Linguistics in Göttingen - A platform for empirical and theoretical linguistics

Paul Marty (Leibniz-ZAS Berlin)

From Maximize to Exhaustify Presupposition!


On standard formulations of Maximize Presupposition! (MP), the presuppositional competitors to a sentence φ are taken to be those alternatives to φ that are (A) equally informative, (B) presuppositionally stronger, and (C) whose presuppositions are already entailed by the context. Scrutinizing this fairly standard picture, I gather in this talk existing and new data suggesting that the logic at stake in the classical MP effects extends beyond the range of cases covered by MP, and that each of (A), (B) and (C) is independently too strong to be empirically adequate. Reviewing and amending in turn each of these formal components, I will start by sketching out a new descriptive generalization, dubbed Exhaustify Presupposition! (EP), that can take on the challenges raised throughout the talk. 


Exhaustify Presupposition! (EP) — A sentence φ is infelicitous in a context set C if there is an alternative ψ ∈ Alt(φ) such that: 

  1. ψ's presuppositions together with φ entail ψ, and

  2. ψ's presuppositions are non-weaker than φ's, and

  3. C does not entail ψ but C together with φ does. 

The new picture that will emerge is one on which speakers are required to make the presuppositional content of their conversational contribution not just as strong as possible, but as exhaustive as possible, regardless of presupposition satisfaction. I will provide further conceptual and empirical motivations for the formulation of EP. First, EP will be shown to be the natural counterpart of a similar constraint on felicity that operates on assertive contents (Exhaustify Assertion!); second, the broader logic behind such constraints will be shown to be derivable at minimal costs from existing approaches to implicature-reasoning. In this regard, I will attempt to clarify (some of) the key assumptions on which the MP/EP logic can be captured both on a Neo-Gricean and on a grammatical approach to scalar implicatures. If time remains, I will discuss possible ways to compare and tease apart the predictions made by these two explanatory approaches to MP/EP focusing on the derivation of uncertainty/ignorance vs. scalar implicatures and the question of the absence/presence of a so-called 'epistemic step'.