How Do Biologists Support Functional Claims?

The press has reported extensively on a new (and frickin’ cool) paper out in Science documenting gears on the back legs of a little insect (Issus). Ed Yong over at National Geographic has a piece on this work, including embedding the little movie of the gears in action, which is, as I say, 100% totally and completely cool.

594px-Issus.coleoptratus.nymph.with.Dryininae.larvaBecause the press has covered this work, I’m not going to talk about the work per se.

Rather, I want to address an issue which I’ve discussed from time to time, which is how biologists draw inferences about function. What kinds of evidence do they use when they make a claim that a particular structure is for something?

I’m not interested in this for purely philosophical reasons regarding the epistemic commitments of biologists as a general matter.

I’m interested in this because of critiques regarding functional claims made in the context of evolutionary psychology. In particular, critics have made various suggestions about what sort of evidence is used by biologists to make claims regarding function, suggesting that similar sorts of evidence ought to be used when evolutionary psychologists make claims about function.

So, how do Burrows and Sutton, the authors of the paper in question – “Interacting Gears Synchronize Propulsive Leg Movements in a Jumping Insect” – provide evidence for their claim?

First of all, note the functional claim in the title. The crucial word there is “synchronize.” The function of the bits of the body in question – the little gears – are to synchronize the legs. If the legs weren’t in synchrony, then the insect’s jump would be asymmetric, and the insect would tilt to the left or right instead of jumping straight ahead. (I watched some videos of frogs jumping, and it looks to me that frogs use asymmetrical leg motions to (intentionally?) get some yaw into their jumps. Anyone know anything about that?)

Second, note the evidence that they present. Figure 1(D) is a picture, showing the interlocking gears. Figure 1(E) is a diagram of the gears. The Supplemental Material includes a series of movies. These materials also indicate the methods used: mostly making the movies and taking the photographs. The authors are not shy about their conclusions, writing:

The mechanical gears in Issus enhance the synchrony between leg movements to the level of microseconds…The gears in Issus… demonstrate that mechanisms previously thought only to be used in manmade machines have evolved in nature. Nymphal planthoppers have interacting gears that play an essential functional role in a natural behavior.

To be clear, then, the authors infer the function – synchronization of the legs, which in turn allows for jumping without any yaw – from the images. The shape of the structures they found, together with the behavior they observe, affords the inference regarding function. This is because those particular shapes – interlocking gears – are complex elements of the phenotype that are improbably well designed to execute the function in question, synchronization. As George Williams articulated so well in Adaptation and Natural Selection, functional complexity is the hallmark of adaptation. It’s worth quoting Williams, who specifically discussed cases like the one in the present paper, in which there is an analog with human artifacts:

A frequently helpful but not infallible rule is to recognize adaptation in organic systems that show a clear analogy with human implements. There are convincing analogies between bird wings and airship  wings, between bridge suspensions and skeletal suspensions, between the vascularization of a leaf and  the water supply of a city. In all such examples, conscious human goals have an analogy in the biological  goal of survival, and similar problems are often resolved by similar mechanisms. Such analogies may  forcefully occur to a physiologist at the beginning of  an investigation of a structure or process and provide  a continuing source of fruitful hypotheses. At other  times the purpose of a mechanism may not be apparent initially, and the search for the goal becomes  a motivation for further study. Adaptation is assumed  in such cases, not on the basis of a demonstrable  appropriateness of the means to the end but on the indirect evidence of complexity and constancy.

In short, both Williams and the present authors infer function from form. But note that there are important dogs that don’t bark here. The authors don’t provide any data regarding the genetics of the structures in question. They don’t provide any data regarding heritability. There aren’t any experiments. And that’s just fine. The point is that the shapes are themselves the evidence. The rows of teeth, interlocking as they do, provide compelling evidence for the function of synchronization.

Is there room for doubt? Of course. I find myself persuaded, but I suppose it’s possible that these teeth arose as a side-effect of some other aspect of the phenotype. That possibility seems wildly unlikely to me, but I suppose it could be. Or maybe the gears are actually some sort of sense organ. Again, that seems wildly improbably to me, but I suppose there’s some room for doubt. (Similarly, the emotion of jealously might be a side-effect of some other computational system, rather than a mechanism to deter infidelity. Could be.)

