Field of Science

Inevitable Spandrels on a Biology Blog

Nearly thirty years ago, a paper was published that almost every student of evolutionary science will end up reading at some point in their career. Despite being only eighteen pages long and containing no original research, many people see it as marking something of a revolution in biology. Like many a revolutionary document, it says little of substance that is not completely obvious, but entire books have been written that derive their subject matter directly from it. I speak of Gould and Lewontin's 1979 paper "The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme".

In architectural terminology, a "spandrel" is the space formed where two arches meet each other, or where an arch meets a wall. The title of Gould and Lewontin's paper refers to such structures spaced around the dome in the Basilica di San Marco in Venice which sits on four arches (technically, the Venice structures are pendentives rather than spandrels, but this terminological error is irrelevant to Gould and Lewontin's argument). The iconography decorating the dome radiates towards those four pendentives. An observer completely ignorant of the requirements of architectural stability might feel that their focal significance in the iconography indicates that the pendentives were specifically designed to support the iconography, but this is simply not the case. Rather, the pendentives are simply a side-effect of building a stable dome, and the iconography has been designed to take advantage of their presence rather than vice versa. Gould and Lewontin used the Basilica di San Marco as a metaphor to criticise the over-reliance in evolutionary biology on adaptation as an explanation for characters of organisms. Wasn't it possible, claimed Gould and Lewontin, that at least some features of organisms were not selective traits in their own right, but merely biological "spandrels", architectural side-effects of the development of other traits?

Let me give a couple of examples. One that Gould and Lewontin refer to themselves involves the famous midget arms of Tyrannosaurus rex. Much speculation has taken place on what the function of these relatively tiny appendages could have been - props in raising the animal up from a lying position, grappling hooks in mating, etc. but as pointed out by G & L these "explanations" all rather missed the point. Tyrannosaurus did not evolve its tiny arms de novo, but inherited them from a proportionately longer-armed ancestor. While Tyrannosaurus arms could well have fulfilled any or all of the functions ascribed to them, none of the suggestions actually explained why Tyrannosaurus arms became so small in the first place. This may well have been a consequence of the rest of the animal increasing in size faster than the arms did, in which case the question is not "why did the arms become small", but "why did the rest of the dinosaur get so big?". Gould later brought up another example - giant pandas have an enlarged protruding wrist-bone on their forelimbs that functions as a crude thumb in manipulating the bamboo they eat. They also possess a similar enlarged ankle-bone on their hindlimbs that serves no obvious purpose, and indeed may be something of a nuisance. Gould suggested that the enlarged bone on the hindlimbs was a side-effect of the development of that on the forelimbs, due to the same processes underlying patterning in the development of both sets of limbs.

Unarguable as this all might seem, a certain frustration tends to set in towards the end of a reading of "Spandrels", as one eventually feels the need to rejoin with a "So what's your point, Vanessa?" Gould and Lewontin claimed, probably accurately, that over-atomising organism traits might lead to the failure to recognise linkage between features - looking at the spandrels only without considering the entire dome - and called for a whole-organism approach. But in practice, a certain degree of atomisation is necessary if our understanding is to go anywhere at all. It would seem ridiculous to claim that we cannot explain the evolution of the eye without also understanding the evolution of the big toe. Also, at the risk of stating a truism, we cannot possibly know everything about an organism, otherwise we would not be conducting the research in the first place.

Also, there is no way of distinguishing a priori a trait that has arisen as a result of being selectively advantageous from one that has "piggy-backed" on something else. It could well be argued that selective function for a given trait is a more fruitful base assumption that non-function because that is the basis that leads to further research, while assuming non-function is a bit of a show-stopper. As a result, while the concept of biological "spandrels" has caused a great deal of debate over the years on a theoretical basis, it is arguable whether it has had much direct practical effect. A similar fate befell another related Gould neologism, when Gould and Vrba (1982) coined the term "exaptation" for an organismal trait that had been co-opted for another use, just as the pendentives had been co-opted for use in decoration, in contrast to "adaptations" that had evolved specifically for the function they currently fulfilled. The distinction never caught on because, while it might seem theoretically significant, in practice it was fairly pointless. After all, evolution does not create things entirely from whole cloth but works through the alteration of pre-existing features, so all "adaptations" are at some level "exaptations".

