Expression of epistemic modality in argumentative writing: A study based on learner corpora

This thesis investigates into how Turkish student writers employ epistemic modality devices in order to make argumentation in their writing. The research concentrates on understanding the ways in which the use of epistemic modality devices by Turkish student writers compare with the use of these same rhetorical features by American student writers in argumentative writing. The study also attempted to investigate how the degrees of epistemic strength differ (in terms of weak, medium and strong categories) between and among the student writers. With regard to the aims of this study, three sets of databases (one for the American students, the other two for the Turkish students) were analyzed in order to compare the epistemic modality features between and across the three groups of student writers in terms of frequency of use, diversity of use, grammatical categories of epistemic devices, co-occurrence patterns of notable epistemic devices and degrees of epistemic strength or commitment. The quantitative aspects of the study were mainly based on frequency counts of epistemic devices, Log-likelihood tests to determine significant differences of epistemic modality use between and across the three groups of student writers. The qualitative aspects relied mainly on a close examination of concordance lines to make comparisons between and across the three databases. Findings of the study revealed that the Turkish student writers are slightly more restricted in their choice of epistemic devices. It was found out that they use smaller set of epistemic devices than the American student writers do. Another striking finding is that this smaller set of epistemic devices is used with higher frequency compared with the American students. This is obvious especially in the case of epistemic lexical verbs, where the Turkish student writers rely frequently on the 'I think' expression. Concerning the distribution of epistemic devices across grammatical categories, it was observed that the Turkish student writers principally use lexical verbs and adverbs in their expression of epistemic modality, but the American student writers appear to prefer modal auxiliary verbs more frequently. The Turkish student writers' preferences of particular modal verbs are also different from those of the American student writers. The Turkish students use will in relatively higher frequencies than would, but the American students use would more frequently than will. Furthermore, the Turkish student writers prefer a more personal style of argumentation than the American student writers. In view of the high frequency of expressions such as I think (that) and the frequent use of of course and maybe, it might be assumed that the Turkish student writers exhibit a more personal and straightforward writing style. Another interesting finding concerns the use of epistemic adverbs 'perhaps' and 'certainly'. These adverbs are not detected in the AELT database by the Turkish student writers while they are frequently used by the American student writers in the US-ARG database. Lastly, it has been observed that the American student writers prefer to use more number of epistemic devices indicating medium and weak level of epistemic strength, while the Turkish students in both databases tend to prefer more number of epistemic devices expressing strong level of epistemic commitment. Keywords: Epistemic modality, Epistemic devices, Argumentative writing, Epistemic strength, Degrees of commitment/ strength