Article révisé par les pairs
Résumé : Several studies on auditory word recognition indicate that word processing is influenced by phonological similarity with other words. We describe a lexical database, VoColex, which provides several statistical indexes of phonological similarity between French words. Phonological similarity is computed according to two distinct principles. According to the first principle, phonologically similar words share initial phonemes with the target word. According to the second principle, phonological neighbours correspond to any words which can be derived from the target by a single phoneme change (substitution, addition, or deletion) whatever the position of the modified phoneme. The statistical data provided by VoCoLex allow the control and the empirical manipulation of various measures of phonological similarity, as well as quantitative descriptions of the auditory lexicon.