This study attempts to advance corpus-based exploration of sound iconicity, i.e. the existence of a non-arbitrary relationship between forms and meanings in language. We examine a number of phonesthemes, phonetic groupings proposed to be meaningful in the literature, with the aim of developing ways to validate their existence and their semantic content. Our first experiment is a replication of Otis and Sagi (2008), who showed that sets of words containing phonesthemes are more semantically related to each other than sets of random words. We augment their results using the British National Corpus and the Semantic Vectors package for building a distributional semantic model. Our second experiment shows how the semantic content of at least some phonesthemes can be identified automatically using WordNet, thereby further reducing the room for intuitive judgments in this controversial field.