The Subjunctive and Indicative da-complements in Serbian: A Syntactic-Semantic Approach
thesisposted on 13.12.2012, 00:00 by Natasa Todorovic
The Subjunctive and Indicative da-complements in Serbian: A Syntactic-Semantic Approach Natasa Todorovic, Ph.D. Department of Slavic and Baltic Languages and Literature University of Illinois at Chicago Chicago, Illinois (2012) Dissertation Chairperson: Colleen McQuillen Although in Serbian traditional grammars it is often generalized that da+present complements are on par with infinitives in that they occur in complements where an infinitive is also possible, some scholars have recognized (Browne 1986, Ivić 1970, Gołąb 1964) that not all da+present complements exhibit same properties. Ivić (1970) notes that there are two different types of da+present complements: mobile and non-mobile. She claims that mobile da+present complements could be replaced by any other tense while non-mobile da+present complements could be replaced by infinitives and potential only. In her (1973) work, Ivić suggests that there is no clausal boundary between the volitional matrix verb and its da-complement by proposing that the two create one clause. Gołąb (1964) extensively analyzes da-complements by looking at the relationship they form with the matrix verb and the contribution aspect of the complement verb plays in that relationship. He emphasizes that complexity of da-complements can only be understood if the syntactic-semantic relationship between a da-complement and matrix verb is not disregarded. Analyzing da-complements from a syntactic-semantic perspective, this project builds on the earlier proposals by incorporating new theories. Adopting Progovac’s (2005) syntactic structure, which suggest that functional projections in Serbian come in two layers (subject and object), and Giannakidous’s (1998, 2009) semantic classification of verbs and notion of (non)veridicality, this work suggests that to properly analyze da-complements, the first step would be to dissolve da+present into da+imperfective present (non-past) and da+perfective present (non-past), and the second step would be to consider semantic properties of the matrix verb which select for da-complements. Within these parameters, I further claim that it is possible to distinguish between subjunctive and indicative da-complements. Indicative verbs select only for indicative da+imperfective non-past complements in addition to other tenses (past and future) while subjunctive verbs select only for subjunctive da+imperfective non-past or da+perfective non-past complements, excluding all other tenses. Additional tests with clitic placement, negation interpretation, and licensing of negative polarity items suggest different syntactic structures for indicative and subjunctive da-complements. Overall, the syntactic-semantic analysis of da-complements suggests that there are two different da(s): indicative da [+veridical] and subjunctive da [-veridical], both of which are associated with different functional projections. I claim that the indicative da is projected in a higher polarity phrase PolSP while the subjunctive da is projected in a lower polarity phrase, in PolOP. This analysis of da-complements particularly highlights that selection of da ([-veridical] or [+veridical]), aspect, tense, as well as the semantic properties of the matrix verb all equally contribute to the interpretation of the indicative or subjunctive mood in da-complements. Furthermore, as noted by Quer (2010: 168) “the verb meaning is not the only factor determining the choice of mood. [Verb’s] interaction with other elements like operators, aspect, or presupposition can be decisive in determining the mood in the embedded clause.” Quer’s observation for Romance clearly extends to Serbian data which suggest that the complexity of mood can only be resolved if both, syntactic and semantic approaches are taken into account.