A comparative usage-based approach to the reduction of the Spanish and Portuguese preposition 'para'
Time: 11:00am - 01:00pm
Place: Ballantine Hall 141
Michael Gradoville (Dissertation defense)This study examines the frequency effect of two-word collocations involving para 'to, for' (e.g. fui para, para que) on the reduction of para to pa' (in Spanish) and pra (in Portuguese). Collocation frequency effects demonstrate that language speakers store the more frequent collocations in memory, thereby providing support for the usage-based model of linguistic representation and evidence against models of representation that cannot account for such effects. This dissertation reports on studies of corpus data from the Spanish of Caracas, Venezuela and New Mexico, United States as well as the Portuguese of Fortaleza and Rio de Janeiro, Brazil. The data were coded for the dependent variable, namely whether the form was unreduced para or the language's reduced form. The independent variables include the frequency of the two-word collocation with the preceding word (e.g. vai para, iba para), the frequency of the two-word collocation with the following word (e.g. para mim, para hacer), the following sound (vowel, coronal consonant, dorsal consonant, non-lingual consonant), the following syllable stress, and the grammatical function of the token of para. The data were analyzed using the statistical program GoldVarb X with supplemental unifactorial linear regressions of the frequency variables.The results from Spanish indicate that the grammatical function of para and the frequency with which para co-occurs with a following word are the most important predictors of the reduction of para. Portuguese, on the other hand, is most affected by the frequency with which para co-occurs with an adjacent word, preceding or following. The results coincide with previous studies of the effect of collocation frequency on phonological variation and thus support the assertion that frequent collocations are stored as chunks with all of the phonetic detail intact, providing evidence against modular models of language that would assert that grammar and the lexicon are distinct units.
In category: Sociolinguistics and pragmatics
JEvents v3.0.9 Stable
Copyright © 2006-2013
Powered by Joomla!®
Copyright © 2008 The Trustees of Indiana University | Copyright Complaints