Sonorant consonants are like vowels in a number of respects. For one thing, like vowels, sonorants are pronounced with a relatively free escape of the airstream through the oral or nasal cavity. Second, they are normally voiced, just like vowels. And third, they are pronounced without friction, again like vowels. As a result of these similarities, we can sing or hum sonorants much as we can vowels.
On the other hand, sonorant consonants share with obstruents the property of being marginal in the syllable: They occupy a position in the coda or the onset, as opposed to the syllable peak. Nevertheless, it is a noteworthy feature of GA sonorants that, like vowels, they can sometimes be in the peak of the syllable.
This is typically the case in the final syllables of button, bottle, banner (/bʌtn̩, bɑːt̬l̩, bænr̩/). Such sonorants are known as syllabic consonants. Within the class of sonorants, some are more ‘consonantal’ than others. Nasals share with stops the feature of having a complete oral closure: [n] has exactly the same oral closure as [t] and [d], and the oral closure for [m] is like that for [p] and [b], while that for [ŋ] is like the closure for [k] or [g].
Approximants (like [j,1,ʋ]) are more vowel-like than nasals (they don’t have a complete oral closure) but notice that when an approximant is devoiced after syllable-initial GA /p,t,k/ (aspiration), it has friction. This is because the open glottis lets through much more air than a vibrating glottis, and the airflow through the stricture for the approximant is much greater as a result.
For a nasal consonant the oral cavity is completely blocked at some point, as for a stop. However, nasals differ in that the soft palate is lowered during the oral closure so that the air can escape freely through the nasal cavity. Stops, on the other hand, are articulated with the soft palate in its raised position, i.e. with a velic as well as an oral closure, so that the air cannot escape but is temporarily trapped in the oral cavity. Nasals are normally fully voiced.
12.2 GA and NA Nasals
The nasal consonants of AN and GA are:
In both languages, the palatal nasal [ɲ] may arise as a result of the assimilation before [j]. It is not considered a phoneme of either AN or GA because its occurrence is predictable by context. In GA this consonant may occur in words like canyon and onion.
Nasals are similar to vowels in that, depending on the context they may be long or short. The duration of GA nasals (and /l/) varies considerably. In a long context, i.e. before lenis obstruents and when final, nasals are fully long. In a short context, i.e. before a fortis obstruent, they tend to be quite short or, more frequently, elided, leaving a nasalized vowel.
There is no instance of /n/ deletion after /ə/ in GA, as in lopen [loˑpə]. So make sure you pronounce a firm /n/ in delicatessen, Staten Island, /ˈstætn̩ ˈaɪlənd/.
In unaccented syllables, final -ing in GA is frequently pronounced /ən/ or /ɪn/ in informal styles, as in trying [ˈtraɪən, ˈtraɪɪn] (This pronunciation is often indicated in informal texts or dialogues by the spelling tryin’, but it must be noted that this is informal speech.)
Following /t, d/ syllabic /n/ may be heard, as in eatin’, ridin’ [ˈiːʔn̩, ˈraɪdn̩]. This substitution is not limited to the present participle ending, but affects all -ing sequences in weak syllables, which would therefore include words like pudding, Washington, lightning, but not shoestring.
As the name suggests, approximants are produced when the articulators are approximated, i.e. brought closely together, without forming a major obstruction of the airstream: there is a light or near contact at some point but the air is allowed to escape relatively freely so that no friction is produced.
The GA approximants are /l,r,j,w/.
12.3.1 Approximants: /l/
GA /l/ has three major allophones: one is called clear l, the other two are traditionally both referred to as dark l.
For clear l, the tip of the tongue articulates with the alveolar ridge and the sides of the tongue with the upper side teeth, allowing the air to pass freely along one or both sides of the back of the tongue. The primary articulation is therefore a lateral alveolar contact.
The traditional term dark l is used to refer to what are in fact — in spite of their auditory similarity — two distinct sounds from an articulatory point of view. These are velarized [l] and vocalized [l]. What they have in common is that in both cases the back of the tongue is raised toward the soft palate, creating a dark [u]-type resonance.
The difference is in what the tongue tip does: it may articulate with the alveolar ridge at the same time, allowing the air to escape freely along one or both sides of the front of the tongue, as in killing. This allophone represented as [ɫ]. If it does not articulate with the alveolar ridge, as may occur in words like kill, milk, filled, where /l/ is in a coda and unisyllabic /l/ is in effect realized as a half- close back vowel: the primary articulation is lost and /l/ is vocalized. A raised [ɫ] is used to represent this allophone.
