Demonstratives have traditionally been analyzed as `pointing words' whose primary function is to indicate relative spatial or temporal distance of a referent from speech participants. Recent research argues that the meaning of demonstratives is not limited to spatial distance and has given alternative accounts for the use of demonstratives that focus on other cognitive and pragmatic meanings (e.g., Gundel, Hedberg and Zacharski 1993, Enfield 2003, Botley and McEnery 2001, OH 2001). This dissertation contributes to research that looks at alternative meanings for demonstratives, focusing on Tunisian Arabic (TA). The goal of the dissertation is two-fold. First, working within the Givenness Hierarchy framework (Gundel et al.), it aims to show how TA demonstratives are used to indicate cognitive status, the assumed memory and attention status of a referent in the mind of the addressee. A combined methodology of questionnaires and corpus analysis is used to test hypotheses formed in a previous study (Khalfaoui: 2004) about proposed correlations between cognitive status and single demonstrative forms in TA and extend the analysis to phrases with double demonstratives. The second goal of this dissertation is to show how other factors can further restrict the choice among certain demonstrative forms that encode the same cognitive status. Specifically, it is shown that when there is more than one activated referent, communicators choose the demonstrative haða as a determiner, but not as a pronoun, although both the determiner and the pronoun encode the same cognitive status. I argue that Relevance Theory (Sperber and Wilson 1986/1995) provides a cognitive explanation for why communicators avoid the demonstrative pronoun in such case. This dissertation also discusses the advantages and limitations of the questionnaire and the corpus analysis as research tools.