced with a closed glottis, only voiceless ejectives are possible. Ejectives are
found in a number of African, North American Indian and Caucasian languages, as
well as elsewhere. It should be noted that there are only below 20 percent of all
world's languages have ejectives.

1. The Production of Stops

There are three identifiable stages involved in the production of all
pulmonic egressive oral stops if produced in isolation. First, there is the closing
stage, when the active articulator is raised to come into contact with the passive
articulator. For example, for the initial sound in 'dog' the blade of the tongue must
be raised to the alveolar ridge. Second, there is the closure stage, when the
articulators remain in contact and the air builds up behind the blockage. Third,
there is the release stage, when the active articulator is lowered, allowing the air to
be released with some force.
However, the sounds are not usually produced in isolation. When oral stops
are produced in ordinary connected speech, the closing stage and/or the release
stage may be missing, due to the influence of neighboring sounds. Only the closure
stage is necessary for all stops in all position. Thus if there is no period of closure,
the sound made isn't a stop anymore.
In connected speech, stop sounds may be produced without the closing
stage when they follow another stop sound with the same place of articulation—
that is, when they follow a homorganic stop. For example, the bilabial [p] in
'shrimp' has no separate closing stage, because the articulators have already been
raised to complete closure for the nasal stop [m]. The change from [m] to [p] is
affected by raising the velum (nasal to oral) and widening the space between the
vocal cords (voiced to voiceless). Another example is the alveolar stop [d] in the
sequence of 'hot dog'.

43
 In the examples of homorganic stops such as 'hot dog' or 'big cat', it is not
only the second stop that lacks a stage; the first stop in each example lacks not a
closing stage, but a release stage. Moreover, the release stage may also be absent in
non-homorganic clusters. In a sequence of sounds such as 'duct', the velar stop [k]
has no release stop. We can compare the velar stop in 'duct' and 'duck'; in the latter
the release of velar stop [k] is likely to be clearly audible, where for 'duct' only the
release of [t] will be heard. The lack of release for [k] here is due to the fact that the
articulators are already in complete closure position at the alveolar ridge for the [t]
before the back of the tongue is lowered at the end of the [k]. Thus the air cannot
escape from the mouth before the release of the second stop [t].

2. The Release Stage

When there is a release stage, it may not always involve a straightforward
lowering of the active articulator. The actual release may depend on the following
sound in a number of ways. So, in a word like 'mutton' the [t] is released not via
lowering of the tongue tip, since this stays in place for the alveolar nasal [n], but the
release of the oral stop occurs when the velum is lowered for the nasal. As a result,
it is allowing the air to escape through the nose. We can compare the [t] in 'mutton'
with that is a careful pronunciation of 'mutt', where the alveolar stop is released
orally.
The release of stop via lowing of the velum is known as nasal release, and
occurs when an oral stop precedes a nasal stop. In a similar way, when the alveolar
stops [t] or [d] precede the lateral liquid [l], in words like 'beetle' and 'badly', the
release is know as lateral release. In this case, the center of the tongue tip remains in
contact with the alveolar ridge for the [l], and the built up air is released when the
sides of the tongue lower.

3. Aspiration
The next important aspect of the release stage of plosives, particularly
associated with voiceless stops, is the phenomenon known as aspiration. Compare
the stops in the pairs 'pie -- spy', 'tie -- sty' and 'core -- score'. For most English
speakers these should sound quite different. When the voiceless stop begins the
word, as in the first member of each pair, there is likely to be an audible puff of air
following the release. When the stop follows [s], as in the second member, there is
no such puff of air. Stops like those in 'pie', 'tie', and 'core' are known as aspirated
stops. Aspiration is indicated by a superscript [h] following the symbol stop, e.g.
[ph], [th], [kh].
Articulatorily, what is happening is that for aspirated stop, the vocal cords
remain wide open after the release of the plosive and into the initial articulation of
the following segment. This means that the first part of the vowel in, say, 'pie' is
actually produced without vibrating vocal cords, i.e. without voicing. Vocal cord
vibration (voicing) thus only begins at some point into the production of the vowel;
the onset of voicing is delayed. For unaspirated stops, such as that in 'spy', the vocal
cords begin vibrating immediately upon the release of the stop; there no delay in
the onset of voicing and the following vowel segment is fully voiced throughout.
44
This difference is illustrated in the following figure. The horizontal line indicates
voicelessness, and a zigzag line indicates voicing.

