I am talking about artificial life systems as Tierra or Archis, not the usual genetic algorithms.
In usual GAs the design is pretty evident in then choice of the fit function and in the genotype --> fenotype map. The result depends on these choice.
For the limited time I have now:
"The May 8, 2002, edition of Nature magazine carried a study showing the terrible dilemma facing the Darwinist account of biocomplexity. Unable to offer any scientific evidence from nature, the fossil record or the laboratory to back up the claim that the complex features in living things were acquired in stages, Darwinists were apparently seeking consolation in the virtual world.
The researchers wrote that on the programming platform called 'Avida' which they had developed they had evolved rival digital organisms by subjecting them to mutation, and that after thousands of (accelerated) generations these came to possess complex features. They then claimed that their results had shown that the complex functions in living things could emerge through natural selection and random mutations.
However, it is erroneous to portray this experiment carried out in silico (in a computer environment) as evidence for Darwinism, because the complexity obtained in the experiment first emerges in a 'programmed' computer program, in other words through conscious intervention. All these simulations in the computer take place in line with the parameters established by the programmers, all of whom are evolutionists.
No matter how much the events in the simulation appear to happen at random, this is actually deceptive. In this research, clearly carried out with the aim of providing support for evolution, the base parameters are clearly set out in accordance with evolutionist preconception. Nonetheless, this appears as a scientific study in some news resources. Space.com carried the story under the heading “Darwin Proved Right by Experiment with 'Alien' Life” ,Yet that interpretation is very definitely incorrect, since the study is based on a directed process, not a random one.
For example, the digital organisms in this experiment are small computer programs 'rewarded' according to their mathematical ability. A digital organism which performs successful functions is allowed more computer time in which to achieve further success by multiplying. Spac.com however ignores the fact that this is a controlled process, evaluates the situation in the light of evolutionist preconceptions, and states,
“Importantly, the experiment found that complex logical ability never evolves unless simpler abilities?foundational mutations?are rewarded.”
It is evident that the question of what change will be rewarded is entirely up to the programmer's initiative. The programmer first sets out the change to be rewarded, and then forms the reward to be enjoyed by the digital organism undergoing this change; in other words he decides which stage it will move on to. In nature, however, there is no conscious system which knows what is a reward and what a punishment and which selects organisms accordingly. Natural selection, upon which evolutionists rely, is a process put forward as an entirely blind mechanism, lacking any consciousness and therefore unable to distinguish between good and bad, useful and harmful, positive and negative. In short, natural selection lacks one thing clearly supported in this simulation; purpose. Richard Lenski, one of the researchers, made the following comment in space.com regarding the study,
“Our work allowed us to see how the most complex functions are built up from simpler and simpler functions."
It is inconsistent to say that a study whose results were evident beforehand, and even set out by the programmer himself, enabled anyone to see anything. Lenski's words are as illogical as saying when an alarm clock rings at the set time, 'Our study enables us to see that an alarm clock will ring at an appointed time.'
Later in the space.com article Lenski says, “We also saw that some mutations looked like bad events when they happened, but turned out to be really important for the evolution of the population over a long period of time." He suggests, in other words, that harmful mutations can have a long-term beneficial effect on a population.
Lenski's mutations may not have had harmful effects on the digital organisms, and may even have been useful enough to give rise to more complex digital organisms. Yet the mutations in Lanski's digital organisms and those in real-life organisms are very different in terms of their effects. That is sufficient to show that Lenski's comments on in silico mutations need to be restricted to the in silico world.
An article by the chemist Steven A. Benner in the January 9, 2003, edition of Nature magazine contained a number of comments on 'synthetic biology' and said that the mutations in simulations were a far cry from the real thing:
“Computer models that simulate replication and evolution in silico are relatively easy to come by. A computer program can suffer mutations and keep on ticking. But real molecules often change their behaviour dramatically upon even a slight change in structure. Chemists have in hand a modest number of chemical systems that can function as templates for their own synthesis. But those that can suffer mutation and still have 'children' are proving harder to find.”
It has been observed that random mutations always have harmful effects on living things. For instance, a fly's head may come out of its chest or a baby may be born with one eye in the middle of its head. Not a single instance is known of these mutations adding information to a living thing's DNA and thus turning it into a more complex life form. In other words, a digital organism subjected to mutation in Lenski's simulation turning into a more complex form by 'surviving' stems not from imitation of the effects of mutations in the real world, but rather from computer parameters set up in the light of preconceptions.
Another inconsistent aspect of this research is the fact that there are no realistic simulations of the digital organisms employed imitating living ones. On space.com Christoph Adami reveals this inconsistency by saying that these have nothing to do with life on Earth but are nevertheless genuine organisms. Organisms bearing no similarity to anything living on Earth are being used in a simulation alleged to produce findings regarding the origin of life on Earth.
In fact, no in silico simulation can realistically imitate the complex structure of life. For instance, the biochemical events which take place inside a bacterium depend on thousands of complex molecules. It is very difficult for a computer to simulate even the packaging of a single protein. Protein folding happens so fast, in units as small as millionths of a second. So computer simulations have been suggested as an alternative but up to this point the massive number of calculations involved have proven to be too much for single computers to handle. Using one computer, "it would take 10,000 days or 30 years just to see one protein fold," the Stanford University physical chemist Vijay Pande, who published such a study in 2002, said.
For that reason, the results of experiments on in silico organisms have no significance for real life, and any experiment based on digital organisms goes no further than being a computer game.
As we have seen, this simulation experiment published in Nature magazine offers no evidence to show that the complexity of life is the result of evolution. Virtual initiatives of this kind stem from the fact that the theory of evolution, based on gradual stages, is quite unable to account for the generally irreducibly complex structure of life. The only explanation of irreducible complexity is that the components were present, fully formed, at one and the same time, in other words that they were created. Our advice to Nature magazine is that it rely on studies based on true facts about the world, not on fantasies allegedly proven in the virtual one."
