Pronoun hypothesis. The objective conjugation suffixes contain incorporated pronouns, which function as the true argument of the verb. – As a corollary, free accusative nominals are really adjuncts, rather than arguments. This section demonstrates that the pronoun hypothesis has quite a few problems that the agreement marker hypothesis does not suffer from. The problems lie in three ar- eas: null anaphora, extraction islands, and the insensitivity of the use of the objective conjugation to any of the properties that govern the use of pronominal clitics. We conclude that the objective conjugation suffixes are agreement markers. We use the term “agreement” in the sense of Xxxxxx (1978): “systematic covari- ance between a semantic or formal property of one element and a formal property of another” (Xxxxxx 1978: 610). In Hungarian, a formal property ([DEF +]) of the object covaries with a formal property (conjugation) of the verb.7 We also mean “agreement” 6 Example (23) is E´. Kiss’s (1987) (26a)). 7 Some scholars are reluctant to call this phenomenon “agreement”. Xxxxxxxxx (1999: 336) writes: “in Hungarian the verbs in the objective conjugation do not actually show agreement with the object, but simply mark it for definiteness”. Siewierska (1999: 244) writes that the Hungarian objective conjugation suffixes “represent a combination of [subject agreement] and what (Xxxxxxx 1992: 49) calls O[bject] reg- istration;” on page 245 she writes: “In view of the fact that the markers of the object conjugation do not in contrast to “incorporated pronoun”. Thus, accusative-marked nominals are true ob- jects, and the affix does not have the ability to refer. In Lexical Functional Grammar (LFG) terms, our thesis is that the objective conjugation affixes do not carry a PRED feature. Moreover, we find that Hungarian objective conjugation suffixes lack even those ‘quasi-pronominal’ properties of certain elements in other languages that may be seen as intermediate between incorporated pronouns and grammatical agreement mark- ers. For example, in many languages a verbal inflection functions as an agreement marker when associated with a nominal argument, but as an incorporated pronoun in the absence of a nominal argument. Bresnan and Mchombo (1987) analyze the sub- ject markers on Chichewˆ a verbs in that way, and model them in the LFG framework as bearing an optional PRED feature. Another instance of quasi-pronominals is found in the phenomenon of clitic doubling, where a clitic co-occurs with an...
