Why exaptation is an unnecessary term in the science of form

December 28, 2015 • 2:30 pm

by Greg Mayer

The most important finding of vertebrate comparative morphology and paleontology is that most of evolution is the gradual, adaptive, modification of pre-existing structures (or, better, pre-existing developmental programs, which result in the structures).  The point about pre-existing structures is very important– the history of evolution is to a great extent the history of making due with what you have– “tinkering”, as François Jacob called it. The current phenotype is where you start, and there is a range of mutationally accessible phenotypes that is determined by the developmental system, the environment, and the genotype. This mutationally accessible region will only rarely include completely new structures.

As mammals, WEIT readers may have wondered why it is so difficult for them to get food to come out of their noses while eating (it usually requires vigorous coughing, laughing, or sneezing), while their pet aquatic turtles can do so effortlessly, small food particles floating gracefully out of their nostrils while they eat. The reason is that mammals have a secondary palate– reach in and tap the roof of your mouth– that’s it, right there. It is a shelf of bone that separates the air passage, which goes from the nostrils to the glottis, from the food passage, which goes from the mouth to the esophagus. In (most) turtles, there is no secondary palate, so the air and food passages are one passage, thus allowing food to exit the nostrils. Where did the secondary palate come from? It is not a new bone(s), but a series of medial processes, off the very same bones found in turtles, that meet in the midline to form a complete shelf. Turtles, having a low basal metabolism, do not need to breathe incessantly, while mammals, with their high metabolism, must generally be breathing and eating all the time– the separated passageways allow mammals to breathe while chewing. The gradual modification of these bones to form a secondary palate (and, somewhat less clearly, the crucially associated soft palate) can be traced in the fossil record. The drawback to the way mammals eat is that when finally you have to swallow the food, the lungs are ventral to (i.e. below) the digestive tract, even though your nostrils are dorsal to (i.e. above) your mouth, and the air and food streams must cross. So, if you sneeze/cough/laugh real hard just as your food is crossing over the air passage, it can be blown into the air passage, and come out your nose.

Mammalian swallowing and breathing is thus based on the same bones, digestive tract, and lungs, arranged in the same basic way, as are found in turtles, and in our common reptilian ancestor. In mammals, the pre-existing structures have been “tinkered” with, in a series of documented steps, to arrive at the current state, which allows for a lot of eating and breathing, which in turn allows for a high metabolic rate (and hence being warm-blooded), although since the lungs remain below the gut, you can choke on your food. (Had engineers designed vertebrates, they never would have crossed the air and food passages.)

What brings all this to mind is Jerry’s mention yesterday of Steve Gould’s concept of “exaptations”, noting that the 7th day of evolution video stated that penguin wings are an”exaptation”, because “not every trait is an adaptation, and they don’t all have a point.” This is surely one of the most unproductive, and, indeed, wrong, ways to look at penguin wings. Penguin wings are in fact adaptively modified pre-existing structures; the earliest known penguins in the fossil record, were flightless swimmers, but not as modified for this as later penguins. The notion that a change in function (flying to swimming, in this case) necessitates a new terminology was argued, quite unsuccessfully, by Gould and Elizabeth Vrba in 1982.

Composite skeleton of Waimanu tuatahi from Slack et al. (2006), via March of the Penguins.
Composite skeleton of Waimanu tuatahi of the Paleocene, one of the earliest penguins, from Slack et al. (2006), via March of the Fossil Penguins.

“Exaptation” is an unnecessary term in the science of form. It confuses more than it clarifies. It ignores the historical component of adaptive evolution (which is curious for Gould, since he frequently, and often correctly, argued for the importance of historical considerations; his animus against natural selection must have overcome his historical instincts in this case).  The great paleontologist George Gaylord Simpson attempted to rehabilitate the term “preadaptation” to cover the not uncommon situation where a pre-existing structure undergoes a change of function; once there is a new function, then “postadaptation” will refine the structure to fit the new function. The morphologist Carl Gans suggested that what he called “excessive construction”– the ability of a feature to perform at least tolerably in circumstances other than the usual ones– could often form the basis for preadaptation. (Gans substituted “protoadaptation” for “preadaptation”, finding the latter term too freighted with unfortunate associations with mutationism, of which Simpson sought to cleanse it, to be used.)

