When people want to identify the causes of an event, assign credit or blame, or learn from their mistakes, they often reflect on how things could have gone differently. In this kind of reasoning, one considers a counterfactual world in which some events are different from their real-world counterparts and considers what else would have changed. Researchers have recently proposed several probabilistic models that aim to capture how people do (or should) reason about counterfactuals. We present a new model and (...) show that it accounts better for human inferences than several alternative models. Our model builds on the work of Pearl (2000), and extends his approach in a way that accommodates backtracking inferences and that acknowledges the difference between counterfactual interventions and counterfactual observations. We present six new experiments and analyze data from four experiments carried out by Rips (2010), and the results suggest that the new model provides an accurate account of both mean human judgments and the judgments of individuals. (shrink)

Recent advances in causal reasoning have given rise to a computational model that emulates the process by which humans generate, evaluate, and distinguish counterfactual sentences. Contrasted with the “possible worlds” account of counterfactuals, this “structural” model enjoys the advantages of representational economy, algorithmic simplicity, and conceptual clarity. This introduction traces the emergence of the structural model and gives a panoramic view of several applications where counterfactual reasoning has benefited problem areas in the empirical sciences.

The classic formulation of causal decision theory (CDT) appeals to counterfactuals. It says that you should aim to choose an option that would have a good outcome, were you to choose it. However, this version of CDT faces trouble if the laws of nature are deterministic. After all, the standard theory of counterfactuals says that, if the laws are deterministic, then if anything—including the choice you make—were different in the present, either the laws would be violated or the distant past (...) would be changed. And as several authors have shown, it's easy to transform this upshot of the standard theory of counterfactuals into full-blown counterexamples to CDT. In response to these counterexamples, I argue here that the problem lies, not so much with CDT's guiding idea—that it's the expected causal consequences of your actions that matter for rational decision-making—but with the fact that the classic formulation of CDT doesn't pay sufficient attention to the context-sensitivity of counterfactuals. I develop a contextualist version of CDT which better accounts for this context-sensitivity. And I show that my theory avoids the problems faced by the classic formulation of CDT in determinstic worlds. (shrink)

Traditionally, epistemologists have distinguished between epistemic and pragmatic goals. In so doing, they presume that much of game theory is irrelevant to epistemic enterprises. I will show that this is a mistake. Even if we restrict attention to purely epistemic motivations, members of epistemic groups will face a multitude of strategic choices. I illustrate several contexts where individuals who are concerned solely with the discovery of truth will nonetheless face difficult game theoretic problems. Examples of purely epistemic coordination problems and (...) social dilemmas will be presented. These show that there is a far deeper connection between economics and epistemology than previous appreciated. (shrink)

In this paper, we compare and contrast two methods for the revision of qualitative beliefs. The first method is generated by a simplistic diachronic Lockean thesis requiring coherence with the agent’s posterior credences after conditionalization. The second method is the orthodox AGM approach to belief revision. Our primary aim is to determine when the two methods may disagree in their recommendations and when they must agree. We establish a number of novel results about their relative behavior. Our most notable finding (...) is that the inverse of the golden ratio emerges as a non-arbitrary bound on the Bayesian method’s free-parameter—the Lockean threshold. This “golden threshold” surfaces in two of our results and turns out to be crucial for understanding the relation between the two methods. (shrink)

It is well known that aggregating the degree-of-belief functions of different subjects by linear pooling or averaging is subject to a commutativity dilemma: other than in trivial cases, conditionalizing the individual degree-of-belief functions on a piece of evidence E followed by linearly aggregating them does not yield the same result as rst aggregating them linearly and then conditionalizing the resulting social degree- of-belief function on E. In the present paper we suggest a novel way out of this dilemma: adapting the (...) method of update or learning such that linear pooling com- mutes with it. As it turns out, the resulting update scheme – imaging on the evidence – is well-known from areas such as the study of conditionals and cau- sal decision theory, and a formal result from which the required commutativity property is derivable was supplied already by Gärdenfors in a different con- text. We end up determining under which conditions imaging would seem to be right method of update, and under which conditions, therefore, group update would not be affected by the commutativity dilemma. (shrink)

In ‘The logic of partial supposition’, Eva and Hartmann present a dilemma for a normative account of partial subjunctive supposition: the natural subjunctive analogue of Jeffrey conditionalization is Jeffrey imaging, but this rule violates a natural monotonicity constraint. This paper offers a partial defence of Jeffrey imaging against Eva and Hartmann’s objection. I show that, although Jeffrey imaging is non-monotonic in Eva and Hartmann’s sense, it is what I call status quo monotonic. A status quo monotonic credal revision rule is (...) monotonic in Eva and Hartmann’s sense if it is conservative in the sense of Meehan and Zhang (‘Jeffrey meets Kolmogorov’), but Jeffrey imaging is in general non-conservative. On the other hand, Jeffrey imaging satisfies a different constraint that I call convexity, and the only rule that is both convex and conservative is Jeffrey conditionalization. To this extent, the real dilemma for a normative account of partial subjunctive supposition is not between monotonicity (broadly construed) and Jeffrey imaging, but between convexity and conservativeness. (shrink)

