This is an investigation into the relation between the two notions. Similarity is the crucial notion of Lewis’s theory of counterfactuals. Cotenability is the crucial notion of the metalinguistic theory (later called the premise semantics or support or derivability theory). Lewis defined cotenability in terms of similarity and explained what he thought to be a connection between his and the metalinguistic theory of counterfactuals.
      This paper has two main parts. The first reminds the reader of the two theories, explains the connection between them as described by Lewis, gives a counterexample to Lewis’s notion of cotenability and shows that the theories are not related in the way Lewis thought they were. All that will not tell us much about the metalinguistic theory, but will point to what I find serious problems for Lewis’s semantics. In the second part I will develop the results from the first part by proposing a counterexample and a pattern for making similar counterexamples to Lewis’s semantics, this time with an intention to make more general conclusions, pertaining not only to Lewis’s theory, but to any possible worlds theory of that kind: I will argue against any total ordering minimal or small change semantics for counterfactuals based on an absolute similarity relation or selection function.

      Subjunctive conditionals entered the decision theory via Causal decision theory (CDT), which has made a name for itself during the last decades, winning the battle against evidential decision theory (EDT). The main arguments in favor of CDT and for using subjunctive conditionals are based on the counterexamples to EDT such as Newcomb's bomb and Common cause cases.
      Recently, Andy Egan reopened the discussion with some interesting counterexamples to CDT. His article made CDT look significantly less attractive, and one could conclude from his arguments that neither EDT nor CDT offered satisfactory solutions for the decision theory (or that EDT and CDT are equally problematic theories). Dorothy Edgington argued that his conclusions could not be applied to CDT in general, but rather to a limited range of versions of CDT; and she made a proposal for the new version of CDT, one which is immune to Egan's counterexamples.
      I present an argument against Edgington's proposal. But, I believe that her conclusion about Egan's counterexamples is nevertheless true. I argue that Egan's counterexamples, together with all those earlier counterexamples to EDT, still lead us to causal decision theory, but the subjunctive conditionals should not be the basis of such a theory. Instead, (counterexamples show us that) it is Skyrms's causal decision theory we should adopt.

A counterlegal is a counterfactual conditional containing an antecedent that is inconsistent with some set of laws. A backtracker is a counterfactual that tells us how things would be at a time earlier than that of its antecedent, were the antecedent to obtain. Typically, theories that handle counterlegals appropriately handle backtrackers poorly, and vice versa. Two cases in point: Lewis’ (1979) ordering semantics handles counterlegals well but not backtrackers. Eric Hiddleston’s (2005) causal-model semantics handles backtrackers nicely but not counterlegals. Taking Hiddleston’s account as a starting point, I offer steps toward a theory capable of handling both counterlegals and backtrackers. The core contribution of this paper is a means for evaluating counterlegals relative to minimally-illegal models.

Timothy Williamson has argued against the principle of conditional excluded middle (CEM) in counterfactual logic. More recently, he has argued that it is necessary what propositions there are. A formal argument will be presented to show that he can't be wrong on both counts: CEM rules out contingency in what propositions there are. The argument uses an abstract possible world semantics which does not presuppose any particular way of modeling counterfactuals. Crucial assumptions in the argument are a restricted comprehension principle for propositions, which Williamson has argued should be accepted even by those who think that it is contingent what propositions there are, and an infinitary agglomeration principle for counterfactuals which is widely endorsed, even though it is not valid according to David Lewis's semantics for counterfactuals. The argument also essentially relies on a modal formulation of CEM which applies to merely possible propositions. It will be argued that endorsing CEM only for propositions there are is not plausible, as the best arguments for CEM extend to its modal formulation.

In The Philosophy of Philosophy Timothy Williamson explores the philosophical significance and modal logic of counterfactual necessity, where a proposition is counterfactually necessary just in case a contradiction would have been true had that proposition been false. In this talk I explore the view that claims of numerical distinctness are not counterfactually necessary. I consider some motivations for this view and then explore some of its implications for the logic of counterfactual necessity. I also respond to an argument of Williamson's for the necessity of distinctness which appeals to certain principles about the logic of actuality. I conclude by considering two implications of the view: one concerning the relation between counterfactual necessity and so called "metaphysical" necessity, and another concerning the role of counterfactual reasoning in philosophical theorizing.

