Processing of nominal compounds and gender-marked determiners in aphasia: Evidence from German

Abstract

The present study tests theories about the representation of compound nouns and grammatical gender in the mental lexicon. Comprehension and production of determiner–compound-noun phrases were examined in three aphasic native speakers of German, a language that marks grammatical gender on definite determiners of nouns.

In picture naming, participants were more impaired in retrieving compounds than matched simple nouns and showed different error patterns. However, retrieving the correct determiner was equally impaired for compounds and simple nouns. Clear dissociations between impaired determiner retrieval in production and relatively preserved processing of determiner–noun phrases in comprehension were observed for existing compounds and simple nouns. In contrast, processing of novel compounds was more impaired in both modalities, and gender-mismatch effects were especially observed for novel compounds. The results support the account of decomposed word forms and holistic lemma representations of compound nouns in the mental lexicon.

• Compound production
• Gender
• Determiners
• Semantic transparency
• Aphasia.

We are indebted to all participants of this study. We thank the Fürst-Donnersmark Haus in Berlin, Franziska Machleb for testing F.W. and H.J., and Eva Henze, Regina Herholz, and Lisa Lütteken who assisted in the development of materials. We are also grateful to two anonymous reviewers for their helpful comments on an earlier version of this paper.

Notes

1 Janssen et al. (2008) employed word familiarity ratings instead of using word frequency measures.

2 In addition to determiners, case and number can also be marked on German nouns by a corresponding affix. In the materials we used (Burchert et al., 2011), the case-marked sentences only include nouns, which are not overtly marked for case, but case is marked on the determiners of masculine nouns here (der_Nom Sohn → den_Acc Sohn = the son). In the number-marked sentences, however, number is marked on the nouns, too (e.g., das_Sg Kind → die_Pl Kinder = the child → the children).

3 The chance level was assumed to be at a proportion of 33.3% correct determiner retrieval.

4 For none of the participants neither interpretability nor the interaction between interpretability and gender match significantly contributed to the model (all p > .1, chi-squared test), whereas gender match had a significant impact for H.J. (χ² = 20.268, p < .001) and a marginally significant impact for M.M. (χ² = 3.195, p = .074).

5 An alternative assumption, put forward by one anonymous reviewer, is that determiner lemma nodes are accessed in addition to grammatical gender coded with the lemma. On the basis of almost perfect determiner decisions in our participants, we can exclude such a deficit as an explanation of the data.