Book Description

The latest developments in the field of computer technology have created new ways to share information without time and space limits. Computer technologies have not only made life easier and more accessible for users, but they have also opened up a new arena for illegal activities. These illegal actions have found an opportunity to spread via e-mails, websites, Internet chat rooms, forum pages, and social networking websites (like Facebook, Twitter, Instagram). Online contributors do not need to provide information such as their real names, the city where they live, age or gender in order to share their opinions, and such feelings of anonymity encourage criminal activities. Thus, disputed authorship cases have become one of the main challenges of the technological era. This research is a corpus-based simulated authorship casework application in Turkish. Texts for the corpora were collected from a collaborative online encyclopaedia – Eksi Sozluk (Sour Times) and Twitter. The corpus consists of 900 texts from 52 authors in total. However, 105 texts belong to seven authors from Twitter. The two methodological approaches that were applied are qualitative and statistical methods, according to Grant’s (2013) approach. Ten different tests were applied, depending on the various parameters that are forensically possible in real-world cases. Accordingly, the role of feature type, size, including the candidate author size, text size and a limited number of texts per author and finally cross-genre application were tested. The analyses revealed that such a combined approach has promising results in some tests in that they attributed authorship in Turkish. The findings of the research indicated that there is the potential to attribute unknown authors in Turkish and it appears that the results have significant conclusions for the broader application of forensic authorship attribution techniques in Turkish texts. Keywords: Authorship Attribution, Turkish, Forensic Linguistics, Authorship Analysis