The terms ‘velarized’ or ‘dark’ /l/ refer both to the vocalized and the nonvocalized variety, provided it has back vowel resonance: A vocalized /l/, then, is a dark /l/ which has lost its primary articulation and is phonetically a vowel. (Note that clear [l], which occurs in unisyllabic onsets, is never vocalized, i.e. always has an alveolar or dental contact.)
Advice for learners:
The /l/ sound is generally rather dark in all contexts. However, speakers from the East of the Netherlands tend to pronounce a strongly palatalized /l/, whose quality differs considerably from the velarized /l/ used in GA. One way to arrive at a velarized /l/ is to pronounce [l] and [ʊ] at the same time, holding the tongue rather tense.
Note that the duration of final /l/, i.e. its vocalized realization, varies depending on the phonological context, as can be seen in the words tell, tells, build, where it is long — certainly by some standards — and else, built, help, where it is short. l-vocalization is increasingly common in many other spoken idioms especially in Western urban speech. An example containing two vocalized /l/-sounds would be a phrase like heel veel [hɛˑo fɛˑo].
12.3.2 l-Velarization and l-Vocalization
All GA speakers velarize /l/ in the coda. Thus, velarization applies not only to unisyllabic /l/ in words like pull, help, killed, but also to ambisyllabic /l/, as in belly, silly, volley, middle, Italy. The rule can be formalized as follows:
l-VELARIZATION: /l/ is velarized when it occurs in the coda
l-VOCALIZATION: /l/ may be vocalized when unisyllabic in the coda
Although in the pronunciation of some GA speakers, /l/ always has some degree of velarization, a clearer variety is used in unisyllabic onsets, as in lean, listen, light, flee, police and lose, look, low, law.
12.3.3 Pre-l Breaking
In Ga, vowels before [ɫ] tend to be diphthongized, so you get a centering offglide between the vowel and [ɫ], which is particularly noticeable afte close (free) vowels and closing diphthongs, as in feel [fiːəɫ], rule [ruːəɫ], foul [faʊəɫ].
The rule can be formalized as follows:
Pre-l Breaking: Insert a schwa between a close tense vowel and a coda-final /l/.
Sometimes pre-l breaking applies before ambisyllabic /l/, but when preceded by /iː/ or /eɪ/, as in feeling, feel it, failing [ˈfiːəɫiŋ, ˈfiːəɫ ɪt, ˈfeɪəɫɪŋ]. You would therefore not have pre-l breaking in falling, killing, kill it, etc.
12.3.4 Syllable Addition
The schwa-glide that is inserted by pre-l breaking in words like rule creates a monosyllabic vowel sequence that is phonetically very similar to the disyllabic sequence that occurs in words like cruel: [ruːᵊɫ] versus [kɹ̥uːəɫ]. Indeed, in the pronunciation of many GA speakers the inserted [ə] may become syllabic through syllable addition, especially after close free vowels and diphthongs. Further examples of such unexpected rhymes are:
rule [ˈruːəɫ] cruel [ˈkɹ̥uːəɫ]
mule [ˈmjuːəɫ] fuel [ˈfjuːəɫ]
foul [ˈfaʊəɫ] vowel [ˈvaʊəɫ]
tile [ˈtaɪəɫ] dial [ˈd̥aɪəɫ]
oil [ˈɔɪəɫ] royal [ˈrɔɪəɫ]
fail [ˈfeɪəɫ] portrayal [pʰɔrˈt̠ɹ̥eɪəɫ]
stole [ˈstoʊəɫ] bestowal [b̥ɪˈstoʊəɫ]
In all of the above examples, glide insertion may apply, especially in the words on the right. The result will be [ˈruːwəɫ, ˈkɹ̥uːwəɫ, ˈtʰaɪjəl], etc.
12.4 The Palato-Velar Approximant: /r/
GA /r/ is articulated in a variety of ways. However, the sounds that are produced by these various articulations are so similar as to be virtually indistinguishable. Indeed, from an auditory point of view, it makes sense to speak of a single /r/-like quality that is shared by all /r/s regardless of the particular articulatory gesture underlying them.
This r-sound is heard not only before and after vowels, as in red, arise, barrel, far, board, but also in words like bird, fur, murmur, where it occurs as a monophthong which functions as syllable bearer.
