A reverse engineering argument

Elliott Sober is a professor of philosophy who has written in support of evolutionary biology.  I’m going through his book “Evidence and Evolution.”

Sober argues for the superiority of the likelihood approach. The  “law if likelihood” states that evidence E favors hypothesis H1 over H2 if and only if the probability of E given H1 is greater than the probability of E given H2; i.e., P(E | H1) > P(E | H2).  Note that this is a comparative approach; it only works when comparing two specific hypotheses.

The surprising thing about this law is that the probability of any hypothesis is irrelevant — it’s the probability of the evidence that counts.  Almost all probability arguments ignore this but Sober thinks Paley’s watch argument is a likelihood argument.  Sober comes close to accepting Paley’s argument but stops short for this reason:  it lacks independent knowledge about what a creator intended.

In other words, it begs the question to say that the creator made the eye to see because we find that the eye sees.  We would need independent knowledge of what the creator intended for the eye before considering whether or not that’s what the eye does.

Christians typically say that God’s intentions are inscrutable or known only generally.  Aristotle tried to discern purpose (final cause) by philosophical means but didn’t get very far, and teleology got a bad name.  Is there another approach?

I see two approaches. One is to find passages of scripture that show specific intentions God has for the creation.  For example, Genesis 1:26 says mankind is to rule over the fish, birds, and livestock on the earth.  What do we observe?  Mankind rules over the fish, birds, and livestock.

However, Genesis assumes the existence of God.  Can we argue without this assumption?  Another approach is to reverse engineer features of the world such as organs like eyes.  What is the design problem that led to the eye being designed as it appears?  For every feature of the world, we could come up with some design scenario.  We would include the possibility that something went wrong and that the design we observe is less than optimal or even perverse.

This would be quite a project, not unlike the evolutionary project of coming up with just so stories about how features could possibly have arisen through evolution.  We would match every evolutionary story with a creation design story.  The likelihood argument then is which hypothesis confers greater probability on what we observe?  The answer is design for several reasons:  (1) even evolutionists agree that life appears designed, (2) complex features such as watches are designed so it would be expected that other complex features we observe are also designed, and (3) there exists a particular design problem for each feature observed.

March 2014