I find preadaptation, shorn of its mutationist connotations, a perfectly serviceable concept. Thus, wings on aquatic birds are a preadaptation for swimming in the water. Behavioral flexibility allows the wings to be used in a new way, which then induces a new selective environment, and postadaptations will then further suit the structure to these new conditions of existence. But even more productive, I think, is the notion of “sequential adaptation”, based on Richard Swann Lull’s notions of primary and secondary adaptations (and related to W.K. Gregory’s related concepts of caenotelic vs. paleotelic and habitus vs. heritage), which provides a much better way of looking at changes of function. (I’ve long attributed the phrase “sequential adaptation” itself to Lull, but after a rereading I can’t find that he used it; it was my term for summarizing his views. Lull, by the way, was Simpson’s doctoral adviser.)

So how do we look at the wings of penguins under this view? Certain dinosaurs’ front legs (they already had front legs, of course– they were preexisting structures) became adapted for flight– that is, there were modifications in the structures that conferred higher fitness on their possessors by virtue of the presence of the modifications.  Are the wings of birds an “exaptation”? No. Wings, qua wings, are an adaptation for flying. The wings of certain birds became adapted for swimming– the “flippers” of penguins. The bones have become solid, more robust, dorsoventrally compressed and knife edged, all of which improve their function for diving (not all of these were fully present in the earliest known fossil penguins). Each of these modifications is an adaptation. In Lull’s terms, the wing is a primary adaptation (for flight), the flipper a secondary adaptation (for swimming). But the wing itself was a secondary adaptation of the primary adaptation of the locomotory forelimb of dinosaurs– and so on back in time. They are a series of sequential adaptations. The error of “exaptation” is to think of traits or features of organisms as unanalysed wholes without a history: penguin flippers are are not merely “flippers”, but a whole suite of features, including many bones, muscles, and behaviors for their use. And the flippers have a preceding history as wings, front legs, and so on ad not-quite-infinitum. If “exaptation” means that later adaptations are based on the pre-existing structures, then the term is vacuous– all adaptations are then exaptations.

A case noted by Simpson as preadaptation involves the predatory behavior of keas (Nestor notabilis), the large alpine parrot of New Zealand. Keas eat sheep (or at least parts of them, since the sheep may survive), by cutting through the skin with their sharp hooked beaks.

Keas are naturally omnivorous, so what we have is an expansion of the dietary range, enabled by the pre-existence of a wicked beak. The beak is an adaptation to wide ranging foraging on and in the ground, and on and in the vegetation (excavating logs and such).  If there arises an innate (as opposed to learned) aspect of sheep feeding, or the bill itself experiences selection for features that enhance the ability to feed on the sheep, there would would then be secondary, or sequential, adaptations. At this point, though, feeding on sheep may well be behavioral flexibility with the available tools (Gans’ “excessive construction”, providing the basis for “protoadaptation”).

We not only don’t need the term “exaptation”, it actually hinders understanding, by suggesting that non-adaptive processes are at work (which was Gould and Vrba’s explicit intention), when in fact we have a series of sequential adaptations.

[And I must give here a strong recommendation to the website March of the Fossil Penguins by Daniel Ksepka, which I discovered while writing this post.]

Gans, C.  1979.  Momentarily excessive construction as the basis for protoadaptation.  Evolution 33:227-233.

Gould, S.J. and E.S. Vrba. 1982. Exaptation- a missing term in the science of form. Paleobiology 8:4-15. pdf

Jacob, F. 1982. The Possible and the Actual. Pantheon Books, New York.

Kirsch J.A.W. and G.C. Mayer. 1998. The platypus is not a rodent: DNA hybridization, amniote phylogeny and the palimpsest theory. Philosophical Transactions of the Royal Society B 353:1221-1237.  pdf

Lull, R.S. 1917. Organic Evolution. Macmillan, New York.  Internet Archive

Simpson, G.G. 1953. The Major Features of Evolution. Columbia University Press, New York.

Slack, K.E., C.M. Jones, T. Ando, G.L. Harrison, R.E. Fordyce, U. Arnason, and D. Penny. 2006. Early penguin fossils, plus mitochondrial genomes, calibrate avian evolution. Molecular Biology and Evolution 23:1144-1155.

39 thoughts on “Why exaptation is an unnecessary term in the science of form

  1. (Gans substituted “protoadaptation” for “preadaptation”, finding the latter term too freighted with unfortunate associations with mutationism, of which Simpson sought to cleanse it, to be used.)

