Aphonological relationship between a prime and a target produces facilitation when one or two initial phonemes are shared (low-similarityfacilitation) but produces interference when more phonemes are shared (high-similarity interference; . Although low-similarity facilitation appears to be a strategic effect (Goldinger, Luce, Pisoni, & Marcario, 1992), this result cannot generalize to high-similarity interference because the two effects are dissociated . In the present study, strategic processing in high-similarityinterference was investigated. The phonological relatedness proportion (PRP) and the prime-target interstimulus interval (lSI) were varied in a shadowing experiment. Low-similarity facilitation was found only with a highPRPand long lSI, but high-similarityinterference was found regardless of PRP and lSI. These results suggest that strategies influence low-similarity facilitation, but high-similarityinterference reflects automatic processing.Several studies (e.g., Goldinger, Luce, Pisoni, & Marcario, 1992;Jakimik, Cole, & Rudnicky, 1985;Radeau, Morais, & Dewier, 1989; Siowiaczek & Hamburger, 1992) have examined the role ofphonology in auditory word recognition by using a priming paradigm (Meyer & Schvandeveldt, 1971) in which a target word is preceded by a prime that shares some of its initial phonemes. Two dissociable effects have been obtained in this area of research: low-similarity facilitation and high-similarity interference . Although low-similarity facilitation involves strategic processes (Goldinger et al., 1992), the influence of strategies in highsimilarity interference has not been investigated.Determining the role of strategic processes in phonological priming is critical for models of spoken word recognition that propose operations relying on the phonology and predict phonological priming under various circumstances. For instance, in cohort theory (Mars len-Wilson, 1987), word recognition begins by activating a cohort of possible lexical candidates whose initial phonemes match the incoming signal. As such, a phonologically related prime could preactivate a target and facilitate responses. Another theory, the neighborhood activation model (NAM; Luce, 1986), suggests that similar-sounding lexical entries compete during word recognition. That is, the probability ofrecognizing a word is a function of the number, word frequency, and phonetic similarity of the word's neighbors. Presenting a phonologically related prime effectively increases its frequency and, thus, increases the competition between it and the target. A connectionist model pro-