Wednesday, July 11, 2018

Minds and Machines on Causality and the Brain June 2018, Volume 28, Issue 2,


This volume of Minds and Machines is the product of a conference, which seems largely to have determined the contributions.  Although purportedly about science, the essays are often principally directed at those philosophers of science who do not understand the banalities of the sciences they write about or are interested in. (Scientists tend to like this kind of stuff, because it is people saying what the scientists know or think. Everyone likes cheerleading.) Only one of the essays, Romeijin and Williamson's, makes any contribution a brain scientist could conceivably use.


Romeijin and Williamson, Intervention and Identification in Latent Variable Modeling


The authors actually do something.  They show that if X, Y, L are binary, and L is the common cause of X, Y, and X, Y are measured and L is unmeasured and there and there are no other causal relations between X and Y, then an exogenous perturbation of the distribution of L allows identification of p(X | L) and p(Y | L) (and of course, p(X,Y | L) for all values of L, without knowledge of the distributions of L before and after perturbation except that the distributions are different. 

Of course, it isn't true if the relation between X, Y and L is linear, or if besides the common cause, X influences Y, or if L has more than two values, etc.

The authors give no empirical example that realizes their result. Still, they did something.

Colombo and Naftali, Discovering Brain Mechanisms Using Network Analysis and Causal Modeling

This is a book--or rather multi-paper--report.  While there is nothing new in it, the essay is sensible and measured.  Unfortunately, it is not up-to-date on search methods for causal signaling relations between brain regions estimated by fMRI or EEG, not even close. Things are happening, fast.

Perhaps philosophers writing what are essentially judgemental review essays ought to talk first with some of the people actually doing the work?

Winning and Bechtel, Rethinking Causality in Biological and Neural Mechanisms: Constraints and Control

Aside from a foray into the "causality power" i (read "oomph") bit of metaphysics, this essay, like many of Bechtel's, is a paradigm of saying the scientifically banal without furthering anything. Banalities, of course, are generally true.

  • Gottfried Vosgerau and Patrice Soom, 
  • Reduction Without Elimination: Mental Disorders as Causally Efficacious Properties
    Here is the upshot:

    "our proposal is to analyze mental disorders as higher-level dispositional properties that cause specific symptoms under specific conditions, and that are token-identical to complex physical states. This proposal secures the causal efficacy of mental disorders and their crucial role in explanations, while specifying the systematic relation to lower levels of descriptions as found in neurology and neurochemistry". 

    That's nice. Thank you Donald Davidson.

    Matthew Baxendale and Garrett Mindt, Intervening on the Causal Exclusion Problem for Integrated Information Theory


    In gyrations through discussions of the mental and the physical, I look for the takeaway. Here is theirs:

    "According to IIT there is an identity between phenomenological properties of experience and informational properties of physical systems…The maximally irreducible conceptual structure (MICS) generated by a complex of elements is identical to its experience… An experience is thus an intrinsic property of a complex of mechanisms in a state."

    Speaking of thoughts of complex mechanisms, I wonder what Pluto is thinking now that it's not a planet but still a complex of mechanisms.

    I did not read the paper. 


    Sebastian Wallot and Damian G. Kelty-Stephen, Interaction-Dominant Causation in Mind and Brain, and Its Implication for Questions of Generalization and Replication


    I am tired, but just in case you want to read it, you will learn again that that lots of variables affect what people do, so generalization in psychology is hard.

