Facilitation and interference from formally similar word primes in a naming task

The naming of a word (e.g., CAVE) is delayed if participants first name a f ormally similar, but nonrhyming, prime (e.g., HAVE). Taraban and McClelland (1987) interpreted this effect in terms of competition between activated p honological codes, while Bradshaw and Nettleton (1974) argued that these in terference effects are due to conflicting output codes and only arise when primes are named. Experiment 1 shows interference effects for nonrhyming pr imes read silently (e.g., HAVE-CAVE), contrary to Bradshaw and Nettleton's claim, but rhyming primes (e.g., NEED-WEED) produced no facilitation, contr ary to predictions from Taraban and McClelland's model. In Experiment 2 par ticipants named both prime and target, and both interference and facilitati on were observed. In Experiment 3 formally dissimilar rhyming prime-target pairs (e.g., EIGHT-HATE) produced no facilitation even when primes were nam ed. Both interference and facilitation effects seem to result from a compli cated interaction of orthographic, phonological, and output codes. (C) 1999 Academic Press.