A Categorial-Modal Logical Architecture of Informativity Dependency Grammar Logic & Information Structure

The dissertation elaborates how a categorial-modal logical framework, Dependency Grammar Logic (DGL), can be used to provide a basic model of certain components underlying the Praguian conception of linguistic meaning. The components dealt with here are head/dependent-asymmetries, “semantic roles” or dependency relations, and information structure. We address the representation, realization, and interpretation of these components, paying particular attention to cross-linguistic modeling. We propose hypotheses that predict cross-linguistic commonalities and differences in realizing dependency relations and information structure, and explain how these predictions can be integrated into multilingual grammar fragments that model these phenomena. Set within the Praguian tradition, the dissertation argues that each of the aforementioned components are necessary to explain sentential form and interpretation. Information structure is a means that the speaker employs to present some parts of the sentence’s meaning as context-dependent, and others as context-affecting. It is therefore an inherent aspect of a sentential meaning. Languages can realize information structure using various means, like variation in word order, tune, or morphological marking. The distinction of heads and dependents, rather than “phrases”, offers a flexible notion of surface structure that is needed for the modeling of word order variation and tune, both typologically and formally. Distinguishing dependency relations like Actor, Patient, or Beneficiary leads to a fine-grained description of sentential meaning. The dissertation argues that named dependency relations are necessary for explaining the realization of information structure and focus projection. Furthermore, they play a role in the interpretation of linguistic meaning. The dissertation explains how causal and temporal dependency relations can effectuate aspectual change, and how interpreting a word as a particular type of dependent can give rise to entailments that must be accommodated in the discourse context for the linguistic meaning to be coherent. Finally, distinguishing named dependency relations offers the means to model the resolution of phenomena such as exempt anaphora. The dissertation consists of two parts. The first part describes the foundations of DGL. DGL produces and represents linguistic meaning by coupling a resource-