Since, in all these cases, we are dealing with what is essentially a single sound quality, we have chosen to use a single symbol /r/ to represent this sound, so that red is /rɛd/, bird /bɪrd/ and banner /bænr/. The problem with this notation is that it fails to reflect the different status of the r-sounds. In red, /r/ functions as a consonant in onset position. In words like bird, banner, it functions as a vowel, i.e. the peak of a syllable. (So some people might argue that instead of /ˈmɜrdər/ for murder we ought to transcribe /ˈmr̩dr̩/.)
What all the articulatorily diverse but auditorily similar /r/’s have in common is that the speech tract is constricted or narrowed at two points. One constriction is formed in the palatal area, the second is due to a retraction of the tongue root toward the back wall of the pharynx.
The narrowing in the palatal area is brought about in one of two ways. In the majority of cases, the tip is held low and the body of the tongue is bunched backward and arched, so that the area where the front and back of the tongue meet is closest to that part of the palate where the hard and soft palate meet. There may be a dip between this narrowing and the one formed by the tongue root. This type of /r/ may be called the bunched or palato-velar /r/.
The narrowing in the palatal area may also arise as a result of a raising of the tip and blade of the tongue in the direction of the area immediately behind the alveolar ridge. This post-alveolar /r/ may, but need not be retroflex, i.e. articulated with the tip of the tongue curled up and pointing backward, and is usually accompanied with lip rounding.
In unaccented syllables, /r/ may be lost when another /r/ follows as a result of r-dissimilation. Examples are surprise [səˈpʰɹ̥aɪz̥], particular [pʰəˈtʰɪkj̥ələr], govemor [ˈg̥ʌvənər] and thermometer [θəˈmɑːməɾər].
Advice for learners:
Remember that GA /r/, in most contexts, is an approximant. This means that the alveolar or uvular contacts that are used for /r/ are undesirable. The easiest articulation to learn is probably the post-alveolar one, involving a raising of the tip and blade in the direction of the post-alveolar region. You can start from a voiced fricative like [z] or English [ʒ]. Now slowly move the contact backward until the tip slides off the alveolar ridge and is tilted slightly backward, making sure it does not touch the palate.
The bunched variety, which is more common in the pronunciation of native American speakers, is more difficult to learn, except possibly for those who use a palatal /r/. Since /r/ is essentially a vocalic sound, detailed descriptions of the configuration of the articulators are not generally very helpful because they are extremely difficult to imitate in the absence of any major articulatory contact. Instead, try to rely on your ability to imitate the sound by ear, and don’t worry too much about exactly how you make the sound. After all, native speakers also use different articulatory gestures to produce GA /r/.
12.4.1 voiced uvular roll and r-sound /r/
Unlike GA, many students cannot be said to have a single r-sound. Even within AN, /r/ varies considerably in articulatory as well as auditory terms. The situation is further complicated by the fact that one and the same speaker may well use several types of /r/ in what may appear to be a rather haphazard fashion.
Initially and between vowels, many speakers use a voiced alveolar flap [ɾ], or rolled [r]. Others use a voiced uvular roll ([ʀ]), a uvular approximant ([ʁ̞]), or a voiced uvular fricative ([ʁ]). Before consonants and silence, a voiceless uvular fricative may be heard, which, incidentally, is identical to the realization of non-Southern AN /x/, so that bocht – bord and maagd – Maart may be homophones.
Increasingly, a palatal approximant is heard, which is similar to [j], and may lead to homophony in pairs like koortje – kooitje, Maart – maait, paartijd – paaitijd. lts high status appears from its rapid spread among young female, and to a lesser extent, male middle class speakers. In view of the similarity of palatal /r/ with GA /r/, it would be possible, though very speculative, to argue that its popularity is a reflection of the high status accorded to American English by these speakers.
12.5 The Palatal Approximant /j/; Yod-Dropping
In many contexts, historic /juː/ became /uː/ in most varieties of English, a change known as yod-dropping.
Since you may be familiar with British varieties of English, it should be pointed out GA /juː/ does not occur after alveolars, so that tune, dune, suit, assume, presume, lunar, new, student have no /j/.
Following other consonants, /j/ remains, as in British English: [pjuːni,, bjuːti, mjuːz, vjuː, (etc.)] for puny, beauty, muse, view, (/ɛtˈsɛtərə/).