Nevertheless, if the "Spandrels" paper had been so pointless as I've just implied, it wouldn't hold the iconic position in modern biology that it does. And agin like many another revolutionary document, this wasn't so much as a direct result of its own propositions, but of the effect it had on how people saw everything else. Gould and Lewontin were protesting against the assumption they saw in many biology studies that an adaptive function for any given trait must be in there somewhere, and if one could not be found it merely meant that we hadn't looked hard enough. As a result, the corollary assumption arose all too often that if an adaptive explanation could be derived for something, then it must be true. Gould and Lewontin argued that mere plausability was not sufficient support in itself for a proposition, because the possibility always existed that an equally plausible explanation had been overlooked - a logical error Voltaire had parodied in 1759 when he had his Dr. Pangloss in Candide, ou l'Optimisme argue that the nose had developed for the purpose of holding up spectacles. It was the recognition that adaptationist hypotheses were hypotheses that required further investigation as much as any other scientific proposition that was Gould and Lewontin's ultimate legacy to biology.


Gould, S. J., & R. C. Lewontin. 1979. The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proceedings of the Royal Society of London Series B – Biological Sciences 205: 581-598.

Gould, S. J., & E. S. Vrba. 1982. Exaptation; a missing term in the science of form. Paleobiology 8 (1): 4-15.


  1. I didn't realize how important this paper was when I first read it as a sophomore in college, but by the time I read it again as a senior taking a course on the history of evolutionary theory it became very apparent just how important it works.

    In a related note, this paper alone still made me discount the entirety of evolutionary psychology.

  2. This is a really great paper in that it makes you consider what it means for something to be an adaptation. One of my biggest beefs with evolutionary biologists is when they use the word 'adaptation' without any reference to whether or not it contributes to fitness or not. I will rant about that sometime later.

    I really just wanted to point out that the text of this article has been made freely available at

  3. @kevin: Apropos fitness, in the technical sense, my wife used to ask me if this or that article of clothing made her posterior look too big, and all I could think was, "Too big for what?"

  4. One of my most favorite papers ever is David Queller's "The spaniels of St.Marx and the Panglossian paradox: A critique of a rhetorical programme". (PDF here)

    The Gould and Lewontin paper was considered such a tour-de-force of scientific writing that an entire book was released with essays on its writing style.

    Queller's paper is an hilarious review of this volume which attempts to take the advice found in this volume seriously. It's full of wonderful puns (e.g. "candide coating") and is actually a really good discussion of the debate.

    My favorite part is the friendly mocking of Gould's fondness for baseball analogies - Queller describes the development of evolutionary theory using the metaphor of a baseball game.

    I really really recommend reading Queller's paper after readin Gould and Lewontin. It's great stuff and a beautiful satire.


  5. I disagree - I read Queller's review (and Borgia's) when I first read "Spandrels" in my first year as a postgrad about eight years ago (eep). Queller and Borgia were quite right that Gould and Lewontin were both prone to overwrought empty rhetoric, but Queller responds with just more and more empty rhetoric, bucketing it in until it all resembles the forcefeeding of a suffragette (see, I can do it to). This is not good satire - satire presents itself seriously, but Queller is just mugging for the camera.

  6. I should have also noted that Queller's implication that Gould and Lewontin's statements (something that, yes, "Spandrels" is actually a lot lighter on than it thinks it is) are somehow inherently suspect because of Gould and Lewontin's socialist leanings just doesn't wash. Conway Morris levelled the same sneering jibe at Gould in The Crucible of Creation - didn't work there, either.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS