Drozdova, P. 1 , van Hout, R. 1 & Scharenborg, O. 1, 2
1 Centre for Language Studies, Radboud University Nijmegen, The Netherlands
2 Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, The Netherlands
Listeners adjust their phonetic categories to adapt to ambiguous sounds using lexical knowledge to interpret the ambiguity. The present study investigates whether words containing an ambiguous sound are processed in the same way as natural words. Using a lexical decision task with an embedded auditory semantic priming task, we show that primes with an ambiguous sound were processed slower and accepted less often as real words than their non-ambiguous counterparts. This difference in acceptance, however, disappeared after an exposure to approximately 15 primes containing an ambiguous sound.
Remarkably, the slower and less accurate processing of the ambiguous items did not have an effect on the processing of the following, semantically-related word. At the same time, lower acceptance rates of ambiguous primes predicted slower reaction times to these primes compared to their natural counterparts. Some ambiguous items were therefore more difficult to process and recognize than others. Supposing that only ambiguous items recognized as real words induce retuning, it suggests an important role of stimulus specific characteristics in triggering learning. Finally, higher acceptance rates to the items with the ambiguous sounds in the lexical decision task led to a marginally larger lexically-guided perceptual learning effect in the subsequent phonetic-categorization task.