If syllables were classification units in speech perception, auditory priming would show it.

Two auditory priming experiments tested whether the final overlap effect relies on syllabic representations. Amount of shared phonetic information and syllabic correspondence between prime and target nonwords were varied orthogonally. In the related conditions, CV.CCVC primes and targets shared the last syllable (e.g. pinclude-viclude) or the last syllable minus one phoneme (e.g. pinclude-viflude); conversely, CVC.CVC primes and targets shared the last syllable (e.g. goltibe-purtibe) or the last syllable plus one phoneme (e.g. goltibe-pultibe). Both experiments required to repeat back the targets, with Experiment 2 including foils. The facilitation induced by related primes increased with the number of shared phonemes, and was by and large independent of syllabic correspondence. There is thus little evidence for pre-lexical syllabic classification