Syllabification is the separation of a word into syllables, whether spoken or written. In most languages, the actually spoken syllables are the basis of syllabification in writing too. Due to the very weak correspondence between sounds and letters in the spelling of modern English, for example, written syllabification in English has to be based mostly on etymological i.e. morphological instead of phonetic principles. English written syllables therefore do not correspond to the actually spoken syllables of the living language. Phonotactic rules determine which sounds are allowed or disallowed in each part of the syllable.
English allows very complicated syllables; syllables may begin with up to three consonants (as in
strength), and occasionally end with as many as four The linking of a word-final consonant to a vowel beginning the word immediately following it forms a regular part of the phonetics of some languages, including Spanish, Hungarian, and Turkish. Thus, in Spanish, the phrase ('the men') is pronounced , Hungarian ('the human') as , and Turkish ('I hated it') as . In Italian, a final sound can be moved to the next syllable in enchainement, sometimes with a gemination: e.g., ('I've never had any of them') is broken into syllables as and ('I go there and she does as well') is realized as . A related phenomenon, called consonant mutation, is found in the Celtic languages like Irish and Welsh, whereby unwritten (but historical) final consonants affect the initial consonant of the following word.
Ambisyllabicity There can be disagreement about the location of some divisions between syllables in spoken language. The problems of dealing with such cases have been most commonly discussed with relation to English. In the case of a word such as
hurry, the division may be or , neither of which seems a satisfactory analysis for a
non-rhotic accent such as in
Received Pronunciation: results in a syllable-final , which is not normally found, while gives a syllable-final short stressed vowel, which is also non-occurring. Arguments can be made in favour of one solution or the other: A general rule has been proposed that states that "Subject to certain conditions [...] consonants are syllabified with the more strongly stressed of two flanking syllables", while many other phonologists prefer to divide syllables with the consonant or consonants attached to the following syllable wherever possible. However, an alternative that has received some support is to treat an intervocalic consonant as
ambisyllabic, i.e. belonging both to the preceding and to the following syllable: . This is discussed in more detail in . Another example could be the nasal velar written in Galician, so the word could be divided or .
Onset The
onset (also known as
anlaut) is the consonant sound or sounds at the beginning of a syllable, occurring before the
nucleus. Most syllables have an onset. Syllables without an onset may be said to have an
empty,
zero or
null onset – that is, nothing where the onset would be.
Onset cluster Some languages restrict onsets to be only a single consonant, while others allow multiconsonant onsets according to various rules. For example, in English, onsets such as
pr-,
pl- and
tr- are possible but
tl- is not, and
sk- is possible but
ks- is not. In
Greek, however, both
ks- and
tl- are possible onsets, while contrarily in
Classical Arabic no multiconsonant onsets are allowed at all. Onset clusters often follow the
sonority sequencing principle, that is, onsets with increasing sonority (/kl/) are usually preferred to ones with a plateau (/ll/) and even stronger preferred to decreasing sonority (/lk/); however, many languages have counterexamples to this tendency.
Null onset Some languages forbid
null onsets. In these languages, words beginning in a vowel, like the English word
at, are impossible. This is less strange than it may appear at first, as most such languages allow syllables to begin with a phonemic
glottal stop (the sound in the middle of English
uh-oh or, in some dialects, the double T in
button, represented in the
IPA as ). In English, a word that begins with a vowel may be pronounced with an
epenthetic glottal stop when following a pause, though the glottal stop may not be a
phoneme in the language. Few languages make a phonemic distinction between a word beginning with a vowel and a word beginning with a glottal stop followed by a vowel, since the distinction will generally only be audible following another word. However,
Maltese and some
Polynesian languages do make such a distinction, as in
Hawaiian ('fire') and / ← ('tuna') and Maltese ←
Arabic and Maltese ← Arabic .