ph ai s p ai
------------- --------

4. Voicing
Plosive may be either voiceless or voiced. This gives us contrasts in English
such as 'lopping' vs. 'lobbing', 'lacking' vs. 'lagging' and (in British and Southern
Irish English, but not North American or Northern Irish English) 'latter' vs. 'ladder',
where there is a difference in the voicing of the medial plosive.
While the difference is clear in these instances, it is not always so obvious.
Voiceless stops remain voiceless throughout their articulation in English, but
voicing is not always constant for voiced stops. Only in instances like those above,
i.e. between two other voiced sounds, is an English stop fully voiced. Elsewhere,
such stops are likely to be wholly or partly devoiced. When in initial position, vocal
cord vibration may not begin until well into the articulation of the stop; similarly, in
final position, vocal cord may cease well before the end of the articulation. This is
indicated in transcription by the diacritic [ •] such as in the [b•]eetle. For some accents
(West Yorkshire, for instance) there is no voicing at all in final position. This is also
true in a number of other languages, such as Danish or German, but by no means
for all; French, for example, has fully voiced final stops.

5. Glottalization and the Glottal Stop

In many kinds of English, voiceless stops may be subject to 'glottalization' or
'glottal reinforcement'. This means that as well as closure in the oral tract, there is
an accompanying (brief) closure of the vocal cords, resulting in a kind of dual
articulation. This glottalization is particularly likely for final stops in emphatic
utterances, such as 'stop that!', where the final [p] and [t] may well be glottalized,
but is common to some degree for many word final voiceless stops. This sound is
often transcribed in IPA by using a superscript [] after the stop symbol: [p] or [t].
Furthermore, voiceless stops may under some circumstances be replaced by
a glottal stop. That is, there will be no oral closure at all, only glottal closure. The
extent to which this occurs will depend on the accent of the speaker, the particular
stop involved and the position of the stop. Thus for many speakers of most kinds of
British English (including RP), a [t] can be replaced by [] before a nasal, as in '
a[n]ight (at night) or 'Bri[n]' (Britain), where the subscript [ ] indicates a syllabic
consonant. Similarly, a voiceless stop may be replaced by [] when preceding a
homorganic obstruent; 'grea[s]mile' (great smile) or 'gra[f]ruit' (grapefruit).
Vowels may also be subject to glottal reinforcement when they occur word-
initially, especially if emphatic, as in 'go []away!' or 'it's []over!', or if there is
hiatus (two juxtaposed vowels in consecutive syllables), as in 'co-[]authors'.

45
6. Coarticulation
So far the discussion of variations in the manner of English stop consonants
has been hade available. Only little have we said about variations in the exact place
of articulation. A major problem in describing speech is that all utterances involve
coarticulations—the overlapping of adjacent articulations. English consonants often
vary their place of articulation so that they become more like the next sound. It is
noted in the previous chapter that /t, d/ are usually alveolar stops but are
pronounced with tongue contact on the teeth, so that they become [t5, d5] when they
occur before dental fricatives as in eighth [eIt5T] and width [wId5D]. (Some people
pronounce the latter word with voiceless consonants at the end, making it [wIt5T].)
Another noticeable change in the place of articulation occurs in the
pronunciation of /k, g/ before a front vowel as in key, geese [k+i, g+is] as compared
with a back vowel as in cart, guard [kA”t, gA”d]. The different forms of /k/—the
allophones—occur because of the influence of the following vowel. The whole body
of the tongue has to be pulled up and forward for [i]. This action begins during the
formation of the closure for /k/, which is consequently farther forward than the
closure in the allophone of /k/ before /A/. In the latter case the /k/ anticipates the low
back position of the body of the tongue in' /A/.
You should be able to feel that the place of articulation of /k/ is much farther
forward before a front vowel. Check that in your own pronunciation of key the
articulatory contact is between the tongue and a point on the roof of your mouth
near the hard palate. Now compare this with your pronunciation of caw, in which
the articulation is at the back of the mouth fairly low down on the soft palate.
Try saying key /ki/ and calm /kAm/. It is easily felt that, before /i/, the stop is
made much farther forward than the stop caw. With intermediate vowels, the
closure is accordingly made at an intermediate point on the velar surface. If we
regard the back variety is basic, it is possible for us to transcribe the fronted variety
with a diacritic [ +], a subscript plus sign; this diacritic means that the sound is
articulated a bit more forward than usual. Thus, [k+] is appropriate for a narrow
transcription of /k/ in key [k+i] or kid [k+Id]. The same variations are found with /g/:
geese [g+is], gill [g+Il]

