The influence of the feature (sonorant) in lexical access in Spanish

Does the feature (sonorant) influence access to lexical entries? In current feature geometry, the feature (sonorant) is a root node feature, distinguishing resonants from obstruents. Marks, et al. (2002) demonstrated its influence in lexical access using the word reconstruction task in Spanish. Participants heard a nonword, e.g., dirmar, and were asked to change it into a real word (firmar) by changing just one consonant. Errors in recovering the correct word (firmar) were much higher when the word contained an obstruent that had been replaced with a resonant (nirmar) than when the word contained an obstruent that had been replaced by another obstruent (dirmar). Obstruent target words were easier to recover when the target segment and the replacing segment matched on the feature [sonorant]. Why do obstruents show this matching effect? The present study subdivided obstruents into stops and fricatives. Two experiments compared them separately to resonants in the word reconstruction task. Are fricatives alone or stops alone responsible for the observed obstruent effect? Results showed greater errors in the mismatch condition for both fricatives and stops, but significantly greater for fricatives than for stops, suggesting that the mismatch effect observed for obstruents is largely based on fricatives.