Ashkenazi and
Sephardi Hebrew may commonly ignore , and , and Arabic forbids empty onsets. The names
Israel,
Abel,
Abraham,
Omar,
Abdullah, and
Iraq appear not to have onsets in the first syllable, but in the original Hebrew and Arabic forms they actually begin with various consonants: the semivowel in , the glottal fricative in , the glottal stop in , or the pharyngeal fricative in , , and . Conversely, the
Arrernte language of central Australia may prohibit onsets altogether; if so, all syllables have the
underlying shape VC(C). The difference between a syllable with a null onset and one beginning with a glottal stop is often purely a difference of
phonological analysis, rather than the actual pronunciation of the syllable. In some cases, the pronunciation of a (putatively) vowel-initial word when following another word – particularly, whether or not a glottal stop is inserted – indicates whether the word should be considered to have a null onset. For example, many
Romance languages such as
Spanish never insert such a glottal stop, while
English does so only some of the time, depending on factors such as conversation speed; in both cases, this suggests that the words in question are truly vowel-initial. But there are exceptions here, too. For example, standard
German (excluding many southern accents) and
Arabic both require that a glottal stop be inserted between a word and a following, putatively vowel-initial word. Yet such words are perceived to begin with a vowel in German but a glottal stop in Arabic. The reason for this has to do with other properties of the two languages. For example, a glottal stop does not occur in other situations in German, e.g. before a consonant or at the end of word. On the other hand, in Arabic, not only does a glottal stop occur in such situations (e.g. Classical "he asked", "opinion", "light"), but it occurs in alternations that are clearly indicative of its phonemic status (cf. Classical "writer" vs. /mak "written", "eater" vs. "eaten"). In other words, while the glottal stop is predictable in German (inserted only if a stressed syllable would otherwise begin with a vowel), the same sound is a regular consonantal phoneme in Arabic. The status of this consonant in the respective writing systems corresponds to this difference: there is no reflex of the glottal stop in
German orthography, but there is a letter for it in the Arabic alphabet (
Hamza (ء)). The writing system of a language may not correspond with the phonological analysis of the language in terms of its handling of (potentially) null onsets. For example, in some languages written in the
Latin alphabet, an initial glottal stop is left unwritten (see the German example); on the other hand, some languages written using non-Latin alphabets such as
abjads and
abugidas have a special
zero consonant to represent a null onset. As an example, in
Hangul, the alphabet of the
Korean language, a null onset is represented with ㅇ at the left or top section of a
grapheme, as in "station", pronounced
yeok, where the
diphthong yeo is the nucleus and
k is the coda.
Nucleus The
nucleus is usually the vowel in the middle of a syllable. Generally, every syllable requires a nucleus (sometimes called the
peak), and the minimal syllable consists only of a nucleus, as in the English words "eye" or "owe". The syllable nucleus is usually a vowel, in the form of a
monophthong,
diphthong, or
triphthong, but sometimes is a
syllabic consonant. It has been suggested that if a language allows a type of consonants to occur in syllable nucleus, it will also allow all the consonant types that are higher in
sonority, that is, a language with syllabic fricatives would necessarily also have syllabic nasals, and that syllabic obstruents will be much more rare than liquids; both statements have been shown to be false. In most
Germanic languages,
lax vowels can occur only in closed syllables. Therefore, these vowels are also called
checked vowels, as opposed to the tense vowels that are called
free vowels because they can occur even in open syllables.
Consonant nucleus Some languages allow
obstruents to occur in the syllable nucleus without any intervening vowel or
sonorant. The most common syllabic consonants are sonorants like , , , or , as in English
bottle or in
Slovak krv [krv]. However, English allows syllabic obstruents in a few para-verbal
onomatopoeic utterances such as
shh (used to command silence) and
psst (used to attract attention). All of these have been analyzed as phonemically syllabic. Obstruent-only syllables also occur phonetically in some prosodic situations when unstressed vowels elide between obstruents, as in
potato and
today , which do not change in their number of syllables despite losing a syllabic nucleus. A few languages have so-called
syllabic fricatives, also known as
fricative vowels, at the phonemic level. (In the context of
Chinese phonology, the related but non-synonymous term
apical vowel is commonly used.)
