This year I’ve been running a reading group I’ve called “Effective Interdisciplinarity.” In my less professional moments I have been describing its concern as this: how does one avoid doing shitty interdisciplinary work? By “shitty” I mean to include not only the more obvious truth-y stuff–e.g., accurately representing the research of a non-native discipline–and methodological stuff–e.g., do philosophers of science read IMRD-style papers the same way as native disciplinarians–but also ethical/moral stuff–e.g., do current ways of pursuing interdisciplinary work exacerbate disadvantages faced by particular disciplines or groups of people? As someone with a variety of interests, I want to make sure I’m avoiding as many of these pitfalls as possible.
So far the group has been focusing on the more truth-y and methodological aspects. One problem we ran into was that there doesn’t seem to be a lot out there on the kind of interdisciplinary research you see in philosophy of science–i.e., very low numbers of collaborators, if any at all. Because of this, the great science of team science stuff out there isn’t straightforwardly applicable. We started with a hodge-podge of stuff, including Strober’s Interdisciplinary Conversations, and we’ve now landed on a “by osmosis” strategy: read a work that’s billed as interdisciplinary and try to identify successes/failures and what (may have) caused them. Obviously this strategy won’t allow for the drawing of strong conclusions–e.g., what we should or should not do in a particular situation–but at the very least it will get us thinking.
We landed on reading Sapolsky’s Behave, and I’ve found myself struck by one feature of his approach: how he frames the question to be answered. Instead of the “usual” framing, wherein a variety of disciplines are expected to answer “why did this behavior occur?” using their own disciplinary tools, he instead frames it in a temporal manner: “what happened during the ___ chunk of time before the behavior that contributed to it?” What he points out–which I tend to believe–is that this inherently involves all of the disciplines in a way that happens much more organically than it does in the usual way.
I’m still going back-and-forth on the “real” significance of this, but here are some thoughts on why it appears to be. Apologies in advance for their lack of clarity.
Asking and answering the natural “why” question
The first benefit seems to come from asking and attempting to answer the more “natural” question: why did such-and-such behavior occur? In particular, respecting the imprecision of the question seems important. Sapolsky’s discussion of his re-framing emphasizes that it keeps one from thinking in terms of “disciplinary buckets” so that one doesn’t come away thinking that “a behavior is caused by a gene, or a hormone, a childhood trauma” (8). Though I’m not sure his argument that this is significantly different from the “usual” way–wherein he points to several famous researches who seem to have suffered from bucket-blinders–is that conclusive, he’s certainly right that some folks think this way and that their work is deeply affected. (To be clear, though, I also don’t think he needs a definitive argument, so I don’t see this as any kind of failing on his part.) Extrapolating from what he says, his suggestion seems to be that re-framing in this way also allows one to bring in all of these various disciplinary perspectives/methods/etc. without unnecessarily committing to any one of them as ultimate.
There are couple ways of thinking about this. One is in terms of what is being embraced with this strategy. The point just made was that the re-framing allows all of the disciplines to have their say without stepping on one another’s toes. But the specific (temporal) way he approaches the problem–what happened just before the behavior? and before that? and before that? etc.–suggests another benefit could come from this: it makes it easier to bring everything together in a way that minimizes conflict. This is because of the natural way in which roles get assigned: duh, we say, of course we don’t need a game-theoretic model of the interaction if all we’re trying to do is explain what happened just before the behavior. Broadly, it gives us a way to determine what is relevant and what isn’t at a particular moment in the explanation. And in bringing everything together in a more natural way, it seems reasonable to take the overall argument to be stronger irrespective of any errors in the nitty-gritty. Moreover, it makes it easier to integrate new discoveries in the nitty-gritty since, in a sense, you’ve set the research agenda to be so broad that just about anyone should be able to contribute.
