Anscombe's commentary on her exchange with Lewis

In 1948, Elizabeth Anscombe, then a student of Wittgenstein and a research fellow at Oxford, publicly challenged C. S. Lewis’s central argument against naturalism. In response to her criticisms, Lewis rewrote the relevant chapter of his book Miracles. Anscombe briefly acknowledged the revision in print as an improvement, but never wrote more extensively about it. In 1985, however, she gave a talk about Lewis’s revised version to the C. S. Lewis Society, discussing its strengths and remaining weaknesses. This chapter is a transcript of that talk.

Here.  I wish this transcript had been available back when I was writing my dissertation chapter on the Lewis-Anscombe Controversy. Now if I can just get my hands on it!

This is Gregory Bassham's summary of Anscombe's discussion. from Church History. 

One reading that will be of special interest to Lewis scholars is Elizabeth Anscombe’s talk on “C. s. Lewis’s Rewrite of Chapter III of MiraclesAs is well-known, Lewis and Anscombe engaged in a famous debate in 1948 over Lewis’s claim that naturalism is self-refeting (his so-called "argument from reason”). Confroversy has swirled over who won the debate and whether Lewis largely abandoned rational apologetics as a result of his perceived defeat. What we know for sure is that Lewis substantially revised and expanded his original argument in the second edition of Miracles (London: Collins-Fontana, 1960), and that Anscombe stated in the early 1980s that Lewis’s revised argument was a substantial improvement over the original formulation, iat we have not known until now is whether Anscombe believed that Lewis’s revised argument was substantially correct. We can now see that she did not. Anscombe examines Lewis’s argument in detail, and finds it to be rife with confiisions, ambiguities, and false assertions (16-22). Lewis argues that naturalism undermines itself, because naturalism can only be justifiably believed if it can be rationally inferred from good evidence, and naturalism excludes the possibility of rational inference by claiming that all human mental processes are wholly determined by nonrational causes. In other words, naturalism can be rationally believed only if reasoning is possible. But if naturalism is true, then reasoning is impossible. So if naturalism is true, it cannot be rationally believed (and neither can much else). Anscombe suggests that this argument rests on a confiision. The fact that a beliefor statement is felly determined by non-rational causes has no bearing on whether it is true or rationally justified. Consider, she says, the analogy of a printed book. Every word in such a book is wholly determined by the machinery that printed the book. Yet no one would suppose that this casts any doubt on whether the things said in the book are true or rationally defensible (16). As Anscombe sees it, the whole issue of causal determination is a red herring؛ even strict, reductionistic naturalists can consistently recognize the existence of reasoning and the possibility of rational beliefs. But Anscombe does find something of real value in Lewis’s discussion. Lewis raises the important and deeply puzzling question of how logical grounds can cause a conclusion to be drawn (18). As Anscombe sees it, Lewis’s “damnably obscure” claim that an inferred conclusion can be "determined only by the truth it knows” (22) does not do much to solve this problem, but Lewis was right in pointing to the deeply puzzling nature of mental causation.

As I noted in the first essay I wrote on Lewis-Anscombe, Anscombe is committed to a divorce between rational justification and the causation of belief that strikes me as implausibly strong, and one that would not have be embraced by most naturalists today, who would follow Donald Davidson in rejecting a strict divorce between reasons and causes (see his essay "Actions, Reasons, and Causes" in which he criticizes the standard Wittgensteinian position on this). Anscombe also has to deal with the ontological restrictiveness of naturalism. Whether an explanation is causal or not, it still has an ontology, and most philosophical materialists and naturalists will not accept reasons into their ontology without some kind of intertheoretic reduction or supervenience relation, and this is not part of Anscombe's critique at all.