A syllable may be defined both phonetically and phonemically. Phonetically (that is, in relation to the way we produce it and the way it sounds), a syllable always contains a center which is relatively loud and which involves a little or no obstruction  to air flow. Such a center may be itself constitute a syllable, but sounds which  produce greater obstruction to the air flow and which  are less loud often precede and/or follow it.

Some terms that are often used to describe the syllable are PEAK for the center, ONSET for less loud sounds preceding the peak, and CODA for less loud sounds following the peak. The peak and coda together are called RHYME. Syllables may have a coda: ‘am’ /æm/, ‘ought’ /o:t/, ‘ease’ /i:z/. some others may have both onset and coda: ‘run’ /rʌn/, ‘sat’ /sæt/, ‘feel’ /fi:l/. a diagram commonly used to illustrate the structure of the syllable is shown in figure 8.1. Onset and coda are shown in parentheses because they are optional elements.

Initial clusters of two-members

Initial clusters in English consists of either two consonants or three. Two consonant clusters may consist of a pre-initial plus an initial, or of an initial plus a post-initial. Two consonant clusters beginning with the pre-initial /s/ may have as initials voiceless plosives /p,t,k/, the fricative /f/ or the nasals /m,n/. For examples:

/sp/     spill                              /sf/     sphere

/st/     still                              /sm/    smoke

/sk/     skill                              /sn/     snail

When /s/ occurs with a set of consonants /l, w, j/, however, it is analysed as an initial whereas the set of cononants is called post-initials. Examples of /s/ as an initial plus a spost-initial are :

/sl/      slow                             /sw/    swat

/sj/      suit

Initial clusters of three-members

Three member initial clusters in English consist of an /s/, a voiceless plosive, and a post-initial. Examples are :

/spl/    split                 /str/    strip                 /skl/    sclerosis

/spr/   spring               /stj/    stupid              /skr/   scream

/spj/    spew                                                    /skw/  squirm

                                                                                    /skj/    skew

Final clusters of two-members

In final position, clusters in English word can consist of two, three or four members. Two-member clusters can consist of pre-final (m, n, ŋ/ l, s) plus a final (any consonan except (h, r , w, j) or of a final plus a post-final  (s, z, t, d, θ). Examples of the first type (pre-final plus final) are:

/mp/   damp               /nt/     mint                 /ŋk/    bank

/lt/      halt                  //sk/   desk

Examples of the second type of final-two member cluster (final plus post final) are:

/ts/     cats                  /gs/     bags                 /kt/     stacked

/gd/    sagged              /dθ/    width

Final clusters of three-members

Three member clusters in final position can consist of a pre-final, a final and a post-final or of a final and two post-finals. Examples of the first type are:

/ndz/  bands               / ŋkt/  ranked                         /lfθ/    twelfth θ

/lpt/    yelped

Examples of the second type are:

/fts/    lifts                  /kst/   boxed               /pst/   lapsed

Final clusters of four-members

Final clusters of four-members can consist of a pre-final, a final and two post-finals, or of a final and three-finals.

/lfθs/   twelfths                       /mpts/            prompts

Examples of the second type are:

/ksθs/             sixths                           /ksts/  texts

We may describe the English syllable as having the following maximum phonological structure.


We learnt that consonants and vowels can be described both from a phonetic point of view, i.e. in terms of how they are produced, and from a phonological point of view, i.e. in terms of where they occur. The same is true of the syllable. Phonetically, a syllable can be described as having a centre, also called peak or nucleuswhich is produced with little or no obstruction of air, and is therefore usually formed by a vowel (either a monophthong or a diphthong). The minimal syllable, then, is typically a single, isolated vowel, as in the words are /ɑ:/, err /3:/, and / /ai/. The few consonants that can occur in isolation, such as the interjections mm /m/ (used to express agreement) and sh l(used to ask for silence), are not regarded as minimal syllables by all linguists.