Figure 3.9. Allophones of /k/ before /i/ and /A/

46
 Now try to find out whether there is as much variation in the place of
articulation when the /k/ occurs after vowels in words such as peak [pik] and hawk
[hçk]. You will probably discover that there is very little difference between these
allophones of /k/. In general, English may be said to be an anticipatory language—
that is, one in which the articulations of the sounds yet to come are anticipated to
some extent. Some other languages, such as French or Italian, are perseverative in
that the articulation of one sound tends to persevere, or continue, into the following
sound.
We have already seen an example of the tendency of English to anticipate
future articulations. We noted that when a stop occurs before another consonant—
as in apt and act—it is unexploded, so that these words are pronounced [Qp*t] and
[Qk*t]. This may be regarded as a case of the articulation of the last consonant being
anticipated during the closure of the previous consonant. The [p] in apt is
unexploded because the closure for the [t] occurs before the lips come apart. In
English an articulator that is not necessarily involved in a given sound will nearly
always start moving toward its position in the next sound in which it is the primary
articulator. This phenomenon is known as anticipatory coarticulation.
We can regard those positions of the vocal organs that are specified for a
given sound as the target positions for that sound. The estimated target for [k] as in
key or caw is the position for the back of the tongue. The soft palate is required to be
raised so as to ensure a velic closure. The position of the rest of the tongue is shown
with a dashed line because in the case of [k] the only requirement is for the back of
the tongue to approach the target shown in Figure 3.10. The targets for [i] and [ç]
are shown in Figures 3.11 and 3.12 respectively. The increase in the thickness of the
line shows that the major requirement for a vowel sound is for certain parts of the
tongue to be in the correct positions so as to cause constrictions in the vocal tract at
the appropriate places.

Figure 3.10. The estimated target position for /k/ as in key or caw

47
 Figure 3.11 The estimated target position for [i] as in key

We can now see why the positions for [k] in "key" and ‘caw’ are different. In
each case there is the same target, but you do not necessarily hit the target for
which you aimed. Even while the back of the tongue is aiming at the target for [k],
other parts of the tongue are already moving toward their targets for the following
vowel. The result is that the articulation achieved in saying [k] will be considerably
influenced by the articulation required for saying the following vowel. '

Figure 3.12 The estimated target position for [ç] as in caw

48
 Ideally, the description of an utterance might consist of the specification of
the targets and a specification of the rules for moving from one target to the next.
We should be able to state the articulatory targets in terms of numerical values of
distances between the vocal organs. One of the objects of any science is to express
things in terms of verifiable measurements. In a sense, we do not really know
anything until it can be stated in terms of numbers. We are now beginning to be
able to do this in the description of English utterances.
The data in Figures 3.10, 3.11, and 3.12 represent an attempt to specify
graphically, and hence measurably, the articulatory targets for three English
sounds. This kind of diagram can be drawn for each of the other English sounds
that have separate targets, making the thickness of the line proportional to the
extent to which the shape of that part of the vocal tract is important. As we have
noted for [k], all that matters is the position of the back of the tongue and the velic
closure. For the vowels what matters most is the position of the tongue in the
region where the vocal tract is most constricted; the other parts of the tongue need
be only roughly in the places indicated. Coarticulation between sounds will result
in the positions of some parts of the vocal tract being influenced largely by the one
target, and others by the other. The extent to which anticipatory coarticulation
occurs depends on the extent to which the position of that part of the vocal tract is
specified in the two targets.
The degree of coarticulation between two sounds also depends on the
interval between them. For example, a considerable amount of lip rounding occurs
during [k] when the next sound is rounded as in "coo" [ku]. Slightly less lip
rounding occurs if the [k] and the [u] are separated by another sound as in "clue"
[klu]. And even less occurs if there is also a word boundary between the two
sounds, as in the phrase "sack Lou" (sQk lu]. Anticipatory coarticulations have
been observed, however, over even longer sequences. In the phrase "tackle Lou"
[tHQkl8 lu] the lip rounding for the final [u] starts in the [k], which is separated
from it by two segments and a word boundary.
There is no simple relationship between the description of a language in
terms of phonemes and the description of utterances in terms of targets and
conjoining rules. Phonemes are certainly not in a one-to-one relation with targets,
even though some allophonic differences (such as those between the [k] in "key"
and the [k] in "caw") can be explained in terms of targets and conjoining rules.
Similarly, we do not have to specify separate targets for the alveolar [n] in "ten" and
the dental [n5] in "tenth." Both are the result of aiming at the same target, but in the
case of [n5] the articulation is influenced by the target required for the following
sound. However, the differences between some allophones are actually the result of
aiming at different targets. For example, the [k] at the end of "back" does not have
the same target as the [k] at the beginning of "cab." A final [k] is longer than an
initial [k]; it is often accompanied by a glottal stop; and it is not followed by a burst
of aspiration. These two [k] sounds must therefore have different targets. Similarly,
the [l] in "leaf" and the [l] in 'feel’ differ in ways that cannot be ascribed to
coarticulation. The initial and final allophones of /1/ have different vocal tract