    I find preadaptation, shorn of its mutationist connotations, a perfectly serviceable concept. Thus, wings on aquatic birds are a preadaptation for swimming in the water…

    I get what you’re saying conceptually, but with creationists around looking to quotemine people and misrepresent biology, I can only look on the term “preadaptation” with horror. Talk about grist for the mill! You are practically handing them an easily abused term. Preadapted…sounds like a preplanned function! Like a designer intended the swimming function when it thought up the wing!

    We not only don’t need the term “exaptation”, it actually hinders understanding, by suggesting that non-adaptive processes are at work (which was Gould and Vrba’s explicit intention), when in fact we have a series of sequential adaptations.

    I would say that the same exact logic applies to the term “preadaptation”: it actually hinders understanding by suggesting nonevolutionary processes are at work.

    To end on a positive note, IMO “a series of sequential adaptations” or just “sequential adaptation” is much better. As I said, I have no problem with the concept you’re trying to promote, its just the word choice that alarms me.

    1. Antievolutionists (broader term that includes both Young Earth creationists and their Intelligent Design kissing cousins) will coopt and quote mine science lit regardless of the terminology used, though certain ways of framing things can make things harder for them. Overall, scientists shouldn’t try to parse their language to make them less quote-mine friendly, but should stick to clearly presenting their positions and let chips fall where they may.

      Whether the exapation concept is unnecessary as seen from certain perspectives, it does seem to engage a lot of practical interest by many investigators, as surveyed recently by Pievani & Serrelli 2011. “Exaptation in human evolution: how to test adaptive vs exaptive evolutionary hypotheses.” Journal of Anthropological Sciences 89: 9-23. http://www.isita-org.com/jass/Contents/2011Vol89/e-pub/21757789.pdf

      1. A major part of “clearly presenting their positions” is knowing what sorts of misconceptions, willful and otherwise, the people you are trying to communicate with bring to the table.

        I think “preadaptation” is a horrible term in that sense. The very first thing it made me think of is of “planning”, of putting something into place in anticipation of some future need. With or without the theistic point of view, this is exactly the opposite of the concept that one would like to convey here. And if someone is religious, this term practically invites them to misunderstand what you are saying.

        It’s an important concept, that evolution tinkers with what exists and that at each point along the way what exists both provides specific avenues for further tinkering and also erects certain barriers. It’s such an important concept that I think it is worth spending considerable time and effort coming up with language that makes wrong ideas about this process less mentally available.

        1. I’d agree that preadaptation is a loaded term, in suggesting a foresight that didn’t apply. Just to put my own perspective clearly, I regard exaptation and spandrel as a conceptually useful spectrum extending outward from the directly adaptive. A feature (anatomical or genetic) that takes on a new role as a result of changes in environment or competitive arms races (surface keratin integuments in maniraptoran dinosaurs ending up being insulatory, aerodynamic airfoils, and sexual displays, for example) puts the point mutation increments involved in those adaptive chains in a broader context. Any one of those three new feather fuctions may be thought of as being spandrels, in the sense that the flying or sexual selection pressures that sync them into fresh adaptive tracks may not have been subject to strong select pressures for some time after the features arose neutrally. (Since we have no time machine to go back and check, hard to resolve of course).

          The fish antifreeze protein case could be another instance where exaptation broadens how we consider the mutation. The adaptive spread of fish with that mutation would not suggest that the initial mutation variant was preserved because of its temperature properties. The fish carrying them could have swam along for a long while before the exaptation led to new niche invasions that then helped preserve and spread the variant more directly adaptively.

          Just for meringue, my working definition for religion is thoroughly Gouldian: “a neotonous spandrel that is sustained as a Scorched Earth Defense” (juvenile cognitive features that spawned as a byproduct an Agent in the Sky way of looking at things, and which persisted because brains that couldn’t come up with imagintive fictions like that probably did less well in the cultural competition department, despite the downsides of dogmatism).

        2. I agree that “preadaptation” will lead to much confusion and misuse. I don’t think it should be promoted as a valid biological term. Otherwise this is a great post. And thanks for the link to the penguin page!

      2. Normally I would agree with you, however since Greg’s entire article is about word choice, I think its legit to bring up science communication-related pros and cons of choosing one set of wording over another.

        This is specifically why I included my second quote. If he’s going to use the ‘hinders (public) understanding’ argument about exaptation, then its fair to argue or question whether “preadaptation” may suffer from the same problem. I believe it does.

