Abstract
The factors that influence English speakers to classify a consonant as ambisyllabic are explored in 581 bisyllabic words. The /b/ in habit, for example, was considered ambisyllabic when a participant chose hab as the first part of the word and bit as the second. Geminate spelling was found to interact with social variables; older participants and more educated speakers provided more ambisyllabic responses. The influence of word-level phonotactics on syllabification was also evident. A consonant such as the medial /d/ in standard is attested as the second consonant in the coda of many English words (e.g. lard), as well as in the single-consonant onset of many others; for this reason such consonants were often made ambisyllabic. This contrasts with the /n/ in standard, which is never the first consonant in a word-initial cluster (e.g. *ndorf) and, therefore, rarely made ambisyllabic in the experiment. Ambisyllabicity was also found more often when the vowel preceding the single medialconsonant was lax, or stressed, or when the medial-consonant was a sonorant rather than an obstruent. The idea that a stressed lax vowel in the first syllable conditions both the ambisyllabicity of the consonant and its geminate spelling is not supported.