49
shapes as the result of aiming at different targets. In my own case, initial [l]
involves trying to make an alveolar lateral; but final [l] involves aiming at a back
unrounded vowel position, with no requirement for alveolar contact. .
To summarize, targets are units that can be used in descriptions of how a
speaker produces utterances. Phonemes are more abstract units that can be used in
descriptions of languages to show how one word contrasts with another. The two
kinds of units overlap to a considerable extent, but they are not completely
identical.

C. Affricates
There is a third and final type of obstruent which is a composite of a stop
and a fricative made in rapid succession, without changing the position of the
articulators. These sounds are known as affricates. The affricates familiar to
speakers of English are the sound found at the beginning and the end of church and
judge.
Davenport and Hannahs (1998) state that affricates are produced like
plosives, in that they involve a closing stage, a closure stage, and a release stage.
The difference lies in the nature of the release: where for a standard plosive, the
active articulator is lowered swiftly and fully, allowing a sudden, unhindered
explosion of air. For affricates, the active articulator remains close to the passive
articulator, resulting in friction as the air passes between them, as for fricatives
Figure 3. 13). Thus, phonetically, affricates are similar to a stop followed by a
fricative. However, they do not behave like a sequence of two segments. Consider
'catch it' and 'cat shit'; the sound represented by 'tch' ([tS]) is noticeably shorter than
the sequence of sounds represented by 't sh' ([t + S]).

Figure. 3.13 Palato-alveolar affricate

English has only two affricates, the voiceless palato-alveolar [tS], as in

'chimpanzee' and its voiced counterpart [dZ], as in 'jaguar'. Both affricates ca

50
appear in all positions; word-initially, word-medially, and word-finally. The
symbols [c&] and [j&] may also be used to replace these two sounds.
Davenport and Hannahs also say that affricates at other places of
articulation are found in many languages; German has a voiceless labio-dental [pf]
in Pferd 'horse', and voiceless alveolar [ts] in Zug 'train'; Italian has a voiced alveolar
[dz] in zona 'zone'.

Nasals
Nasals are a variety of stops; they are formed with complete closure in oral
tract. The difference between nasal and oral stops is that for nasals the velum is
lowered, allowing air into (and out through) the nasal cavity. Unlike oral stops,
nasals are sonorant, and are thus typically voiced only. However, there are few
languages (e.g. Burmese) do contrast voiced and voiceless nasals.
English has nasal stops in the same places of articulation as it has oral stops:
bilabial [m], alveolar [n], and velar [N]. Other languages have nasal stops in other
places of articulation, e.g. dental [n5] in Yanyuwa (Australia) [wun5un5u] 'cooked',
palatal [¯] as in Indonesian tanya [ta: ¯a] (ask).

Figure 3.14 bilabial nasal Figure 3.15 Alveolar nasal

Figure 3.16 Palato-alveolar nasal Figure 3.17 Velar nasal

51
 The following table shows the distribution of nasals across languages:

Place of Articulation Voiced Symbol Example

Bilabial + m meat
Labiodental + M some food
Dental + n5 ninth
Alveolar + n Fan
Retroflex + ˜
Palatal + ¯ Indonesian [ba˘¯a/]
(many)
Velar + N sing
Uvular + N

The bilabial and alveolar nasals [m] and [n] occur word-initially, word-
medially, and word-finally in English: e.g. 'mill', 'tumor', 'ram', 'nil', 'tuna', and 'ran'.
On the other hand, the velar nasal [N] cannot occur word-initially. It can only occur
word-medially and word-finally. It should be noted that this is true for English but
not for all languages with [N]. For example, Burmese has velar nasal word-initially,
as in [Nâ] 'fish'. In some varieties of English, such as North West or West Midland
English, and Long Island American English, [N] is always followed by an oral velar
stop, either [k] or [g].
Positionally, the nasal [N] shows no important assimilation. However, there
is some socio-linguistically governed alternation between [N] and [n] for the
inflection '-ing'. The bilabial [m] may be labio-dental [M] before the labio-dental
fricatives [f] and [v] (so[Mf]un).

