Predicting the judged “similarity of sound” of english words