Thursday, March 11, 2010

Review of Creating Scientific Concepts by N. Nersessian

Together with Hyundeuk Cheon, I have finished a review of N. Nersessian's most recent book for Mind. Some readers may be interested. It can be read here.

Comments and suggestions are naturally welcome! (by e-mail or in the thread)



  1. You say: "Similarly, Nersessian throws her lot with the extended and embodied theory of cognition on the grounds that scientific models involve interacting with extracranial objects. In this case too, a more critical stance would have been called for. As extensive discussions in philosophy (Adams & Aizawa, 2001; Rupert, 2004) and psychology (Wilson, 2002) have made clear, the fact that successful cognizing requires constant interactions with extracranial objects does not entail that these are part of scientists’ mental states or processes."

    But results published earlier this week demonstrate the existence of extended cognition, and empirically verify Heidegger's phenomenology to boot! As WIRED's summary puts it: "An empirical test of ideas proposed by Martin Heidegger shows the great German philosopher to be correct: Everyday tools really do become part of ourselves." !!! (via Experimental Philosophy)

    Of course, this is mostly tongue-in-cheek. It's just a happy coincidence that I came across these rather bombastic claims shortly after reading your review.

    Still, I think you're wrong on this particular issue, in two ways. First, I think you're wrong in your judgment of the value of the extended/distributed cognition perspective in light of the relevant literatures. Second, I think you're being uncharitable to Nersessian by implying that she endorses a simple entailment from "successful cognizing requires constant interactions with extracranial objects" to "these are part of scientists’ mental states or processes." Of course, you're right that she doesn't devote much critical attention to this issue. But Nancy's aim isn't to defend the theory of extended/distributed cognition. Rather, she's trying to explain modeling and conceptual innovation, and she finds that adopting the extended/distributed perspective allows for better explanations of the phenomena of interest.

    Well, it's a complicated issue. But I think you're being unfair to be so dismissive simply because she adopts one position in a controversy when you prefer the opposite, when the controversy in question isn't really the point of the book. Surely you could put this in a more qualified way, e.g., pointing out that the issue is controversial and those inclined to the traditional view may find this element of her work unpersuasive. To the uninitiated, you give the false impression that the debate has obviously resolved, more or less, and Nersessian is blind to the outcome.

  2. Thank you for your comment, Matthew J. Brown.
    The article in PLoS you mentioned seems to be interesting although I just read the WIRED summary. However I'm not sure whether it verifies extended mind thesis.

    One of my worry is that you (and the Book) fail to attend differences in embodied/extended/distributed view. Nersessian calls those as the "environmental view". Actually, I'm sympathetic distributed view, so I agree with the usefulness of distributed perspective, especially for science studies.(I'm not sure whether Edoaurd would agree) While the article can support distributed cognition, I do not think it supports the extended mind.

    Furthemore, we did not simply dismmiss Nersessian's position, while we said she need to take more critical stance. Those are different.

    Finally, I heard that she doesn't accept the extended mind thesis!(just now from email)

  3. Hyundeuk - A few things:

    1. Sorry if I wasn't so clear: I didn't mean to suggest that you were dismissive of her position overall, but I do think you're overly dismissive on this particular point. When you reference "Nersessian's uncritical embrace of some influential, but unsatisfactory developments in cognitive science," that seems pretty dismissive. It is far from obvious that the developments are unsatisfactory, and her approach only looks "uncritical" from the assumption that old-fashioned view is the default. How many philosophers talking about cognitive-psychological states feel the need to critically evaluate their claims vis-a-vis behaviorism? Are they being uncritical? If so, so what?

    2. Actually, with respect to extended mind vs. distributed cognition, it's worth noting that the one thing the "environmental view" doesn't seem to encompass is the Clark & Chalmers "extended mind" hypothesis. Looking back at this part of the book (thanks for reminding me of this point), Nersessian clearly embraces (while keeping distinct) the theses of embodied, enculturated, distributed, and situated cognition.

    3. I think the "extended mind" hypothesis is problematic from the distributed cognition perspective, because it assumes that there the standard case where cognition is in the head, and then there is the possibility of extension. One of the main points of the theory is that the idea of purely internal cognition is a mistake from the beginning. As Hutchins says in the intro to Cognition in the Wild, "I hope to show that human cognition... is in a very fundamental sense a cultural and social process." This is also the point of his "alternative history" of cognitive science. Perhaps that is why Nancy rejects "extended mind." But it seems to me that those who accept d-cog but only as an occasional or derivative phenomena either don't understand or substantively reject the theory.

  4. Here's a point of disanalogy between the dismissal of behaviorism and the dismissal of individualistic cognitivism. Cognitivists can account for the effects of experience on the individual's cognizing; behaviorism thus seems not only empirically flawed, but otiose. In contrast, as much as I like some of Hutchins's work, he seems to have no grounds for asserting that human cognition is cultural and social in the first instance. The only remotely plausible explanation of the individual's being affected by social and cultural forces presupposes the existence of a cognitive system in the individual. (The alternative explanations seem to me not be serious scientific explanations: individuals are magically governed by social norms or a way of life or whatnot.) But once the necessary individual cognitive system is presupposed, we have in our ontology all the cognitive systems we need in order to explain the full range of phenomena, including the social ones.

  5. Rob, I find your comments kind of puzzling. Hutchins does provide grounds throughout CiTW for asserting these things. Perhaps you find them unsatisfying? But saying "he seems to have no grounds" or "seem to me not be serious scientific explanations" doesn't really engage with the argument. I don't recall appeals to magic in his book.

    I suspect that some helpful work could be done by distinguishing different ways of using the word "cognitive." If we look at what an individual human brain does, its got to be something like a connectionist network, and it really works nothing like classical cognitive science thought it did. Nevertheless, human beings do manage to do the things that classical cognitive science was concerned with (e.g., symbol processing), and they manage to do it by exploiting artifacts and social relations. You might go the Churchlandian route and identify cognition with what the brain does, and thus become radically revisionist about what cognition is. Or you might do as folks like Rumelhart and Hutchins did, and be a little more conservative about the kinds of things we call cognition, and thus try to figure out how those things work.

  6. Yes, I mean he has no satisfactory grounds. I don't recall his offering any serious scientific explanation of how people come to be able to follow social norms, use artifacts, etc. I'd like to know how those things are supposed to happen absent an integrated organism-bound cognitive system. But once you've got such a cognitive system in play (the one that travels around learning to use different artifacts, acquiring new interpersonal skills, etc.), then it's gratuitous to posit further cognitive systems. More to the present point, once you include those those local systems in your developmental story as a central part of the explanation of how individuals come to follow social norms, use artifacts, etc., you've abandoned the claim that cognition is, in the first instance, social or extended. As I recall CitW, Hutchins doesn't engage with this kind of concern, which is common among folks who emphasize the sociality of cognition. Thus, the exasperated (and snarky) remark about magic.