Can Biologists Admit We Are Wrong? Dunno. But We Will Say Other Biologists Are Wrong

Because we are human after all. Jason Collins at Evolving Economics, in response to my post about one economist's misunderstanding of biology, asks a very good question:

On the flip side, did Dawkins or Gould (or their respective supporters) ever concede to the other side that they were wrong and substantially change their world view?

I agree with Razib about what happened:

My own attitude is that both Richard Dawkins and Stephen Jay Gould retreated from maximalist positions when it came to the gradualism vs. punctuated equilibrium arguments of the 1970s substantively. But rhetorically they often seemed to downplay their modifications, and assert more that their own positions were a change of degree, while their antagonist in the argument would have to make a change of kind to align with the evidence. This sort of semantic gamesmanship is disappointing, though alas rather conventional. But since I'm not a thorough master of the oeuvre of both men I'd be curious what readers think.

Let's consider two areas where Gould and Dawkins both backed off: punctuated equilibrium and levels of selection (it's interesting that they disagreed on both issues, and I don't think the two issues are necessarily related). So onto punk eek.

I think the most neglected, but important, thing that came out of this debate was the development and refinement of statistical methods to see if, in fact, the fossil record was actually punctuated. In terms of 'maximalism', Gould, at one point, did get into a weird saltationist phase where punctuated patterns (to the extent they exist) were due to events that were similar to Goldschmittian 'hopeful monsters' (i.e., mutations of large effect). He was--or at least, the field as a whole was--disabused of that notion, and I would argue that the field has incorporated punctuated equililbrium into a standard population genetics/evolutionary framework. The other thing to point out is that, at least among biologists, Philip Gingerich was far more influential than Richard Dawkins in critiquing the patterns of punk eek (fossil data are more convincing--though not all were convinced). And now that we have some historical context, Dawkins' claim that punctuated equilibrium is an "interesting but minor wrinkle on the surface of neo-Darwinian theory" is unfair. While it might seem hard to believe with a dinosaur special on TV all the time, in the 1970s, paleontology was derided by many population geneticists as 'stamp collecting' and wasn't incorporated into evolutionary biology. Punk eek went a long way in changing that view.

Onto levels of selection (or units of selection. Or targets of selection). Here, Dawkins really got his ass handed to him. Dawkins' early work (e.g., The Selfish Gene) was full-throttled, unrepentant genic selection. And it was blasted out of the water by many, including philosophers of science (Robert Brandon, Bill Burian) and population geneticists (Lewontin). It just didn't work once you include epistasis--the fitness of one allele depends on the alleles at different genes that allele is associated with. Genic selection reduced to 'genetic bookkeeping', but it didn't say anything meaningful about the targets of selection (Leo Buss, in The Evolution of Individuality, argued for this reason that only entities that reproduce could be viewed as units of selection). There are a lot of good things in The Selfish Gene, but this wasn't one of them (then there's the whole meme thing, which really drives me around the bend*). While introductory biology courses still assign Dawkins--he is a good, passionate writer--I don't think the genic selection ideas have much credence at all within the discipline**.

The other reason the units of selection debate receded has to do with something John Hawks raised:

I do get the feeling that reductionist theories are easier to deal with, and discard, than "world system" theories. Lamarckism comes to a crux on a single kind of observation in a way that Marxism does not. But the kin vs group selection debate seems never-ending precisely because many empirical cases can be reframed from both perspectives, with some vivid dispute about the relevance of some facts.

I used to be heavily invested in units of selection debate, to the point where I conducted some lab experiments using bacteria (and planned many others; they were actually really cool experiments.). But ultimately I moved away from this area--and I think many others did too--because it was less a theoretical issue that could be tested and that helped us understand actual data, and more of a "world system theory", that while interesting, didn't lead to testable biological experiments (except for mine, of course...).

Finally, I want to bring up a conflict Collins didn't raise: the neutralist-selectionist debate (at which point, some of the old hands are screaming, "Why? Why?...."). For a very brief overview of what the crux of the argument was about, see here. But one of the reasons the argument went on for so long--and has been resurrected with the advent of every new type of data--is that the data were conflicting: sometimes the data were in favor of neutralism, sometimes selectionism. I bring this up because of what Collins describes here:

But I am not sure that there is a general lack of empirical work in economics. This hits on one of Russ Robert's favourite issues, which is the use of complex statistical techniques to empirically validate theories. Statistics can be as misused as theoretical models. Take the back-and-forth on "more guns, less crime" or the impact of legalised abortion on crime. The debate is now predominantly about data and neither side has conceded. As Roberts usually asks, how many economists have changed their mind on the basis of an empirical study?

