|Title:||Floating Consonants in French: the need for the skeleton in input|
|Comment:||this is nothing but an undergraduate essay with a suggestion, not the last word on any particular topic|
|Abstract:||OT and GP don't seem to differ on the basic notion that phonology is, in its purest form, the science of how sound structure of the input (UR) is different from the sound structure of the output (SF).
In GP, a floating consonant is a segment which is part of the lexical entry of a word but isn't attached to its skeletal point. If nothing changes the floating consonant will remain unattached and thus un-parsed. Phonology is change, thus a phonological effect, in this area, will result in the uncoupled segment coupling ('epenthesis') just as phonological effects can lead to a coupled segment uncoupling (deletion).
'Epenthesis' is in brackets because there is a conceptual difference between the insertion of phonological material into the output and the phonological material of the input surfacing in the output. The latter will be called surficant in this essay. Surficant segments are lexically specific and variant meanwhile epenthetic segments are stipulated by the grammar of the language and so will be identical in association with all lexical items. In GP this is an unspoken tenet. OT also views surficant consonants and epenthetic consonants as different with epenthetic consonants occurring with a lowly ranked DEP and floating consonants surfacing by a highly ranked PARSE-X (Prince and Smolensky 1993 and Tranel 1995, ms. respectively).
The difference between GP and OT in terms of floating consonants is that GP considers syllabification to be a feature of the lexeme itself and thus a feature of the lexicon. Syllabification in GP isn't a phonological process it is a phonological status quo. Conversely, in OT, syllabification is imposed on the lexical material (input) by constraints such as NO CODA. This difference of opinion is relevant to this essay in as much as these hypothesis help or hinder our understanding of floating consonants. The scope of this essay is not to compare and contrast GP and OT, rather, in order to understand floating consonants in OT we must understand two things, the nature of input and the nature of output, especially with respect to syllabification.
Section one will show how French floating consonants are problematic for a correspondence theory (McCarthy and Prince 1995) which makes two strong claims: there is only one step of derivation and the input is not syllabified. If input is not syllabified, the assumption that floating consonants are treated differently from non-floating consonants with regards to PARSE X has to be a consideration made after consonants have (or have not) been attached to skeletal points in the first place (in line with Tranel's assumptions (1992, 1993, 1995, ms.)). Section Two will show that the ideal one step derivation can still be maintained in containment theory (Prince and Smolensky 1993) but that what allows this approach to work is no different from Tranel's practice of introducing syllabification, at least partly (the skeleton) into the input, the significance and consequences of which are not at all discussed in Tranel (ms.). Section three will then ask itself the next logical question: if syllabification up the skeleton is in the realm of the lexicon and Goldrick (2000) shows how metrical structure can also be part of the input then should we understand (as Government Phonology does) that all basic syllabification is a characteristic of the lexicon and thus not a process but a status quo. Arguments promoting this are presented which include a brief discussion of Kenstowicz's Base-Identity constraint and its application to Spanish' /-ito/ vs. /-cito/.