Information about word class is both semantically and lexically represented: evidence from an advantage for verbs in two speakers with aphasia

ABSTRACT

Background

There has been a debate over how a word’s syntactic class is represented in the mind. One model claims that information is represented solely at the semantic level while another account holds that syntactic class information could be represented both at a semantic and an independent lexical level.

Methods

Two aphasic participants with a significant advantage for naming verbs in comparison to nouns took part in the investigation. They were assessed with several picture naming tests, two comprehension tests as well as an object naming test comparing animate and inanimate objects.

Results

In one participant, the pronounced deficit for nouns resulted from a semantic impairment: there was a comparable dissociation in comprehension of words, a nonverbal semantic deficit, and performance was mainly affected by semantic variables (imageability; animacy). In the other participant, the difference between nouns and verbs arose from a lexical impairment since no comparable comprehension deficit was observed. In addition, performance was affected, mainly, by lexical variables (word frequency).

Discussion

Results from one participant suggest a semantic locus of noun–verb differences and provide empirical support for a specific model of objects’ and actions’ semantic representations. In contrast, the other participant’s performance strongly suggests the word class effect to arise from a lexical impairment. Thus, the study provides support for two independent loci of noun–verb differences thereby contradicting one account’s strong claim that noun–verb differences arise solely from semantic impairments. The results also speak against recent cognitive accounts rejecting the idea of independent lexical representations.

Acknowledgments

We would like to thank the participants for their considerable time and patience. We gratefully acknowledge Gerhard Blanken’s extensive support and discussion at all stages of this study. It was his humble choice not to be included as co-author. Christin Scheidler was supported by a Martin-Wieland- scholarship of the University of Erfurt. We acknowledge the helpful comments of an anonymous reviewer.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. In set 1 of Experiment 1 (72 nouns vs. 72 verbs), only eight of the unimpaired control participants could be included.

2. In our experimental investigations, age of acquisition (AoA) was controlled for, but effects of AoA were not evaluated separately due to the ambiguous status of AoA (cf. Belke et al., 2005).

3. The same argument applies to RL’s advantage for inanimate over some animate nouns although the raw scores (0 vs. 3 correct out of 20) are quite low and there was no effect at all in a second analysis.