In many syllables, the centre is preceded by an onset [Kopf], which is produced with greater obstruction of air, and is therefore always formed by one or more consonants. Such syllables are exemplified by words like bar /ba:/, stir /stɜ:/, and my /mai/. A syllable that ends in a vowel, i.e. one that ends with the centre, is commonly referred to as an open syllable [offene Silbe]. In many other syllables, there is no onset, but the centre is followed by a coda [Koda], which is also produced with greater obstruction of air, and is therefore also formed by one or more consonants. Such syllables are exemplified by words like art /a:t/, urge /ɜ:/, and ice /ais/. Most syllables, however, have both an onset and a coda, like bath /ba:θ/,perk /pɜ:k/, and mine /main/. (If they do not, we also speak of a zero onset and a zero coda.) A syllable that ends in a consonant, i.e. o ne that ends with a coda – irrespective of whether it has an onset or not – is commonly r e ferred to as a closed syllable [gescblossene Silbe]. It is sometimes also termed a checked syllable, and the vowel forming the centre is then a checked vowel. The centre and the coda (if there is one) together account for the rhyming potential of a syllable, as can be illustrated by word pairs like mine/fine and err Istir, and they have therefore collectively been referred to as the rhyme.

From the point of view of auditory phonetics, which was defined in Lesson One as investigating the perception of speech sounds by the listener, the centre of a syllable is perceived to be more prominent than the margins, i.e. onset and coda. The prominence of a syllable centre (and the prominence of a particular sound in relation to its surrounding sounds in general) can be attributed mainly to a combination of greater loudness, higher (or sometimes lower) pitch, and greater duration, but also to sound quality: Each phoneme has, besides the absolute loudness of its phonetic realisation, an intrinsic, relative loudness (or “carrying-power”), referred to as its sonority, and, according to a common sonority hierarchy, vowels are more sonorous (i.e. they “carry further”) than consonants.

On the basis of what we have just learnt, we can say that the word knowing consists of two syllables: the open syllable /nao/, ending in a diphthong, and the closed syllable /in/, without an onset. Here, syllable boundary and morpheme boundary clearly coincide (in other words, syllables and morphemes in this example are identical), but this is not usually the case. The word standing, for example, is commonly divided into the syllables /staen/ and /din/ (although its morphemes are stand and -ing). This analysis is probably guided by the desire to have two evenly balanced syllables, namely two closed syllables with an onset, but  the word could theoretically also be divided into the syllables /staend/ and /in/ if that were “felt” to be more appropriate. (Incidentally, syllable boundaries and morpheme boundaries are the two competing criteria for hyphenating a word at the end of a line, with American English favouring the former, and British English favouring the latter.) The example shows, then, that the division of words into syllables, referred to as syllabification or syllabication, is based to a considerable extent on intuition – certainly to a greater extent than is desirable for a science that aims to be exact and objective.

The difficulty of determining the exact boundaries between consecutive syllables can be further exemplified by words like master and extra. Most English speakers feel that master consists of two syllables, but should the word be divided as /ma:-sta/ (with two evenly balanced open syllables in non-rhotic accents), /ma:s-tar/ (with two evenly balanced closed syllables in rhotic accents), or maybe even /ma:st-a(r)/ (because the word mast consists of one syllable when it occurs alone)? The word extra is also felt to have two syllables, but should it be divided as /ek-stra/, /eks-tra/, /ekst-ra/, or /ekstr-a/? While it is simply not possible to say which of these syllabifications is “correct”, /ek-stra/ and /eks-tra/ are certainly the most common.


Concerning the possible positions of phonemes within a syllable, we already know that one or more consonants can occur at each margin, and that either a monophthong or a diphthong usually occurs at the centre, which could mean in the middle, but also at the beginning (if there is no onset), at the end (if there is no coda), or alone (if there is neither onset nor coda). As far as the possible positions of particular phonemes and their combinations are concerned, there are many more restrictions than can be outlined here. In this section, we shall therefore look at the particular phonemes and phoneme combinations that can occur at the beginning of a word, or more precisely of a word-initial syllable, in some detail, and shall touch on those that can occur at the end only briefly.

