Sunday, May 18, 2008

'The Theologian's Fallacy'

Alan Rhoda has a post at "Alanyzer" called The Theologian's Fallacy arguing that it is incoherent to hold that "the authority of the Bible always trumps human reason when the two come into conflict". The argument is that if the person who holds this also holds that it is possible that one could be mistaken about the status of Scripture (or whatever else) "as a source of absolute and infallible Truth" there is some sort of contradiction:

At this point the Trumper is caught in a bind. On the one hand, his own aspirations to intellectual humility encourage him to answer with a 'yes'. Human fallibility is too familiar for us to dismiss such possibilities out of hand. On the other hand, the Trumper's preparedness to use his Trump as a Trump, if necessary against even the best that human reason and inquiry can muster, requires him to say 'no'. After all, once a person genuinely admits that it is possible for him to be mistaken about a Trump, it ceases to function for him as a Trump. But how could one honestly answer 'no' to this question without committing the sin of intellectual pride? There is a way - the Trump must be self-authenticating in the highest possible degree, such that its status as Truth is at least as obvious (to those who are sufficiently prepared) than the most secure deliverances of human reason and inquiry.


The key movement here, "once a person genuinely admits that it is possible for him to be mistaken about a Trump, it ceases to function for him as a Trump," is very obscure, but there is clarification in the comments:

Sure, the epistemic possibility that one is wrong about a Trump does not imply that he is in fact wrong about the Trump. But to acknowledge that possibility as a genuine epistemic possibility and to continue to use that source of authority as a Trump is performatively inconsistent. To use something as a Trump one has to act as if the epistemic probability that the Trump is false is zero, whereas to acknowledge the possibility that the authority is mistaken is to take the epistemic probability that the Trump is false to be non-zero.


The argument sounds superficially good -- after all, who wants to go around supporting the "Trumping" of reason and inquiry? But the more I think about it, the more I think the argument is obviously wrong, and will not stand close examination.

What Rhoda calls "Trumping" is in fact simply a tendentious way of saying "correcting one's own reasoning on the basis of authority"; and the Trumper Rhoda particularly has in mind is someone who says that on matters where Scripture speaks plainly and "indubitably opposes our understanding" we should, in fact, correct our own reasoning on the basis of that authority. Much of Rhoda's argument appears to be an equivocation between "Trumping" in this sense and "Trumping" in some other sense that's never quite defined. Because if we consider "Trumping" in this non-tendentious sense, we find that the argument rejects far, far too much.

Suppose that I am reasoning about quantum physics. The argument looks flawless to me. And someone I recognize as an authority on quantum physics hears me out and tells me that my argument, however clever, is wrong, and simply overlooks some key features of quantum physics, or confuses some key features with other things entirely, or what have you. We would normally say that it would be irrational for me not to correct my reasoning light of that authority, unless we had clear, positive reason for doing so -- i.e., clear, positive reason for thinking that either our authority has misunderstood our argument, or has put forward a view that we know to be rejected by many authorities on quantum physics, or some such. Now along comes someone who has read Rhoda's post, we'll call him Rhoda2, and he puts the following questions to me:

R2: (1) Do you value truth?

A: Of course.

R2: (2) Do you value your authority as an authority?

A: Of course.

R2: (3) Is it possible that you could be mistaken about the value of your authority as an authority?

A: Of course.

R2: Then you are caught in a bind. Your preparedness to use your authority as an authority to correct even your best reasoning on the subject requires you to say 'no'. Once you admit that it is possible for you to be mistaken about an authority that suffices to correct your best reasoning on a subject, it ceases to function as such an authority.

A: The fact that I may be mistaken about such an authority does not imply that the authority is, in fact, not an authority.

R2: Of course, not. But to acknowledge that possibility as a genuine epistemic possibility and to continue to use that authority as an authority sufficient to correct your best reasoning on a subject is performatively inconsistent. To use something as an authority sufficient to correct your best reasoning, you have to act as if the epistemic probability that the authority is false is zero, whereas to acknowledge the possibility that the authority is mistaken is to take the epistemic probability that the authority is false to be non-zero.


It is clear at this point that Rhoda2's argument has gone completely off the rails. For one thing, there is no performative inconsistency with acting as if something is simply certain even when it is known that there is some probability that it is false; as even Locke knew, it is unreasonable not to act as if the sun will rise tomorrow even though you know that there are genuinely possible scenarios in which it won't. The only question here is whether it is reasonable to reason and act on the basis of a presumption known not to be perfectly certain; and obviously it is. But even setting that aside, it's clear that Rhoda2's argument would require that we never correct our reasoning on the basis of authority; and we have a name for people who do this: crackpots. They are the sorts of people who will not correct their best reasoning on, say, quantum physics, or biology, or whatever else, no matter how many authorities, no matter how eminent, no matter how qualified, tell them they have the wrong answer. And an argument that requires us all to be crackpots is an argument obviously to be rejected.

I think Rhoda has the idea that somehow it is the strength of the authority that is key here (it is emphasized a number of times in the argument). But the strength of the authority is simply its quality; and there is certainly no reason to think that as our assessment of the quality of an authority goes up our recognition of it as an authority should go down. On the contrary, our recognition of it as an authority should go up as our assessment of its quality goes up; and this is a correlation unaffected by the probability that we could be wrong in our assessment unless that probability shows that we are probably wrong in our assessment (in which case we would presumably change our assessment). Merely being possibly wrong is not enough: no one in their right mind will reject a conclusion, or refuse to act on a conclusion, merely because it is possible that they are wrong. Even Descartes refused to countenance that possibility.