More Recent Comments

Saturday, October 06, 2007

The Spandrels of San Marco and the Panglossian Paradigm

This week's citation classic on The Evilutionary Biologist is "The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme" by S.J. Gould and R.C. Lewontin [This Week's Citation Classic].

This is John Dennehy's best choice by far. It's a classic paper and everyone interested in evolution must read it carefully. Whether you agree with Gould & Lewontin or not, you can't participate in the debate unless you've read and understand this paper. I'm pleased that John appreciates it, although I'm a little upset over some of the things he says about Gould. Clearly, he needs some remedial indoctrination re-education ....

[There's a link to an online version of the paper from John's article so nobody has any excuse not to read it.]

13 comments :

Anonymous said...

What? Is traffic slow? :)

Cheers,

Divalent
(not gonna get sucked into this again :)

Anonymous said...

'Spandrels of San Marco' is indeed a classic that has been good for evolutionary science. While offering a touchstone critique of Panglossianism and Just-So Stories it also promoted the rise of what I call Gouldian Fables. (Hairlessness through global neoteny, phylogenetic inertia in kiwi eggs, giant robust australopithecines, origin of language as side effect etc.) In other words, nonadaptive evolutionary scenarios purporting to be more rigorously scientific than adaptive alternatives - with the claim that they ought to be accepted as the default explanation, even if they are equally untested.

As a discussion of non-adaptative evolutionary phenomena, Rupert Riedl's book Order in Living Organisms is much more sophisticated than Gould.

Tupaia

John Dennehy said...

I figured you'd like this one.

SJG, say what you want about the man, but he was a Yankees fan! He couldn't have been much of a scientist if that was the case...

Andrew said...

This should have been required reading in BIO150.

Larry Moran said...

anonymous says,

... it also promoted the rise of what I call Gouldian Fables. (Hairlessness through global neoteny, phylogenetic inertia in kiwi eggs, giant robust australopithecines, origin of language as side effect etc.) In other words, nonadaptive evolutionary scenarios purporting to be more rigorously scientific than adaptive alternatives ....

You won't find very many people who make that claim. For the most part, the pluralists just offer alternative explanations to offset the adaptationist bias. They don't usually claim that their non-adaptationist explanations are any more scientifically rigorous than the adaptationist ones. That's the point.

In some cases, neoteny for example, there are people (I am one) who think the data favors the non-adaptationist explanation but that's not the same thing as advocating that all non-adaptationist explanations are superior to adaptationist ones.

... - with the claim that they ought to be accepted as the default explanation, even if they are equally untested.

That's a different point. There is legitimate debate over the best way to approach evolutionary problems. Adaptationists argue that the default explanation is adpatation and the onus is on the pluralists to disprove the adaptationist story before postulating any other explanation.

The pluralists, on the other hand, argue that the default explanation should be neutral with respect to adaptation—in other words other mechanisms should not be ruled out a priori when forming a hypothesis. According to this methodolgy, if you're going to propose an adaptationist explanation then the onus is on you to demonstrate that there's something that confers significant benefit on the population.

A. Vargas said...

Adaptationists usually point to the adaptive value of a trait to mean that trait has evolved by selection for that function.
What they mean is that function, or the "selective advantage", was CAUSAL in the origin of this adaptation by providing directionality, accumulating several genes with similar effects.

Adaptationists simply don't live up to the standard of the evidence implied by their own thinking. They point out the adaptive value here and there as if it where in itself an evolutionary explanation of the origin of the trait,. Many times, this is not the case. The adaptive value can be a mere consequence, not cause, of the abrupt origin of a trait. Exaptation shows how selection for current function was NOT involved in the origin of an adaptation.

This makes it simply stupid to go around thinking that natural selection has been crucial to the evolution of every adaptation, as if adaptation is in itself evidence for selection. Bad, bad stuff.

Anonymous said...

Larry Moran: "the default explanation should be neutral with respect to adaptation—in other words other mechanisms should not be ruled out a priori ... the onus is on you to demonstrate that there's something that confers significant benefit"

I have no substantive disagreement with that.

