![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
Abstract
Background: Recent research on the comprehension of subject and object wh-questions in English Broca's aphasic speakers has shown the existence of a specific pattern: subject and object who questions, as well as subject which NP questions, are comprehended above chance, while object which NP questions are comprehended at chance. Since English is a strict word-order language, the subject–object asymmetry is identified as an asymmetry both in word order and in thematic role assignment. The present study investigated the comprehension of wh-dependencies in Croatian, which is a language with relatively free word order that does not necessarily associate thematic roles with specific syntactic positions.
Aims: The goal of the study was to determine whether the pattern of comprehension of wh-dependencies that has been reported for English Broca's aphasic speakers is also exhibited by Croatian aphasic speakers. We hypothesised that due to the role of morphological features such as case in sentence comprehension in a free word-order language like Croatian, aphasic speakers of that language would exhibit a different pattern.
Methods & Procedures: We tested the comprehension of six types of wh-dependencies formed with tko “who” and koji “which” (direct, embedded, long-distance with za, long-distance without za, and passivised questions, and relative clauses with koji) in six Croatian aphasic speakers. All participants had had a lesion to the left hemisphere: three of them were Broca's aphasic speakers and three were mixed cases of non-fluent aphasia.
Outcomes & Results: The main finding of the study is that Croatian Broca's aphasic speakers do not exhibit the same subject–object asymmetry that has been reported for English Broca's aphasic speakers. Some asymmetry was found in the three patients with mixed aphasia, whose comprehension was better on object- than subject-extracted structures.
Conclusions: The asymmetric pattern found in the cases of mixed aphasia runs in the opposite direction to the asymmetry typically reported in the literature. Relying on cross-linguistic findings from the processing-based research on subject–object comprehension asymmetries in neurologically intact populations we assume that, in Croatian, thematic roles are assigned directly to case-marked arguments without the mediation of syntactic information, and propose a morphological account of the Croatian data based on the validity of case cues in MacWhinney and Bates' competition model.
We would like to thank Professor Dr Vida Demarin of the Neurological Clinic at Bolnica Sestara Milosrdnica, Zagreb, Dr Marija Paškvalin at the Polyclinic Suvag in Zagreb, and Višnja Akrap-Kotevski and Zvjezdana Trifunovi1 for their assistance in finding the participants for this study. We are also grateful to the Croatian participants who allowed us to investigate their comprehension of their language. Thanks are also due to Marie-Odile Junker, Jack Kelly, Jo-Anne LeFevre, as well as to David Caplan, the audiences at the Science of Aphasia VIII, and Auditory and Audiovisual Language Processing in Aphasia 2008 workshop for helpful comments and suggestions on earlier ideas and versions of this paper. Finally, we thank the two anonymous reviewers for their questions and comments.
Notes
The present manuscript is based on a study that was conducted by the first author and defended by her as a doctoral dissertation under the title “Comprehension of wh-dependencies in Broca's aphasia” at the Institute for Cognitive Science, Carleton University, Ottawa, in 2005. The authors contributed equally to the present manuscript.
1Alongside linguistic explanations for the subject–object and who–which NP asymmetries in agrammatic comprehension are processing-based accounts of these asymmetries in normal comprehension (see, for example, Fiebach, Schlesewsky, & Friederici, Citation2001, 2002; King & Just, Citation1991; King & Kutas, Citation1995; Gibson, Citation1998). These studies also provide evidence, using a variety of methodologies such as reading times, ERPs, and fMRI, that object-extracted structures are more difficult to process than subject ones. Processing explanations for this asymmetry are based on the increased demand on syntactic working resources in object structures because of the greater distance between the wh-element and its associated gap. As for aphasic speakers, it is predicted that structures that are difficult for normal speakers to comprehend would also be difficult for aphasic speakers due to their compromised processing resources.
2One exception is Thompson, Tait, Ballard, and Fix's (Citation1999) study, in which they attempted to replicate Hickok and Avrutin's (Citation1996) results on subject and object who and which NP questions. They also included passivised wh-questions in order to test the predictions of Grodzinsky's (Citation1995) trace-based account. Thompson et al.'s results revealed that only one of their four participants (Subject 3) exhibited Hickok and Avrutin's who/which NP pattern. As for passive questions, none of the participants performed as predicted by the trace-based account. Unconvinced by theories that differentiate who and which NP questions syntactically in terms of Binding chains or referentiality, Thompson et al. suggested an alternative account of agrammatic comprehension based on the integration of syntactic and semantic representations: which NP questions are semantically more difficult to process because they require two computations to interpret, whereas who only requires one.