A philosopher in the NY Times, again!

Where O where is habituation?  I mean, after repeated exposure to something, isn’t one’s reaction supposed to get less?  Well, maybe the fact that this article has me depressed rather than angry is a sign of habituation.  In any case, it is a well know philosopher on the real science of mind. 

The real science has nothing to do with fMRI experiments, of course.  And here’s why not:

 it provides little insight into psychological phenomena.  Often the discoveries amount to finding stronger activation in some area of the brain when a psychological phenomenon occurs.  As if it is news that the brain is not dormant during psychological activity!  The reported neuroscience is often descriptive rather than explanatory.

And that’s just false, I say, having skimmed through the first 70 articles Academic Search brought up, besides having some background here.  The “it’s just about brain location” argument has been floating around for about 10 years, and it certainly seems problematic.  Even a cursory glance shows that the claims are often implicitly or explicitly comparative.  And they also invoke some knowledge of the functions of the areas of the brain in question.

You can see both characteristics in a study in Science that we commented on here.**  The studies may provide very rich hypotheses.  For example, a comparison between borderline and non-borderline people looked at the insula, which apparently becomes very active as one senses norms are being violated.  Interestingly, seriously borderline people seemed able to register when they transgressed against someone else, but not when someone transgressed against them.  That’s quite the opposite of what their behavior suggests, and so may indicate a more global way in which they are missing cues. 

A lot of work is being done on the differences between people who have had strokes, do have alzheimer’s, and more.  Here again there are comparisons.  If one wants to find the cause and then the cure of some condition, a comparison between the with and the without conditions seems like a very good idea.

This is not to say there are no just complaints.  There are studies with too few samples, along with, it seems safe to say, very sloppy ones.  And disability theorists may indict a normalizing that goes on with the result that the neuroatypical are seen as inferior.  The latter is very serious. 

Burge starts off talking about science journalism’s reports of brain studies.  There’s certainly a lot to complain about there.   But it isn’t clear that the philosopher has understood the difference between the journalism and the science.


***My apologies to anyone who tried the link before thurs pm and went to an entirely different study on the sense of self.  The studies are equally telling against Burge; you can find the sense of self  here.

22 thoughts on “A philosopher in the NY Times, again!

  1. In my experience, so-called borderline people are very unskillful at reading when they are transgressed against and tend to attribute aggressive intentions when they do not exist.

    Generally, when they “blow it”, they realize it, but the next day.

    Is that what you are referring it?

    I have no idea about the brain studies, but I do have some experience with so-called borderline personalites.

  2. No, YES, Amos, the brain scans suggest just the opposite what you are saying. That can sounds totally wrong, until one focuses on “tend to attribute aggressive intentions when they do not exist”. I think more research is needed, but one thing we might be learning is that their seeming hypersensitivity to insults actually is the result of a sort of scatter bombing they do because they can’t really see what is going on.

  3. Amos, I misread your question. My answer should be “yes, that’s it exactly.” Good for you spotting this, though I hate to think about the experience that gave this to you.

    Sorry for the misreading!

  4. That’s ok.

    New ways of looking at things are helpful.

    It’s hard to keep one’s center when the fits, if that is what they are, occur, but the best way to deal with them is not react to outbursts of anger with anger nor to argue with someone who is “out of control”.
    At the same time, one has to respond, because the person demands a response and a lack of a response just makes things worse. It’s an art which I’m trying to learn, but haven’t learnt yet.

  5. I’m more sympathetic to Burge on this than you.The behavioral evidence trumps the neuroimaging evidence, typically. We know about the functions of the brain, to the extent we do, from behaviour/brain correlations. of course we might – and one day will – have enough evidence about neural functioning that we can be confident from fMRI evidence alone what is going on functionally, and reinterpret the behavioral evidence in that light. But we are in that position with regard to few areas of the brain as yet. Insular activation is a case in point. Yes some people think it is a sign of norm violation detection, but it is safer to say it is a sign of negative affect (and a few other things as well).

  6. Oh, and about the patient comparison data: that tends to be more useful, because there is quite a lot of evidence from lesion studies in particular for double disassociations, and that’s good evidence for function. But Burge’s complaint was about neuroimaging and not neuropsychology.

