Van den Bussche and Reynvoet (2007, Experiment 1 ) report unconscious priming of comparable magnitude from novel words belonging to small and large categories, evidence that they interpret as demonstrating independence from category size of priming that involves semantic analysis. Three experiments raise the possibility that the findings in Experiment 1c of Van den Bussche and Reynvoet reflect subword processing, not semantic analysis. In Experiments 1 and 2, priming was obtained from primes and targets that shared approximately the same degree of subword features as in Experiment 1c of Van den Bussche and Reynvoet, but no priming occurred when sharing of features was minimized. Experiment 3 demonstrated priming driven by subword features when those features were set in opposition to whole-word meaning. These results indicate that orthographic overlap must be considered a potentially important confound in findings that ostensibly support priming mediated by semantic analysis.
Category Size and Category-Based Induction