Yod-dropping did not affect syllables after strong syllables, as illustrated by /ˈmɛnjuː:/ menu and /ˈvæljuː/ value.
After /t, d/ in weak syllables, /j/ tends to be assimilated, as in /ˈsætjəˈreɪʃn̩/ saturation, situation, education, which have /t͡ʃə/ and /d͡ʒə/, respectively. The same applies to initial /j/ in the weak form of the pronoun you, as in don’t you (or, informally, doncha) [ˈdoʊnt͡ʃə] and did you [ˈdɪd͡ʒə].
12.6 Syllabic Consonants
The sonorants /l, n/ and, to a lesser extent, /m, ŋ/ may be syllabic, i.e. form the peak of a syllable, as in middle /ˈmɪdl̩/, wooden /wʊdn̩/, open, bacon (frequently /ˈoʊpən/, /ˈbeɪkən/ , but also /ˈoʊpn̩, ˈbeɪkŋ̩/). Syllabic /l, n, m, ŋ/ frequently alternate with /əl, ən, əm, əŋ/, as in whistle, station, prism, Golden Gate.
12.6.1 Syllabic Consonants
Although syllabic /l, m, n, ŋ/ are quite common in conversational-style in words and phrases like tafel, openbaar, op z’n tenen’, op zn kant’, their occurrence is more restricted than in American English.
12.6.2 Syllabic /l/
After /t,d/, as in subtle, muddle, the voiced tap is laterally released, unless /l/ is vocalized. The brief closure required for the voiced tap is turned into a lateral contact by prolonging the tip contact and allowing the air to escape abruptly along the sides of the tongue. This is called lateral plosion.
Words like babbling cuddly, settler, startling, and wobbly, may be pronounced in two ways: with two or with three syllables. The latter pronunciation, in which /l/ is syllabic, would seem to be more common. Note that in this case we would expect the /t/ in startling and settler to be voiced and flapped rather than glottaled and /l/ to be dark rather than clear. Compare [ˈstɑɾʔlɪŋ] and [ˈstɑrɾɫɪŋ].
babbling – [ˈbæblɪŋ]
babbling – [ˈbæbəɫɪŋ]
cuddly – [ˈkʰʌdli]
cuddly – [ˈkʰʌɾəɫi]
settler – [ˈsɛʔlər]
settler – [ˈsɛɾəɫər]
startling – [ˈstɑɾʔlɪŋ]
startling – [ˈstɑrɾəɫɪŋ]
wobbly – [ˈwɑːbli]
wobbly – [ˈwɑːbəɫi]
12.6.3 Syllabic /n/
Syllabic /n/ occurs after GA /t,d/, as in cotton, certain, wooden, garden. Recall that the preceding /t/ is normally glottaled in these words, e.g. [ˈkɑːʔn̩] and [ˈsɜrʔn̩].
The /d/ in cotton, certain, wooden, garden is pronounced with nasal plosion: the alveolar closure is maintained throughout the [dn]-sequence, and the release of the stop is effected by a sudden lowering of the soft palate, so that the air compressed behind the closure escapes abruptly through the nasal cavity. After other stops, as in open, urban, taken, organ, as well as after approximants, as in melon, lemon, common, singin’, foreign, /ən/ is normally heard.
After fricatives, usage of syllabic /n/ varies. Some speakers clearly pronounce a transitional vowel in words like station, cousin, dozen, and especially after the fricative /v/, as in seven.
station – /ˈsteɪʃən/
cousin – /ˈkʌzɪn /
dozen – /ˈdʌzən/
seven – /ˈsɛvən/
Others use syllabic /n/, as reflected in the transcription that is used here: /ˈsteɪʃn̩, ˈsɛvn̩/.
station – /ˈsteɪʃn̩/
seven – /ˈsɛvn̩/
A preceding weak syllable inhibits some of these reductions. For instance, instead of [ʔn̩], we get /tn̩/ (nasal plosion) in words like skeleton, Puritan, Sheraton.
skeleton – /ˈskɛlətən/
Puritan – /ˈpjʊrətən/
And when /n/ is followed by /t/, it does not even become syllabic, as illustrated by accident and competent /ˈæksədənt, ˈkɑːmpətənt/.
accident – /ˈæksədənt/
competent – /ˈkɑːmpətənt/
Similarly, /ən/ is more common than /n̩/ after fricatives when a weak syllable precedes as in elephant, Sullivan.
elephant – / ˈɛləfənt /
Sullivan – /ˈsʌləvən/
After /s/, as in innocent, Madison, syllabic /n̩/ is always used, though: /ˈɪnəsn̩t, ˈmædəsn̩/.
innocent – /ˈɪnəsn̩t/
Madison – /ˈmædəsn̩/
In consonant + plosive sequences, /ən/ is common. So, we would expect /ən/ in words like Stockton, Ashton, Ogden, London, sunken, dampen.