A word begins either with the centre or with the onset of its first syllable. If it begins with the centre (as in art / ɑ:t/), that centre can be formed by any vowel (although /o/ occurs rarely in word-initial position). If it begins with the onset, that onset can be formed either by one consonant or by a cluster consisting of two or three consonants. If the onset is formed by one consonant (as in bar /bɑ:/), that consonant can be any consonant except / ŋ / (although / ʒ / is also rare word-initially). If the onset is formed by two consonants (as in stir /st3:/), the two-consonant cluster must be one of the following: /pl, pr, pj, bl, br, bj, tr, tj, tw, dr, dj, dw, kl, kr, kj, kw, gl, gr, mj, nj, fl, fr, fj, vj, θr, θw, sp, st, sk, sm, sn, sf, si, sj, sw, ∫r, hj, lj/, plus a few unusual clusters in some proper names, in some archaic or otherwise rare words (such as /vr/ in vroom), and in certain pronunciation variants of words (such as /sr/ in syringe when that word is pronounced /srindʒ/ instead of /sirindʒ). We note that, with the exception of clusters beginning with /s/, the second (or last) element of a word-initial two-consonant cluster is always one of the four frictionless continuants, /l, r, j , w/. If the onset is formed by three consonants (as in street /stri.t/), the three-consonant cluster must be one of the following: /spl, spr, spj, str, stj, ski, skr, skj, skw/. Here, we note that the first element is always /s/, the second element is always one of the three fortis plosives, /p, t, k/, and the third (or last) element is again always one of the four frictionless continuants, /l, r, j , w/. No English word, and no English syllable, for that matter, begins with more than three consonants and many consonant combinations, such as /tl/, /fs/, and /spm/, are not possible in word-initial position.

A word ends either with the centre or with the coda of its last syllable. If it ends with the centre (as in my /mai/), that centre can be formed by any vowel except /eæʌɒ / , or it can be formed by a syllabic consonant, which will be explained in the next section. If it ends with the coda, that coda can be formed either by one consonant or by a cluster consisting of up to four consonants. If the coda is formed by one consonant (as in mine /main/), that consonant can be any consonant except /h/, the semi-vowels, /j, w/, and in non-rhotic accents Ixl. If the coda is formed by two, three, or four consonants (as in bc[p_ /help/, next /nekst/, and glimpsed /glimpst/), there are numerous restrictions similar to those applying to clusters at the beginning of a word-initial syllable. No English word, and, again, no English syllable, ends with more than four consonants, and, as before, many consonant combinations are not possible in word-final position.

Phonologically, the English syllable can thus be described (with a capital ‘C’ representing a consonant, and a capital ‘V’ representing a vowel) as having the maximal structure CCCVCCCC (as in strengths /streŋgθs/), the minimal structure V (as in are / ɑ:/), or any structure in between, such as CCVC (as in stop /stop/), CVCC (as in cats /kæts/), and CCCVCC (as in streets /stri:ts/).


In the previous sections, we said that the centre of a syllable is “usually” formed by a vowel. This is true, without exception, in words consisting of only one syllable, like art /o:t/, my /mai/, help /help/, and most of the other example words given in this lesson so far. As the tentative wording suggests, however, there are syllables whose centre is not formed by a vowel, but by a consonant instead. Such syllables contain no vowel at all, and the consonant forming the centre is termed syllabic consonant. This is the case in some words consisting of two or more syllables, as we shall see shortly.

Before we proceed, however, a brief note on terminology: A word that consists of a single syllable is referred to as a monosyllabic word, or simply as a monosyllable. One that consists of two syllables, like clever /klev- Ə / and delay /di-lei/, is referred to as a disyllabic or bisyllabic word, or as a disyllable or bisyllable, and one that consists of three syllables, like compulsive /kƏm-pʌl-srv/ and delicious /di-li∫-Əs/, is referred to as a trisyllabic word, or as a trisyllable. Less specifically, a word that consists of two or more syllables is also called a polysyllabic word, or a polysyllable.

