THE NEED FOR A METAPHYSICAL THEORY OF CAUSALITY

Thus, the idea that causal explanation is a non-circular deduction has won the day. It is common to William Whewell, Sir Karl Popper, and Carl G. Hempel, among other thinkers. But what constitutes non-cir- cularity or non-ad-hocness? For Whewell, non-circularity amounts to empirical verification, much as for Newton; for Popper non-circularity amounts to refutability; for Hempel it amounts to confirmation.

Assume, then, that these theories offer necessary and sufficient condi­tions for a theory to be an explanation. It may be impossible to offer a counter-example to Whewell’s theory if we insist that verification is im­possible. But if we use Whewell’s conditions for verification instead - i.e., that a theory stand up to an independent severe test - then, surely, there can be a counter-example to Whewell’s theory. Indeed, my example of the touchstone is a counter-example to all three theories, Whewell’s, Popper’s, and Hempel’s. The touchstone hypothesis should count as explanatory by all three, yet I dare say they will all agree that intuitively this is defective.

What the counter-example shows is that causality or causal explanation cannot be fully captured by sheer methodology. I have explained, in my paper in the first volume of the Philosophical Forum, how the method­ological idea of causal explanation broadens the classical concept (e.g. to include statistical theory); as we see now, it broadens matters too much.

So much for causal versus accidental theories. We also have, we remem­ber, philosophical versus mathematical, where philosophical corresponds to causal but where mathematical does not even correspond to accidental but more to the fictitious. All three models of explnation - Whewell’s, Popper’s, and Hempel’s - cover equally well both electromagnetics and the theory of elasticity. Indeed, formally the resemblance between the two theories is astounding even to this day. Somehow, we know, electromag­netic field theory was meant as an explanation proper of electromagnetic phenomena; and with good reasons, we feel. Somehow, we know, the theory of elasticity was meant not as an explanation but as a mathemati­cal tool for description of the phenomena of elasticity; and for good reason, we feel. Hence, all three theories of explanation are either inade­quate or false.

So much for my exegesis on Bromberger’s critique of Hempel. Why only of Hempel, and why he does not elaborate, I do not know. I have found the elaboration necessary. I do not yet know what Whewell, Popper, or Hempel, would say about these criticisms; I dare say there are a few ans­wers to them. I feel, somehow, these answers are not very good. The very idea of Whewell and his followers, including Wittgenstein of the Tractatus, is to reduce causality from metaphysics to methodology. Popper even expresses this idea - whenever possible, and especially in the case of causality, reduce metaphysics to methodology - openly in his classical Logic of Scientific Discovery. I find this idea interesting, important, and useful. But on causality, for example, it breaks down.

A few more words about the modern instrumentalism of Duhem and his followers. They deny that any theory is explanatory, and claim only deductive or descriptive force for all theories. This abolishes the distinc­tion between the theories of elasticity and electromagnetism. This, equally, makes nonsense of the dissatisfaction commonly expressed about the eightfold way theory which, admirable as it surely is, is regrettably more descriptive than explanatory. The dissatisfaction proves that intuitively even scientists who endorse instrumentalism mean to do so merely in order to reject essentialism. They simply respond too violently to a past mistake.

X.