Is the claim that these structures are for synchronizing leg movements a “Just So” story? Does a structure’s shape, in itself, allow you to increase your confidence in a specific functional hypothesis? If you’re working in the realm of behavior, what is the equivalent of measuring and documenting the shapes of structures?

How do (these) biologists support their functional claims?

17. September 2013 by kurzbanepblog
Categories: Blog | 21 comments

Comments (21)

  1. Yeah, but what about reproductive success? Do individuals with more gear teeth have more offspring? Otherwise it’s all meaningless.

  2. To be fair, I’m not sure that the target of these criticisms of EP is really about the issue of inferring function from form pe se. I suspect most such critics would not have a problem with the idea that the function of jealousy is to defer infidelity, for example. Instead, what gets their panties in a bunch is the inference (or even the hypothesis, it seems) that this function is an ADAPTIVE function: i.e., that jealousy (e.g.) is and operates as it does because it is an adaptation designed by natural selection, the recipe for which is encoded in our genes. Their point, I think, is that there exist many alternative explanations for why jealousy works the way it does other than being an adaptive product of evolution: for example, we might all just “learn” how and when to be jealous from our parents or from our “culture,” etc.

    I think it’s also fair to acknowledge that this problem of empirically ruling out alternative, non-adaptationist hypotheses is arguably the single most challenging aspect of doing evolutionary psychology well, because humans’ abilities for creative problem-solving, cultural transmission, and so forth really do often provide fodder for plausible alternative explanations that aren’t always easy to falsify empirically.

    In contrast, it seems to me that there really just aren’t very many plausible explanations for bug gears. It isn’t plausible that the bugs “learn” to have them, or build them because they figure out how useful they would be or get the idea from their local bug culture. It seems certain that the recipe for building bug gears is encoded in the species’ genome. And it’s really hard to imagine what other functions such gears could possibly provide other than the one inferred by the authors, or even how it could be a byproduct of another adaptation. So if there is only one plausible explanation on the table, maybe a photograph or video is sufficient to make the case, at least until someone comes up with a viable alternative.

    Your Friendly Devil’s Advocate

    • FDA, IMHO, this is a really edifying reply to the blogpost…I would only add that we should note the qualified nature of the Williams quote…

    • Thanks for the comments, Lee. I’m not quite sure what
      distinction you’re drawing between “function” and “adaptive function.” Can you
      give me a sense, or an example, of a (biological) function that is not adaptive? As for the part about jealously, my point here is not about that issue as much as the
      broader question of what counts as evidence of function in papers such as this
      one. As I say in the post, it’s not impossible that jealousy is a side-effect.

      • OK, how about this. I think all parties would agree that the function of tooth-brushing is to promote healthy teeth and gums (as well as, perhaps, to enjoy and share minty-fresh breath). Now suppose an evolutionary psychologist were to propose that the reason we engage in this behavior (and obtain the consequent benefits) is that humans possess an evolved toothbrushing mechanism or module, the recipe for building which is encoded in the human genome; specifically, he or she reasons, toothbrushing genes were favored over non-toothbrushing genes in ancestral environments because people with healthy teeth and gums were better able to eat a variety of foods and, buttressed further by the minty-fresh breath, were more attractive to members of the opposite sex and thus more successful in attracting quality mates. The theory would surely (and deservedly, I think we’d all agree) come under heavy criticism, but not because the theorist failed to adequately demonstrate that toothbrushing has those particular functions. Instead, the target of the criticism would be the claim regarding the existence of an evolved mechanism, designed by natural selection, the recipe for which is encoded in the human genome. They would (reasonably) ask what the evidence was for this claim, and point to plausible alternative hypotheses related to “learning” and “culture” (i.e., that the practice of toothbrushing emerged over many generations, in some parts of the world, via some kind of cultural evolutionary process, and is maintained today by parents teaching their kids to do it).

        As you well know, of course, you and I are on the same team here. My motivation here is not to defend the critics — although I would side with them in this imaginary toothbrushing-theory scenario — but rather to understand what exactly it is they are objecting to and why.

