A recent posting [Darwin Still Rules, but Some Biologists Dream of a Paradigm Shift] raised the issue of adaptationism. The controversy is over the main mechanism of genetic change in evolving populations. Adaptationists tend to attribute as much as possible to natural selection while pluralists emphasize the important role of other mechanisms of evolution, like random genetic drift.
There seems to be little doubt that most of the fixed alleles at the molecular level are probably neutral in their effect. Thus, they have been fixed by random genetic drift. This includes many amino acid substitutions in proteins. Even though these substitutions change the structure of a protein by a small amount, it does not seem reasonable to assume that they have all been selected.
Most adaptationists are content to concede this point (although there are holdouts). However, they draw the line at more "visible" mutations. According to this group, the vast majority of "visible" mutations are subject to natural selection and therefore most fixed alleles with a "visible" phenotype are adaptations. The argument seems to be that once a mutation produces a "visible" phenotype then it is not appropriate to suggest that it might be neutral with respect to natural selection. The line seems to be drawn somewhere above differences in the amino acid composition of proteins but it's not clear exactly where.
p-ter is one of those who are very reluctant to admit that a visible character could have been fixed by accident. He has posted a short article on Gene Expression [Do phenotypes evolve neutrally?]. I recommend that you read the comments to see examples of the extreme version of adaptationism. Most of these adaptationists will even argue that human blood types are adaptive. The idea that most native North Americans have type O blood is due to some undefined selective advantage and not to accident.
This argument has been going on for several decades. As usual, it's not about the existence of natural selection or random genetic drift. It's about their relative importance in evolution. To reiterate, the adaptationists believe that almost all mutations with a visible phenotype have been fixed by natural selection. The pluralists think that many of them are neutral and have been fixed by accident. The adaptationists make a distinction between what happens at the molecular level and what happens at the "visible" level while the pluralists think the same mechanims are operating at both levels.
Richard Lewontin uses the example of the Indian and African rhinoceros to focus the debate. The African rhinoceros has two horns while the Indian rhinoceros has only one. The question is whether this difference is due to natural selection—is two horns better than one in Africa? Or, is it just an accident of evolution that one species has two horns while the other has only one?
I don't understand why the adaptationist camp is so reluctant to admit that some visible characters can be fixed by random genetic drift. The idea that every feature of an organism has to be an adaptation seems so out of touch with our modern understanding of evolution that I'm really puzzled by the vehemence with which adaptationists defend their orthodoxy. It seems as though admitting that visible phenotypes might be non-adaptive is a major threat to their worldview.