Julia Galef (She of the Straw Vulcan discussion) vlogged about the skeptic (rationalist) take on traditional practices. It's well worth a watch.
She discusses a number of things, including various biases. She also touches on the natural selection analogy. She briefly discusses, edging toward meme theory, that a behavior can be beneficial without being chiefly beneficial to the organism that expresses it.
There are a number of examples of this.
First: An ant climbs to the top of a stalk of grass and waits there until it gets eaten. The ant is expressing a gene that is clearly not beneficial to itself, but to that of a parasite whose life cycle requires spending time in an ant's brain and in a bird's gut. Similarly, when you get sick, you sneeze, which is more to the benefit of your infection than to you.
Second: The Westermarck Effect is that which prevents close kin from doing the nasty. Any humans which grow up in close proximity from a young age imprint on one another such that they are actively disinclined to have sex with one another. This confers little benefit to them, but definite benefit to their descendants, as it encourages outbreeding and reduces the chance of inbreeding depression. This may be countered by the genetic sexual attraction hypothesis, which suggests that relatives are more attractive to us than complete strangers. This can be used to combat outbreeding depression. In both cases, the benefit is to your genes rather than to you.
Galef was responding to her listener's hypothesis: traditions aren't completely groundless. That they exist implies that they probably exist for a pretty good reason. That is, their existence is evidence that they should continue existing. Galef linked it to natural selection; cultural traditions and mores that exist do so because they benefit individuals or groups; she gives a decent example. She follows up with an analogy by G.K. Chesterton of a fence/gate across the road; don't destroy it without learning why there's a fence, that it's part of a larger structure of traditions.
I want to take this further: the good reason a tradition exist might have nothing to do with our benefit, but for the benefit of the tradition (it exists so that it can continue to exist: like a lot of little superstitions), or another tradition (reflexive deference to religion), or other people (the ingrained obsequiousness that supports class structures; "yassa massah"), and so on.
Traditions and cultural mores exist within a world that contains other traditions and mores and each has to be considered within that context. Take for example the tradition of dueling in early modern Europe. Any insult or minor blow had to be answered with a duel, usually to the death. This was a particularly barbaric practice that faded by the end of the nineteenth century, but to try and remove it entirely in 1775 would have been absurd, because it was part of a larger complex of "honor", whereby any insult or blow had to be answered or the insulted's honor was compromised and he was weak, cowardly, or conniving. This caused him material harm as people had less to do with him, were less charitable, less friendly, and so on. That dueling was in fact a much more liberal and cultured response than would have previously been the case is important. Before the careful dance of the duel (arranged by seconds with standards and rules), the response would have been an immediate fight to the death, without call for apology, that probably would have turned into a brawl with multiple injured/dead parties that would have lead to the clans a feudin' down the generations.
Galef points out that her reader's natural selection argument is flawed: a bad gene dies out quickly, whereas a negative cultural more is actually quite sticky. To quote the Declaration of Independence, "Indeed, experience hath shown that mankind are more disposed to suffer, while evils are sufferable, than to right themselves by altering the forms to which they are long accustomed." We stick with the devil we know, occasionally even in the face of an obviously better alternative. Galef argues that this conservatism is in two parts, the other being that there isn't much randomization (mutation) in social mores.
While I can't help but agree with the first, I have to disagree with the second. Cultures and traditions, being untethered in many cases, to reality are free to shift and do so with remarkable rapidity. Music, clothing, speech, how we treat one another, all can change in a heartbeat. She also said that there's not much selection pressure because of this; in an increasingly global and diverse community, I think that's increasingly untrue.
An important point I think Galef failed to make was that while a tradition's existence is evidence of its positive value, it's not necessarily valuable to us, either individually or as a group. A tradition might exist for the benefit of another group, another tradition, or just to itself. Futher, given the very valid point she raised about the status quo bias, that greatly reduces the strength of the evidence the tradition's existence provides.
In the end, though, I agree with her conclusion, that experimenting with social norms is good for all of us. It's why diversity is a good thing.