        • I like the toothbrush example, though I think it can be countered quite easily. Toothbrushing is a behavior, not a psychological system, and evolutionary psychologists apply adaptationist reasoning to the latter. If you were to frame toothbrushing as a system (e.g. it takes parental instruction and repeated goading during childhood as input and produces toothbrushing behavior as output), it’s functionality as a “dental health maintenance” system no longer seems so compelling. On the other hand, when jealousy is framed in terms of its architecture (e.g. it takes a highly targeted set of social and contextual cues as input and produces a variety of strategic and sexually dimorphic behaviors as output) it looks far more compelling from an adaptationist standpoint. The point that is often lost on critics of evo psych is that when you zoom in to the level of psychological systems, distinctions between evolutionary and nonevolutionary explanations cease to make sense.

          • …OH!…I get it now…thanks…

          • To be clear, my example wasn’t intended to present an argument that needed to be “countered.” I’m just trying to help figure out what exactly is at the root of (some) biologists’ criticisms of EP. I interpreted Rob’s original post to imply that they are accusing EP of inferring function from form without insufficient evidence, in which case the bug-gears example suggests that they are being hypocritical. I’m merely suggesting that I don’t think facile-function-inferring is really what they are objecting to, and I tried to illustrate what I think is their main objection (whether a valid objection or not).

            That said, I think you are right that the deeper problem has to do with the conceptualization of psychological mechanisms/systems/modules at a cognitive/psychological level of analysis — the very idea that most clearly distinguishes EP from behavioral ecology and sociobiology. The depth of this problem is, or at least one aspect of it, is illustrated pretty clearly in the blog post to which “James” linked at the bottom of this page; he makes explicit an argument that I think might implicitly lie behind many biologists’ criticisms of EP.

      • The attached Buss paper names a number of topics studied by “evolutionary” psychologists…some of these are “biological functions”, e.g., aggression, jealousy, sex, language…we might agree that, originally, these responses enhanced fitness…however, are all of the uses to which humans can put these responses “adaptive”?…for example, “evolutionary” psychologists often study “mate choice”…is non-random mate choice induced by advertising “adaptive” or is it a function of media manipulation designed to accrue resources by a corporate entity?…most would agree that sex was originally “adaptive” and probably still is in many conditions…, but are paraphilias or fetishes “adaptive”?, even if partners are selected non-randomly…how do we tell non-random “adaptive” functions from non-“adaptive” functions in the human species…humans often cheat & deceive…which of these manifestations of behavior are “adaptive”, which, not?…how do we tease these apart…to take the title of a recent paper in your journal: “Blind dates & mate preferences”…mate preferences occur non-randomly but does mate preference in this condition serve “adaptive” ends?…related, language was almost certainly “adaptive” in the original, but are all of the [non-randomly expressed] uses to which language can be put “adaptive”…how about reasoning [another of Buss’ traits]…probably originally adaptive [unless, say, reasoning came along as a corollary to other brain processes that were adaptive]…are all of the [non-random] uses to which reasoning can be put adaptive?…how do we differentiate “adaptive” from non-randomly expressed but not “adaptive” responses?…for example, when & under what conditions are humans cheating or being cheated or detecting the expectations of the researcher?…when are humans exploiting or being exploited?…when are humans deceiving or being deceived?…when are humans “testing” [Zahavi] or being tested…when are humans “tinkering” [phenotypic plasticity literature] or when are they the objects of “tinkering”…and when, if ever, does any of this activity [including thinking & its byproducts] have a relation to fitness…finally, we haven’t yet figured out why some classes of humans in a society have fewer, some more offspring than others…i.e., how is “adaptive” defined for the human species & “Qui bono?”, if anyone?

