Letter search and relatedness proportion: Further electrophysiological evidence for the automaticity of semantic activation

Abstract

Automatic semantic activation was recently called a myth when the behavioral semantic priming effect was found to be subject to task demands. Semantic priming effects as measured by lexical decision times are typically reduced to the point of being absent when a letter search has to be performed on the prime word. It seems premature to draw any definite conclusion from this behavioral result as ERP recordings show that the same prime task leaves N400 priming effects unaffected. It is still a matter of debate whether N400 priming effects are generated chiefly by automatic or by cognitively controlled priming processes. In the present study we therefore recorded both lexical decision times and ERPs while we varied the ratio of related to unrelated prime–probe pairs (25% vs. 75% related) and thus manipulated the degree to which controlled mechanisms contribute to the semantic priming effect. Behavioral measures revealed reliable semantic priming effects only when RP was high and the influence of controlled processes hence increased. The modulation of the N400, however, provided clear-cut evidence of semantic processing at both RPs, indicating that semantic activation occurred even when RP was low and the involvement of controlled mechanisms was thus minimal.