Contextualism for Conditionals
Consider the following sentences:
(1) If Sarah has the measles, she will have a fever.
(2) If Sarah has a fever, then she has the measles.
(3) If Julia asks nicely, Bill still won-t help us.
Intuitively, (1) would be used to communicate that the present case is of the kind in which having fever follows causally from having measles, (2) would be used to communicate that the case is one in which having the measles follows evidentially from having a fever, and (3) would probably be used to communicate that the case in question isn-t the kind in which asking nicely elicits helpful action. My question is through what processed these intuitive messages concerning causal and evidential relations are encoded and decoded given the conventional contribution of the -if P, Q- form.
Standard philosophical theories of indicative conditionals - the material implication theory, the conditional probability theory, and versions of the possible world theory - take these intuitive messages to be pragmatically inferred from the utterance of a sentence the conventional meaning of which gives it more abstract truth- or acceptability conditions. Elsewhere, I have defended an alternative model: relational contextualism. It agrees that the intuitive messages are determined pragmatically, but denies that we identify the intuitive messages by first identifying more abstract truth- or acceptability conditions. Instead, the conventional contribution of if-clauses to the meaning of conditionals is that if-clauses introduce a proposition without asserting it so that the main clause can be understood in relation to it. When we decode the conditional form, our primary task is to identify the relevant relation between the content of the if-clause and the main clause. In the case of conditionals like (1), (2) and (3), that relation will typically be one of causal or evidential consequence or independence.
There are several reasons to think that something like relational contextualism is correct. It gives a unified account of a wide variety of conditionals: apart from consequence conditionals like (1) and (2) and independence conditionals like (3), it covers relevance conditionals (-If anyone cares to ask, I do have views on celebrity couples-) as well as conditionals expressing conditional commands, questions and bets (-If it is snowing hard, stay with your car-). It explains both why conditionals embed systematically and why some embedding constructions nevertheless seem unintelligible. And it promises a unified account of indicative and subjunctive conditionals, as well as a straightforward explanation of the well-known context-relativity of the latter.
My concern in this talk is to clarify how relational contextualism explains a phenomenon that appears to undermine standard theories of conditionals. Standard theories all take it that when the consequent is highly probable independently of the antecedent, we have good reasons to accept what the conditional says. But not all such conditionals seem acceptable. I am reasonably confident that both clauses of the following conditional are true, but I find the conditionals nonsensical rather than acceptable:
(4) If Berne is the capital of Switzerland, John Lennon was killed in 1980.
I am also reasonable confident that it will rain tomorrow, independently of what I do, but my first intuitive verdict about the following is false:
(5) If I go to the movies tonight, it will rain tomorrow.
Verdicts like these are very common, especially among people without background in logic and the theory of conditionals, and relational contextualism explains them easily: (4) comes off as unintelligible because no relation between antecedent and consequent has been identified; (5) comes off as false because it has been taking to convey a relation of causal consequence between