There are numerous studies demonstrating that people’s judgments about meanings of words can sometimes derive from their sound – a phenomenon often referred to as sound symbolism. A recent comprehensive assessment of English demonstrates that some small amount of systematicity exists between form and meaning. Is this small level of systematicity in language sufficient to drive the observed behavioral effects of sound symbolism? In this study we first tested the extent to which similarities amongst the sounds of words was sufficient to drive sound symbolic effects. We then tested whether a computational model that learned to map between form and meaning of English words better accounted for the observed behavior. We found that phonological similarity alone was sufficient to account for several effects of sound symbolism (without reference to meaning at all), but that the form-meaning mapping model was able to reproduce additional key behavioral effects of sound symbolism.