Neural Differentiation of Lexico-Syntactic Categories or Semantic Features? Event-Related Potential Evidence for Both

Abstract

Event-related potentials (ERPs) were used to investigate whether processing differences between nouns and verbs can be accounted for by the differential salience of visual-perceptual and motor attributes in their semantic specifications. Three subclasses of nouns and verbs were selected, which differed in their semantic attribute composition (abstract, high visual, high visual and motor). Single visual word presentation with a recognition memory task was used. While multiple robust and parallel ERP effects were observed for both grammatical class and attribute type, there were no interactions between these. This pattern of effects provides support for lexical-semantic knowledge being organized in a manner that takes account both of category-based (grammatical class) and attribute-based distinctions.

To watch, to see, and to differ: An event-related potential study of concreteness effects as a function of word class and lexical ambiguity

“…However, strikingly, no frontal concreteness effects emerged for syntactically and semantic ambiguous items used as verbs (AA-V). Thus, similar to the Kellenbach et al (2002) findings in German, a morphologically rich language in which word class ambiguity (and associated semantic ambiguity) is unlikely, we find that verbs in English do manifest concreteness effects, as long as they are not semantically ambiguous. The response to the semantically ambiguous verbs in the present study was, however, similar to that observed more globally to Chinese verbs in Zhang et al's (2006) study, in that the AA-V items elicited concreteness effects that were limited to posterior channels in the N400 time window.…”

“…Given that there are suggestions of multiple subcomponents of the ERP concreteness effect, it is further possible that different aspects of the response (e.g., the N400 and the slow frontal negativity) may be differentially sensitive to these factors, allowing these subcomponents to be more definitively separated. Finally, the differences in the results of the Kellenbach et al (2002) and the Zhang et al (2006) studies for concreteness effects on verbs, seen across languages that vary in degree of lexical ambiguity, leads to the specific prediction that concreteness effects for verbs in the present study may be more sensitive to ambiguity than concreteness effects for nouns.…”

“…In contrast, effects for verbs were quite small and limited to posterior electrode sites. However, word class did not interact with concreteness in a study using a recognition memory task with German words (Kellenbach et al, 2002). In that study, for both nouns and verbs, words with rich visual or motor semantic attributes elicited more negative ERP responses between 300 and 750 ms over anterior electrode sites.…”

“…Thus, the results of this study concur with the behavioral findings in suggesting that concreteness effects are absent-or greatly reduced-for verbs. On the other hand, however, Kellenbach, Wijers, Hovius, Mulder, and Mulder (2002), using a recognition memory task in German, found that concreteness had similar effects on both nouns and verbs, with more negative responses between 300 and 750 ms over anterior electrode sites for words with rich visual or motor semantic attributes. The authors did notice a tendency for this effect to be reduced for verbs in comparison with nouns, but the difference did not reach statistical significance.…”

Electrophysiological evidence for a natural/artifactual dissociation

“…Several studies using ERP methodology with a focus on the N400 component have revealed wave-form related differences associated with differences in semantic or grammatical category membership or with differences in orthographic properties, distinguishing between concrete versus abstract words (Paz-Caballero and Menor, 1999), nouns versus verbs (Pulvermüller et al, 1999;Kellenbach et al, 2002), items having high versus low numbers of orthographic neighbors (Holcomb et al, 2002). On the specific subject of the natural/artifactual dissociation, Kiefer (2001), using a superordinate object categorization task, found that the topographic distribution of the N400 component varied as a function of domain.…”