Regarding the neutralist-selectionist controversy, the strongest proponents of each school probably haven't backed off (like I said at the outset, we're human), but the field of evolutionary biology has reached a loose consensus that whether genes evolve following a neutral or selectionist model depends on the gene of interest (and the organism). It's context-dependent. So, for the issue of gun control, it's not clear to me why there should necessarily be a single yes or no outcome: some communities will gain by strict gun control laws, while others might not. This variation in reported effects could be real.

I don't want to be pollyanish about this: biologists disagree, sometimes intractably so, even when that's a result of being a stubborn idiot and not driven by data. But there has been some consensus, even if it's loose, on many issues, in large part, because the maximalist positions have been confronted by data.

So I don't know if Gould or Dawkins would admit they were wrong (Gould is gone, and Dawkins is more than capable of speaking for himself), but a lot of other biologists would be willing to admit that they were partially (or completely) wrong on some issues, to the point where I think there is a loose consensus.

Whether that's a good thing though, remains to be seen....

*If you mean idea, then just use that word instead. Same number of characters.

**Ironically, during the whole levels of selection debate, which not only dealt with genic versus individual level selection, but also examined whether other levels of biological organization, such as kin-related groups or social groups, could be plausible targets of selection, the prokaryotes (eubacteria and archaea) were completely ignored, even though, in simple systems in the lab at least, genic selection is viable. But at the time, microbiology wasn't even considered biology by many evolutionary biologists.

More like this

Gould, at one point, did get into a weird saltationist phase where punctuated patterns (to the extent they exist) were due to events that were similar to Goldschmittian 'hopeful monsters' (i.e., mutations of large effect).

This is not correct. It's a myth promulgated by many of Gould's opponents. Gould has repeatedly denied that saltation had anything to do with punctuated equilibrium.

Here's what Gould says in The Structure of Evolutionary Theory (p. 1005).

I did speak extensibelyâoften quite criticallyâabout the reviled work of Richard Goldschmidt, particularly about aspects of his thought that might merit a rehearing. This material has often been confused with punctuated equilibrium by people who miss the crucial level of scaling, and therefore regard all statements about rapidity at any level as necessarily unitary, and necessarily flowing from punctuated equilibrium. In fact, as the long treatment in Chapter 5 of this book should make clear, my interest in Goldschmidt resides in issues bearing little relationship with punctuated equilibrium, but invested instead in developmental questions that prompted my first book, "Ontology and Phylogeny" (Gould, 1977b). The two subjects, after all, are quite separate, and rooted in different scales or rapidityâhopeful monsters in geniune saltation, and punctuated equilibrium in macroevolutionary punctuation (produced by ordinary allopatric speciation). I do strive to avoid the label of homo unius libri.

Even Daniel Dennettâwho wrote the definitive attempt at Gould's character assassinationâhad to admit that he was wrong about Gould on this point (Darwin's Dangerous Idea, p. 289).

Please, let's try and put an end to this false story.

Regarding the neutralist-selectionist controversy, the strongest proponents of each school probably haven't backed off (like I said at the outset, we're human), but the field of evolutionary biology has reached a loose consensus that whether genes evolve following a neutral or selectionist model depends on the gene of interest (and the organism).

The fight between Gould and Dawkins came out of the Gould & Lewontin Spandrels paper. It was decidedly NOT about neutralism vs. adaptationism. It was pluralism vs. adaptationism. The question was whether it was/is right to view everything through an adaptationist perspective or whether one should entertain many different explanations of evolution.

To the extent that we now recognize the importance of non-adaptive change, the pluralists (Gould) won and the adaptationists (Dawkins) lost.

I would argue that the field has incorporated punctuated equililbrium into a standard population genetics/evolutionary framework

More precisely, of the very few cases where we can be reasonably sure we're seeing cladogenesis* in the fossil record, most but not all are punctual; the remaining ones are gradual. This evidently depends on such things as geography -- punctuation is easiest with geographic isolation, and that's difficult to achieve when the habitat of a population is the top layers of the entire tropical Pacific Ocean, for instance.

* The same as speciation under some species concepts; this, not necessarily speciation, is what the debate was about.

By David MarjanoviÄ (not verified) on 25 Mar 2011 #permalink

'It just didn't work once you include epistasis--the fitness of one allele depends on the alleles at different genes that allele is associated with.'

In what sense does it 'not work'? What alternate approach is there that does 'work'?
Marios

By Marios Richards (not verified) on 10 Jun 2011 #permalink