Wednesday, 7 December 2016

Is Constructive Empiricism's Use of Counterfactuals Illicit?

Bas van Fraassen

Possible Worlds

At first glance it may seem odd that James Ladyman (2000) argues that constructive empiricist talk of observability may well require a commitment to possible worlds. Thus it would also seem - again on the surface - that any kind of empiricist must shun possible worlds. 

Possible worlds - to state the obvious – are neither empirical entities and nor are they observable (not even in principle). Thus, as a consequence, constructive empiricists would require unobservable entities in order to legitimise or justify their talk of observable entities.

Let's put the observability-in-principle position thus:

x is observable iff observers were in a suitable place, then they could observe x.

Now that's a bone fide counterfactual. Thus an old-style empiricist may now ask:

What are the truth-conditions of the statement “x is observable” (i.e., the statement above)?

Perhaps the constructive empiricist would say:

x is observable-in-principle because it has truth-conditions(-in-principle).

Why should the old-style empiricist accept either observables-in-principle or indeed truth-conditions-in principle? The whole point of observables is that they can be seen, smelt, heard, touched, etc. at the present moment in time (or at least they can... in principle!).

The constructive empiricist, thus, may require his truth-conditions to exist at possible worlds. Nonetheless, the truth-conditions can't exist at this moment in time because possible worlds don't exist at this or at any moment in time – at least not according to the empiricist.


I'm not sure if I understand Bas van Fraassen's reply to such points. It involves a strong use of the notion of “context-dependence”. In basic terms, when we “fix the context” of a counterfactual claim about observability, that somehow stops the claim from being modal in nature.

What is fixed is the epistemic community – all the “suitably constituted observers” who're relevant to the counterfactual claim of observability. This, itself, is supposed to make the counterfactual conditional non-modal in nature. That is, once we explicate the nature of (ideal) observers in (ideal) situations, it is these things which we can empirically investigate (Monton and van Fraassen 2003, 413-414). Nonetheless, we still have the non-Humean move from what's empirically observable at this present moment in time and in this place at this time, to that which is not empirically observable at this moment in time or in any place at this moment in time.

Thus statements about what's true at this moment in time - and at this place at this time - slide into what would be the case at other times and at other places. There's still a non-Humean jump that's not eased with these technical additions.

Van Fraassen adds extra detail to this.

He argues that the principle of observability can be cashed out in terms of the objective properties of the world. Moreover, we can use our best scientific theories in order to determine the truth (or content) of “x is observable” (Monton and van Fraassen 2003, 415-416). But, again, there are hidden modal assumptions in all this technical detail and even, as James Ladyman argues, hidden commitments to possible worlds.

Take this statement:

If Bertrand Russell's teapot in Andromeda showed itself to observers (suitably prepared, etc.), then they could observe it.

It seems like a sleight of hand to say that to understand the statement above is - even though a counterfactual - entailed by the facts or the phenomena of the observable (empirical) world. Presumably that must be a reference to teapots and observers which and who exist at this moment in time and can be observed at this moment in time. However, wouldn't Hume have argued that a bone fide empiricist couldn't jump from teapots in our solar system to teapots in Andromeda without begging a few questions or assuming a few facts?

I mentioned facts in the last sentence. Ladyman (2004, 762) talks, instead, about “laws”. That is, he says that

unless we take it that the specification by science of some regularities among the actual facts as laws … is latching onto objective features of the world”.

Wouldn't that mean, in our case, that the laws and objective features of Andromeda are assumed to be like the laws and objective features of the earth and our solar system? Yes, the laws of Andromeda are the same as the laws of the earth and our solar system. And, I assume, Russell's teapot would behave in a pretty similar way to how it would behave if it were floating near the moon or even in the sky above us.

Nonetheless, Ladyman does go on to say that only objectively-existing laws (rather than “pragmatically selected empirical regularities”) can justify (or warrant) our claims about the nature of Russell's teapot or any other phenomena of Andromeda. So it's not the constructive empiricist's claims about the nature and behaviour of Russell's teapot in Andromeda that are problematic per se. It's that in order to justify (or warrant) those claims the constructive empiricist would need to commit himself to entities which aren't kosher from an empiricist's point of view: viz., objective laws. Again, the only thing that a constructive empiricist can rely on are pragmatically-selected empirical regularities, not objective laws. Indeed the acceptance of objective laws commit one to a metaphysics that's not empiricist in nature.


Monton, B., and van Fraassen, B., (2003) “Constructive Empiricism and Modal Nominalism”, British Journal for the Philosophy of Science, 54: 405–422.
Ladyman, James. (2000) “What's Really Wrong With Constructive Empiricism? Van Fraassen and the Metaphysics of Modality”, British Journal for the Philosophy of Science, 51: 837–856.
- (2004) “Constructive Empiricism and Modal Metaphysics: A Reply to Monton and van Fraassen”, British Journal for the Philosophy of Science, 55: 755–765.

No comments:

Post a Comment