tag:blogger.com,1999:blog-4456348657596914237.post1333241820694555455..comments2024-03-29T01:19:46.849-04:00Comments on Eco-Evo Evo-Eco: Non-adaptive explanations as null hypotheses in Evolutionary EcologyBen Hallerhttp://www.blogger.com/profile/17875404974157070805noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-4456348657596914237.post-41969671485088396342013-09-09T09:17:52.348-04:002013-09-09T09:17:52.348-04:00Great comments guys, keep 'em coming. It's...Great comments guys, keep 'em coming. It's good to hear from other EvoEcos, and I'm interested to find out whether I'm in the minority on the parsimony issue.Njal Rollinsonhttps://www.blogger.com/profile/13621792646309807020noreply@blogger.comtag:blogger.com,1999:blog-4456348657596914237.post-23952983126449856792013-09-05T11:22:09.234-04:002013-09-05T11:22:09.234-04:00Very nice discussion! Thanks guys. Here are some t...Very nice discussion! Thanks guys. Here are some thoughts related to an article I read some time ago on the topic: I find that part of the problem lies in the modus operandi of science. The first thing science does is to 'take apart things' (interestingly this is reflected in the syllable 'sci', as in 'scissors' or 'schizophrenia'). Then, after observing particular phenomenons and variables in this very detailed way, science has to reverse the process and generalize these observations to build theories that might explain the bigger, broader picture.<br />I think the problems surrounding parsimony in scientific reasoning are somehow located in this transition process from the specific to the general. To me it seems that parsimony is a very good first heuristic to guide our approach to specific observations and our model explanations for them. However, if we zoom out of the details and try to generalize, parsimony should by no means be seen as a general principle. Parsimonious approaches in particular cases do not have anything in common and parsimony should not be seen as a shared property of different cases. It is in fact one out of many heuristics that are needed to gather information to build hypotheses and it has to be rejected (of course) once we figure out that more complex interactions are at work than originally thought. Simplicity always has to depend on the context and is therefore not general. An interesting and more in-depth discussion on this can be found in this article by Elliot Sober: Let's razor Ockham's razor.<br />Looking forward to read more contributions. Gregor<br />gregorhttps://www.blogger.com/profile/07799185650144409927noreply@blogger.comtag:blogger.com,1999:blog-4456348657596914237.post-56963829257672473142013-09-04T11:25:44.608-04:002013-09-04T11:25:44.608-04:00Njal, very thought-provoking post! I think that I...Njal, very thought-provoking post! I think that I disagree with you, although you raise many good points.<br /><br />You say "Isn’t life-history theory rooted in the idea of trade-offs, such that there is an expectation that not all traits can be simultaneously optimized in an individual?" I would certainly agree with that (although empirical evidence for such trade-offs is surprisingly thin, as far as I could find in a recent lit search). But if constraints and trade-offs are universal, then won't that necessarily mean that many differences between species are not, in fact, adaptive, but are merely the result of selection on other traits, plus ubiquitous pleiotropy and trade-offs?<br /><br />At the other end of the spectrum, you argue that within-population variation is unlikely to be adaptive. But it seems to me that negative frequency-dependent selection is probably very common in nature (after all, it can be generated by processes ranging from competition to predation to parasitism to sexual selection, and more), and that it can often provide a compelling adaptive explanation for within-population variation. This hasn't been tested empirically very often, but when it has been (e.g. Schluter, Bolnick), positive results have been common. And the alternative position seems hard to argue: that some of the individuals in the population are substantially more fit than others, but that stabilizing selection is somehow unable to weed out the losers. That would make sense in high gene flow source-sink situations, but otherwise, mutation is just too weak to maintain the within-population diversity that we typically see, it seems to me.<br /><br />Now, I make no claim that the above arguments are airtight. I only claim that they are plausible: it is plausible that differences among species are commonly non-adaptive, and that differences within populations are commonly adaptive. That can be reasonably argued. And this is the problem with parsimony-based arguments: reasonable people can disagree as to which explanation is the more parsimonious, except when parsimony is defined in some rigorous mathematical way, as in information theory.<br /><br />I think rejecting theoretical models based on parsimony alone is deeply problematic for that reason. The role of theory is to show what is possible, not what is probable. To determine what is probable, empirical data is needed. When empirical data is hard to get, that can make it difficult to choose among competing theoretical models; but I don't think that ought to be taken as justification for weeding out models using ad hoc parsimony-based arguments. Instead, it should be taken as a challenge for the empiricists! :->Ben Hallerhttps://www.blogger.com/profile/17875404974157070805noreply@blogger.com