Syntactic Control of Letter Detection: Evidence From English and Hebrew Nonwords

According to the unitization account, letters are more often missed in function words (e.g., the) than in less common content words because their higher familiarity allows access to their whole-word representations. The present study, however, replicated this pattern with nonwords. For both Hebrew and English, nonwords produced more detection errors when placed in function slots than in content slots. A similar effect was found for Hebrew prefix nonwords, where the initial letter could be interpreted as a function morpheme or as part of the stem. The results were seen as support for a structural model in which function morphemes are initially utilized to define the structural frame of a phrase but recede into the background as meaning is uncovered. Several interactive patterns illuminated the details of frame extraction.