20
back
5 / 06
Image of birds flying. Image of birds flying.

#631 Would’ve, Could’ve, Might’ve

May 19, 2019
Q

Dear Dr. Craig, thank you for your ministry that has blessed me tremendously.

While exploring the coherence of Molinism, I ran into a problem that does not seem to have a solution. Can you help explain the following syllogism?

1. A choice can be free only if it is true that it is possible for person X to choose A or not-A in circumstance B.

2. If it is possible for person X to choose A or not-A in circumstance B, then it must be true that person X might not choose A and might not choose not-A in circumstance B.

3. If it is true that person X might not choose A or not-A in circumstance B (required for free will) then it cannot also be true that person X would certainly choose A or not-A in circumstance B (required for middle knowledge).

4. Both of these contradictory statements must both be true in molinism, which affirms free will and middle knowledge.

5. Therefore, molinism cannot be true.

In other words, it seems nonsensical to say that’s it’s true that person X MIGHT NOT choose A in circumstance B (required for free will) AND also to say that’s it’s true that person X WOULD choose A in circumstance B (required for middle knowledge). For example, the following two statements cannot both be true at the same time: (1) I might not kiss my wife if she is wearing red tomorrow AND (2) I would certainly kiss my wife if she is wearing red tomorrow; yet, they both would have to be true for molinism to be true. Thank you again for your ministry and service to the faith!

Blessings,

Rosser

Flag of United States. United States

Photo of Dr. Craig.

Dr. craig’s response


A

Although I think that (1) is false—libertarian freedom requires only that the agent is not causally determined to do what he chooses—, the fundamental failing with the argument is (2).

To see why, consider the standard semantics for counterfactual discourse (see chapter 2 of Philosophical Foundations for a Christian Worldview). Imagine the actual world to be like the sun in our solar system and different possible worlds to be like planets arranged in concentric spheres around the sun. The most similar worlds to the actual world occupy the innermost sphere, and as one goes outwards, the worlds in each sphere become less and less similar to the actual world.

Now a would-counterfactual (like “If I were rich, I would buy a new car”) is true just in case in the all the worlds most similar to the actual world in which the antecedent clause (viz, “I am rich”) is true, the consequent clause (viz., “I buy a new car”) is also true.

If in some of the worlds most similar to the actual world in which the antecedent clause is true, the consequent clause is also true, then a might-counterfactual is true (like “If I were rich, I might buy a new car.”  It is immediately obvious that would-counterfactuals logically imply the relevant might-counterfactuals. For if in all antecedent-permitting worlds I buy a new car, then obviously in some antecedent-permitting worlds I buy a new car. So if it’s true that I would buy a new car, then it’s also true that I might buy a new car.

Such a semantics has the odd implication that if I would buy a new car, then it is false that I might not buy a new car. For if in every antecedent-permitting world most similar to the actual world I buy a new car, then there is no antecedent-permitting world in that sphere in which I fail to buy a new car. But for the counterfactual “If I were rich, I might not buy a new car” to be true, there needs to be an antecedent-permitting world in that sphere in which the consequent is false.

If in some of the antecedent-permitting worlds the consequent is true and in some it is false, then it is true that “I might or might not buy a new car.” So in counterfactual logic “If it were the case that p, then it might not be the case that q” is the contradictory of “If it were the case that p, then it would be the case that q.”

If in all the most similar antecedent-permitting worlds I buy a new car, does that imply that it is impossible that I not buy a new car?  Not at all! In spheres of worlds less similar to the actual world than the one we are considering, there may be lots of antecedent-permitting worlds in which I do not buy a car. So in the envisioned circumstances I could buy a new car even though it is false that I might not buy a new car. The reason this sounds odd is that in ordinary language, as your argument reveals, we typically take “could” to be a synonym for “might.” Since it is true that I could or could not buy a new car, we think that I might or might not buy a new car. That’s incorrect. “Could” does not imply “might.” All that needs to be kept in mind here is that the semantics for counterfactual logic is an artificial, technical semantics in which “might” doesn’t have its ordinary language meaning.

So you’re quite right that “it seems nonsensical to say that’s it’s true that person X MIGHT NOT choose A in circumstance B . . . AND also to say that’s it’s true that person X WOULD choose A in circumstance B.” But you err in thinking that A’s freedom entails that he might not choose A in B. There’s no problem in saying that even though X would choose A in B, he could instead freely choose not-A in B. Thank goodness, Molinism doesn’t require that if your wife were to wear red tomorrow, you both would kiss her and might not kiss her!

- William Lane Craig