I have been mostly defending the Trilemma argument against various objections for the past few weeks, so I have not spent much time thinking about how to refute it. I reject the conclusion, of course, on the basis of various other reasons unrelated to the Trilemma.

God, as understood in Western theism, is an all-knowing, all-powerful, perfectly good person, and Jesus was none of those things:

God is, by definition, a perfectly good person.
Jesus was not a perfectly good person.
Therefore,
Jesus was not God.
God is, by definition, all-knowing.
Jesus was not all-knowing.
Therefore,
Jesus was not God.
God is, by definition, all-powerful.
Jesus was not all-powerful.
Therefore,
Jesus was not God.
I’m not going to argue here for the factual premises, but I feel confident that those premises are true and that I could make a good case for each of those premises.

For these and other reasons, I do not accept the conclusion of the Trilemma. But the fact that there are strong counterarguments to the conclusion of the Trilemma does not demonstrate that the Trilemma is a bad argument, and it certainly does not show how or why the Trilemma argument goes wrong.

Here is my preliminary evaluation of the Trilemma:

The Trilemma is a valid deductive argument; the logic, as I have interpreted the argument, is good. But there are problems with at least some of the premises:

Premise (1) is probably false.
Premise (2) is true (or true for the most part).
Premise (3) is true (or true for the most part).
Premise (4) is probably true.
Premise (5) is probably true.

Given this assessment of the premises, the argument fails, because premise (1) is probably false (I am building a case against this premise on my own blog). However, whenever possible, I like to present believers with an objection in the form of a dilemma, in order to broaden the scope of the attack on an apologetic argument, and to make it more difficult for the believer to dismiss my objection. This sort of move can be made with the Trilemma.

Premise (1) is probably false, and I would argue strenuously for this objection. If my analysis of the evidence is correct, then the Trilemma is a bad argument. But what if, despite my best effort to accurately assess (1), I am mistaken, and McDowell is correct concerning this issue? Suppose that Jesus did claim to literally be the God of Western theism. If was persuaded that this was the case, then I would change my view on the probability of the truth of premises (4) and (5).

Probability assessments are made on the basis of background information. If my assumption about what Jesus did or did not claim about himself changes, then that would be a legitimate basis for changing my assessment of the probability of (4) and (5). If Jesus really did claim to be God, in the sense of literally being the God of Western theism, then it seems probable that either (4) or (5) would be false.

I don’t know of any solid reason to believe that Jesus was mentally ill or a liar, and furthermore, as a matter of fairness, I would presume him to be innocent until proved guilty (either of being mentally ill or of being a liar), but if we reason hypothetically and suppose that Jesus claimed to literally be the God of Western theism, then the game has changed and we would then have a good reason to believe that we were dealing with either a liar or a mentally ill person.

Most people are sane and most people are generally honest, but people who go around claiming to literally be the God of Western theism are not most people. Once premise (1) is granted, all bets are off on Jesus’ sanity and honesty. The presumption of innocence is overturned by this supposition, and the burden of proof shifts to McDowell (and to anyone who would defend the Trilemma) to show that, contrary to appearances, Jesus was actually sane and honest.

Now there is evidence in the New Testament to support the sanity and honesty of Jesus. So, there would be no decisive refutation of the Trilemma on this second horn of the dilemma. However, the information that we have about Jesus is sketchy and questionable, so no claim about the mental health or honesty of Jesus can be shown to be more than probable at best.

One of the most solid claims that can be made about the historical Jesus is that he often taught or preached using parables about the Kingdom of God, but even that is not known for certain. That claim might reach a probability of .9 (9 chances in 10), but most specific claims about the historical Jesus cannot be shown to be that probable. So the claim that Jesus was not mentally ill might be shown to be probable in the range of .8 to .9 (8 or 9 chances out of 10), setting aside the supposition that he claimed to literally be the God of Western theism. The claim that Jesus was not a liar might be shown to be probable to a similar degree, again setting aside the supposition that Jesus claimed to literally be the God of Western theism.

When we add into the mix the supposition that Jesus claimed to literally be the God of Western theism, then the probabilities of (4) and (5) drop significantly. I would say the probability that Jesus was not mentally ill could not be any higher than about .7 to .8 at best, given this additional information. Similarly the probability that Jesus was not a liar would be significantly reduced.

All it takes is for one of these premises to be false and the argument fails. Both (4) and (5) must be true for the argument to succeed. So, the best that the Trilemma can do, even granting the highly dubious assumption that Jesus claimed to be God (in the sense of literally being the God of Western theism), is to give the conclusion a probability of approximately .5 to .6 (.7 x .7 = .49 and .8 x .8 = .64).

So, if my evaluation of premise (1) is correct, then the argument is clearly no good and must be rejected, on the other hand, if my evaluation of premise (1) is wrong, the most that could be derived from the Trilemma (best case) is to make the conclusion somewhat more likely than not.

Furthermore, this best-case probability would not take into account the powerful counterarguments mentioned in the opening of this post, so the all-things-considered probability would end up somewhere significantly south of .5, unless additional reasons equally as weighty as the skeptical arguments above could be produced to counteract them.