This month’s letter — from Lewis to Tony Coady — is a bit of light relief from abstract philosophising. Lewis is worrying about testimony, and in particular worrying about whether or to what extent someone’s known past deception affects, or should affect, whether we believe what they tell us later. Lewis seems to think that it doesn’t but it should: ‘the testimony of a past perjurer is apt to be taken more seriously than it deserves to be’ because we ‘simply take the testimony we’re how hearing on trust’ — that is, without taking into account the speaker’s past record.

Of course, this is an issue that’s of immense interest just at the moment, what with the role in the Mueller investigation against Trump apparently revolving around the testimony of two people — Michael Cohen and Michael Flynn — whose past relationship with the truth is known to be somewhat flaky. Though the fact that Trump supporters have latched on to this — at least in the case of Cohen (‘it’s laughable that anyone would take a convicted liar like Cohen at his word’, Sarah Sanders said) — is of course beyond irony. Apparently being convicted of lying is totally different to just, you know, clearly and repeatedly demonstrably saying things that are false outside a court of law.

When it comes to the Milgram experiment, Lewis really does seem to think it’s just as likely that (a) Milgram was duped by the subjects of the experiment as it is that (b) they themselves were being duped — and it’s equally likely that (c) Milgram ‘deceived his readers’. This last possibility connects with what Lewis says at the start of the letter: Milgram is in effect claiming (in his published work) that he lied before (to the subjects), and so he’s either telling the truth now and lied then, or vice versa. (So the suggestion here seems to be that Milgram might have lied in print not just about the results of the experiment but — since the deception is a crucial part of the experiment — about the existence of the experiment itself!)

That all strikes me as really odd — but I guess it all depends on your priors. Lewis evidently finds it really hard to believe that the subjects of Milgram’s experiments genuinely thought they were being asked to administer potentially fatal electric shocks to people — since surely everyone knows psychologists aren’t permitted to do that kind of thing! Me, I don’t find that very hard to believe at all. I also find it pretty hard to believe that people who are not trained actors, and who have no prior warning, might be good enough actors to fake their responses. (Well maybe a few of them. But I’d say most people are pretty bad actors. So I guess maybe Lewis thinks acting is a lot less hard than I think it is.)

Lewis says he ‘wouldn’t be surprised if people’s ‘trust [in Milgram’s findings, despite the acknowledged lying to the subjects] is somehow reasonable, presumably because of the difference between truthfulness in one role and in another’. ‘But’, he adds, ‘I don’t really understand what’s going on’. Well I don’t either, but that idea strikes me as a much more plausible starting-point than Lewis’s alternative — that possibilities (a)-(c) all merit the same credence!