    Friday, June 29, 2018

    Dinner at Princeton Comeuppance


    N.B. This is a true personal memory about philosophers, not a review,


    Dinner in Princeton, or How Bas van Fraassen Shut Up a Bigot, Albeit in a Politically Incorrect Way--He Had It Coming



    There were things—most things—about Princeton my wife did not like in 1970, and one thing she did, working at a women’s health clinic in inner city (read:, black) Trenton with Rennie Hampshire. In her fifties, Rennie was slim and energetic with the bones and eyes and skin that make for that rare type, a beautiful English woman. She had once been the wife of a famous English philosopher, A.J. Ayer, but was at the time the wife of a still more famous English philosopher, Stuart Hampshire, who was the Chairman of my department. She was a busy person in occupation and manner, never still, and she had a ferocity of spirit that set her apart from anyone else I can remember in my life. Stuart often remarked with pride—and I assume with truth—that she had been the first woman to ride a motorcycle across the United States, but now her energies were directed to a poor black community in a depressed state capital. (In those days, as you crossed the Delaware River into Trenton, you were welcomed by the most self-piteous sign: Trenton Makes, the World Takes. And doesn’t give much back.)

    The clinic, in the basement of an undistinguished building, provided gynecological services, birth control, pregnancy testing and check-ups, abortion referrals (to New York—abortion was illegal in New Jersey then), advice and help of many kinds. It was funded, so far as I could tell, almost entirely by the efforts of Rennie Hampshire. Her husband’s position put her often at dinner parties and other social gatherings with the well-heeled and the immensely rich, and she used these chances shamelessly to extort contributions for the clinic. I recall watching her with admiration at a dinner party in her own home, first telling some well-to-do guest about the clinic and it’s needs, and then, after dinner, pursuing him until they retired to the kitchen, one to give a check and one to receive. She was not just a fund raiser. She worked at the clinic as did any non-medical volunteer, doing whatever she could that needed doing, from counseling patients to filing records to scrubbing. My wife, Anita, helped her and loved doing it, and she loved Rennie. Rennie alone in Princeton showed Anita that Anita’s world—the world of poor, simple people, the world of personal charity—counted, and counted more than the hot house of Princeton refinements, where Anita never felt welcome or at home or free to breathe. Anita was never so happy as when leaving for Trenton, and never so exhilarated as when she returned.

    Rennie had her eccentricities. She would leave small gifts—tea, or chocolates—on our porch, but on the several occasions Anita or I caught her at it, she refused to come into the house. When she and Stuart once came to dinner, she insisted on washing the dishes. I think she was determined not to be fussed over, but really did not quite understand the peculiar texture of lower middle class American formality and informality.

    Of all things Princetonian, Anita most despised faculty dinner parties. Her prejudice was confirmed one evening when, after I had begged her on behalf of my career, we attended a party at the home of an eminent (about as eminent as an academic gets) professor of history of science, Thomas Kuhn. The professor’s wife, whom Anita found particularly cold, greeted us at the door and threw her arms around my wife in a grand hug, exclaiming how delighted she was that we—and my wife in particular—had come. Anita reddened, not in embarrassment, but in pleasure. I followed them into the house, our hostess walking with her arm wrapped around my wife’s waist. Half way to the other guests, she leaned her mouth to Anita’s ear and I heard her whisper: Do remind me, dear, what is your name? An understandable lapse, surely, but Anita never attended another faculty party, except once.

    Bas van Fraassen is a man to envy: handsome, elegant, charming, original, brilliant, sensitive and European; even nowadays, past seventy, he can pull off dressing like a rock star and can fly the trapeze. We were friends, and his warmth and charm made Anita forgive him his occupation (he professed philosophy at Yale then). He liked her too. He was visiting us in Princeton one weekend when, on Saturday morning, Anita answered a phone call from a colleague of mine, Margaret Wilson, inviting us to dinner that evening, with apologies for the short notice.  Anita had the perfect excuse—we had a guest for the weekend—but she erred in revealing the guest’s name. Margaret, who was a straightforward person, immediately said to bring him, and Anita was stuck. So, with Bas, we went to dinner at the Wilson’s.

