Science and Randomness

Jim Manzi has a post at NRO which I mostly agree with, but I think he makes a fundamental error regarding randomness.  So I sent him this:

Greetings, Jim!

I'm going to take issue with a point in this post at The Corner, Science and Religion in The New Republic. For the most part, I agree with you, but I think you have a few points wrong.

Coyne's demonstration that we know that humanoid consciousness could not have been inevitable because evolution is "contingent," and depends on "unpredictable changes" and "random physical events" is an attempt to sweep a lot under the rug. In fact, even the "random" elements of evolution—for example, mutation and crossover—are really pseudo-random. For example, if a specific mutation is caused by radiation hitting a nucleotide, both the radiation and its effect on the nucleotide are governed by normal physical laws. Human uncertainty in describing the evolutionary path that results, which as a practical matter we refer to as randomness, is reducible entirely to the impracticality of building a model that comprehensively considers things such as the idiosyncratic path of every photon in the universe compounded by the quantum-mechanistic uncertainty present in fundamental physical laws that govern the motion of such particles. As a practical matter, we lack the capability to compute either a goal or the path of evolution, but that is a comment about our limitations as observers, not about the evolutionary algorithm itself.

I think it's an error to describe the events in evolution as "pseudo-random". The term "pseudo-random" shows up in computer science, especially with respect to programs designed to generate "random" numbers. An example of this is the RAND() function available in Microsoft Excel. This function generates a random-looking number every time it is called. The number is not really random, because the underlying computer program is deterministic. As a result, among other things, the sequence of numbers the function generates will eventually repeat. A good pseudo-random generator is one where the numbers generated pass various statistical tests for randomness. That is, until they do start repeating, they "look random enough".

Other phenomena in nature actually are random. In particular, quantum events -- any which involve quantum mechanical effects -- cannot be predicted, even in principle. If you build a quantum mechanical random number generator (very simple), the numbers generated will be random, and the sequence will never repeat. Ever.

Taking up the example of a mutation caused by radiation hitting a nucleotide, yes, it's true that radiation obeys physical laws. And yes, it's true that a nucleotide altered by radiation obeys certain physical laws. However, the interaction of radiation with a nucleotide is a quantum-mechanical event and whether it happens at any given time to any given nucleotide depends on the probability that a quantum-mechanical wave function will collapse when radiation meets molecule. In effect, when radiation passes by a nucleotide, dice are rolled. If the dice come up one number, the radiation passes on without having any effect. If another number comes up, the radiation interacts with the nucleotide. And because of the laws of quantum mechanics, there is no way, even in principle of predicting which number will come up.

And believe me, lots of very determined and very resourceful people have looked.