Writing and speaking two bare nouns from pictures : The issue of dependency

Nathalie Malardier, Patrick Bonin, Alain Méot and Michel Fayol
LAPSCO/CNRS (UMR 6024), University Blaise Pascal (Clermont-Ferrand), France

Participants had to write down or to speak aloud two bare nouns from pictures presented side-by-side starting with the left one. Multiple regression analyses were performed on the spoken and written latencies. In the first kind of regression analyses, some characteristics corresponding to the two pictures (in first and in second position) and their names were introduced as independent variables. In the second kind of analyses, the latencies to name individually the pictures corresponding to the pairs were introduced as independent variables. The first kind of analyses have revealed that, for the targets in second position, the characteristics indexing object identification and semantic processing in writing, and object identification in speaking, were reliable determinants of naming latencies. For the targets in first position, the characteristics indexing object identification, semantic processing and lexical access were reliable determinants of naming latencies in both production modes. The amount of unique variance accounted for by the set of characteristics corresponding to the second targets was small in both production modes. The second kind of analyses have shown that the latencies to name the second target pictures accounted for a small part of unique variance in predicting the latencies of the pairs. The findings suggest that, in both writing and speaking two bare nouns from pictures, naming is initiated when the processing of the first target is quite complete and the processing of the second target is restricted to initial levels of processing involved in picture naming.