Round Table with Igor Douven: The Epistemology of Conditionals


University of Konstanz, Rooms D431/D434, 10:00-18:00





10:00-11:00Arthur Paul Pedersen (University of Regensburg)
11:00-12:00Eric Raidl (University of Konstanz)
What Are Epistemic-state-sensitive Conditionals?
12:00-12:15Coffee Break
12:15-13:15 Hans Rott (University of Regensburg)
On the Acceptability of Indicative Conditionals
13:15-14:30Lunch Break
14:30-15:30Niels Skovgaard-Olsen (University of Freiburg)
Relevance Differently Affects the Truth, Acceptability, and Probability Evaluations of 'and', 'but', 'therefore', and 'if then'
15:30-15:45Coffee Break
15:45-16:45Wolfgang Spohn (University of Konstanz)
From Expressivism to Dynamic Semantics
16:45-17:00 Coffee Break
17:00-18:00Round Table with Igor Douven
19:00 Dinner at Restaurant Wessenberg




Arthur Paul Pedersen: tba


Eric Raidl: What Are Epistemic-state-sensitive Conditionals?

Igor Douven, in chapter 3 of his book, points to an empiric-theoretical tension. On the one hand, Lewis’ triviality result(s) seem to suggest that Stalnaker’s thesis (ST) that the probability of a conditional is the conditional probability, cannot be maintained in pain of trivialising conditional probabilities and the language. On the other hand, psychological findings seem to suggest that real people do judge the probability of a (simple) conditional by the conditional probability. Douven proposes to reconcile the tension, by appealing to Van Fraassen’s idea that a conditional is sensitive to the epistemic state. With sensitive conditionals (even if closed under conditionalization) ST does not lead to triviality. I will provide a closer look at Van Fraassen’s construction, it’s philosophical assumptions, and ask whether it can be generalised. But most of all: what does it mean for a conditional to be epistemic-state-sensitive?

Hans Rott (joint work with John Cantwell and Paul Egre): On the Acceptability of Indicative Conditionals

According to Ernest Adams' (1965, 1975) influential account, indicative conditionals do not have truth conditions, but they have assertability conditions, which we better ought to think of as acceptability conditions. Adams uses a probabilistic semantics to account for these acceptability conditions. Recently, Igor Douven (2016, Ch. 4, based in part on Douven and Verbrugge 2010) has given a detailed discussion of Adams' thesis and of various ways of articulating it. Douven claims to have refuted the principal version of Adams' Thesis on empirical grounds. Our goal in this paper is to have a close look at those various formulations, and more generally to clarify what we take to be the relation between the acceptability and the probability of conditionals and non-conditionals. We present yet another version of Adams' thesis that we think stands a much better chance of being normatively and empirically correct than the ones considered by Douven. Our main contention is that acceptability is function of probability, but also of other factors, such as utility, and that the dependence between them should likely not be linear.

Niels Skovgaard-Olsen: Relevance differently affects the truth, acceptability, and probability evaluations of ‘and’, ‘but’, ‘therefore’, and ‘if then’

According to the Frege-Grice tradition, differences in the meaning of ‘and’, ‘but’, and ‘therefore’ do not affect their truth conditions. In Experiment 1 we tested this assumption, along with the assumption that the truth conditions of ‘if then’ are not affected by relevance manipulations, employing a new version of the truth table task. In a follow up study, we tested for relevance effects on these target sentences in evaluations of probability and acceptability, employing a novel experimental task in which the participants supply rank orders of sentences according to their acceptability. A dissociation was found between evaluations of truth and acceptability/-probability: The relevance manipulation did not affect the truth evaluation of ‘and’, ‘but’, and ‘therefore’ differentially, but it strongly affected evaluations of their probability and rank ordered acceptability. Moreover, it was found that neither the material implication, nor the influential de Finetti table, nor truth conditional inferentialism could account for the truth conditions of ‘if then’ across relevance levels.


Wolfgang Spohn: From Expressivism to Dynamic Semantics

According to expressivism the meaning of an utterance consists in the mental state or attitude expressed by the speaker. According to dynamic semantics, the meaning of an utterance consists in its contribution to the common ground. One aim of the talk is to explain the relation between the two conceptions of meaning. This builds on a somewhat different understanding of the common ground, which is traditionally understood as a set of facts that are common knowledge and which will now be more generally understood as a common mental state, i.e., a mental state commonly known to be had by all communicants. Another aim is to explicate the standard threefold distinction of presupposition, linguistic meaning, and implicature within the framework introduced. The hidden motive is, of course, to translate my expressivistic account of conditionals into the more common framework of dynamic semantics.