Utjecaj kognitivne složenosti zadatka na samoispravljanja

Abstract

This paper investigates the influence of the cognitive complexity of a certain task type on the distribution of different categories of error self-repairs and appropriacy repairs. A recorded speech sample, in the Croatian language, lasting for approximately eight hours has been transcribed on a speech sample of 101 students at the Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture in Split. The classification of self-repairs is based on Levelt’s model of speech production, as the empirically best supported theory of monolingual speech processing. The students have been individually tested by performing five speech tasks and their speech has been recorded. The tasks included: a) narration of the chronological order of events on the example of a cartoon, b) room description, c) repeated room description with different furniture arrangement, d) semantically unrelated utterance formulation based on pictures, and e) story telling based on a sequence of pictures. The retelling of the chronological order of events resulted in a higher frequency of syntactic and morphological error repairs compared to other tasks, whereas the frequency of lexical and phonological error repairs was not influenced by the task type. Furthermore, different information repairs occurred more frequently in the cartoon retelling task, compared to the description of rooms and the formulation of semantically unrelated utterances.