There are languages that have consonants that are a bit longer in duration i.e. the same as long vowels (e.g. like /iː/). So a long consonant is represented by writing /ː/ after it: long L = /lː/, long S = /sː/ etc.

Now English syllable structure is believed to be CCCVCCCC as in /streŋkθs/ "strengths".

If some language has a a word in which there is a long consonant before/after the vowel, how do linguists represent it; C or CC? For example, suppose that the initial /s/ in "strengths" is long: [treŋkθs], would it be CCCVCCCC or CCCCVCCCC?

  • 2
    I think there's an underlying assumption that one would immediately know whether a consonant is short or long (single or geminate) just by encountering it, which is not the case. Some consonants are longer than others, some longer than a nucleus. I think we usually regard a consonant geminate only if it contrasts phonologically with a shorter one. – Nardog Apr 19 at 10:37

How they're best analysed depends on the language

In Korean, syllables are maximally CGVC where G is a glide, and either C can be tense, with the tense consonants sometimes viewed as long

Meanwhile, in Arabic, syllables are maximally CVC (or CVCC in pausa, or word-finally in most vernaculars) and long consonants only occur across syllable boundaries (or in pausa) and so are best analysed as a sequence of two consonants that happen to be identical


There are many theories of what a "long consonant" is: here are the main theories. One is that every consonant has a "timing slot", which can be notated as C (vowels are notated as V), and a long consonant has two C's. This is the CV theory of Clements & Keyser. An alternative is that there is no difference between C and V at the level of the timing slot, everything is just X, and a long consonant (or vowel) has two X's. This is the X-bar theory of Levin (→Blevins). A version of this is Hyman's weight theory which has a single unit (he uses "x"), but with the special proviso that segments can merge under a single x, thus [pa] might start with two x's, but end up with a single x that both segments associate to. This then morphs into moraic theory, where vowels and some consonants are associated to a "mora" notated as μ, and a long consonant is somehow connected to a mora. The details of that theory are not particular uniform, because it doesn't handle long consonants in the beginning of a syllable very well, but there are a number of proposal. Morén's (dissertation) proposal is simply that "is moraic" is how you represent long consonants, but whether or not there is obvious "length", thus sometimes a final moraic consonant is said to be "long" and sometimes it is said to be just "moraic" (being moraic is the fundamental linguistic variable).

There is, incidentally, a potential for a contrast between a long consonant and two adjacent identical consonants, the latter often being called "fake geminates". Arabic is a language that has "real" geminates (a consonant is lengthened in a certain verb form) but also "fake" geminates, where a stem ends with /t/ and a suffix begins with /t/ so you get /t+t/. I haven't seen any evidence that the t-sequence /falat-ta/ or similar form is pronounced differently from a second-measure derived geminate. The question of whether, in some dialects, /falat-t/ "I escaped" is phonetically [falat] or [falatt] (in pre-pausal or pre-consonantal position, in a non-epenthesis dialect) is not clearly established, but it is plausible that there is a phonetic difference.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.