Mandarin Chinese allows such sounds in at least some of its dialects, for example the
pinyin syllables
sī shī rī, usually pronounced , respectively. Though, like the nucleus of rhotic English
church, there is debate over whether these nuclei are consonants or vowels. Languages of the northwest coast of North America, including
Salishan,
Wakashan and
Chinookan languages, allow
stop consonants and
voiceless fricatives as syllables at the phonemic level, in even the most careful enunciation. An example is Chinook 'those two women are coming this way out of the water'. Syllabic
obstruents used to be considered very rare, but surveys have shown that they are relatively common and might even be more common than syllabic
liquids. Other examples: ;
Nuxálk (Bella Coola) : 'you spat on me' : 'he arrived' : 'he had in his possession a bunchberry plant' : 'seal blubber' In Bagemihl's survey of previous analyses, he finds that the Bella Coola word 'he arrived' would have been parsed into 0, 2, 3, 5, or 6 syllables depending on which analysis is used. One analysis would consider all vowel and consonant segments as syllable nuclei, another would consider only a small subset (
fricatives or
sibilants) as nuclei candidates, and another would simply deny the existence of syllables completely. However, when working with recordings rather than transcriptions, the syllables can be obvious in such languages, and native speakers have strong intuitions as to what the syllables are. This type of phenomenon has also been reported in
Berber languages (such as Indlawn
Tashlhiyt Berber),
Mon–Khmer languages (such as
Semai,
Temiar,
Khmu) and the Ōgami dialect of
Miyako, a
Ryukyuan language. ; Indlawn Tashlhiyt Berber : 'you sprained it and then gave it' : 'rot' (imperf.) ; Semai : 'short, fat arms' Languages with long sequences of obstruents pose a problem to several models of the syllable.
Coda The
coda (also known as
auslaut) comprises the
consonant sounds of a syllable that follow the
nucleus. The sequence of nucleus and coda is called a
rime. Some syllables consist of only a nucleus, only an onset and a nucleus with no coda, or only a nucleus and coda with no onset. The
phonotactics of many languages forbid syllable codas. Examples are
Swahili and
Hawaiian. In others, codas are restricted to a small subset of the consonants that appear in onset position. At a phonemic level in
Japanese, for example, a coda may only be a nasal (homorganic with any following consonant) or, in the middle of a word,
gemination of the following consonant. (On a phonetic level, other codas occur due to elision of /i/ and /u/.) In other languages, nearly any consonant allowed as an onset is also allowed in the coda, even
clusters of consonants. In English, for example, all onset consonants except are allowed as syllable codas. If the coda consists of a consonant cluster, the sonority typically decreases from first to last, as in the English word
help. This is called the
sonority hierarchy (or sonority scale). English onset and coda clusters are therefore different. The onset in
strengths does not appear as a coda in any English word. However, some clusters do occur as both onsets and codas, such as in
stardust. The sonority hierarchy is more strict in some languages and less strict in others.
Open and closed A coda-less syllable of the form V, CV, CCV, etc. (V = vowel, C = consonant) is called an
open syllable or
free syllable, while a syllable that has a coda (VC, CVC, CVCC, etc.) is called a
closed syllable or
checked syllable. These are not the same definitions of "open" and "close" as in
open and
close vowels, but are defined according to the phoneme that ends the syllable: a vowel (open syllable) or a consonant (closed syllable). Almost all languages allow open syllables, but some, such as
Hawaiian, do not have closed syllables. When a syllable is not the last syllable in a word, the nucleus normally must be followed by two consonants in order for the syllable to be closed. This is because a single following consonant is typically considered the onset of the following syllable. For example, Spanish ("to marry") is composed of an open syllable followed by a closed syllable (
ca-sar), whereas "to get tired" is composed of two closed syllables (
can-sar). When a
geminate (double) consonant occurs, the syllable boundary occurs in the middle, e.g. Italian "cream" (
pan-na); cf. Italian "bread" (
pa-ne). English words may consist of a single closed syllable, with nucleus denoted by ν, and coda denoted by κ: • i
n: ν = , κ = • cu
p: ν = , κ = • ta
ll: ν = , κ = • mi
lk: ν = , κ = • ti
nts: ν = , κ = • fi
fths: ν = , κ = • si
xths: ν = , κ = • twe
lfths: ν = , κ = • stre
ngths: ν = , κ = English words may also consist of a single open syllable, ending in a nucleus, without a coda: •
glue, ν = •
pie, ν = •
though, ν = •
boy, ν = A list of examples of syllable codas in English is found at English phonology#Coda.
Null coda Some languages forbid codas, so that all syllables are open. Examples are
Hawaiian from the
Austronesian family, and
Swahili from the
Atlantic–Congo family. ==Suprasegmental features==