Thomas Pechmann & Dieter Zerbst
University of Leipzig
In a series of picture-word interference experiments (Pechmann & Zerbst, 2002), subjects named common objects presented as simple line drawings on a CRT screen by producing simple noun phrases. Either slightly before, simultaneously, or slightly after the picture appeared on the screen they were presented with distractor words which they were told to ignore. In the critical conditions distractor words were either nouns or non-nouns (closed-class words or adverbs). In the first experiment in which subjects produced bare noun descriptions we did not obtain any word class effect. Naming latencies were not affected by the grammatical category of the distractor words. In the following experiments we used a new variant of the picture-word interference paradigm: The naming of the pictures was embedded into the production of a simple sentence. Subjects first saw a proper name and a verb in its infinitive form (e.g., <Peter beschreiben>, Peter describe). They were asked to overtly pronounce the name and the verb in its inflected form ("Peter beschreibt", Peter describes). Shortly afterwards they saw the picture of an object and a distractor word. Their task was to start naming the picture as fast as possible together with the definite article thereby completing the sentence fragment (e.g., "Peter beschreibt den Apfel", Peter describes the apple). Using this method we obtained highly significant word class effects. Noun distractors interfered more strongly with picture naming than did non-noun distractors. This held for both visual and auditory presentation of the distractor words. The interference effect showed up in a time window where semantic interference can usually be observed, supporting the claim that at an early stage of lexical access semantic and syntactic activation processes overlap. In a final experiment we compared the time course of activating semantic and syntactic features of words in a within-subjects design. The data demonstrate that the activation of word class information precedes semantic effects. Furthermore, semantic activation could still be observed when syntactic activation had already declined. These findings are problematic for the two-stage theory of lexical access as proposed by Levelt, Roelofs & Meyer (1999).
References
Levelt, W.J.M., Roelofs, A., & Meyer, A. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22, 1-38.
Pechmann, Th. & Zerbst, D. (2002). The activation of word class information during speech production. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28 (1).