Another way to think about it is in terms of what is being avoided. He’s made it very difficult for himself to avoid criticism. Since he’s hitting on so many different perspectives, the strategy of defining away criticism doesn’t work. That is he can’t claim, for instance, that he is merely concerned with the social-interactional aspects of altruism, thank-you-very-much, so the specifics of the projections between the frontal cortex and the amygdala are irrelevant.; these specifics must be relevant at some point in the explanation because of its scope. And so it should be easy to identify errors, unlike in that talk we’ve all been to where we have a vague feel that something is amiss but we can’t quite pinpoint it. Being so open to correction seems to have a huge benefit, however: whatever else is the case, his work is not very open to the deep intellectual insult of being misguided. Wrong? Too ambitious? Sure. But misguided? Unlikely.
This strikes me as a subtle, if profound, way of re-framing debates. For instance, this seems analogous to the way Maddy has approached the “logical must.” A common approach, of course, is to understand the logical must by bringing out the buckets–say, metaphysics and epistemology–and posing questions and answers (or portions thereof) that cleanly drop into one or the other. Instead, Maddy asks and answers a more natural question (The Logical Must, 2): “why is it that ‘if it’s red or green, and it’s not red, then it’s green’?, and what’s the added force of the ‘must’ in ‘it must be green’?” The full answer, presumably, would trace from socio-cultural and developmental factors, through basic features of cognition and perception, to features of the world we inhabit. To be sure, such a re-framing could be said to “miss the point” of the metaphysicians’ and epistemologists’ questions. However, it seems to me, at least, to evade characterization as misguided in the same way Sapolsky’s argument does.
Model sobriety: dynamical rather than kinematical approach
The second benefit seems to show up in the attendant approach to the biology of behavior, namely, a sober appreciation for the complexity of the link between behavior and biology. I’ve been thinking of this by a rough analogy with dynamics and kinematics. I’m wary that it’s rather strained, particularly given the regularity with which other sciences are given the physics-goggles treatment, but at least at the moment I see it as useful.
On the one hand you could take a kinematical approach. Here the aim is to specify the possible states the system can take (e.g., states are regions of activation or patterns of activation) and the possible phenomena to which these states correspond (e.g., behaviors or propensities for behavior). The analogy here is with kinematics, where a phase space specifies all of the possible combinations of position and momentum. This is then hooked up to the observable phenomena of motion, hence, in a sense, reducing the observable phenomena. Over-and-above the usual applied-math caveats that complicate this picture, the problem with carrying the analogy over to the present case is that behaviors just aren’t like position: we have a pretty darned good idea what it means for something to be in a position, but what it means to be happy, anxious, afraid, etc. just isn’t that fixed. Sure, there are some common “behaviors” that more-or-less cleanly map to regions (e.g., temperature regulation or moving your index finger), but in many others it seems we may end up modifying/removing common concepts (e.g., happiness) as we learn more, thus significantly modifying the structure of the “phase space”.
On the other hand, you could take a dynamical approach. Here the aim is, roughly, to explore the communication of / relations among various factors (again, I’m being rather unsophisticated here) that give rise to particular phenomena. This could even allow one to “sneak up” on a more stable kinematics-y understanding without fixing either the phenomenological aspects or the relevant reduction. Essentially, this allows you to more freely wiggle things on either the behavior side or the biology side to learn something new about the other.
I take the latter to be Sapolsky’s strategy (80):
Thus, studying behavior is useful for understanding the nature of the brain–ah, isn’t it interesting that behavior A arises from the coupling of brain regions X and Y. And sometimes studying the brain is useful for understanding the nature of behavior–ah, isn’t it interesting that brain region A is central to both behavior X and behavior Y. For example, to me the most interesting thing about the amygdala is its dual involvement in both aggression and fear; you can’t understand the former without recognizing the relevance of the latter.
I also take the latter to be the strategy taken even in physics, to nail down a good notion of “quantity of motion” which allowed for the kinematics of, e.g., 19th Century.
- Who cares about Haag’s theorem? - April 19, 2024
- The Book Report Problem - March 29, 2024
- Announcing the PSA DEI Caucus Affinity Group for Philosophy of Physics - August 24, 2023