Phonological and Visual Mapping of Letters of a Foreign Alphabet for Educational Software

Abstract

Similarity coding is dominant in the human lexicon. The educational game MindSort is based on this idea and therefore constitutes a psychologically plausible game for learning vocabulary. Not only words, also letters are coded based on similarity. MindSort could therefore be extended to learning foreign alphabets if a distance measure for the letters is found. The present paper established a distance measure for the Indian script Devanagari by fitting phonological and visual features on similarity judgement data obtained by ten native Hindi-speakers. The analysis showed that for both sets of features, all but seven features constituted a significant contribution to the distance measure. An individual distance measure has also been found for each individual subject. For almost all subjects, the phonological features were a good explanation of their similarity judgements and for the visual features it was a good explanation for most of the participants. As the weights for the features varied per subject, an individual arrangement should be obtained from MindSort users after they familiarized themselves with the letters such that the game suits the individual learner’s brain best.