Doctoral Thesis

Moskovsky conducted a pioneering study of pronominal binding in Bulgarian within the framework of the then influential Government and Binding Theory. Because Bulgarian has a richer pronominal system than English, including a full set of pronominal clitics, as well as reflexive possessive pronouns, the Bulgarian language contains a wealth of interesting new data yielding both empirical and theoretical insights which the English data alone do not have the capacity to provide. Based on data analyses from Bulgarian, as well as a number of other mostly Slavic languages, Moskovsky came to the conclusion that pronominal binding is at the intersection of syntax and pragmatics, in which structural and discourse factors play complementary roles in determining the distribution and interpretation of referentially deficient elements like pronouns. The dissertation draws attention to some interesting differences in the distribution of reflexive and non-reflexive pronouns in 1st and 3rd person. Here are a few English examples.

John thinks highly of him/himself. (co-referential reading of 'John' and 'him' is not acceptable)
I think highly of me.
Sarah was not amused by Lucy's joke about her/herself. (co-referential reading of 'Lucy' and 'her' is not acceptable).
Sarah was not amused by my joke about me.

The unacceptability of the non-reflexive pronoun (in the co-referential reading) in the 3rd person sentences cannot be attributed to structural factors, otherwise the 1st person sentences would have been equally unacceptable. Such data rather suggest that, because 3rd person pronouns can have multiple referents, in some 3rd person sentences the use of a reflexive pronoun is driven by reasons of disambiguation (rather than by structural factors)—in other words, in such instances discourse factors (e.g. "avoid ambiguity") come into play and influence the distribution and interpretation of an anaphoric element.

Moskovsky's thesis argues that such facts have significant theoretical implications. It is worth noting that the vast majority (if not all) of the existing studies of pronominal binding are based on 3rd person data. But given the evidence provided above, in at least some 3rd person sentences a pronoun's binding is the product of both structural and discourse factors. Because of that, 3rd person data may not be well-suited for insights into structural aspects of pronominal binding. This is particularly relevant to the formulation of locality constraints on pronominal binding: in most extant formulations, the use of the non-reflexive pronoun in the 1st person examples above would be in locality violation and would be ruled out as ungrammatical. The thesis suggests that it would be worthwhile revisiting the issue of locality constraints exclusively on the basis of 1st and 2nd person data.