The word student /stju:dnt/, then, could theoretically be analysed as a monosyllable with a vowel as its centre, a three-consonant cluster as its onset, and another three-consonant cluster as its coda. Most English speakers feel that student consists of two syllables, however, and the word is therefore commonly analysed as a disyllable with /stju:d/ as the first syllable, and /nt/ as the second, and with /u:/ and /n/ as the syllable centres. One just ification for such an analysis might be that student is pronounced /stjuidant/, with a schwa, when it is pronounced very slowly, in which case it is the schwa that forms the centre of the second syllable. In “normal” speech, when the schwa is omitted, the centre simply shifts to the following sound.

In more general terms, we could say that a syllabic consonant can occur in certain phonetic environments where, in very slow speech, there would be a schwa, or where we imagine there could be a schwa, as a syllable centre. The time needed to pronounce the (real or imaginary) schwa is then added to the duration of the following consonant, thus transforming that consonant into a syllabic consonant. As a syllabic consonant always forms the centre of a syllable, it has the phonological characteristics of a vowel, but, of course, it retains the phonetic characteristics of a consonant.

A syllabic consonant and a corresponding non-syllabic consonant cannot usually distinguish meaning, or differentiate words, which means that they must be regarded as allophones of the same phoneme. We learnt in Lesson Five that allophones are not recorded in a phonemic transcription, and the transcription /stju:dnt/ is therefore correct and sufficient for that degree of accuracy. In a phonetic transcription proper, on the other hand, and, to a very limited extent, in a broad phonetic transcription, allophones are recorded. In those types of transcription, a syllabic consonant is indicated by a small vertical line, [, ], under the relevant symbol, as in [stju:dnt]. There are five consonants that can be transformed into syllabic consonants. They are, roughly in order of frequency: /l, n, m, n, r/.

The syllabic [1] also occurs when it is represented in the spelling by the letter sequences <al> or <el> at the end of a word, and is preceded by a consonant, as in proposal [prƏpƏuzl], pedal [ped-ld], and shovel [∫ʌv-I ] . A S before, it forms the centre of the syllable in which it occurs, and it usually remains syllabic even when a suffix is added, as in proposals [pra-pau-zlz], pedalled [ped-ld], and shovelling [∫ʌv-l-irj]. In most cases, the syllabic [1] is obligatory. In other words, pronouncing the word table as [teibal], with a schwa, would be plain wrong, except perhaps in very slow speech. It is optional mainly in those cases where it is followed by a suffix with a vowel, as was seen above, and in some less common or more technical words, such as acquittal [akwitl / akwital], missal [misl / misƏl], and shrapnel [∫raepnl / ∫raepnal].

(2) [n]. The syllabic [n] often occurs when Iwl is preceded by a plosive or fricative (not by an affricate) in unstressed syllables, but not usually in word-initial position. It is especially frequent when preceded by an alveolar plosive or fricative, as in button [bʌtn] and horizon [horaizn], and in contracted negations like hadn’t [haednt] and isn’t [iznt]. It is also frequent when preceded by a labiodental fricative, but here the /n/ can also be pronounced as [an]. The words hyphen and seven, for example, are usually pronounced [haifn] and [sevn], but less commonly also [haifan] and [sevan]. Otherwise, however, the syllabic [n] is largely obligatory in this environment. Pronouncing a schwa here would usually sound odd or, at best, overcareful.

In other environments, the syllabic [n] is relatively rare. After bilabial and velar consonants, [n] and [on] seem to be virtually interchangeable, or in free variation, as in happen [haepn / haepan], ribbon [nbn / nban], and thicken [θikn /θikƏn]. After velar consonants, when it is represented in the spelling by the letter sequences <an> or <on>, as in wagon, and after the phoneme sequence nasal or /s/ or /l/ plus plosive, as in London, the syllabic [n] may be acceptable, resulting in such pronunciations as [waegn] and [lAiidn], but pronunciations with [an], as in [wægƏn] and [lʌndƏn], are certainly much more common.

