The time course of lexical involvement in phonetic categorization

Abstract

Two experiments in Dutch examined how the influence of lexical knowledge on phonetic decisions changes over time. A [f]–[s] continuum (edited natural speech) was placed in initial position in monosyllables to make word–nonword and nonword–word continua: flauw (dull)–slauw and flaap–slaap (sleep). Syllable-final continua were made in the same way: maf (silly)–mas and jaf–jas (coat). Materials were low-pass filtered at 3 kHz. In experiment 1 there was a significant lexical effect for both initial and final fricatives: throughout the continua there were more [f] responses in the word–nonword than in the nonword–word continua. Lexical involvement was strongest in the listeners’ fastest responses but disappeared in their slowest responses, both for the final fricatives (in line with previous findings) and the initial fricatives (contradicting earlier studies showing the strongest effects in the slowest responses). Experiment 2 tested whether listeners would still use lexical knowledge in their fastest responses under severe time pressure. They were asked to respond before a tone, presented 500 ms after fricative offset. The results replicated experiment 1. Lexical knowledge appears to be used in phonetic decision-making only within a limited time frame.