To: Neil M.
From: Geoffrey Klempner
Subject: Counterfactualism in history
Date: 14 April 2001 09:56

Dear Neil,

Thank you for your letter of 3 April, with your Associate Award essay, 'Counterfactualism in history'.

You have succeeded in convincing me that there is a problem for the philosopher of history in grasping the role of counterfactuals in historical explanation. I feel that there is more work to be done, however.

First, we need to clear out the way a confusion between counterfactuals and subjective conditionals. I accept the point that when in the process of deliberation we consider certain subjunctive conditionals, 'If I were to choose X then Y would happen' 'If I were to choose not-X then Z would happen', this ability to consider different possible alternatives or different ways things might happen is intimately connected with the ability to make counterfactual judgements. I choose X, and Y happens, confirming my judgement. In so doing, I might see this as corroborating, although not confirming, the judgement, made at the time, 'If I were to choose not-X then Y would happen'. What was a subjective conditional judgement is now a counterfactual judgement. A conditional is, or becomes a counterfactual only when the antecedent is *contrary to fact*!

However, there is an equally important distinction between counterfactual judgements of the form, 'If X had occurred then Y *would* have occurred', and would one might term 'counterfactual speculations' of the form, 'If X had occurred then Y *might* have occurred.' A book of counterfactual history, as you describe it, is largely (though not wholly) made up of interesting, entertaining and historically relevant counterfactual speculations.

A few years ago a book was published on the theory of evolution, which described the strange and fabulous creatures that might have evolved, had the worlds climate etc. been different in subtle ways. All illustrated in gorgeous colour. A book like this certainly has entertainment value. It serves as an interesting contrast to books about mythological beasts such as the dragon or the griffin, which bear little relation to the constraints of biological possibility. It has a heuristic value in teaching the basics of evolutionary theory. But it is not science, nor is it intended to be.

My impression is that at least some the complaints of professional historians against counterfactual history are based on the distinction between counterfactual judgements and counterfactual speculations. There is enough real history to do, why waste time with mere speculations and fancies?

What I find more gripping is the thought that some historians reject counterfactual judgements altogether. Is it really possible to do this? Your argument that the agents whose actions the historian describes engaged in subjective reasoning is not relevant. What we require is an argument to the effect that historical explanation necessarily involves counterfactual judgements. Is it possible to write a work of history that does not contain counterfactual judgements?

My main reason for scepticism comes from the point that you correctly raise, concerning the close connection between judgements of cause and effect and counterfactuals. There is no way to identify 'the' cause of an event without raising counterfactual questions about what would have happened had certain conditions been different. Any historian, in other words, who makes a claim that can be construed as a causal claim is implicitly making a counterfactual judgement.

It would seem unlikely, however, that the 'idealist' historians whom you cite would say, when this was pointed out to them, 'Oh, sorry, we didn't realize. Of course we make counterfactual judgements all the time!' So what are they saying. What is really at issue here? This is the central question of your essay. This is the thing that the reader will be looking for.

One more point. There is a lacuna in your essay regarding the analysis of counterfactuals. What are the truth conditions for a counterfactual? Or, how do we assess the truth value of a counterfactual? Theories have been proposed (e.g. by David Lewis and Robert Stalnaker) which you should at least be aware of. The question concerns the way we choose amongst different possible worlds in which the antecedent of the conditional is conceived as being true. Lewis and Stalnaker's basic idea is that possible worlds can be ranked in terms of relations of 'similarity' to the actual world. So in considering the truth of a counterfactual judgement we are concerned with that possible world, or set of possible worlds in which the antecedent of the conditional would have been true, which are most similar to the actual world.

To see that there are serious problems with this idea, consider the counterfactual, 'If Oswald had not assassinated Kennedy, then he would have lived out his term of office'. If we look at the possible worlds in which Oswald does not assassinate Kennedy which are most similar to the actual world, obviously these will be worlds in which Kennedy was assassinated! The result is that the conspiracy theory automatically comes out true. Obviously, the time factor is relevant here. The problem is finding a way to include it in a theory of counterfactuals.

Although this is not the main topic of your essay, you should show some awareness of this problem: "For the purposes of this essay, I take a counterfactual to be such-and such. There is a philosophical question concerning the truth conditions for counterfactuals. It is sufficient for the purpose of this essay that we judge the truth of counterfactual statements in the following way....".

All the best,

Geoffrey