Exploring Register Variation in Turkish Web Corpus

Abstract

<p><br></p><p>In linguistics, web registers are language varieties occurring on the web such as news reports and editorials. Most of the previous web register research has been done for Indo-European languages. Additionally, previous research has mainly focused on the restricted corpora with pre-determined registers. This article describes Turkish web registers on the web. The data is Turkish web register corpus which consists of 2601 web texts. A taxonomy was adapted to register label these texts. The manual annotations of the texts were done with the adapted taxonomy, and the registers were defined accordingly. Text dispersion keyword analysis was used to generate the keywords of the registers and examine the basic linguistic characteristics of them. The results display the web registers existing for Turkish, and the linguistic characteristics associated with the news report and editorial registers. Keywords: Turkish web registers, manual annotation, text dispersion keyword analysis.<br></p>