ABSTRACT
The consonants which occur word-finally in words such as those listed above pose no immediately obvious problem for synchronic phonology. Given the phonetic string:
we can, without too much difficulty (setting aside for the moment the question of the status of some of the vowels in these words), posit a phonemic string:
(2) Ilasdrizru3 syrlatablor03/3
and can identify the words Ila/, ISdriz/, lru3/, Isyr/, Ila/, Itabll, lor03/. The consonants at the ends of the words cerise, rouge, sur, table and orange are all fixed, that is they are always pronounced. The problems arise where this is not the case, that is when the same word is pronounced sometimes with, and sometimes without, a final consonant.4 For example, if the red cherry of (1) is a small one, we might encounter the phonetic string:
It would be dangerous, however, to conclude that the new word was Ipdtit/, since we know that there is a word with the same meaning and function in the string:
The examination of further occurrences of the word petit will show that there are two singular surface forms: [pdti], which occurs before masculine nouns beginning with consonants and at the end of a rhythmic group, and [Pdtit], which occurs before masculine nouns beginning with vowels and in agreement with all feminine nouns, whatever the position of the adjective. It is possible to offer an historical explanation of the variable presence of [t], but given that speakers of a language do not have an innate knowledge of the history of their language, a synchronic account of the phonology of this word is problematic: where does the final consonant come from, and how do speakers know when to pronounce petit with a final [t] or without? In the
case of [pgtitsgKiz] versus [pgtiga~s3], the difference is clearly a grammatical one; in the case of [pgtiga~s3] versus [pgtit:lm], the selection of one form rather than the other is phonologically determined, depending on whether the following word begins with a vowel or a consonant.
