This study reports an analysis of confusion data in Cutler, et al.  designed to probe interactions between distinctive features in English consonant identification by English and Dutch native listeners. While both listener groups exhibit extensive interaction between features, the Dutch listeners' interactions deviate systematically from the English listeners'. In the original analysis, coda voicing neutralization in Dutch was invoked to account for the lower identification accuracy and information transmission rates for coda voicing contrasts in Dutch listeners . The present study augments these findings, analyzing consonant pair similarity measures, finding evidence for different laryngeal feature structure in both language groups in both onset and coda positions. This is not accounted for by a general neutralization rule.