A binary articulatory production classification of English consonants with derived difference measures

English consonants were classified according to six binary articulatory variables. Differences between the members of each possible consonant pair were quantified and the overall difference between each consonant and all other consonants was calculated. The consonant difference measures were found to be significantly related to children's substitution errors and to account for 56% of the variance in Templin's (1957) data on normal consonant acquisition. Similar measures calculated from the distinctive feature classifications of Chomsky and Halle (1968) and Singh and Singh (1976) accounted for less than one third of this amount. These findings suggest the potential utility of the present classification for the analysis of articulated speech.