Consider the following scenario.

A computer has been set to work on a large series of mathematical optimization problems. Either they are not solvable analytically or the computer does not know how; it solves them (to a specified degree of precision) by means of numeric methods.

An expert mathematician-*cum*-computer-scientist is told that the computer has been working on the optimization problems for a single day, and he is told that the computer has solved the first *x* optimizations in the series. He is provided with a cross-section of all the computer's operations for one millisecond. From these he can compile a list of the algorithms that the computer is currently running, but he does not know all the algorithms available to the computer. There may be latent algorithms that are not being implemented at the time.

Having compiled this list, the scientist is then asked: What is the probability that the computer was using *only the algorithms observed and recorded in the list* to solve the optimization problems?

If, as is common, the algorithms the computer is using involve some amount of random search, the scientist probably cannot determine *with certainty* whether the listed algorithms would have been sufficient to solve the optimization problems. Even if the algorithms seem radically inadequate to solve *x* optimizations in one day, the computer *might *have benefited from a string of lucky guesses.

What the scientist could do, in principle, is assign an *ex ante* probability to the occurrence of the computer solving *x* optimizations in one day using the given algorithms. He could deduce a probability distribution for time to solve for each optimization, and then add these distributions to calculate a probability of time to solve *x* optimizations. In a typical case the scientist might find that the chances of the computer solving at least *x* optimizations in one day is 99.9999%, or 1E-30%. The scientist might also be able to calculate the odds of the computer solving *no more than x* optimizations.

If the scientist assumes that anything that is infinitesimally unlikely did not occur, he may draw one of the following three conclusions: (a) the list of algorithms plausibly explain the computer's degree of success, (b) the computer could not have solved *x* optimizations in a day with those algorithms and must have used other, more effective ones for part of the time, or been supplied with some answers by an outside source, or (c) the computer would certainly have solved more than *x* optimizations using the set of algorithms observed, and it must have been interrupted or have been using less effective algorithms for part of the time.

If, on the other hand, the properties of the algorithms or of the optimization problems are not sufficiently well understood, the scientist may profess himself unable to ascertain whether the algorithms are sufficient, or super-sufficient, to explain the computer's success or not.

Evolutionary biology is in a position similar to that of the computer scientist in the example. Life exhibits a certain amount of order. Geology and the fossil record suggest a time scale during which this amount of order is supposed to have emerged. Biologists observe certain processes at work in nature which would tend to alter gene pools over time, and thus to produce order.

But unlike the scientist in the scenario, evolutionary biologists do not understand the properties either of biological order or of the mutation and natural selection processes sufficiently well to estimate the *ex ante* probability that life as we know it could have emerged in a few billion years. If they did, they might find that it's infinitesimally unlikely that life as we know it could have emerged so quickly, or, on the contrary, that sentient beings would have emerged with near certainty within 100 million years at most. In either case, they would have to conclude that the theory of evolution as taught today can't be true; that some other processes we have not observed or conceived of must account for the existence of life; or perhaps, that some other processes must have been slowing evolution down.

As it is, we're just not in a position to know whether the Darwinist claim that all life on earth emerged through evolution is possible or not. If we did know that it was possible, we still would not know, of course, whether it really occurred that way.

The whole above argument assumes that life is reducible to a material basis. I think we know by the experience of free will that human life, at least, is not. In the past I have also presented a proof of "the necessity of agnosticism about mind-brain supervenience." If the task of a theory of life is not merely to explain its material organization but also how at least one organism, man, possesses a supernatural soul, then it seems clear right away that Darwinian evolution is not a very promising theory. But even if we set that aside, the radically inadequate evidentiary basis for Darwinism makes it, not a *bona fide* scientific theory, but simply the latest of the creation-myths man has believed in from time to time.

## Recent Comments