Abstract
In the field of spoken word recognition, psycholinguists investigate the way lexical items are stored in memory and the nature of perceptual processes that allow access to these representations. “Full listing” models propose a direct access to lexical items from the phonological form, while “full parsing” approaches implement a pre-lexical morphological decomposition procedure of complex words. Prefixed words are thus detected similarly to words beginning with a stem morpheme in one model, but differentially in the other.
In our study a word-monitoring task was used in which listeners reacted to a specified target word. Their reaction times were analysed with respect to type of the target word (prefix/stem) and the segmentation mode (carrier sentences were broken into syllable-sized fragments that either respected morpheme boundaries, or disregarded the morphological structure, maximizing syllable onsets).
Czech as an inflectional and prosodically unique language offers a possibility to test the degree of universality of the morpheme category effects, and investigate the role of morphemes in syllable boundary assignment. The results suggest a statistically significant difference between stem and derivative morphemes (the former eliciting faster reactions) but were inconclusive as regards the segmentation mode.
The findings are discussed in the light of current theories and relevant experiments in other languages.