[Japanese | English]
It has often been suggested that the subjective probability of an indicative conditional 'If A then C' is the conditional probability of C, given A. In earlier work (Kaufmann, in press), I have extended this idea to counterfactual conditionals and proposed a simple account of the intricate semantic relationship between the two classes of conditionals: An indicative and its counterfactual counterpart are equivalent, without therefore being ``equiprobable.'' Central to the treatment of counterfactuals is a probabilistic interpretation that is sensitive to causal independencies.
In this talk, I will start by showing that this proposal, making plausible predictions about counterfactuals, is seemingly at odds with the central premise of the probabilistic account of indicatives: The probabilities it assigns to them are not equivalent to the corresponding conditional probabilities. However, I will argue that while this may appear to be a problem at first, it actually leads to a deeper insight about the inference involved in their interpretation. There are indeed conditionals whose probabilities are intuitively not the corresponding conditional probabilities; such judgments are predicted by the account. What is more, these conditionals generally have an alternative, less prominent reading under which their probability is indeed the conditional probability. Thus these facts point to a hitherto rarely discussed systematic variability in the interpretation of indicative conditionals. This phenomenon was discussed in Kaufmann (2004), where it is also argued that the two readings are related to each other in a straightforward way via an abductive inference step.
If this correct, then the apparently contradictory consequences of Kaufmann's (in press) proposal give way to a significant step towards a unified account of indicative and counterfactual conditionals. In this talk, I will report preliminary experimental results supporting this conclusion.
References:
Kaufmann, S. 2004. Conditioning against the grain: Abduction and indicative conditionals. Journal of Philosophical Logic 33(6):583-606.
Kaufmann, S. In press. Conditional predictions: A probabilistic account. Linguistics and Philosophy.
<Transportation>
Last modified: 2005-01-19 16:51:29 JST