Retour vers liste

Détail de la contribution

Auteur: Ekaterina TOMAS

Co-Auteur(s): Katherine DEMUTH

The Effects of Phonological Complexity on the Use of Verbal and Nominal Morphemes in Children with SLI

Abstract/Résumé: It has long been observed that typically developing (TD) children are variable in their production of grammatical morphemes [2, 3], and that phonological complexity might explain some of this inconsistency [4, 6, 7, 9, 14]. Children with Specific Language Impairment (SLI) can also be variable in their use of verbal [5, 8, 11] and nominal [10] morphemes. However, it is not yet known whether TD and SLI children have similar underlying processes of morpheme omission. Of particular interest is the reported later acquisition of syllabic morphemes (e.g. buses) in TD children [1, 3]; are these also acquired later by children with SLI? The data were drawn from speech samples collected using the Grammar Elicitation Test [13]. Participants were 20 children with SLI aged 4;10–5;11 years, who showed difficulties with morphemes of tense/agreement (past tense -ed; 3rd person singular -s), and/or possessive marker -s. All of the children were able to produce the relevant consonant clusters in a non-morphemic context. The test consisted of question/answer elicitation involving picture props. The experimenter described a picture and asked a question: “This man loves running. What does he do every day?” The child was then encouraged to give an answer like “He runs.” All items were piloted on age-matched TD children, who performed at ceiling. Analysis included a set of 30 stimuli for each morpheme, plus any additional spontaneous-like responses elicited during the session. The total number of analysed tokens included 1525 words (past tense: 707; present tense: 393; possessives: 425). These were grouped according to the phonological complexity of the coda: simple (e.g. cried/cries/May's), complex (e.g. jumped/jumps/Brett's), and syllabic morpheme (e.g. added/watches/horse's). For all three suffixes the test of within-subjects effects using a GLM Repeated Measures showed that phonological complexity had a significant effect on morpheme production (past: F(2,20)=43.964, p<.001; present: F(2,8)=16.185, p=.002; possessives: F(2,12)=30.123, p<.001). In particular, production of the syllabic form (e.g. added/watches/horse’s) was significantly lower in both nominal and verbal morphemes. This result might have several explanations. Firstly, adding a new syllable to the stem lengthens the word, increasing phonological working memory demands. Alternatively, it could be that there are increased challenges of producing similar segment types separated by only a schwa (e.g. horse’s [hɔ:səz]). Finally, the syllabic allomorphs are less frequent, and therefore children have less practice with producing them. The theoretical implications of these findings are discussed.