  7. Burge is certainly not alone in his critique, and I very much doubt if his actual views are as simple as this article (in the NYT!) suggests. Gregory Miller (UIUC Psychology) details, in almost excruciating detail, a variety of conceptual blunders made by neuroscience-enamored psychologists in a lead article to a special section on fMRI in Perspectives on Psychological Science. You can find the article here: http://pps.sagepub.com/content/5/6/716.abstract

  8. Neil, you say, “The behavioral evidence trumps the neuroimaging evidence, typically.” “Typically” is, of course, very vague, but I think that what you are saying is just false. The reason for that is that fMRI researchers are often asking questions that go beyond the behavioral evidence and getting results that support distinctions themselves beyond the behavioral evidence. The study of borderline behavior I referenced is a case in point, as is the other case I cited.

    About the insula: the fact that the insula supports X, Y and Z does not show it doesn’t support X, so I am not sure of the point of your claim. However, the role of the insula in reacting to norm violations was the subject of large and repeated studies of the “trust game” in Montague’s labs. Here they hyperscanned interacting individuals playing a form of betting. The borderline people were looked at in that context. A description of the prior research on norms and the trust game can be found here: http://www.hnl.bcm.tmc.edu/articles/Read/MontagueLohrenz2007.pdf
    The borderline study itself can be found here:

    Click to access King-Casas.pdf

    Another really nice study going way beyond behavioral evidence was conducted in the same labs. They looked at high functioning autistic young people who had undergone intensive behavioral therapy and knew how to act. Again, trust games were used. What they looked at were remaining problems the autistic subjects have and what they found looks to be a problem/difference with a significant aspect of one’s sense of self; in particular, owning an action’s results (I am relying on memory here I should say, and you could check it out by searching under autism and montague).

    You seem to suggest that MRI tells us nothing that lesion work hasn’t done already. I’m finding that hard to take in, since of course lesion research in human beings is severely restricted by the fact that one can’t induce them to study them.

  9. Charles, perhaps his views are more complicated; I was just looking at the article in the NY Times. I don’t think, however, we should assume that philosophers are all that capable of developing general critiques of a scientific field, however good they are as philosophers. You cite Miller’s article, and though I don’t have time now to analyze it thoroughly, the abstract suggests that his targets are actually different from Burge’s. I also suggest that Miller’s abstract makes us aware of the important, underlying problems. Unfortunately, it’s hard from his article to see how Burge could get more subtle about his major claim that neuroscience cannot inform us about psychological explanation.

    Miller is talking about mistakes made in claims about localization. Burge, rather, is claiming that localization claims are not relevant. That’s very different. Further, Miller seems to think that the underlying problems concern understanding something called psychological/biological causation and the relations among neural and psychological explanations. Here again these are not the problems Burge sees as the central ones.

  10. Right, the fact that x has been claimed to support functions 1, 2 and 3 doesn’t show that it doesn’t support 1. But it does invalidate the inference from x shows activation, therefore it is playing function 1 (or 2 or 3). Evidence in the cognitive sciences is convergent when it is worth anything. That is, when the behavorial evidence supports fMRI data and vice versa. Have a look at Max Coltheart’s paper, “Can Cognitive Neuroscience Tell Us Anything About the Mind?” His negative answer is far too strong, but the evidence he adduces supports the claim that so far it hasn’t told us much.

  11. Neil, I agree that the inference you describe is invalid; I wasn’t making that inference, however, though it may have looked as though I was. What I was relying on were prior studies that argued for a particular interpretation in the same context.

    I think there are actually big questions about fMRI, and sometimes staggeringly fallacious reasoning. I didn’t think Coltheart’s paper was all that well taken, at least with respect to current research.

    I also agree with your remarks about convergence.

  12. Now I’m confused. You said the original paper was evidence that borderline people knew that they were violating moral norms. That requires you to make the inference, given that there are a variety of functions attributed to the insular.

  13. I agree with Neil that the basis for much of what we know about the functions of the brain come from behavior/brain correlations. However, I think that neuroimaging studies often constitute (or contribute to) some of the latest forms of behavior/brain correlations. It seems a plausible conjecture that much good work in many related areas involves studying both behavior and various kinds of neuroimaging. Although I do not defend (epistemological) reductionism – either from the mental to the neurological, or from the neurological to the microphysical, or from the mental to the microphysical – I do defend a more optimistic view of what all forms of nonhistorical studies of both the mind and the brain can tell us about the mind and the brain, including neuroimaging work. Interested readers can check out section 3.2 of my Synthese paper “Confusion and dependence in uses of history” (in particular, what I call the ‘four parts’ of ‘exhibit B’ in section 3.2 of that paper). By the way, Springer provides free access to Synthese through the end of December 2010! To check out my two cents, see the following Synthese “Online First” webpage for my aforementioned paper:

    Alternatively, readers can just skip to the download here:

    Click to access fulltext.pdf

    Here is another link that will track the article wherever Synthese puts it, from the current “Online First” location to the eventual journal issue:

    Although I could have added many more examples to constitute more parts of ‘exhibit B’ in section 3.2, in that part of this paper I just list some examples that I have found especially interesting and telling. Please also note that although the thesis of and the agenda for this paper are related to Burge’s claims, Burge and I are not explicitly addressing the same issues, or approaching those issues with regard to the same concerns.

