All morphemes are not the same: accuracy and response times in a lexical decision task differentiate types of morphemes

Background Evidence indicates that fluent readers automatically decompose morphologically complex words. However, few studies have directly compared processing of stimuli comprising different types of morphemes, particularly bound and free morphemes. Methods Eighty fluently reading young adults participated in a lexical decision task with word and nonword stimuli composed purely of bound stem (e.g. discern, predict; disject, percern) or free (e.g. cobweb, earring; cobline, bobweb) morphemes, along with comparison control stimuli (e.g. garlic, minnow; gartus, buzlic). We used both traditional analyses of variance and modern mixed-effect modelling to investigate morphological effects on accuracy and response times for words and nonwords. Results In both ANOVAs and mixed-effect analyses, bound stem words were responded to more quickly and accurately than free morpheme words, indicating that behavioural responses in a lexical decision task can differentiate between types of morphemes. Moreover, in mixed-effect analyses, accuracy and response times were similar for bound stem words and control words. In contrast, in the mixed-effect analyses, accuracy was worse, and response times were longer for both bound stem and free morpheme nonwords as compared with control nonwords, but accuracy and response times were similar for bound stem and free morpheme nonwords. Conclusions That it was easier to process words composed of bound stem morphemes, as compared with free morphemes, indicates that all morphemes are not processed similarly. That this pattern of results for word stimuli was not seen with nonword stimuli suggests that the processing advantages for bound stem words are related to the semantic familiarity of real words. Overall, these findings have implications for learning and teaching morphology.