B07 Sound patterns in Human Language: Syllables revisited
Most of the material we have covered has dealt with
segments. Phonologically, these were
phonemes which were composed of groups of
features. Phonetically, we classified segments along
such dimensions as place and manner of articulation etc. We discussed
syllables briefly, but did not go into any detail about the
relationship between syllables and segments. Here we will look at how
syllables are built from segments and how syllables organize segment
sequences.
Phonotactics
Our starting point will be to consider the restrictions which govern
possible sequences of sounds. As a competent speaker of English, you
have internalized all sorts of complex knowledge about which sound
sequences can occur in English and which can not. For each word in
the following table, decide whether it is a possible English word or
not.
| flabble | prznk | spronk | bmil |
keladulance | trozzit |
| squirthy | prlauiop | stroimpt |
treh | ztreet | flampidator |
The knowledge you bring to bear here is the set of
phonotactic constraints on English sound sequences.
Phonotactics refers to the set of restrictions which determine all
possible sound sequences within a particular language. The
phonotactics of Polish or German is thus different from the
phonotactics of English or Japanese.
A first important observation is that almost anything one cares to say
about phonotactic restrictions must make reference to the syllable.
In English, the sequence /
/ is
not well formed. That is, it is not a possible word as it stands.
Yet it does occur, as in bootleg. In this latter case, the /t/
occurs at the end of one syllable, and the /l/ occurs at the start of
the next. Likewise, the syllables /
/, /
/ and /
/ are all clearly disallowed in English, yet they all
occur (in the words merchandise, submit and collapsing,
respectively). Where they occur, a syllable border serves to break up
the initial consonant clusters. In English, the sequence /tl/ is
allowed if and only if the /t/ and the /l/ belong to different
syllables. The syllable is the basic phonotactic unit.
Syllable structure
The nucleus
We already noted that a syllable has a nucleus. The
nucleus is the energy rich middle of the syllable, usually occupied by
a vowel. Every well-formed syllable has a nucleus, and a syllable may
have nothing more. The word eye, the first syllable in
about and the last syllable in bottle all have only a
nucleus. In bottle, the last syllable consists only of the
sound /l/. Other consonants which can occupy the nucleus include the
nasals (the /n/ of button, the /m/ (in some pronunciations) of
captain, the // at the start of inquire), and /
/ (terror).
The onset
The most common sort of syllable found in the World's languages
consists of a single consonant and a single vowel. These are referred
to as CV syllables. The position occupied by the consonant is called
the syllable onset. The table below gives the
consonants of English. One or two of these cannot occur in the onset
of a CV syllable. Which ones are they?
English allows more complex onsets than simple CV structure. We have
CCV in words like flee, glue, pray, spy. Here the set of
restrictions becomes more complex. The first consonant in a CCV
syllable cannot be an affricate (/
/) ,or a sonorant (/,,,,,
), or an /h/. If the first
consonant is an /f/, what phonemes can occur as the second consonant?
Try to answer this by defining a class of sounds, rather than a list.
What if the first consonant is a /p/? What if it is an /s/?
If you answer these questions, you will have found that /s/ allows
many more consonants in the second position than either /f/ or /p/, or
any other consonant for that matter.
English also has CCCV syllables, e.g. in strew, straw.
These always begin with /s/.
Observations like these allow us to start to outline the possible
forms of English syllable onsets. The onset may be absent, may be a
sonorant, an obstruent, an obstruent followed by a sonorant, or any of
the above preceded by an /s/. This is approximately correct for
English.
The coda
Many (but by no means all) languages allow consonants at the end of a
syllable as well. This position is called the coda.
An entirely different set of restrictions apply to coda consonants.
Which of the English consonants are not found in simple VC or CVC
codas? Is it the same set as those not found in onsets?
The coda position can be empty (cow, bee, aye etc). If there
is only one consonant in the coda, almost any consonant can occur
(exceptions: /w,j,h/). If there are two, the second must be a
obstruent (wax, waft, adze etc). Any of the above
possibilities can be followed by an /s/.
Note the quite remarkable role of /s/ in the above restrictions! IF
we have a maximally complex syllable, which in English is CCCVCCC, the
first and last consonants must be /s/, as in strengths.
The rhyme
Together, the nucleus and the coda constitute the
rhyme of the syllable. Indeed, this is the part that
should match if two syllables are felt to rhyme
(trounce/bounce, stew/you etc). This gives us a picture
of the syllable which looks like this:
The above presentation has simplified things somewhat, but the sketch
is intended to reveal the type of constraints found which depend on
syllable position.
This page has been developed by Fred Cummins, Department of
Linguistics, 2016 Sheridan Road, to whom all questions should be
referred.
Copyright (1998) Northwestern University
Goto top of page
Goto index
