engleski

Challenges in the development of written corpus of adult speakers

In two projects Adult language processing (CSF – ALP-2421) and Computer assistant supporting text input for individuals with language disorders (EU – Structural fund ; RC.2.2.08- 050), being carried out by the group of experts in linguistics, language pathology and computer sciences, the final goal is to create the computer application that will support individuals with language impairments in writing process. To achieve this goal, it was necessary to collect written samples of children and adults, both with and without language impairments, and to develop the Croatian corpus of written language (CCOWL). The CCOWL presents a comprehensive database for further research and it is intended to be an online available database. The CCOWL consists of approximately 5600 samples (texts) written by 401 participants in the age range from 11 to 89+ (gender and education level controlled, as well). Moreover, the participants differ in terms of the existence of language impairment ; 134 of them are healthy, 91 have aphasia (acquired impairment) and 176 have dyslexia (developmental impairment). Written samples differ in the level of text structures and the writing media (pen and keyboard). While designing a plan on how to create a comprehensive corpus suitable for contemporary research and while collecting the written samples, a lot of questions came up: which participants to include (considering age, gender, education, etc.) ; how to collect written samples, by handwriting or typing ; how to create tasks for different text structure levels ; how to adapt these tasks to children, etc. The main aim of this paper is to present and discuss the greatest challenges in creating CCOWL since here presented challenges are prerequisites for development and analysis of written corpora in general.

written corpus ; comprehensive corpus ; creating corpus

nije evidentirano