    The party was small: the three of us, Margaret, her husband, and another colleague, George Pitcher. George was tall, slender and athletic—the only faculty member who hit home runs at the annual faculty versus student softball game—with a full head of brushy hair that fell over his forehead in a boyish cowlick but seemed never to be out of place, altogether an extraordinarily handsome man who had at least three loves, his housemate, Ed, and a pair of stray dogs that had taken up residence with them and about which he later wrote a very sentimental book. George was a polite and to appearance a gentle man, but, as I discovered when serving with him as the junior member of a committee of two that made minor personnel decisions, he was privately the kind of bigot that once flourished in Princeton, the kind for which there is only a neologism, a classist.

    We sat in the Wilson’s dining room, next to a cool bower, around a circular table centered with a bowl of fruit, and had drinks and nibbles and get reacquainted talk before dinner. The talk, inevitably, turned to academic gossip, who is doing what, going where, with whom, the sort of thing that fascinates academic neighbors and bores everyone else, not least Anita, who sat quietly with the bland look that I knew hid an interior woe: not this again. Professors, philosophy professors anyway, at parties scarcely ever talk about ideas (that’s business), or politics (they all have the same, or none), or religion (what’s to talk about?), or sex (not done), or money (not done), or sports (intensely not done). Its gossip, travel and high culture.  Somehow, Rennie Hampshire’s name came up, and I think Anita mentioned her work in Trenton. George, in his quiet but forceful and authoritative voice, began a rant: What was she doing there, among the refuse of the city? Why not leave those people to themselves? The trouble with Rennie is that she does not respect her own class.

    The denunciation did not stop, and as it continued Anita changed in color, her smooth face lined in anger, and, I knew, frustration. Her good manners conspired with her sense of social unease; she could not speak without demonstrating her anger. I objected ineffectually, but George talked over me. The Wilson’s looked uncomfortable and said nothing while George went on into the evils of crossing the borders of social strata that he seemed to think should be guarded with machine guns and barbed wire, Berlin walls of Class.

    Noting Anita’s distress and my clumsiness, Bas quietly reached an elegant arm to the centerpiece and removed a banana. Smiling at George, he carefully peeled back half the skin of the fruit, and then, delicately, began to lick the tip. Eyes still full on George, still silently, he pushed the end of the banana into his mouth and, moving it back and forth, began to suck. George reddened, then blushed, and fell silent for the rest of the evening, shamed, if not for the right reason, then, at least, for the wrong one.



    Hayley Clatterbuck, The Logic Problem and the Theoretician's Dilemma

    While surfing around I ran across Hayley Clatterbuck's, essay,  The Logical Problem and the Theoreticians Dilemma, Philosophy and Phenomenological Research doi: 10.1111/phpr.12331 in a journal I usually don't read. It is almost good, just tiptoes up to goodness and gets no farther, leaving some oddities and bad arguments along the way and huge opportunities untouched. It reminded me of two occasions. On one, Judea Pearl asked a dinner party of UCLA philosophers "Why don't you guys do anything?" On another, after hearing two hours of lectures by Alvin Goldman on the difference between "hard wired" and "soft-wired" capacities, Allen Newell asked: "So what has your laboratory discovered about hard wired capacities?"

    Clatterbuck's problem is the warrant for attribution of understanding to creatures that are not human. She rightly sees that claims of behavioral evidence for such attributions come face to face with the behaviorist version of Hempel's Theoretician's Dilemma.  She first proposes that understanding can be established by having independent observable stimuli with correlated responses, inviting explanation by a mediating variable. She represents this by a graphical causal model, roughly

    E1 :  S1                         R1 
                          U
    E2: S2                           R2

    with arrows S1 -> U; S2 -> U; U -> R1; U -> R2. Citing the Causal Markov Condition, which she modestly says she does not fully understand, she claims the graphical model above implies that R1 and R2 are correlated.  (That is correct but since S1 and S2 are mutually exclusive, they are associated, so perhaps they should be collapsed into a single variable with 2 values; but Clatterbuck does not want the experimental treatment to be a variable common cause of the results.) She rightly goes on to object that nothing says the mediating variable U has to be some state of understanding; it could be a lot of different things.  So she goes on to suggest that designs are needed in which U -> R1 is a positive association and U -> R2 is negative, or vice-versa.  I guess the idea is that understanding would produce positive associations in some circumstances and negative ones in others that were perceptually similar.  So here is a good reason for positing a mediating variable, essentially using Reichenbach's common cause principle, and an argument I don't fully understand for it's interpretation.