In ‘The logic of partial supposition’ (Analysis vol. 81), Benjamin Eva and Stephan Hartmann investigate partial imaging, a credence-revision method which combines the partiality familiar from Jeffrey Conditioning(The Logic of Decision, 1983) with the formal notion of imaging familiar from Lewis’s ‘Causal decision theory’ (1981). They argue that because partial imaging is non-monotonic, it ‘fail[s] to provide a plausible account of the norms of partial subjunctive suppositions’.In this note, I present a notion of partial imaging that does satisfy monotonicity, and (...) discuss some of the applications and ramifications. The account frames conditioning as a form of imaging, and rejects Gärdenfors’s principle of linearity in ‘Imaging and conditionalization’ (1982). (shrink)

The present paper deals with the tools that can be used to represent causation and to reason about it and, specifically, with their diversity. It focuses on so-called “causal probabilities”—that is, probabilities of effects given one of their causes—and critically surveys a recent paper in which Joyce argues that the values of these probabilities do not depend on one’s conception of causation. I first establish a stronger independence claim: I show that the very definition of causal probabilities is independent of (...) one’s conception of causation. Second, I investigate whether causal probabilities indeed take the same values under their different possible definitions. (shrink)

According to orthodoxy, there are two basic moods of supposition: indicative and subjunctive. The most popular formalizations of the corresponding norms of suppositional judgement are given by Bayesian conditionalization and Lewisian imaging, respectively. It is well known that Bayesian conditionalization can be generalized (via Jeffrey conditionalization) to provide a model for the norms of partial indicative supposition. This raises the question of whether imaging can likewise be generalized to model the norms of ‘partial subjunctive supposition’. The present article casts doubt (...) on whether the most natural generalizations of imaging are able to provide a plausible account of the norms of partial subjunctive supposition. (shrink)

This paper formulates a new argument against the received view in the philosophy of poetry. The received view consists of three tenets: the unity of poetic form and poetic content; the impossibility of paraphrasing and translating poetry; and the hyperintensionality of poetry. We will explore the same detour via direct quotation that has been used by proponents of the received view. We will argue that the hyperintensionality and unity of quotation do not guarantee its untranslatability, and thus that the inference (...) of untranslatability from hyperintensionality and unity is not universally valid. As a consequence, the claim that poetry cannot be translated is left without much support, since it relies on the same kind of inference. However, this does not mean that poetry is translatable; it means only that if it is not translatable, this is due to something other than the combination of hyperintensionality and unity. (shrink)

Richard Bradley’s landmark book Decision Theory with a Human Face makes seminal contributions to nearly every major area of decision theory, as well as most areas of formal epistemology and many areas of semantics. In addition to sketching Bradley’s distinctive semantics for conditional beliefs and desires, I will explain his theory of conditional desire, focusing particularly on his claim that we should not desire events, either positively or negatively, under the supposition that they will occur. I shall argue, to the (...) contrary, that permitting non-trivial desirabilities for events whose occurrence is known or assumed is both more intuitively plausible and more theoretically fruitful than Bradley’s approach. In the course of the discussion I will contrast Bradley’s broadly evidentialist picture of decision theory with my own more orthodox causal approach. (shrink)

It is well known that aggregating the degree-of-belief functions of different subjects by linear pooling or averaging is subject to a commutativity dilemma: other than in trivial cases, conditionalizing the individual degree-of-belief functions on a piece of evidence E followed by linearly aggregating them does not yield the same result as rst aggregating them linearly and then conditionalizing the resulting social degree- of-belief function on E. In the present paper we suggest a novel way out of this dilemma: adapting the (...) method of update or learning such that linear pooling com- mutes with it. As it turns out, the resulting update scheme – imaging on the evidence – is well-known from areas such as the study of conditionals and cau- sal decision theory, and a formal result from which the required commutativity property is derivable was supplied already by Gärdenfors in a different con- text. We end up determining under which conditions imaging would seem to be right method of update, and under which conditions, therefore, group update would not be affected by the commutativity dilemma. (shrink)

The technique of imaging was first introduced by Lewis, in order to provide a novel account of the probability of conditional propositions. In the intervening years, imaging has been the object of significant interest in both AI and philosophy, and has come to be seen as a philosophically important approach to probabilistic updating and belief revision. In this paper, we consider the possibility of generalising imaging to deal with uncertain evidence and partial belief revision. In particular, we introduce a new (...) logical criterion that any update rule should satisfy, and use it to evaluate a range of different approaches to generalising imaging to situations involving uncertain evidence. We show that none of the currently prevalent approaches to imaging allow for such a generalisation, although a lesser known version of imaging, introduced by Joyce, can be generalised in a way that mitigates these problems. (shrink)