Encoding topic, focus and contrast: Informational notions at the interfaces

The dissertation deals with the encoding of discourse notions in sentence structure. Its aim is twofold: a) to elaborate a semantics for ‘topic’, ‘focus’ and ‘contrast’; b) to investigate how these three notions are coded in grammar. As for the former issue, it has been attempted to formulate definitions which could be able to identify each of the three informational units unequivocally. More specifically, after discussing some difficulties inherent in Reinhart’s definition of topic in terms of ‘aboutness’, I will identify topics with constituents appearing in syntactically marked structures and interpreted as necessarily d-linked. On the other hand, the definition of focus will be couched within the Alternative Semantics framework (Rooth, 1992). Most of the theoretical effort, however, will be directed toward showing that contrast is an autonomous informational unit, contrary to those authors who claim that it should be reduced to focus, according to the observation that they both evoke alternatives in discourse. In particular, I will introduce plenty of data in which the semantic contribution of contrast can be easily distinguished from the one of focus. Instead of complicating the ontology of informational notions, the hypothesis about the autonomy of contrast offers new insights into the linguistic nature of both contrastive topics and contrastive foci. Contrary to those authors that treat the latter as independent categories of information structure, I will show that their behavior could be explained in terms of the interaction between contrast, on the one hand, and topic and focus, on the other hand. As for the latter issue, the main purpose of the dissertation is to understand how the different components of language interact with each other for the expression of the abovementioned information-structure related notions. In particular, I will provide theoretical and empirical evidence against the idea that informational notions are coded as discourse-features heading their own functional projections in the syntactic structure (hence, against the analyses of information structure couched within cartographic frameworks). With this regard, the analysis of the linguistic encoding of contrast will be the most revealing one: it will be shown that the expression of this informational category correlates with variations in the prosody of the associated constituent, which is hardly explainable in terms of operations involving the underlying syntactic structure. It will be claimed that the analysis of the experimental material is more compatible with the line of investigation (recently pursued by authors like Tania Reinhart and Ad Neeleman), according to which informational notions are coded at the interface of syntax with the external systems of use, i.e., outside the computational system. This assumption is grounded on the more general hypothesis that language is an optimally-designed system, so that the interfaces are able to code information on the basis of visible properties of the syntactic derivations. Languages express informational notions through a large variety of linguistic means, either syntactic (e.g., movement to the left periphery exhibited by English topics and base-generation in the C-domain of their Italian counterparts) or phonological (e.g., main sentence stress assignment on focused constituents and increase of the phonetic prominence associated with contrastive phrases). In this dissertation, I will not account for this cross-linguistic variability. Rather, I will deal mainly with Italian data, extending the analysis to other languages whenever the comparative analysis will lead to a deeper understanding of the phenomena of stake each time. For example, my hypothesis on the semantics of Italian contrastive foci will be inspired by the observation of the linguistic behavior of their Hungarian counterparts or the comparison between Italian and Spanish will shed new light on the linguistic nature of postfocal topics.