LECTURE #14: Proving God's Existence/Teleological Arguments


I. Why prove God's existence?

A. Because otherwise our belief in God would be irrational (strong evidentialist)

B. Because it can be useful in converting skeptics (weak evidentialist).

C. Because it can support the existence of a civil religion (ethical monotheism) that is needed as a foundation for morals and law and an inspiration to art & culture.

II. Levels of Provability

A. Absolute proof: any reasonable person must believe the premises and must draw the inferences leading to the conclusion that God exists.

B. Person-relative proof: some reasonable persons could believe the premises and draw the inferences leading to the conclusion that God exists, and they could reasonably believe that God exists on the basis of these proofs.

C. Paralogism: no reasonable person could both believe the premises and draw the inferences needed to reach the conclusion.

D. Enlightenment proof: any sane person would believe the inferences and draw the inferences leading to the conclusion.

E. Cartesian proof: any person would be unable to resist believing the premises and drawing the inferences leading to the conclusion.

III. Kelly Clark's critique of "classical apologetics"

A. Clark seems to confuse A with D or E. Aquinas, Geisler, Sproul & Gerstner all claim that God's existence can be proved absolutely (in sense A) -- not that it can be proved by the Enlightenment or Cartesian standard.

B. So, the mere existence of sane and knowledgeable atheists does not refute the position of classical apologetics. The question is: are these sane and knowledgeable atheists reasonable in their unbelief?

IV. Gilson's Arguments for the need to prove (demonstrate) God's existence.

A. If the existence of God were evident (knowable without proof), then there would be no atheists. Makes a mistake similar to Clark's: assuming that "evident" means "irresistibly believable without proofs" instead of "reasonably believable without proofs".

B. Aristotle and Aquinas prove God's existence: you can't prove what is already evident. Why not? The proof might be of help to people who are atheists despite the evident character of God's existence.

C. The only things directly accessible to our knowledge are sensible things. How does Gilson know this?

D. Theological reason: for God's existence to be evident to us, we would have to be able to grasp God's essence. However, we obviously cannot do so. Grasping God's essence is equivalent to the beatific vision enjoyed by the saints in heaven, which we clearly do not enjoy.


I. William Paley's 1803 Natural Theology: the argument of the watch.

Three major construals of the argument: (1) an argument by analogy/similarity (Hume), (2) theory confirmation by Bayes' theorem (Swinburne), (3) argument by statistical elimination (Dembski's "design inference")

II. Various Evidences appealed to by Teleological Arguments

A. The existence of simple, uniform laws of nature (or, equivalently, the existence of a small number of basic kinds of fundamental particles).

B.The fine-tuning of the fundamental constants and the character of the Big Bang (the initial conditions): the so-called "anthropic coincidences".

C. The fitness of the solar system and the earth as an environment for life.

D. The origin of life (at almost the earliest possible moment, within about 10 million years)

E. Irreducibly complex biological mechanisms (Behe's Darwin's Black Box)

F. Pre-adaptations. Primitive life w designed for the sake of leading to more complex forms.

III. The Argument from Similarity. Hume's Cleanthes in Dialogues concerning Natural Religion.

The cosmos resembles a human artefact.

Therefore, the universe was created by something like a human artesan. Objections:

A. There are many dissimilarities as well (e.g., huge difference in scale).

B. There are many apparent defects in the design of the universe -- death & suffering.

C. A dilemma: (1) if the analogy is very strong, then we would be forced to an anthropomorphic view of God (God is very much like us -- corporeal, finite, etc.), (2) if the analogy is not very strong, then we cannot draw any conclusion with certainty.

IV. The Argument from Bayes' Theorem

A. The theorem enables to calculate the posterior probability of a hypothesis, given some new piece of evidence: P(h/e).
P(h/e) = P(h)P(e/h)/P(e)

B. We need three inputs: (1) the prior probability of h, P(h), (2) the likelihood of the evidence, assuming the truth of the hypothesis, P(e/h), (3) the prior probability of the evidence. For P(h/e) to be high, we want P(h) and P(e/h) to be fairly high, and P(e) to be very low.

C. We can compute the probability of P(e) by the formula:

P(e) = P(e/h)P(h) + P(e/-h)P(-h)

Since we want P(e/h) to be high, P(e/-h) must be very low. That is, the evidence e must be something that would be very unlikely (surprising) if the hypothesis were false.

D. If P(e/h) > P(e/-h), and P(h) 0, then we say that the evidence e confirms hypothesis h. That is, e raises the probability of h: P(h/e) > P(h), the posterior probability of h is higher than the prior probability of h.

E. Swinburne's h = the existence of the God of classical theism (an infinite, uncaused, perfectly good & reasonable person). Swinburne argues that this hypothesis is extremely simple, so the prior probability P(h) is fairly high.

F. Swinburne argues that the existence of simple natural laws and the presence of the anthropic coincidences are both evidences for which P(e/h) is fairly high and P(e/-h) is extremely low.

V. Dembski's Design Inference (from The Design Inference and Intelligent Design)

A. Dembski claims that his model more accurately captures Paley's argument. Dembski agrees with Thomas Reid that we have the innate capacity to recognize the effects of intelligent agency through a certain kind of mark: specified complexity.

B. For Dembski, the hypothesis h = the claim that the observed phenomenon is the product of intelligent design. Note: this does not by itself establish that the designer is God.

C. According to Dembski, we do not have to estimate P(h), the prior probability of h, or P(e/h), the likelihood of the evidence on the assumption of intelligent design.

D. Instead, we need only two things: (1) P(e/-h), the probability that the phenomenon was produced by undirected natural forces, must be astronomically small (something like 1 in 10 to the 50th power, and (2) the evidence e must be specified (it must conform to some real, independently describable pattern).