Repetition, semantic priming, and stimulus quality: implications for the interactive-compensatory reading model.

Recognition latencies of single words were manipulated by repetition, degradation, or both, and the effects of context were observed. In both lexical decision and pronunciation tasks, repeated words were recognized faster than nonrepeated words yet were not any less affected by semantic context. Both inserting asterisks between a word's letters and masking slowed word recognition in comparison with a clear presentation, but only the masking manipulation showed contextual inhibition.

On obtaining episodic priming in a lexical decision task following paired-associate learning.

In four experiments, subjects made lexical (word-nonword) decisions to target letter strings after studying paired associates. In this lexical decision test, word targets previously studied as response terms in the paired associates were preceded at a 150-ms and/or 950-ms stimulus onset asynchrony (SOA) by one of various subsets of the following six types of primes: a neutral (XXX or ready) prime, a semantically unrelated word prime episodically related to the target through its having been previously studied in the same pair, a semantically related word prime previously studied in a pair with some other unrelated word, a semantically unrelated word prime previously studied in a pair with some other unrelated word, a nonstudied semantically related word prime, and a nonstudied semantically unrelated word prime. At the 950-ms SOA, facilitation of lexical decisions produced by the episodically related primes was greater in test lists in which there were no 150-ms SOA trials intermixed, no previously studied semantically related primes, and no studied nonword targets.