        • p.s. …students of behavior, including brain operations, have not come to a consensus yet on whether “play” or “maintenance activities” are or were “adaptive”…in human or non-human animals…sometimes a thought is just a thought…sometimes a game is just a game…sometimes a bath is just a bath…sometimes a signal has no [adaptive?] effect unless it occurs in a multi-component display…are “e”Pers confusing “adaptive” with “effective”…

          • …finally, consider…we might have a mutation that is deleterious to reproductive success [inclusive fitness] that is expressed on the phenotype because it is linked to a gene whose beneficial phenotypic effects far outweigh the deleterious effects of the mutation…here we would need to know something about thresholds of response and differential costs & benefits of expressed traits…

  3. It’s an old refrain, but I’ll say it anyway… part of the problem being discussed is one of language and the shadowy meanings behind words. Saying a particular body part is ‘for’ a particular function implies teleology; it implies a narrative, or a ‘just so story’.

    If we said instead ‘behavioural complexity is the hallmark of adaptation’ then you could work backwards from an organism’s fitness, through intermediate anatomy and biochemistry, to (if it can be demonstrated) whatever variation of genetic components were necessary to enable that behaviour.

    In the case of Issus you could look at their behavior compared to related lesser or ungeared bugs and see if the geared bugs were fitter. Not an easy task, I know, but at least you could work out if the geared legs provided a mechanical advantage, or were merely a preferred sexual characteristic, or maybe mimicry of the sound of a predator bug. Same with humans. Start with behaviour and work backwards – except you have further intermediate stages of individual learning to unpick.

    But at least you minimise any false talk of ‘design’ or ‘ the function of X is…’.

    • …problem is that where behavior is “plastic” or where responses [& their proteins] have been co-opted for a variety of functions [as happens all the time in humans], even to “work backwards” may not say anything reliable about the adaptive [i.e., original] function[s]…further, as one commenter already pointed out, “adaptive” requires us to demonstrate that a response enhances fitness…finally, as FDA says, we, also, need to identify the nature of alternative hypotheses and how many there might be…a careful consideration of Hopi Hoekstra’s work on the digging responses of voles might be instructive to “evolutionary” Psychologists, including, the next steps of demonstrating with gene ontological analyses what the causal pathways are…

      • I never said it would be easy 🙂 and I take the points you make seriously. What I hoped to work around was the type of phrase “adaptive i.e. original function(s)” because it is so easy to fall into ‘adaptations for a purpose’ when natural selection works more like free form sculpture – chipping away the consequences of genetic variation which contribute to less fit behaviour.

  4. …are “evolutionary” [“cognitive”] Psychologists really prepared to say that “airship wings”, “bridge suspensions”, and “the water supply of a city” are evolutionarily convergent to “bird wings”, “skeleton suspensions”, and “the vascularization of a leaf”, respectively?…aren’t these “analogies” representative of the utilities of mechanical designs rather than the adaptive natures of the analogies?…my plumbing system is analogous to “the vascularization of a leaf”, however, it is unlikely to represent a human adaptation…humans have co-opted natural designs for their own benefit, and even the originator’s fitness would not have been enhanced if plumbing resulted from persuasion, coercion, or force…probably, the adaptive significance of a plumbing design can be traced back to the fitness of a tree…

  5. Some thoughts from an ethological perspective at my blog:

  6. Pingback: Biology’s Just-So Stories | Incredulous

  7. “To be clear, then…. The shape of the structures….”
    1. …what “structures” are “Evolutionary” Psychologists referring to when they use the word, “adaptive”…what brain structure is associated w a particular behavior of interest (say, jealousy, or mate choice, or “rape”)…
    2. …surely, you would not claim that the gear structure of Issus is comparable to anything that an “E”P study has identified…
    3. …related, the function of the Issus gear can be MEASURED, it is not presumed…the gear mechanism cum structure is an observable THING…
    4. …surely you would not claim that there is a “toothbrushing” “module”…
    5. …on the other hand, if “E”Pers were to measure, say, spatial memory, they could point to the hippocampus…but we really don’t know how this structure works…
    6. …this is my penultimate comment on this blog, and thank you for permitting me to participate…
    7. …final comment: the phenotype is not located inside of the organism…

    • …1 modification to “final comment” #7: The phenotype UPON WHICH SELECTION ACTS is not located inside of the organism…the phenotype upon which selection acts is exposed to the abiotic and biotic [including, social] environments external to the whole organism…

  8. Pingback: Biology’s Just-So Stories | Skeptic Ink

Skip to toolbar