  2. It’s appropriate to note that the diverse alcids [murres, puffins, auks, auklets, etc] swim in the same manner as penguins, while retaining flight [save the Great Auk, the lamented ‘penguin’ of the north Atlantic].

    Of course for birds to ‘fly’ in liquids of very different densities, their wings are optimized for neither flight or swimming. Their wings lack the swimming-specific characters of modern penguins, while being rather too small for optimal flight. Quite comical to see the wind-milling flight of auklets.

  3. There is a widespread urban legend here in NZ that Kea target sheep kidneys. They don’t, but the fact that sheep kidneys are within the general area of attack, the back, means they must get hit from time to time. Kea also deal to cars with the same beak – they can do thousands of dollars worth of damage to rubber seals, aerials etc and will also hop into the car if a window is left open. South Island rural sight-seeing stops often have signs up warning tourists of the problem.

    1. As a touchy-feely human, I was horrified to see the Kea tearing into the poor sheep. The bird seems so…harmless!

  4. You could also point out that the dinosaurs’ forelimb evolved from a fish’s fin, so that the appendage has returned to its original function of directing movement in water (though propulsion isn’t a function of a pectoral fin).

  5. I think the quote that “not every trait is an adaptation, and they don’t all have a point.” is a definition of spandrels, and s not about exaptation. In any case, I cannot find the quote in the previous posting from Jerry.

    I am not at present on board with this criticism of the term ‘exaptation’. I will try hard to be persuaded as I read this again, and the readers’ comments.
    For now, let me point out a well known issue with the term ‘preadapation’. I have read on more than one occasion that the problem with that term is that it implies that structures evolve with some foresight about their future use.

    1. The post is not a critique of Jerry: the quote “not every trait is an adaptation, and they don’t all have a point” is from the 7th day of evolution video. It is in the video that penguin wings are said to be exaptations, followed by the quote.


  6. “Preadaptation” and “excessive construction” seem to correspond to what David Deutsch calls “reach”: the range of problems a given solution is able to address.

    His prime example of a solution with reach is the genetic code, which evolved to aid the reproduction of the simplest protocells, but turns out to be able to specify the construction of multicellular forms of (as far as we can tell) arbitrary complexity.

  7. I stile beleave that Goulds use of exadaptation is a very explicative term which helps in clarifying evolution and that term is better and diferente than preadaptation. As Darwin stated that sutures in mammalian skulls are heldfull in birth, but it is something young birds and reptiles also have bacause of laws of growth but do not need. A nonadaptation is coopted for a current use, mammalian birth. Exadaptation is a feature coopted for a current utility following an origin for a different function or no function at all. Wings where not made for flying, wings where feathery appendages functioning for insulation, that in some birds they function for flight or swimming. We can say flight is an exadaptation. I think that Gould laays out his views very well in his book The Structure of Evolutionary Theory. especially pag 1233.

    1. I agree with you and many others here that ‘preadaptation’ reeks confusingly of ‘predestination’ and should be avoided.
      I don’t like the term exaptation, confusing too, why could we not use the word co-option (particularly if the earlier function is still used)or even adaptive co-option? It is the term I use most in describing the process.
      I note you use ‘coopted’ more than once, it pops up somehow naturally 🙂

      1. Yes, “coopted” is a good term because it more clearly points out what the mechanics of natural selection are doing: taking something that already exists and applying it to a new problem. It has the connotation of short-termism that such a concept needs, making the most of what’s there, which is leagues better than some implied pre-planning.

      2. I’ve generally thought of the term ‘co-option’ as synonymous with exaptation when used to discuss evolutionary biology, and it has the advantage of being a much more familiar term for laypeople.

        It has a little sense of active-mode intentionality about it (who is doing the co-opting, and why?), but not much, and in some cases that intentionality may actually be legit. Who is doing the co-opting in the case of a puffin diving into the water? The puffin is. Why? Because using wings to swim helps them feed. It may be a less useful term, however, when you talk about co-option in bacteria or something like that. ‘Sequential adaptation’ avoid these pitfalls and avoids the baggage of implied preplanning that ‘preadaptation’ carries. The downside is its less understandable to laypeople than co-option.