D. Liquids
Sonorants can also be articulated exclusively in the mouth, with no nasal
component. These non-nasal sonorants still have a simultaneous continuant and
non-continuant articulation: during their production one part of the oral channel is
blocked, while another part remains unobstructed and allows the air to escape
freely. Such sounds are commonly referred to as LIQUIDS (perhaps because they
sound fluid), and we now turn our attention to them.
Liquid is a cover term given to many 'l' and 'r' sounds (or laterals and
rhotics respectively) in the languages of the world. In a broad sense, what liquids
have in common is that they are produced with unhindered airflow (which
distinguishes them from obstruents) but nonetheless involve some kind of
obstruction in the oral tract (unlike glides and vowels, which are articulated with
open approximation). However, the exact nature of the obstruction, particularly in
the case of those sounds grouped together as rhotics, is a complicated matter cross-
linguistically.

52
 Liquids are sonorants, and are typically voiced. Voiceless liquids do occur
(Scottish Gaelic has [r8], for example), but often voicelessness in 'l' and 'r' sounds
also involve friction, as in the Welsh voiceless alveolar lateral [¬] in llan 'church'
and, as such, these sounds are obstruents rather than liquids proper.

1. Laterals
Let us compare the middle consonant in mellow with its counterpart in
meadow. The articulation of both these sounds is alveolar, that is, it involves placing
the blade of the tongue on the upper alveolar ridge. Both sounds are also voiced.
Last, they both involve a complete closure at the upper front alveolar area. Given
these striking similarities, what is it that makes these two sounds different?
If you pay close attention to the articulation of the two sounds in question,
you will notice that, in the case of [d], the tongue presses firmly against the upper
teeth all around, not just at the front, but also on the sides, to prevent any air from
escaping.
By contrast, for the sound found in the middle of mellow, represented by
the phonetic symbol [l], the sides of the tongue (only one side in some speakers) do
not touch the complete set of upper teeth, and air comes out continuously through
the resulting gap. Because the air flows over sides of the tongue, these sounds are
known as LATERALS: Latin lateralis means 'of the side(s)', from latus 'side'
(compare such English expressions as lateral thinking, collateral, etc.).

Figure 3.18 Articulation of [l] and [d]

As regards place of articulation, [l] is defined as an alveolar sound, exactly

like [d] (figure 3.18). The definition of [l] and [d] as alveolar obviously suggests
that the gesture made by the blade of the tongue is regarded as primary, and the
gesture made by the sides of the tongue (closing for [d] and opening for [l]) as
subsidiary. Indeed, primary articulation is commonly related to the median line of
the tube through which the air flows out: to the mouth's median plane, in more

