| Summary: | Introduction
People with aphasia (PWA) presented sentence processing deficits (Berndt et al, 1997; Caplan & Hildebrandt, 1988). However, linguistic features associated with sentence processing deficits might differ across languages. In verb-final languages such as Korean, case markers are one of the critical factors to denote thematic roles in a sentence. However, there are few studies, which specifically tested abilities of case marker assignment for PWA using a verb-final language. The current study investigated 1) whether PWA presented greater difficulties in a case marker assignment task (CMAT) compared to their controls and 2) which syntactic structure best predicts aphasia severity.
Methods
Twenty-five individuals with aphasia and 25 age- and education-matched controls participated in the study. The diagnosis of aphasia was based on the administration of the Korean Version of Western Aphasia Battery (K-WAB) (Kim & Na, 2001). Aphasia quotients (AQs) from the K-WAB ranged from 6.6 to 98 (mean AQ = 59.79, SD=28.72). The normal group showed normal range of the Korean version of the Mini-Mental State Examination (Kang, 2006). All participants were native speakers of Korean.
The CMAT consisted of active-2place, active-3place and passive syntactic structures with either canonical or noncanonical word-order: 1) active 2place-canonical, 2) active 2place-noncanonical, 3) active 3place-canonical, 4) active 3place-noncanonical, 5) passive-canonical, and 6) passive-noncanonical. In the CMAT, each noun phrase and verbs were displayed below the target picture, and participants need to put each case marker next to the noun phrases.
Results
1. ANOVA analyses
Three-way mixed ANOVA (sentence type x canonicity x group) revealed main effects for Group (F(1, 48)=58.505, p<.0001), and Canonicity (F(2, 96)=16.57, p<.0001), indicating that PWA and noncanonical word-order elicited worse performance on the CMAT than their counterparts (Figure 1). The main effect for Sentence type was significant (F(2, 96)=14.956, p<.0001). Bonferroni post-hoc comparisons revealed that passive sentences were more difficult than active-2place (p=.001) and active-3place sentences (p<.0001), whereas there were not significant differences between the 2- and 3-place active sentences. No other effects were significant.
2. Regression analyses
Stepwise regression analyses were performed with each condition as independent variables for dependent measures for overall aphasia severity (AQ), and scores from comprehension, repetition, fluency, and naming domains. Passive-canonical condition was selected as a significant factor for AQ, F(1, 23)=11.82, p=.002, repetition, F(1, 23)=8.427, p=.008, comprehension, F(1, 23)=7.044, p=.014, and naming, F(1, 23)=9.512, p=.005. Passive-noncanonical structure significantly predicted performance on Fluency F(1, 23)=12.408, p=.002.
Discussion
Individuals with aphasia demonstrated greater difficulties in the case marker assignment compared to their normal control group. Furthermore, noncanonical word-order and passive sentences elicited more errors on the task than canonical and active sentences. Passive sentences were the significant predictors for overall aphasia severity. The results suggested that PWA using a verb-final language with well-developed case-marking systems presented deficits in case marker processing. The syntactic structure and canonicity of word order need to be considered as critical linguistic features in testing their performance on dealing with case markers.
|
|---|