Utterances of counterfactual conditionals, unlike utterances of indicative conditionals, tend to impart the information that their antecedents are false. In a body of recent work I have been arguing that this information arises as a special sort of implicature. The current paper will first review that position, and then describe an objection to the view raised by Moritz Schultz. In developing a response to Schulz' objection we find a novel argument for a unified semantics of indicative and counterfactual conditionals and an argument that the interpretation of counterfactuals are sensitive to the epistemic states of their utterers.

Sobel Sequences are consistent sequences of counterfactuals like the following:

    (1a) If Sophie had gone to the parade, she would have seen Pedro dance.

    (1b) But of course, if Sophie had gone to the parade and been stuck behind someone tall, she wouldn’t have seen Pedro dance.

But reverse the sequence, and it does not sound so good at all. This observation — that order makes a difference to the consistency of the sequence — motivated Kai von Fintel and Thony Gillies to abandon the classic Lewis-Stalnaker semantics and adopt a dynamic semantic account of counterfactual conditionals. Subsequently, Sarah Moss defended the classic Lewis-Stalnaker semantics against the charge that it need be abandoned because of Reverse Sobel Sequences, arguing for a pragmatic account of the infelicity of the sequences. I argue that ultimately both the dynamic semantic account and Moss’s account are untenable, but that seeing what is good about each account points the way to the right positive story. Finally, I defend a positive view that attributes the effect of counterfactuals on the context to pragmatics, but treats the effect of the context on counterfactuals as semantic.

Counterfactual reasoning plays an important role in human cognition; it is frequently used in everyday discourse and enables us to evaluate how past events could have turned out differently. However, research reveals a general lack of knowledge about how children come to understand counterfactual conditionals in a way adults do. In this talk I will present developmental data showing that children understand and process counterfactual questions differently from adults. I will also discuss what children may need to understand to be able to answer counterfactual questions in an adult-like manner. 

Alan Hájek has been arguing that the counterfactuals we take for granted in practical deliberation, ordinary discourse and philosophical analysis are actually false. The aim of this talk is to respond to this skeptical challenge by suggesting a strong reading of a recent multidimensional possible world semantics for conditionals. This semantics has it that in additional to ordinary facts, there are counter-facts, that are truth-makers for counterfactual claims. I will suggest that these counterfacts are part of the fundamental structure of reality but are not in all cases entailed by the ordinary facts. The upshot is a theory of counterfactuals according to which most counterfactuals that we ordinarily accept do indeed come out true. But the cost is that we have to abandon the popular thesis of Humean supervenience.

I toss an indeterministic coin, and whilst the coin is in mid-air, I offer you good odds that it will come up heads. You decline the bet and the coin lands heads. The following then seems true

1.      If you had bet heads, you would have won

Although the truth of (1) has been embraced by many, its acceptance has not been universal. In this paper argue that conditionals like (1) are indeed true, even in indeterministic contexts. First, I provide a positive argument for the claim, which builds on the work of Edgington. Second, I show that attempts to explain away our judgement that (1) is true by appeal to error or pragmatic considerations are unconvincing. Third, I provide a direct response to arguments that such conditionals are in fact false. Finally, I reject recent suggestions that semifactuals are distinctive type of conditional and are not to be assimilated to a class comprised of genuine counterfactuals and subjunctives with true antecedents. The truth of (1) is, then, a non-negotiable aspect of our thought about subjunctives.

Angelika Kratzer’s benchmark modal semantics builds on the following biconditional analysis: ‘It must be that p’ is true iff in all of a certain set of possible worlds, p is true. A dozen years ago, I drew attention to a fundamental Problem with Kratzer’s semantics: it makes any instance of the schema ‘if p, then it must be that p’ a logical truth. My assessment was: to avoid the Problem, the biconditional analysis has to go. While the Problem has since received extensive attention from linguists and philosophers, few of them were convinced by my assessment. In my presentation, I will explore some of the responses to the Problem, in particular, double-modalization-based strategies — including Kratzer’s recent proposal in her (2012) book Modals and Conditionals — arguing that their application to Problem cases is ad hoc and too narrow. These considerations bring to the fore new arguments for my initial assessment: the biconditional analysis, specifically, its right-left direction, is not worth its keep.