53
technical parlance. In addition, [l] is voiced, also like [d], unsurprisingly so, since
we stated above that sonorants (of which class liquids, and thus [l], are members)
are characteristically voiced across languages. Finally, [l] is also oral, since the
velum remains raised during the whole of its production, just as it does with [d].
The difference between [l] and [d], therefore, lies exclusively in the respective
status of these two sounds with regard to LATERALITY, the term referring to the
lowering of the sides of the tongue during articulation: [l] is lateral, while [d] is
not.
In the opening paragraph of the section we said that [l], like [d], involves
air stoppage at the front. We now know that during the articulation of [l] air
continues to flow out of the mouth through the gap formed by the sides of the
tongue and the upper teeth on the sides of the mouth. We came across a similar
situation earlier for nasals, characterized by air stoppage in the mouth but
continuous airflow through the nose, and we attributed their conventional
description as stops to the fact that the articulation in the mouth is regarded as
primary.
Phonologically, nasals always function as stops, and therefore their
classification as such is uncontroversial. The tendency for [l] is also to function as a
stop, but in a few languages it appears to pattern with the continuants.
Consequently, we must leave the classification of this sound somewhat flexible.
Articulatorily, however, [l] is always considered alveolar: this confirms the
privileged status of the central region of the mouth in the identification of the
primary articulatory gesture.
An additional lateral sound, a palatal lateral, bears a similar relationship to
[l] to that which [¯] bears to [n]. The phonetic symbol for the palatal lateral is [¥],
an inverted "y". The articulation of this sound, spelled gli in Italian (zabaglione) and
ll in Spanish (paella) is a bit exotic for the English speaker, and we will accordingly
put some extra care into its description. Pronunciation guides tend to suggest the
lateral sound in million as a close English equivalent, but this correspondence is
subject to a number of caveats. To produce [¥], the body of the tongue (that is, the
area behind the blade) must be raised to the roof of the mouth to block the exit of
air, as was the case for the nasal [¯]. However, for [¥] the sides of the tongue must
be allowed to hang free, to let the air flow out through the resulting gap, in the
familiar lateral gesture.
The correspondences in place of articulation between nasals and laterals are
completed with the velar laterals that exist in a handful of languages. In the Mid-
Waghi language of New Guinea, for instance, the word [a;a;e] 'dizzy', with the
velar lateral [;], contrasts (although not minimally) with the word [alala] 'speak
incorrectly', with the ordinary alveolar [l]. The sound [;] is also reported to occur
in some English accents in some contexts (for instance, before labial or velar
consonants), but it is otherwise rare. The velar lateral [;] must not be confused with
the velarized alveolar lateral […]. The sound […] occurs allophonically in English in

54
word-final position, and syllable-finally generally, as in pill, mole or cool, although
some accents only have plain, or "clear", [l]
In languages like Russian, […] functions phonemically: mo… 'pier' and po…ka
'polka', with […], contrast with mol 'moth' and polka 'shelf', without. The
articulatory difference between the velarized l, […], and its plain counterpart [l] lies
in the additional bunching of the body of the tongue at the back that characterizes
[…].
The English lateral [l] can appear word-initially, word-medially, and word-
finally, as in 'louse', 'bullock', and 'gull'. For many accents of English there is
considerable variation in the articulation of [l] according to position. For most
speakers [l] devoices if it follows a voiceless obstruent, as in the word play [pHl8eI].
There is also a noticeable difference for many speakers between the lateral
in 'loot' compared to that in 'tool' or 'milk'. The lateral that occurs initially and
medially before a vowel (pillow) is known as 'clear l'. The lateral that occurs
elsewhere is known as 'dark l'. The symbol for 'dark l' is […], such as in the word
'fi[…]' and 'fi[…]m' (see fig. 3.19)

Figure 3. 19 Clear and dark /l/

2. Rhotics
"Rho" is the Greek name for the letter 'r', and the label RHOTICS therefore
refers to a class of sounds that are "r-like". It will soon become apparent, however,
that the members of this class do not necessarily have much in common with each
other phonetically: their common grouping as "rhotics" is grounded on similarity of
phonological behavior, rather than on shared phonetic substance.
The sound represented by the letter r in most accents of English is very
different from its counterpart in many other languages. In turn, the r sound typical
of Scottish English is different from its common English equivalent, as we will see
below. The phonetic symbol for the common English r is [®], an inverted r. The

55
articulatory gesture for [®] is almost the opposite of the articulatory gesture for [l],
hence the tongue twister "red lorry, yellow lorry".
In the previous section we explained that for [l] we blocked the air in the
central part of the mouth by pressing the blade of the tongue firmly against the
alveolar ridge, while letting it flow freely down the sides. By contrast, for [®] the
sides of the tongue touch the back teeth, while a fairly wide gap is created in the
centre of the mouth for the air to pass through without causing friction (figure
3.20).

Figure 3.20 Articulation of [®] and [l]

There are two ways in which the blade of the tongue may be positioned for
[®]. The chances are that speakers from Britain will keep the blade flat, leaving a
channel open at the front for the air to escape (figure 3.21). American speakers, on
the other hand, are more likely to curl back the blade towards the roof of the mouth
(without of course touching it or drawing the tongue too close to it). Fortunately,
the effect of this articulatory difference on the resulting sound is not great. The
action of curling back the tongue is known as retroflection, and the sounds thus
produced are referred to as retroflexes.

