Detecting plagiarism in the forensic linguistics turn

  • Rui Sousa Silva

    Student thesis: Doctoral ThesisDoctor of Philosophy

    Abstract

    This study investigates plagiarism detection, with an application in forensic contexts. Two types of data were collected for the purposes of this study. Data in the form of written texts were obtained from two Portuguese Universities and from a Portuguese newspaper. These data are analysed linguistically to identify instances of verbatim, morpho-syntactical, lexical and discursive overlap. Data in the form of survey were obtained from two higher education
    institutions in Portugal, and another two in the United Kingdom. These data are analysed using a 2 by 2 between-groups Univariate Analysis of Variance (ANOVA), to reveal cross-cultural divergences in the perceptions of plagiarism. The study discusses the legal and social circumstances that may contribute to adopting a punitive approach to plagiarism, or, conversely, reject the punishment. The research adopts a critical approach to plagiarism detection. On the one hand, it describes the linguistic strategies adopted by plagiarists when borrowing from other sources, and, on the other hand, it discusses the relationship between
    these instances of plagiarism and the context in which they appear. A focus of this study is whether plagiarism involves an intention to deceive, and, in this case, whether forensic linguistic evidence can provide clues to this intentionality. It also evaluates current
    computational approaches to plagiarism detection, and identifies strategies that these systems fail to detect. Specifically, a method is proposed to translingual plagiarism.
    The findings indicate that, although cross-cultural aspects influence the different perceptions of plagiarism, a distinction needs to be made between intentional and unintentional plagiarism. The linguistic analysis demonstrates that linguistic elements can contribute to finding clues for the plagiarist’s intentionality. Furthermore, the findings show that translingual plagiarism can be detected by using the method proposed, and that plagiarism detection software can be improved using existing computer tools.
    Date of Award12 Jun 2013
    Original languageEnglish
    Awarding Institution
    • Aston University
    SupervisorTim Grant (Supervisor)

    Keywords

    • intention and intentionally
    • punitive turn
    • computational forensic linguistics
    • translingual plagiarism
    • linguistic evidence

    Cite this

    '