|Title:||Re-examining Spanish 'Resyllabification'|
|Authors:||Timothy L. Face|
|Comment:||Phonetic symbols in IPA Garamond font, available as "Other" file|
|Abstract:||Re-examining Spanish 'Resyllabification'
Timothy L. Face
The Ohio State University
Revised version of paper presented at the 29th Linguistic Symposium
on Romance Languages, University of Michigan, April 8-11, 1999.
This paper reexamines Spanish syllabification, focusing on
"resyllabification", within Optimality Theory (Prince and
Smolensky 1993) and Correspondence Theory (McCarthy and Prince 1994).
The first part of the paper presents the data relevant to
syllabification in Spanish, showing preferences for a consonant
to be an onset, both within a word and across word boundaries,
for the formation of complex onsets within words but not across word
boundaries, and for the interaction of syllabification and aspiration
in those dialects of Spanish where /s/ aspiration is not limited to
coda position. It is also seen that syllabification across prefix
boundaries shows exactly the same pattern as syllabification across
word boundaries. This leads to the proposal that prefixes have a
different status than do suffixed morphemes, and specifically that
they are phonological domains (PDs). A PD is defined as a grouping
of morphemes that is the input to the phonology. A PD may consist of
one or more morphemes, but only PDs (and not individual morphemes
within a polymorphemic PD) can be referred to by the phonology.
Taking into consideration the morphological status of prefixes, a new
analysis of Spanish syllabification is provided using universal
constraints independently motivated in previous studies, and
by making minor modification to the Uniform Exponence constraint
proposed by Kenstowicz (1995) and used by Colina (1997) in her attempt
to account for Spanish "resyllabification".
Previous analyses of Spanish syllabification have had trouble dealing
with prefixes, or have ignored them altogether. Previous accounts are
reviewed, pointing out where they fall short. The account offered in
this paper is superior to previous accounts in two ways: (1) It
recognizes the morphological distinction between prefixes and suffixed
morphemes, and (2) it is able to account correctly for data about
which previous analyses make false predictions.