Wednesday, July 15, 2015

Is a necessary being inconceivable?

Consider this argument:

  1. Obviously necessarily, if N is a necessary being that exists, it is impossible that N doesn't exist.
  2. It is conceivable that N doesn't exist.
  3. So it is inconceivable that N exists.
For this argument to work, we need to be able to make the inference from:
  1. Obviously necessarily, if p, then necessarily q.
  2. Conceivably not q.
  3. So, not conceivably p.
Suppose that p just is the statement that necessarily q. Then (4) is uncontroversial. If the above argument form is good, then so is this one:
  1. Conceivably not q.
  2. So, not conceivably necessarily q.
But why can't we conceive both of not q and of necessarily q? Why should the ability to conceive of one thing, viz., the necessity of q, preclude the ability to conceive of another, viz., not q?

The principle that conceivability is defeasible evidence of possibility may seem relevant, but I don't think it establishes the point. That I can conceive of necessarily q is evidence of the necessity of q. That I can conceive of not q is evidence if the possibility of not q. So, if both, then I have evidence for two contradictory statements. Nothing particularly surprising there: quite a common phenomenon, in fact!

Suppose A and B are contradictory statements. It may be that evidence for A is evidence against B. But is evidence for A evidence against there being evidence for B? If it is, it is very weak evidence. Likewise, even given the principle that conceivability is evidence for possibility, the argument from (7) to (8) is very weak, much weaker than the inferential strength of this principle.

To summarize: The strength of the inference from (1) and (2) to (3) in the original argument is about equal to the strength of evidence that the existence of evidence for A provides against the existence of evidence against A. But the existence of evidence of A provides very little evidence against the existence of evidence against A. So the original argument is a very weak one. It would be improved if the conclusion were weakened to the claim that it is impossible that N exists, and then I would focus my attack on (2).

1 comment:

Walter Van den Acker said...

Dr Pruss

First let me tell you that I don't think conceivability really is evidence for possibility.
But I do think that some poeople who argue for a necessary being tend to underestimate what exactly they are arguing for.
A necessary is a being for which there exists no possible alternative. So, I wouldn't state (3) as "It is conceivable that N doesn't exist,so not inconceivable that N exists" but instead I would say, "There is a possible alternative A for N, so N doesn't exist".

Of course it would be very difficult for me to prove that A is possible, but keep in mind that A doesn't have to possess a lot of properties, in fact it simply has to be something very simple (or maybe even nothing at all), whereas N is much more complex as a hypothesis than A.
The question then becomes, why should we favour the more complex hypothesis over the simpler one?
Notice that I am not claiming that N is necessarily a more complex being than A.