    Notice that adding the “Previous View” and the “Next View” buttons to the Adobe Toolbar (under “Page Navigation” in the toolbar settings) makes it much easier to use the links that Springer and other journal publishers are increasingly providing to notes and references inside individual articles. Users can do this by right clicking on the Adobe toolbar, selecting “More Tools”, and then just clicking away the toolbar buttons you wish to add/delete (especially such as “Previous View” and “Next View”).

    Critics with different views, please either go gently with me here and/or at least consider reading the whole paper while formulating criticisms; individual parts of the paper make much more sense, I think, in light of where they stand in the paper as a whole.

  14. Neil, sorry for not being clearer. I think the first paper linked to in #9 provides good arguments for saying that an area of the anterior insula reacts to violations of norms; the second in effect describes the different betweeen BP and non-BP in the reactions of the insula.

    David, “correlation” often means “mere correlation” and if that is what you mean, I’d dispute that. Further, it one correlate is what we can be conscious of, then I think it’s important that a lot goes on of psychological importance that we probably can’t be conscious of. In addition, there are general trends – for example, it looks as though the initial awareness of rewards gets shiften down into early vision; I don’t know how I’d phrase that in terms of correlation.

  15. JJ,

    I am no big defender of correlation – “mere” or otherwise.

    As regards your post and Burge on neuroimaging, I find that we are in near total agreement.

    An attempt to clarify: I mean to speak of correlation both in terms of noticing that many people have cognitive changes when they bump their heads, and in the terms used when we study/discover that various mental phenomena correlate with various neuroimaging data. Many such studies and discoveries do not merely present correlation data. In many cases, they allow us to understand and predict what is going on in people’s heads by looking at appropriate neuroimages. In my Synthese paper, I give some examples of such work (involving, but not limited to, what many people call ‘correlations’) that allow researchers to understand the mind, the brain, and behavior to such an extent that they can and do develop brain machine interfaces so that people can control machines with their minds, and so that researchers can read people’s minds with machines [and so that machines can read people’s minds…] (all of this based on what I meant above by contemporary forms of studying correlations). Unless that sort of thing counts as “mere correlation”, I think we might agree a bit more than some might think.

  16. David, I’ve just had a quick look at your paper; I’m afraid I don’t have the time today to read it carefully. But it looks very interesting and very well argued; I’ll definitely come back to it.

  17. Thanks jj. Just to be clear, your post that started this comment thread is of course up to your usual high/excellent standards and insightful points, from which I always learn a great deal whether we agree or disagree – but in this case I really think we agree (on the claims in my comments #14 and #16 above, but not necessarily about everything in my Synthese paper provided/referenced in those comments).

  18. I didn’t get the impression Burge is claiming that neuroscience cannot inform psychological explanation, or that any neuroscientific evidence is irrelevant to psychological explanation. He just seems to be saying that neuroscience cannot replace psychology; that the latter cannot be wholly reduced to the former. Perhaps the difference here is slight, but I think Burge is being relatively safe in his assertions.

  19. Jason, I’m having a bit of computer woe at an airport. let me start to reply by quoting Burge. The quote is consistent with what I said in the post. Do note that the topic is not the whole of neuroscience. It’s about brain scans.

    First, it provides little insight into psychological phenomena.  Often the discoveries amount to finding stronger activation in some area of the brain when a psychological phenomenon occurs.  As if it is news that the brain is not dormant during psychological activity!  The reported neuroscience is often descriptive rather than explanatory.  Experiments have shown that neurobabble produces the illusion of understanding.  But little of it is sufficiently detailed to aid, much less provide, psychological explanation.

  20. jj,

    I was responding more directly to your comment #10. As for the main post, I am sure you (and others here) are in a better position than I am to challenge Burge’s claim that most neuroscientific advancements have not done much to further our psychological explanations. But I’m not sure Burge’s error here is profound. Perhaps he just makes the case a little stronger than he should.

Comments are closed.