The consonant that precedes the plosive must not be approximant or vocalic: if the first of two preceding consonants is /r/, as in carton, harden, or vocalized /l/, as in moulten, golden, or elided /n/, as in mountain, sentence, syllabic /n/ is the rule, since we are essentially dealing with a single consonant rather than with two. Thus, [ˈkɑrʔn̩, ˈhɑrdn̩, ˈpɜrsn̩, ˈmoʊɫʔn̩, ˈmãʊ̃ʔn̩, ˈsɛ̃ʔn̩s].
Stockton – /ˈstɑːktən /
Ashton – /ˈæʃtən/
carton – /ˈkɑrʔn̩ /
Here are some examples to illustrate the use of syllabic /n/ and the realization of a preceding /t/ or /d/:
Glottaled t + syllabic /n/ button, certain, important, Hilton, fountain
button – / ˈbʌʔn̩/
Nasal plosion of /t,d/ sudden, ardent, skeleton, Sheraton, harridan
sudden – /ˈsʌdn̩/
Non-syllabic /n/ pagan, demon, commitment, militant, abandon
pagan – /ˈpeɪgən/
Word-medially, /ən/ alternates with /n̩/ after fricatives, as in reasonable, passionate: /ˈriːzənəbl̩ ~ ˈriːzn̩əbl̩/; /ˈpæʃənət ~ ˈpæʃn̩ət/. In more rapid styles, /n̩/ is apt to lose its syllabic status, so that reasonable is pronounced /ˈriːznəbl/ and personally /ˈpɜrsnəli/. After /t,d/, syllabic /n/ is the rule, as in matinee, botany, ordinary, hedonist, as well as in adjectives and present participles like fattening, maddening, fastening.
On the other hand, nouns ending in –ing have non-syllabic /n/. Compare lightning (noun) /ˈlaɪtnɪŋ/ and lightening (verb) /ˈlaɪtn̩ɪŋ/, fastening (noun) /ˈfæsnɪŋ/ and fastening (verb) /ˈfæsn̩ɪŋ/.
lightning – /ˈlaɪtnɪŋ/
lightening – /ˈlaɪtn̩ɪŋ/
Word-initial /ən/ may also become /n̩/ after fricatives and /t,d/, as in he had an affair /ˈhædn̩/, it’s an error /ɪtsn̩/, a hundred and ten /ˈhʌndrədn̩/, it’s enormous /ɪtsn̩ˈɔrməs/. Syllabic /n/ may also represent the suffix –ing, as in huntin’, eatin’, shootin’ [ˈhʌ̃ʔn̩, ˈiːʔn̩, ˈʃuːʔn̩].
he had an affair – /hi hædn̩ əˈfɛr/
it’s an error – /ɪtsn̩ ˈɛrər/
fastening – /ˈfæsnɪŋ/
fastening – /ˈfæsn̩ɪŋ/
huntin’ – /ˈhʌ̃ʔn̩ /
eatin’ – /ˈiːʔn̩ /
Advice for learners:
For syllabic /l/, concentrate on the context after /t,d/, as in little, middle, where you should practice lateral plosion: [ˈɪɾl̩], etc.
For syllabic /n/, concentrate on /tn/ and /dn/ sequences, as in written, hidden. Remember that /t/ tends to be glottaled (or glottalized) before syllabic /n/ if the preceding vowel is accented. Note that wouldn’t /ˈwʊdnt/ has nasal plosion for /d/.
After fricatives, try to use /n̩/ rather than /ən/, especially after /s,z,ʃ/, as in mustn’t, dozen, patient.
12.6.4 Syllabic /m/ and /ŋ/
In words like prism, bosom, rhythm, syllabic /m/ alternates with /əm/. Syllabic /m/ and /ŋ/ may also arise as a result of assimilation to a following consonant, as in I can go /ˈaɪkŋ ˈgoʊ/. Pronunciations like ribbon /ˈrɪbm̩/, seven /ˈsɛbm/, reckon /ˈrɛkŋ/ are less likely.