http://www.adequacy.org/stories/2001.9.30.172813.212.html
I must clarify my position before I start. This is will not be yet another creationist argument. My essay will not push any alternative theory as to why the universe is the way it is. I will merely point out problems, and grave ones.
One of the reactions I expect will come up can roughly be stated as follows: "Hey, the theory is not perfect, but it is scientific, and it's the best we have!" To which my answer, let it be known beforehand, is the following:
If Evolutionary Theory truly is the best theory we have to explain the diversity of species, then it can only mean that all our theories suck.
With this out of the way, let's dive into the suckiness.
Is Natural Selection a tautology?
The short answer is: it depends on what you take the relationship between mathematical and logical truth to be. Of course, this only reveals that the question is the wrong one to ask. The correct question is rather the following: is Natural Selection an empirical theory? And the answer is a clear and resounding no.
Of coruse, defenders of evolutionary theory, by presenting a refutation of a couple of weak arguments that Natural Selection is a tautology, leaves its readers thinking that the principle has some actual empirical content. But this is plainly false, as follows from their own argument.
Let's quote their own words:
The current understanding of fitness is dispositional. That is to say, fitness is a disposition of a trait to reproduce better than competitors. It is not deterministic. If two twins are identical genetically, and therefore are equally fit, there is no guarantee that they will both survive to have equal numbers of offspring. Fitness is a statistical property. What 'owns' the fitness isn't the organism, but the genes. They will tend to be more often transmitted so far as what they deliver is better 'engineered' to the needs of the organisms in the environment in which they live.
The mathematical model being invoked implicitly here is a generational model, where each gene is marked with a fitness value, and this value determines its expected number of sucessors in the following generation. Within such a model, it logically follows from the laws of probability that the genes with the higher tendency towards leaving offspring will command a higher share of the total population in each successive generation. There is absolutely nothing whatsoever empirical in this argument-- it is a purely mathematical proof. Natural Selection is mathematically true. Thus, it is impossible to present empirical evidence against it, and this disqualifies it as a scientific theory.
Please note what this argument is not. It is not an argument that specific hypotheses about how a species evolved are unfalsifiable. Such arguments invoke contingent hypotheses or facts in addition to Natural Selection, and thus are subject to refutation. For example, the famous early peppered moth experiments fall clearly within the range of empirical science. But, contrary to what is claimed, they never provided any sort of empirical support to Natural Selection nor did the principle run the risk of being refuted if the experiment failed. Had the experiment failed, the theorists would have had to reject some assumption(s), but rejecting Natural Selection is tantamount ro rejecting the laws of probability and statistics themselves!
Can Natural Selection really change organisms as extensively as claimed?
A very common complaint of the skeptic against Evolutionary Theory, and in particular to the very peculiar claim that all life is derived from a common ancestor, is to question the possibility that Natural Selection can actually produce all the change in populations over time that the theory would require to account for the current diversity in organisms. In layperson terms, can Natural Selection really take some claimed ancient lizards, and through a series of mutations that inflict gradual changes on its offspring, through an extended period of time, turn the lizards into dinosaurs, birds, and human beings, and all through the process, keep the intermediate forms well adapted to their environments?
This has to do with the difference between microevolution and macroevolution. One can accept the first without having to accept the second. Indeed, the first one is supported by facts quite extensively and convincingly; even creation scientists accept it. But can we question macroevolution?
A typical evolutionist answer is to laugh away the question by claiming that the person asking it "does not grasp the enormity of the time scale involved", or by subtler arguments:
Antievolutionists argue that there has been no proof of macroevolutionary processes. However, synthesists claim that the same processes that cause within-species changes of the frequencies of alleles can be extrapolated to between species changes, so this argument fails unless some mechanism for preventing microevolution causing macroevolution is discovered. Since every step of the process has been demonstrated in genetics and the rest of biology, the argument against macroevolution fails.
I'm sorry, but this is clearly among the worst arguments ever made in support of anything. Instead of answering the challenge with an actual model of how gradual microevolutinary change can result in macroevolutionary change, not only do the evolutionists decline responsibility for doing so, but to top it off place the responsibility of showing that their unsupported process can't happen on the people who point out that it's unsupported!
This is simply ridiculous. If you want people to believe you that X happens, you actually have to offer evidence that it does happen. You can't claim that unless the skeptics can prove that it doesn't happen, then it happens. It is intellectually dishonest, and logically inconsistent.
Sure, the talk.origins crowd has put forward more detailed arguments about why you should believe in macroevolution. But don't bother reading that document just yet. Not until you read my next section, which outlines the fallacy underlying all evolutionary reasoning, and will enable you to refute it point-by-point yourself.
Do greater than chance similarities prove common origin?
One of the classic arguments in favor of the evolutionist claim that all species have a common origin is that species show all sorts of similarities which can't be due to chance. Indeed, the enormous mass of odd similarities that he found in his studies of nature around the world is what convinced Darwin of the common origin of all species, and he didn't know half of the similarities that we know today. Thus, the standard evolutionist argument is: If all those species are much more similar than mere chance would allow, then they must be related.
However, a greater than chance similarity between organisms only proves precisely that: greater than chance similarity between organisms. There exist other possible non-evolutionist explanations for the same facts, the most popular of which is that species were designed by some kind of ultra-powerful being. However, we don't need to go to such extremes to put the argument under stress.
An aside: Historical Linguistics
Let's start out by discussing a different but methodologically related field: Historical Linguistics. This is the original branch of modern linguistics; it started with the discovery in the late 18th century of systematic (way greater than chance) sound-meaning correspondences between the vocabularies of Latin, Greek and Sanskrit. This led to the theory that the three languages are descended from a common ancestor language, Proto-Indoeuropean. As the topic was explored during the 19th century, more and more languages were shown to belong to the family, additional families were discovered, and a powerful method arose for proving such relations: the Comparative Method.
If you strip away the linguistic details, the comparative method looks suspiciously alike to the methods of evolutionary biology. This is no surprise, since Darwin actually was inspired and influenced by the methods of Historical Linguistics, and mentions their methods and findings as an important analogy to his theory.
But there is a question that remains unasked here: how do linguists actually know that the systematic correspondences observed in the relevant languages are actually due to common descent, and not due to some other reason? For example, could it just not be the case that the correspondences are due to some undiscovered natural relationship between certain sounds and meanings?
The answer to this is three-fold:
1. We actually have records of one language diverging over time and geography, with several distinct languages as a result. We have ample documentation of this happening in Latin, to give only one example.
2. Linguists ahdere to a theoretical principle called the abritrariness of the linguistic sign, due to Ferdinand de Saussure, which states that the relationship between sound and meaning is arbitrary.
3. More crucially, we have thousands of languages for which one can't show common descent. We even know about at least one language that is not related to any other, Nicaraguan Sign Language
This last point is very important. Having languages that are very different from the languages of one family shows indeed that the similarities in the related languages can't be due to chance. If we had no languages that differed wildly from the Indoeuropean ones, we would have no empirical basis for the arbitrariness principle. The fact that such languages actually outnumber the Indoeuropean languages boosts the likeliness of the common descent explanation.
What does this mean for Evolutionary Theory?