Evidentiality Strategies as Distinguishing Markers in Forensic Authorship Analysis

The current research starts from the premise that authors form idiosyncratic habits in how they use evidentiality strategies in academic writing. The goal of the research is to examine whether the frequency of evidentiality strategies is authorspecific and if lexico-grammatical expressions used to denote it could be useful distinguishing markers in authorship analysis. Evidentiality is, here, understood in a narrow sense, as a discourse function with a primary meaning to denote source of information. A modification of Gurajac’s (2010) classification is used to identify and classify evidentiality. The corpus for this research consists of 19 samples coming from 5 authors, which results in approximately 15,000 words total. The qualitative aspect of the research implies identification and classification of lexicogrammatical expressions used to denote evidentiality, as well as recognizing certain author-specific tendencies in the usage of given expressions. The frequency of evidentiality strategies is expressed per 1000 words and in percentages. One-way analysis of variance is used to test whether between-author variability is higher than within-author variability, that is, whether the tested parameters are useful distinguishing markers in authorship analysis. The results show that the percentage of assumptive inferred evidentiality is rather constant across samples of each author. In addition, there is evidence that the relation of assumptive inference to deductive inference and general knowledge are also idiosyncratic.