engleski

Croatian Discourse Corpus of Speakers with Aphasia (CroDA): overview and first data

Corpus-based research is a common practice for studying language variations in the field of linguistics (Svartvik, 1992). Over the past few years, it has also become the most prominent method in speech-language pathology (Ferguson et al., 2009). Although corpora are objective and comprehensive resource of language data collected in different modality (written vs spoken) and in various situations (formal vs informal), there is still a considerable lack of the ones that provide data about language of speakers with aphasia (SwA). Croatian Discourse Corpus of Speakers with Aphasia (CroDA), developed within the project Adult Language Processing (HRZZ-2421-UIP-11-2013), is the first specialized corpora which contains discourse samples of Croatian SwA. By creating such a database, we are now able to gain different findings about language and communication of SwA. Currently CroDA includes discourse samples of 14 speakers with fluent (7) and non-fluent (7) aphasia, with all the demographic data registered, as well. The aim of this presentation is to provide a general overview of the CroDa and to present first data extracted from it. The sampling protocol has been done according to the AphasiaBank Protocol which includes the following discourse tasks: free speech samples, picture descriptions, story narrative and procedural discourse. The analyses were completed using basic language measures. The preliminary analyses show that SwA have the biggest production during their spontaneous speech (T: M=762, 64 ; SD=405, 32), which is also a discourse in which their syntactic complexity seems to be the biggest (MLU: M=5, 37 ; SD=1, 76), but on the other hand, then their lexical diversity is the smallest (T/T ratio: M=0, 25 ; SD=0, 08). This measure seems to be the biggest in the procedural discourse task (M=0, 51 ; SD=0, 19), in which these persons produce the least words (M=60, 07 ; SD=38, 53). Even though a certain production pattern is obvious, quantitative analyses are not sufficient and informative enough when it comes to language of SwA. Corpora are therefore a great tool from which researchers and practitioners can extract data regarding qualitative characteristics of production (i.e. articulation or syntactic errors, neologisms, paraphasias...). According to these kinds of data, researchers can possibly gain better insights into the underlying mechanisms of the condition, and practitioners can plan their therapy sessions in order to make them more effective.

aphasia ; corpus ; language abilities ; discourse

nije evidentirano