Cognitive status and ra-marked referents of nominal expressions in Persian discourse

University of Minnesota Ph.D. dissertation. May 2014. Major: Linguistics. Advisor: Jeanette K. Gundel. 1 computer file (PDF); xi, 171 pages, appendices A-C.This dissertation examines the referential status of the morpheme ra in Persian discourse. A nominal expression in Persian can occur with or without ra as illustrated below: a. nævar ra zæbt kærd tape ra record did `s/he recorded the tape' But not * `s/he recorded a tape' b. nævar zæbt kærd tape record did `s/he recorded a tape' But not * `s/he recorded the tape' a. nævar zæbt shod tape record became two possible interpretations: `the tape was recorded', or `a tape was recorded' But not b. * nævar ra zæbt shod. However, the exact nature of the conditions under which ra occurs is a source of disagreement in the literature on this morpheme. Various, and sometimes contradictory, assertions about the functions of this Persian morpheme have been made in the literature. While most researchers agree that 'ra' primarily marks direct objects (albeit there is evidence that it can also potentially mark other non-subjects), they differ in whether these nominal expressions are marked for definiteness (Phillott 1919, Lambton 1953, Ghomeshi 1997), or specificity (Browne 1970, Karimi 1989, Windfuhr 1990). In defining the referential status of ra-marked nominal expressions, I use a framework, the Givenness Hierarchy (Gundel, Hedberg, and Zacharski (1993)), that makes it possible to provide a more precise definition of `definiteness' and `specificity'. This framework allows for more precise and fine-grained primitive notions than the traditional concepts of specificity and definiteness. Distinctions in the Givenness Hierarchy are made in terms of cognitive statuses (the location of referents in the memory and attention of a hearer) at the point in the discourse just before a particular linguistic form is used. In this dissertation I examine naturally occurring data from forty (40) different sources, as well as constructed examples. The purpose for choosing a variety of sources of data for this study is to ensure that the results reflect language in use in a variety of genres and contexts. As such, I examine data that includes short stories, newspaper articles, children's stories, and a telephone conversation. The data examined contains a total of 17,611 words, yielding 408 instances of ra-marked DPs. Each expression containing `ra' is examined and coded for the highest cognitive status of the referent of the nominal expression. Using the cognitive statuses outlined in the Givenness Hierarchy, I examine the referent of each nominal expression containing `ra' to determine if it is type identifiable, referential, uniquely identifiable, familiar, activated, or in focus. I conclude that the referents of DPs marked with only ra are uniquely identifiable or higher. However, a more nuanced situation arises when ra co-occurs with another referential marker, namely -i (whose referential status remains unclear)