Semantics and Pragmatics of Conditionals: from Sentence to Discourse



PI: Prof. Dr. Maribel Romero, Department of Linguistics, University of Konstanz

Postdoc: Dr. Brian Leahy, Department of Linguistics, University of Konstanz


P2 investigates the division of labor between semantics and pragmatics in the interpretation of conditionals. We explore this question in three domains: at the sentence level, at the discourse level, and at the interface with epistemology.

At the sentence level, continuing the research initiated in the first funding period, we will examine the morphological composition of the full paradigm of conditional constructions, moving from the two-way distinction studied in the first phase (indicative vs. single layer subjunctive) to the complete three-way paradigm (indicative vs. single layer subjunctive vs. double layer subjunctive). We will investigate how the observed surface differences lead to the observed differences in truth and/or felicity conditions and why some kinds of conditional cancellably / uncancellably generate the information that their antecedents are false or unlikely.

At the discourse level we engage the debate over strict and variably strict analyses of conditionals. There are two arguments in favor of a strict analysis: (i) Sobel sequences and (ii) Negative Polarity Item (NPI) licensing. The strict conditional explanation of Sobel sequences has recently been challenged by a large body of data. We will pursue an explanation for conditional sequences in terms of implicatures, rhetorical relations and information structure which has the potential to explain all of the available data. von Fintel’s NPI argument for a strict analysis of conditionals relies on a widely successful generalization that explains NPI-licensing in other domains. However, the key ingredient that is added in extending the account to conditionals threatens to undermine the success of the generalization in some domains: non-adnominal only and focus-sensitive regret and surprise. We investigate how the focus-sensitivity of these constructions can be modelled while at the same time licensing NPIs as proposed in von Fintel’s analysis.

At the interface with epistemology, we study the interaction between conditionals, epistemic modals, and evidentials to develop tools that will help bridge a gap at the border between semantics and epistemology. Epistemologists are concerned with modelling how individuals should change their beliefs in the light of new information. Conditionals play a special role in such theories. However, existing theories of the dynamics of belief by epistemologists have taken conditionals to lack truth values. This raises the embedding problem: if conditionals lack truth conditions, how do we explain the interpretability of complex sentences with conditional components? We expect fruit from studying conditionals together with evidentials and epistemic modals because the members of all three classes of lexical items (i) contribute information about the likelihood of the propositions they are attached to and (ii) are syntactically embeddable, though in an interestingly restricted way.