I have no problem with presenting reasonable hypotheses of any kind, as long as their testing status is made clear. Even better, alternatives are also presented, along with their respective predictions. Better still, the researcher is candid about his or her theoretical bias - adaptationist, structuralist etc.

Sanders: ""selective advantage", was CAUSAL in the origin of this adaptation ... accumulating several genes with similar effects. ... Many times, this is not the case."

Which means that it sometimes IS the case. And over the entire history of the biota, that's a lot of cases.

Tupaia

A. Vargas said...

That's right! I'm giving you the chance, if you provide the evidence. We know for SURE that it's not the only thing, so it is silly to assum it is, like adaptationists do.

However, to be completely honest, I must say that I seriously suspecting that all this idea of directional selection perfecting a function exists more in the adaptationists imagination than anywhere else...

Anonymous said...

Of course, Sanders, you also have an oddly strained concept of adaptation via selection, disallowing cases studies brought up already by myself and others. Gene duplication? Dismissed. Gene duplication and modification? Probably the same response. Several genes acting on a functionally integrated trait complex? Not allowable because of a eccentric, self-styled definition of micro changes in a single metric measurement of a single trait, rather than an developmentally interactive functional complex, no matter how modular or integrated. Preadaptation? Exaptation? Can't be an adaptation shaped by selection then. Multiple physiological changes acting on the same function? No. Developmentally plastic reaction norm? Dismissed. Two or three genes rather than scores of genes? Nearly a single mutation. A significant effect? Maybe that's a macromutation. What a bizarre hermetically self-justifying view. You're demanding a case of adaptation through selection that somehow is independent of developmental and phylogenetic histories. A total straw man of adaptation through natural selection.

Have you in fact read or heard of Tingergen's four questions? What about SJ Gould's triangle of causal factors? Does that ring a bell?

The good news is that it doesn't matter what either of us think. In the coming years and decades, the empirical database of the relationship of genetic change to change in form and function will be vastly greater, and much of these questions will be settled.

Tupaia

A. Vargas said...

What we are tired of is adaptationists just pointing to anything that is an adaptation and blurting "natural selection". This premature "explanation" seems so satisfactory to the adaptationists they forget they must have empirical rootings for such an affirmations (and what are they, exactly).

"You're demanding a case of adaptation through selection that somehow is independent of developmental and phylogenetic histories. A total straw man of adaptation through natural selection"

I am asking for something pretty simple thing, no matter what tupaia may think. For instance, we can greatly exaggerate a trait by artificial selection, no matter what the phylogenetic and developmental constraints. I just want an example of how directional selection has had a similar causal effect in the origin of ANY adaptation.
The smart person will realize I am NOT asking for the impossible. If the effect of selection is indeed powerful in adaptaion, "the essence of the adaptive process" and blablabla, it should be easy to find tevidence where selection, and not developmental or phylogentic history, is the main explanation behind the origin of an adaptation. Does this sound "eccentric"? It's just "what for" according to selective theory itself.

Many adaptationists believe that phylogenetic and developmental history have minimal roles in explaining the origin of adaptations. Well, at least those "adaptationists" that still sound genuine. Some people think they can just call themselves "adaptationists" and then flip-fop all the time.

Anonymous said...

Try this on for size, Sanders - I'll meet you partway.

One possible evolutionary process: genetic assimilation of macrophenocopies.

1) Environmental change/insult (abiotic parameter, toxin, pathogen etc.) induces major disruption in developmental program.
2) Facilitatory developmental processes, including somatic selection (selection!), accommodate changes, resulting in viable organism.
3) If the new phenotype is fitter than the non-induced type, selective forces (conventional selection!) canalize the developmental response so it appears even in the absence of inducing environmental factor.
4) Later, modifying genes appear that reduce the costs of the phenotypic novelty and are positively selected (more conventional selection!).

See, it's almost Goldschmidtian, yet wholly Darwinian. And selection is crucial at several steps in the process.

Tupaia

A. Vargas said...

do you think that's my "paradigm"? Nope.

A. Vargas said...

plus, it's BS at many points (somatic seection my ass...plus "if it has greater fitness" is NOT a requisitie, something taht highlights the fact selection is NOT so important.

Let's just stick with field examples, OK? I really don't care for your particular theoretical framework if they are not attached to real cases.