Replies

8. Mathematically, it is inconceivable that anything as complex as a protein, let alone a living cell or a human, could spring up by chance.
Chance plays a part in evolution (for example, in the random mutations that can give rise to new traits), but evolution does not depend on chance to create organisms, proteins or other entities. Quite the opposite: natural selection, the principal known mechanism of evolution, harnesses nonrandom change by preserving ‘desirable’ (adaptive) features and eliminating ‘undesirable’ (nonadaptive) ones.

But the raw material on which natural selection acts is random copying errors (mutations). If evolution from goo to you were true, we should expect to find countless information-adding mutations. But we have not even found one.

As long as the forces of selection stay constant, natural selection can push evolution in one direction and produce sophisticated structures in surprisingly short times.

An example would have been nice.

As an analogy, consider the 13-letter sequence ‘TOBEORNOTTOBE.’ Those hypothetical million monkeys, each pecking out one phrase a second, could take as long as 78,800 years to find it among the 2613 sequences of that length. But in the 1980s Richard Hardison of Glendale College wrote a computer program that generated phrases randomly while preserving the positions of individual letters that happened to be correctly placed (in effect, selecting for phrases more like Hamlet’s). On average, the program re-created the phrase in just 336 iterations, less than 90 seconds. Even more amazing, it could reconstruct Shakespeare’s entire play in just four and a half days.

These computer programs have been widely popularized by the atheist Richard Dawkins, but are a lot of bluff. Such simulations as Dawkins, and now Rennie, propose as ‘simulations’ of evolution work towards a known goal, so are far from a parallel to real evolution, which has no foresight, hence a ‘Blind Watchmaker’. The simulations also use ‘organisms’ with high reproductive rates (producing many offspring), high mutation rates, a large probability of a beneficial mutation, and a selection coefficient of 1 (perfect selection) instead of 0.01 (or less) which parallels real life more accurately. The ‘organisms’ have tiny ‘genomes’ with minute information content, so are less prone to error catastrophe, and they are not affected by the chemical and thermodynamic constraints of a real organism.

For more information, see this refutation of Dawkins’ book Climbing Mt Improbable, Weasel Words and Dawkins’ weasel revisited. Also, in the coming issue of TJ (16(2)), we will have an article about a more realistic computer simulation, which will be downloadable (i.e. the program) from the AiG Web site, which shows that the goal is NOT reached if realistic values are programmed, or it takes so long that it shows that evolution is impossible. For a refutation of the whole idea of computer simulations of evolution, particularly in the guise of genetic algorithms, see Genetic algorithms—do they show that evolution works?—all these problems also apply to the simplistic ‘simulation’ Rennie writes about.

Also, when it comes to the origin of first life, natural selection cannot be invoked, because this requires a self-reproducing entity. Therefore chance alone must produce the precise sequences needed, so these simulations do not apply (see Q&A: Probability). And a further problem with the alleged chemical soup is reversibility, intensifying the difficulty of obtaining the right sequence by chance—see Could Monkeys Type the 23rd Psalm.