The author proposes that syllable weight is driven by considerations of phonetic effectiveness and phonological simplicity, and that the phonetically best distinctions are those which divide syllables into groups that are phonetically most distinct from each other. Phonologically complex distinctions are those which exceed an upper threshold in the number of phonological predicates to which they refer. It is claimed that languages adopt weight distinctions that are phonetically most effective without being overly complex phonologically. Syllable weight thus reflects a compromise between phonetic and phonological factors. The proposed model of weight further suggests that phonological weight distinctions are ultimately predictable from other basic phonological properties, such as syllable structure.