LECTURE #15: Objections to Design

I. God of the gaps objection:

"Design inferences are simply appeals to ignorance or incredulity. Just because we haven't yet found a naturalistic, undirected cause, doesn't mean we never will. Strictly speaking, the God of the gaps error is made whenever P(e/~h), the probability of the event's occurring without design, is simply unknown. If we know that P(e/~h) is low, then an appeal to design is perfectly reasonable. Admittedly, it can be hard at times to determine the value of P(e/~h).

II. Some Objections to Swinburne's Bayesian Argument

A. The prior probability of theism, P(h), might be very low (Clark makes this point). All of the intelligent agents we know are finite and corporeal, but God is supposed to be infinite and incorporeal. In addition, all the persons we know consist of highly improbably arrangements of molecules: God would have to consist of a highly improbable arrangement of "spiritual" components.


  1. Classical theism is an extremely simple hypothesis (all of God's attributes take values zero or infinity). Good scientific practice gives high probability to simple hypotheses.
  2. An infinite person might be much simpler in construction than a finite one. We need so many parts precisely because our knowledge is limited and mediated by physical processes. God has immediate access to all facts, and so needs no internal complexity at all.
  3. As we've seen (Plantinga), assigning theism/design a low probability leads to epistemological breakdown.
B. Swinburne is just guessing when he assigns a low value to P(e/~h). We haven't observed a large sample of undesigned universes, so we don't know how likely it is for a random universe to be governed by elegant, anthropic laws.


  1. In the absence of God, the anthropic values and the uniformity of natural law would constitute an extraordinary series of coincidences. It is reasonable to assign such coincidences a low probability.
  2. In the absence of actual observations of undesigned universes, we should assign probability to P(e/~h) on the basis of some kind of symmetry consideration: use a smooth, flat distribution. When doing this, the very narrow range of anthropic values will end up with a very low probability, no matter which specific distribution you use. (I.e., the result is robust.)
C. We can't possibly assign a high probability to P(e/h), since we have no idea what sort of world God would be likely to create.

Response: given God's rationality and goodness, we can reasonably assume that certain kinds of worlds (those containing sentient and intelligent life) are especially appropriate objects of God's choice.

III. The Many-Universes Hypothesis: Observer Selection

Consider this alternative hypothesis: there are an infinite number (or astronomically large number) of universes like our observable universe, existing in other, inaccessible dimensions. The laws and fundamental constants of these universes vary randomly from one to the other. Observers like us can exist only in universes with uniform laws and anthropic values.


[See my web site.]

IV. Design Arguments Fall Short of Establishing Theism

As David Hume pointed out, the designer of the universe might be finite, a committee, fallible, corporeal,... It's hard to establish the infinite power and intelligence of the creator, given a finite creation. Swinburne: an infinite creator is the simpler explanation. However, that doesn't make the infinite-creator hypothesis more likely than all the finite-creator hypotheses put together.

V. An Infinite Regress Problem?

A. We infer the existence of God from the existence of specified complexity (Dembski): an improbably by highly functional or significant arrangement of matter. Wouldn't God's existence be another example of specified complexity, requiring another designer to design God, ad infinitum?

B. See IIA2 above, and my svsu.html web page: an infinite mind might be extremely simple. God needs no representations and no sense organs: everything (including every possibility) is immediately present to His mind. God needs no inference engines, because God never has to infer anything -- there are no gaps in His knowledge needing to be filled.

Some other useful web sites:

Reasons to Believe (scientific evidence of fine-tuning)
Access Research Network (intelligent design movement)
Critique of Behe
Robert C. Newman's "A Designed Universe"
Robin Collins
Walter Bradley
William Lane Craig