    However that works out, her discussion is lexically odd.  She says that the graphical model shown, which produces (with Faithfulness) an association between R1 and R2 is "syntactical" but the revised model that produces a negative association is "semantic."  Associations are syntactic but negative associations are semantic?Since "syntactic" is a term of abuse in contemporary philosophy of science, I wonder at the rationale for her terminology.  But on to something more serious.

    She argues, I think, that the schema illustrates a way round the Theoretician's Dilemma. Following John Earman, Clatterbuck argues that prior evidence, call it E,  provides “inductive support” for some theory t, and t entails (and hence predicts) some new phenomenon N which thus would not have been predicted without recourse to t.  Earman puts the argument in Bayesian terms, as does Clatterbuck.  But if t entails N then Pr(N, E) ≥ Pr(N, t, E) hence Pr(N | E) ≥ Pr(N | t, E) Pr(t | E) = Pr(t | E): the novel phenomenon is at least as probable on the prior evidence as is the theory on the same evidence.  The Bayesian could skip the theory and go directly to the predicted phenomenon.  So that doesn't work.

    Towards the end of her essay, she reformulates the idea in a way that I think she takes to be just an elaboration of Earman's (bad) argument, but is not: “In a case of within‐domain extrapolation, an empirical regularity within a domain is redescribed in terms of a theoretical relation which is then extrapolated to unobserved cases of that same domain. In a case of cross‐domain extrapolation, empirical regularities in domain A and B are redescribed in terms of the same theoretical relation, and it is induced that what is true of A is also true of B.”

    I read this suggestion this way: given data from cases in one domain, find "theoretical" features of those cases that hold in another domain, and use those invariant features to predict about data in this second domain.  That would be a reason for theories. 

    Ok, how can that be done...what is the script, the recipe? How can such invariant theoretical features be found? Nada. Full stop in Clatterbuck's essay. My graduate student, Biwei Huang, has an illustration for neuroscience. Taking fMRI images from subjects in one laboratory, she identifies strengths of some neural causal relations ("effective connections" in contemporary neuropsychology jargon) that separate autistic subjects from normals. She then uses the presence (or absence) of these connections and their strengths inferred from  fMRI scans from another laboratory (another "domain")  to predict autistic versus normal in the second laboratory (actually, in each of several other laboratories). 

    This is a pretty neat illustration of the invariance strategy whose suggestion I attribute to Clatterbuck. And it does defeat the Theoreticians Dilemma: you couldn't make comparably accurate predictions by, say, comparing the correlations among fmri signals in different brain regions in subjects in one lab with those in another lab. People have tried.  

    Huang's example is just that, not a general procedure for learning theoretical invariants for cross-domain classification.  My colleague, Kun Zhang, has developed one, explicitly in those terms.  Of course, that leaves a lot of room for work on the description of general procedures for other kinds of cross-domain invariants, of which physics provides many examples, e.g., classical thermodyanmics.

    Afterword: Goldman replied to Newell that there is a division of labor: philosophers help science by posing the problems and distinctions; psychologists investigate them in the laboratory. Newell said thanks, but psychologists have no trouble doing both jobs. 

    References

    B. Huang, Diagnosis of Autism Spectrum Disorder by Causal Influence Strength Learning from Resting-Stae fMRI Data, M.S. Thesis, Department of Philosophy, Carnegie Mellon University, 2018.

    Gong, M., Zhang, K., Huang, B., Glymour, C., Tao, D., & Batmanghelich, K. (2018). Causal Generative Domain Adaptation Networks. arXiv preprint arXiv:1804.04333.