کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
932040 | 923061 | 2011 | 19 صفحه PDF | دانلود رایگان |
This study examined the lexical representation of words with two pronunciation variants. We tested whether both the schwa and reduced variants of French words are stored as lexical entries. The results of four experiments in which speakers named pseudohomophones and pseudowords show an advantage for pseudohomophones over matched pseudowords for both variants. As the pseudohomophone effect is assumed to reflect the activation of phonologically matching stored phonological representations, these results suggest that both variants of schwa words are stored. Importantly, the pseudohomophone effect is found for alternating words (Experiments 1 and 2) and for non-alternating words when the non-produced variant corresponds to the word’s spelling (Experiment 3) or is frequently encountered in the speech of speakers of other regiolects (Experiment 4). These findings extend the scope of our previous proposal that words with two variants are stored with two lexemes. This conclusion needs to be integrated in word production models.
Research highlights
► Among the first studies to examine the representation of non-canonical word forms.
► Focus on (non-)canonical productions of French schwa words.
► Uses the pseudohomophone effect in naming to examine which variants are stored.
► Pseudohomophone effect found for canonical and non-canonical forms.
► Result taken to suggest that canonical and non-canonical forms are lexically stored.
Journal: Journal of Memory and Language - Volume 64, Issue 4, May 2011, Pages 424–442