(3,4) [m, n]. Although the syllabic [m] and [ŋ] are not uncommon, they can occur only as a result of phonetic processes such as assimilation and elision, which will be discussed in detail in Lesson Nine. When they occur in words like happen and thicken, for example, these words are pronounced [haepm] and [Bikn], but, as we saw above, the pronunciations [haepn / haepan] and [0ikn / 9ikan] are also possible.

(5) [r]. While the syllabic [r] is very common in many rhotic accents, it is rare in non-rhotic accents, where the hi phoneme has disappeared almost entirely (except before a vowel). The words particular and perhaps, for example, are often pronounced [prtikjalr] and [prhaeps] in General American English whereas in RP they are usually pronounced [patikjala] and [pahaeps]. There are only two environments in which the syllabic [r] can occur in RP, and even in those environments it is usually optional. Firstly, it can occur when III is preceded by one consonant (and followed by a vowel) in unstressed syllables, as in flattery [flaetri] and watering [wo:triŋ]. Here, the syllabic [r] is usually interchangeable with [ar], and the example words would then be pronounced [flaetƏri] and [wa:tƏriŋ]. Secondly, the syllabic [r] can occur when hi is preceded by two or more consonants (and followed by a vowel) in unstressed syllables, as in history [histri] and wanderer [wDndra]. Here, the syllabic [r] is usually interchangeable with a non-syllabic [r], without a schwa. The example words would then be pronounced [histri] and [wDndra], the only difference being that a syllabic consonant always has a greater duration than its non-syllabic counterpart.

As was mentioned earlier in this section, a syllabic consonant and a corresponding non-syllabic consonant cannot usually distinguish meaning, i.e. there are few minimal pairs in which they appear to be contrasting sounds. One such minimal pair seemed to be coddling [kɒdlŋ] / codling [kDdlŋ], but we said that coddling is occasionally also pronounced [kDdlŋ], in which case the two words have the same pronunciation (and are then termed homophones). All in all, there is not enough reason to regard a syllabic consonant and its non-syllabic counterpart as two separate phonemes. They are, as mentioned before, allophones of the same phoneme.

Finally, it should be noted that neither the choice between the various pronunciation variants discussed here nor their exact pronunciations are usually as clear-cut as the description of the environments and a neat transcription might suggest. The word veteran, for example, can be pronounced [vetaran] with [ar], [vetran] with a syllabic [r], less commonly also [vetrƏn] with a non-syllabic [r], [vetarn] with a syllabic [n], possibly even [vetrn] with two consecutive syllabic consonants, or some way in between. When we transcribe spoken language, the choice is therefore often arbitrary, and the representation through IPA symbols is often only approximate.

Stressed and unstressed syllables vs. strong and weak syllables

We learnt earlier in this lesson that, at the level of auditory phonetics, the prominence of a sound can be attributed mainly to a combination of loudness, pitch, duration, and sound quality. Just as there are more prominent and less prominent sounds within a syllable (the most prominent being the syllable centre), there are more prominent and less prominent syllables within a (polysyllabic) word. Furthermore, we learnt in Lesson Two that, at the level of articulatory phonetics, the same four features mentioned above – loudness, pitch, duration, and sound quality – are also the main components of stress. Prominence in the perception of speech, then, results from stress in its production, and we therefore speak of a stressed syllable and an unstressed syllable. These terms seem to be largely self-explanatory (which is the reason why we have been able to use them from Lesson Four onwards), but when we look at stressed and unstressed syllables more closely, we see that they need more explanation than their simple labels might suggest. One aspect in particular that must be explained is the sound quality of the vowels in stressed and unstressed syllables