        On the whole, my personal choice would be ‘sequential adaptation’ or ‘co-option’ being about equally valuable in terms of clear communication of the concept. They have different pros and cons; the latter is slightly more intentional than the former, while the former is less familiar to laypeople. But they are both IMO better than exaptation, and far and away better than preadaptation, which I would frankly rank worst of the bunch in terms of clearly communicating the concept to the public.

      3. I see that the debate is over the words to describe the idea, and then each person can use the word they like as long as they define it. I personally like exadaptation.

  8. If there is a problem with saying that ‘all adaptations are then exaptations’, then there would equally be a problem with saying that ‘all adaptations are preadaptations’, right? The problem I sense with both statements is that it implies that all past and present adaptations were, or are, slated to be modified and re-purposed in their respective futures.

  9. I would not reject ‘exaptation’ so definitively. I don’t see that the term excludes adaptive selection and, contrary to ‘sequential selection’, it keeps the notion of change of function – e.g. flight certainly was not the primary function of the feathers. And, by the way, ‘sequential selection’ is practically synonym with ‘evolution’…
    As for ‘preadaptation’ I agree with Mark and other commenters: it has a taint of teleology.

  10. Preadaptation has the risk of making people think of evolution in teleological terms, that is, to believe that evolution anticipates selective pressures that will occur in the future.

  11. Very interesting, and a new perspective to me.

    I wouldn’t normally correct people, but given this is about terminology – it’s ‘make do’, not *make due.

  12. Do you think the term “preadaptation” might also confuse more than clarify? To me, preadaptation sounds suspiciously like a teleological statement, and I can imagine creationists running with it as proof that evolution requires some sort of intervention in order to explain life. Exaptation, for all its limitations, seems to me to convey the reuse inherent in evolutionary processes.

  13. I am predisposed to agree with Dr Mayer. (Not ‘exposed’ to agree with him, because that doesn’t even make sense.)

    Both terms, in their intended usage as descriptive shorthand, are pretty much synonymous. Apparently Simpson attempted to clarify preceding usage and reclaim a term that might potentially have been misunderstood or misused. (Despite the omnipresent caveats, I don’t know of any instances of the term ‘preadaptation’ being used in a ‘gotcha!’ by creationists… only by antiadaptationists.)

    Gould & Vrba, on the other hand, erected a strawman and didn’t even cite Simpson in their paper. ‘Exaptation’ only makes sense in a terminology that uses ‘aptation’ as its core concept, which has been found unnecessary.

    1. I’m not entirely sure what the etymology of “exaptation” is. I avoid using it mostly because it’s such an ugly neologism, with none of the clarity of, say, “co-opting” or the like.

  14. “not every trait is an adaptation, and they don’t all have a point.”

    Although this comment may apply to some traits (as JAC suggested this is one possible interpretation of nipples on mammalian males) anyone who has watched film of penguins swimming (or been lucky enough to witness it directly in the wild)could not possibly conclude that their wings don’t have a point!

    Great post, Greg.

  15. Quite the contrary, adaptation is an illusion. In the end survival is always caused by exptatition. Imagine a bacteria under stress. It has been shown this sometimes results in active increase of mutation. Adaptation, yes but only made possible since the possibility to speed up mutation rate was already present in the first place. I believe it is a crucial idea in understanding evolution.

  16. I agree that it seems plausible that penguin wings are or are composed of adaptations for their lifestyle. However, it seems to me from what I remember about exaptations is that the notion still has use, even if (like adaptation) it is difficult to knowhen it applies.

    The way I understood it from my biology courses and reading since it is an organ (or I suppose part of one) whose *current use* within an organism is not the one for which it was selected for. For example, perhaps the famous “panda’s thumb” started that way. (I remember looking all of this up years after my last biology class because I was concerned about people using software in non-designed ways and needed a very roughly cognate term and borrowed this one.)

    This is different from a spandrel, which is a part of an organism which has no use at all presently. (Think the web between toes in humans, perhaps.)

  17. Exaptation, preadaptation, sequential adaptation seem to be all more or less like “critical thinking”, that is meaning whatever the writer has in mind. I have never seen criticial thinking distinguished in any reliable way from good old thinking. Likewise, there are adaptations, and every one of them will have its own unique combination of history behind it. The vagueness of “exaptation” applies with equal force to “preadaptation”, and the others. The clear contrast is with a spandrel, which is a structure that arises as a by-product of other (selected-for) processes. Not an adaptation of any kind.

Leave a Reply