Figure 3. 21 British [®] and American [”]

56
 A wide variety of articulations are subsumed under the general heading of
rhotics, even within English. Rhotics include:
• the alveolar trill [r], in which the tongue blade vibrates repeatedly against the
alveolar ridge (this sometimes heard in Scottish accents)
• the alveolar tap [R], a single tap of tongue blade against the alveolar ridge
(heard more commonly in Scotland)
• the alveolar continuant [®], produced with the tongue blade raised towards the
alveolar ridge and the sides of the tongue in contact with the molars, forming a
narrow channel down the middle of the tongue (heard in many kinds of
English, including RP)
• the retroflex [”], produced in a way similar to [®] but with the tongue blade
curled back to a post-alveolar position (heard in many North American and
South West English)
• the uvular roll [{] or fricative [“], respectively produced with the back of the
tongue vibrating against or in close approximation to the velum (heard in rural
Northumberland and parts of Scotland; this is also the kind of rhotic often
heard in French and High German)

E. Glides
In articulatory terms, glides are rather more like vowels than consonants,
since there is no contact of any kind between the articulators; indeed, an alternative
term for such sounds is semi-vowel. They behave like consonants, however, in that
they do not form syllabic nuclei; rather, they appear at the edge of syllables, as in
the first sound of 'yes'. They are included here then for reasons more to do with
their phonology than their phonetics; that is, their behavior with respect to the
other sound of the language, rather than the details of their articulation.
English has two glides: the palatal [j] as in 'yes' and the labial-velar [w] as in
'weigh'. The palatal [j] involves an articulation similar to that for the vowel [i]
(where [i] is a vowel sound like in 'beat'), with the front tongue close to the palate.
The labial-velar [w] is similar to [u] (where [u] is a vowel sound like that in 'shoe'),
with rounded lips and the back of tongue raised toward the velum. These two
glides are by far the most common cross-linguistically, though other glides are
occasionally found. French, for example, has a labial-palatal [Á] (similar to the front
round vowel [y]) in words like lui [lÁi] 'him'.

1. Distribution
English [j] appears freely in word-initial position before a vowel; [j]ield,
[j]es, [j]ak, [j]awn, and [j]ou. In a word-initial cluster, [j] is restricted to appearing
before the vowels [u:] and [u´] as in 'm[j]ute' and 'p[j]ure, except for many speakers
in East Anglia, who have no [j] at all these words. In non-word-initial cluster, [j]
may also appear before [´], as in 'fail[j]ure'. The exact range of consonants [j] may
follow will depend on the variety of English. Many forms of North American

57
English have more restricted set than British English varieties in that [j] cannot
follow the alveolars [t], [d], [s], [z], [n] and [l], and the dental [T] in the words like
'tutor', 'dune', 'assume', 'resume', 'new', 'lute' and 'enthuse'.
The labial-velar [w] appears freely word-initially as in 'we', 'wag', 'watch',
'woo', etc. As part of a cluster, there are no restrictions on the following vowel (twit,
twenty, quarter, etc), but English does not allow [w] after consonants other than [t],
[d], [k], [s], [T] and the sequence [sk]. For example the words twin, dwarf, quit, sway,
thwart, and squat show the use of [w]. The sound [w] may also follow [g], but only
in loanwords like the proper name 'Gwynneth'.
The question of whether glides appear following vowels is to some extent
again a phonological question. The word my contains a vowel sequence, known as
diphthong, which my represented either as a sequence of two vowels [ai] or as a
vowel + glide [aj]. For some speakers, words such as 'here' or 'lower' may involve
an inter-vocalic glide; [hij´] and [low´] (as opposed to RP [hI´] and [l´u´).

2. Variation
The articulation of [j] varies according to the following vowel; the tongue
blade is higher before high vowels (as in 'yeast'), lower before low vowels (as in
'yak'). Following voiceless obstruents, [j], as with other consonants, is subject to
devoicing (p[j]ewter). Particularly following voiceless stops in stressed syllables,
this may lead to friction, resulting in the palatal fricative [C] rather than a devoiced
glide. This is especially noticeable with the sequence [h] + [j], which may well
coalesce, giving rise to pronunciations like '[C]uman' (human).
As with [j], the articulation of the labial-velar [w] will vary according to the
height of the following vowel; the tongue is higher before high vowels (we), lower
before low vowels (was). Furthermore the degree of lip rounding will also vary
according to the following vowel; the lips are more rounded before round vowels
('woo'), less rounded before unrounded vowels ('wept').
Following voiceless obstruents [w] devoices, and as with [j], this may result
in friction being audible, especially after voiceless stops: 't[w]it' or 't[„]it' (voiceless
labial-velar fricative).
In some varieties, particularly Scottish, Irish and North American English,
the voiceless labial-velar fricative [„] occurs as a speech sound in its own right,
since these varieties have contrasts between words such as 'witch' and 'which',
'Wales' and 'whales', 'weather' and 'whether', etc., with the first member of each
pair having the glide [w] and the second member having the fricative [„]. For other
speakers, these words are homophones, both having the glide.

