The distribution of pronoun case forms in English
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
This thesis investigates the influence of linguistic factors on the distribution of pronoun case forms in Modem English and argues that the alternation between nominative and objective pronoun forms is a surface phenomenon best captured in a probabilistic constraint-based approach, where constraints are weighted and the combined weight of constraint violations determines the probability of occurrence of a particular variant. I propose that the distribution of both weak and strong pronoun forms in English is affected by the interaction of two structural case constraints: Argument Case, which restricts the overt case form of structural arguments of a predicate; and Positional Case, which constrains the form of pronouns that appear as the specifier of an agreement-related functional head at Spell-Out. Pronouns that occupy surface positions not covered by the Positional Case constraint are further influenced by a Default Case constraint that calls for objective pronoun forms. A survey of data reported in existing studies suggests that all instances of pronoun case variation that cannot be given a purely case-based account occur in strong pronoun contexts. The consistent nominative/objective case distinction found with weak pronouns is due to their syntactic deficiency and the increasing importance of Positional Case in English. Unlike strong pronouns, weak pronouns must be licensed by an agreement-related functional head at Spell-Out, which means that they will generally be subject to the Positional Case constraint as well as the Argument Case constraint. Strong pronouns, on the other hand, tend to occur in positions not covered by Positional Case, which leaves them open to other influences. I present results from a written survey of 90 speakers of English, which indicate that strong pronoun forms no longer merely identify the structural case of a pronoun, but also code its position within a syntactic construction, and identify its morphosyntactic status as a strong pronoun. These additional functions of strong pronoun forms are captured in two Relative Positional Coding constraints and a set of Invariant Strong Form constraints. Variation occurs where the demands of the case constraints clash with the requirements of Relative Positional Coding and the tendency towards invariant strong pronoun forms. The case trends reported in existing studies suggest that Relative Positional Coding and the tendency towards invariant forms affects not only personal pronouns but also wh-pronouns. For personal pronouns, the emerging invariant forms are the objectives me, him, her, us, them, but for wh-pronouns, the emerging invariant forms are the nominatives who and whoever. As a result, the Invariant wh-form constraints clash with the three case constraints in different environments than the remaining Invariant Strong Form constraints. Discrepancies between the grouping of pronoun forms associated with stmctural case and the grouping of pronoun forms associated with Relative Positional Coding are largely responsible for the distributional differences between strong Isg (l/me) and non-lsg forms (he/him, she/her, we/us, they/them, who/whom). For the purposes of stmctural case, I groups with the non-lsg nominatives he, she, we, they, who, and me groups with the non-lsg objectives him, her, us, them, whom. For Relative Positional Coding, on the other hand, I patterns with him, her, us, them, whom, and me patterns with he, she, we, they, who. All of the trends identified in this study point to an increasing influence of surface position on pronoun case choice, which can be seen as a correlate of the shift from morphological to positionallicensing at the end of the Middle English period.