Listeners learn to preprocess their native language in a way that provides the most useful information for word recognition, and they may also use their native strategies when listening to a second language (Cutler, Mehler, Norris and Segui, JML, 1986; Cutler and Otake, JML, 1994). Dutch and English are both lexical stress languages, but in English, stress information covaries with vowel quality information, and stress is not used by English listeners in lexical access (Fear, Cutler and Butterfield, JASA, 1995; Culter and Clifton, 1984). The stress/vowel-quality relation is less strong in Dutch, and experimental evidence shows that Dutch listeners do use lexical stress to constrain lexical access in their native language (Koster and Cutler, Eurospeech, 1997; van Donselaar and Cutler, 1997). Can Dutch listeners also process stress cues in English?

Two fragment priming experiments investigated Dutch and English listeners' processing of lexical stress in English. Pairs of words were found with the same unreduced first syllable and following phoneme but mismatching stress pattern, eg MUsic, muSEUM. Visual lexical decisions to one of these words (eg MUSEUM) were measured following auditory presentation of 1) a phonemically- and stress- matched fragment (muS); 2) a phonemically matched- but stress-mismatched fragment (MUs); 3) a control fragment (MONs). Both English and Dutch listeners' responses after phonemically matching fragments (MUs/muS) were significantly faster than after control fragments, and both showed sensitivity to stress information, giving faster responses after fragments with correct stress (muS) than after fragments with incorrect stress (MUs). The facilitatory effect in the stress mismatching condition is surprising for both language groups, first because a similar experiment with Dutch listeners attending to Dutch stimuli showed an inhibitory effect following stress mismatches (MUs), and secondly because previous research with English listeners suggested that English listeners are not sensitive to stress cues in English. To investigate whether the difference betweeen the current results and the previous English results could be explained by differences in unconscious and conscious levels of processing, an off-line two-alternative forced choice task was given to English listeners to determine if they could consciously select which word of the stress pair each fragment came from. The fragment priming results will be discussed in the light of data from the forced choice experiment.