Description
The current study examines the nature of homophonous lexeme pairs which result from linguistic conversion (specifically noun-verb lexeme pairs). Noun-verb lexeme pairs were extracted from the Buckeye Corpus (a corpus of naturally occurring speech data between native English speakers from Ohio) and then re-tagged by hand to ensure accuracy. Once the noun-verb lexeme pairs were determined, their durations were extracted (the Buckeye Corpus includes the durations of the speech data). Semantic analysis was also performed for each lexical item in the data set using wordnet online to identify homomorphs (phonetically and semantically identical lexemes pairs). Once the necessary data was extracted, statistical analysis was performed using a linear mixed effects model. This analysis revealed that noun-verb lexeme pairs which result from linguistic conversion are not phonetically identical: verbs are shorter in overall duration. Additionally, various other covariates allow for several other conclusions to be made: there is no evidence of a zero-affix present, homomorphs and homophones are stored identically in the mental lexicon, and this is not the morphologization of a phonetic effect. Further research needs to be developed to examine the psycholinguistic nature of these effects, and it would be extremely beneficial to examine an agglutinative language under this same premise.
