When morphological structure overrides meaning: evidence from German prefix and particle verbs

A key question in the study of lexical processing has been whether the semantic transparency of multimorphemic words affects processing. Previous studies of English and French prefixed words have found that words with greater semantic transparency show greater morphological constituent activation. Studies of German, however, have shown constituent activation that is unaffected by semantic transparency. We report two cross-modal priming experiments that examined the underlying causes of this between-language difference. Experiment 1 investigated German particle verbs, in which the particle can be separated from the verb stem. Experiment 2 investigated prefix verbs, which (as in English and French) cannot be separated from the verb stem. Results showed that semantically transparent and opaque prefix and particle verbs induced equivalent priming relative to form controls. This indicates structured representations for all categories. We discuss these findings in relation to a language user’s experience with the particular properties of word types.