F. Syllabic Consonants
Ordinarily every syllable contains a vowel as its central part. Sometimes,
however, a syllable contains a sonorant nasal or liquid instead of a vowel.
Pronounce the word sudden [s√dn`J at an ordinary rate. You will notice that there is

58
nasal release between the /d/ and the /n/, that is the tongue stays at the alveolar
ridge throughout the /d/ and /n/. Thus, there can be no intervening vowel. The /n/
itself forms a syllable; we call it a syllabic nasal, transcribed [n`], with a short
subscript stroke as the diacritic to indicate syllabicity.
In like manner, you can observe that the final sonorant in ladle can be
pronounced with a syllabic consonant: [leId…`]` . Words which have syllabic
consonants can alternatively be pronounced with a vowel: [s√d´n], [leId´…].
However, such pronunciations with the vowel are typical of very careful, slow
speech and may sound stilted and artificial; the pronunciations with the syllabic
consonant are much more common. Syllabic consonants also occur occasionally
after fricatives: prison [p®Izn`], prism [p®Izm`], rhythm [®IDm`]. In rapid speech,
further combinations may occur: [beIkN`] bacon, [b√mpm`] bump 'em, [®Ibm`]
ribbon. In General American English (GA), syllabic [m`] is normal after /t/ (= [R]):
bottom [bARm`]. The test for a syllabic consonant is whether or not a vowel
intervenes; a syllabic consonant is never next to a vowel. In a word like gambolling
[gQmbl`lIN], note that the syllabic consonant is in the second syllable, and the third
syllable begins with an [l].

Further Reading
Ladefoged (1993) is an accessible textbook for greater details on the
production of consonants and vowels (see also further reading on chapter 2).
Davenport and Hannahs (1998) provide the detail explanation about the English
consonants, and Roca and Johnson (1999), especially the early chapters, provide
more phonological explanation on consonants. For a reference book on the
articulatory and acoustic details of the sounds of a large number of languages, see
Ladefoged and Maddieson (1996).

Exercises

1. Describe the articulation of the following sounds. Be sure to include

information about the path of the airflow, the state of the vocal cords, and the
position of the velum and any obstruction in the oral cavity.
a. [b] b. [N] c. [tS] d. [s] e. [T]
f. [g] g. [t] h. [v] i. [Z] j. [p]

2. Assuming the consonants of English, indicate the symbol representing the

sound described by each of the following:
a. voiceless alveolar stop
b. voiced dental fricative
c. voiced labial-velar glide
d. voiceless velar stop
e. voiced alveolar nasal

59
 f. voiced palatal nasal
g. voiceless dental fricative
h. voiceless bilabial fricative

3. Describe each of the following symbols in words. Example: [d] = voiced alveolar
stop
a. [q] b. [B] c.[X] d. [g] e. [“]
f. [ˇ] g. [M] h.[Z] i. [f] j. [C]

4. Identify the difference in articulation between the following groups of sounds.

For example, [p b t g] differ from [f s S T] in that the sounds in the first set
are all stops and the sounds in the second set are fricatives.
a. [p t s k] Vs. [b d z g]
b. [b d g] Vs. [m n N]
c. [n l ®] Vs. [t d s]
d. [p b f v m] Vs. [t d s z n]
e. [w j] Vs. [m S]

5. The following exercises are designed to help you become more familiar with the
shapes of the vocal tract connected with the production of different consonant
sounds. For each drawing presented below there is only one sound of English
that could be produced by a vocal tract positioned as shown; you are to figure
out which consonant sound is represented (either by referring to the
descriptions of different sounds or by experimenting with your own vocal
tract). Write the phonetic symbol for that sound between the brackets below the
appropriate drawing. Note that voicing is shown by two wavy or "bumpy" lines
where the larynx would be, while voiceless sounds are represented by two lines
shaped as an ellipse, indicating an open glottis at that point. Take care also to
note whether the air passage to the nasal cavity is open or closed (i.e., if the
velum is raised or lowered). The first drawing is labeled to start you off.

60
61