Semantic processing in auditory lexical decision: Ear-of-presentation and sex differences

Abstract

In this study we examined the interplay between appetitive (approach) and defensive (avoid) responses in spoken word recognition. Ninety-two right-handed participants (half women) took part in an auditory lexical decision experiment in which speech was presented to only one ear. The danger and usefulness of the word referents interacted in predicting RTs, as in previous (binaural) studies with poorer control of psycholinguistic covariates. Specifically, higher danger ratings were associated with faster RTs for words rated low on usefulness; but higher danger ratings were associated with slower RTs for words rated high on usefulness. In addition to this primary finding, men showed more lateralised performance, as indicated by significant interactions of sex and ear of presentation with word frequency, and with the animacy of the word referents. For both sexes, word frequency had a stronger effect on accuracy for speech presented to the right ear. Finally, men's but not women's RTs were related to the danger dimension. This last finding provides an intriguing avenue for future research in the area of sex differences and emotion.

Acknowledgements

We would like to thank Mark Pluymaekers and an anonymous reviewer for making very helpful comments on an earlier version of this manuscript.

Notes

¹Linguists may argue that danger and usefulness are pragmatic rather than semantic variables. Our treatment of them as semantic is consistent with the (perhaps technically incorrect) terminology in the psycholinguistics literature.

²Baayen, Feldman, and Schreuder (2006) demonstrate that family size is a variable rooted in semantics. In fact, while it is not central to the current proposal, Baayen et al. (2006) have argued that the word frequency effect can be thought of as semantic, too, because it reflects an underlying conceptual (rather than form-based) familiarity.

³For consistency with our previous work, we measured RTs from the UP of each word. It might be argued that a more theory-neutral approach would be to measure RTs from item onset, but in fact it makes little difference. In an onset analysis, the UP effect changes sign, so that words with later UPs are recognised later. All effects that are not significant in our presented analysis are also not significant in the onset analysis; and all of the effects that are significant in our presented analysis are also significant in the onset analysis, except for the main effects of stress and morphological family size.