Paul Meehl on "Genes and The Unchangeable Core"
Anticipating the limits of genetic essentialism
All of Meehl’s papers are available HERE.
The third class in my graduate psychopathology seminar is all about Paul Meehl. I have noted here before that there is a gulf in modern attitudes about Meehl. When I was in grad school, he was treated as a Godlike figure. Thinking theoretically about scientific clinical psychology was mostly a matter of coming to grips with what Meehl had said about it. Meehl was present at the birth of the scientist-practitioner model in 1954 (?), then ran the clinical program at the University of Minnesota, which was then (it is still very influential) the iconic program in the US.
The University of Texas clinical program I attended in the 1980s was sometimes referred to as “Minnesota South Campus.” The whole Psych Department had been built around the recruitment of Gardner Lindzey from Minnesota. Reading Meehl’s “Why I Don’t Attend Case Conferences” was an early ritual in getting acculturated as a young clinical psychologist. So when I started teaching my own graduate seminar I naturally figured I would continue the tradition and assigned it in the second class.
The students hated it, for a whole variety of reasons. Some aspects of Meehl’s written personality have not aged well. He often assumes the role of the smartest guy in the room, talking down to the students or clinicians he is addressing. The issues he is discussing in “Case Conferences” have also drifted somewhat out of date. In those days R1 clinical psych programs were still staffed by faculty who were primarily clinicians, often of the old-fashioned, “science doesn’t have any relevance to what I do” way of thinking. Many of them (like Meehl, more on that in a second) were psychoanalysts. Those attitudes are foreign to today’s graduate students, who know nothing of psychoanalysis, and if anything are prone to believe that there is nothing to clinical psychology except the routine application of science. “Case Conferences” is also very long, really a monograph.
Anyway, I no longer assign it. Instead, we read “Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progress of soft psychology” because…. well because it’s my favorite Meehl paper and I get to pick. I also ask them to browse through the website and find a paper to present, just for fun. I still think Meehl is hugely relevant, the most important clinical psychologist of my lifetime, and I am glad to introduce young people to his work. If nothing else, he showed me and a lot of people like me how to be a psychologist, combining empirical psychology, behavior genetics, quantitative methods, philosophy of science, and (for a few of us) psychoanalysis.
That’s a too-long introduction to the random Meehl paper I found: Genes and the Unchangeable Core (1975). The paper seems to have been part of some workshop about the intersection between psychoanalytically oriented psychotherapy and empirical science. Meehl set himself the task of thinking about whether behavior genetics had shown that some people had a genetic core making them the way they are, and therefore out of the reach of psychotherapy. He decides that there is. Quoting him at Meehlian length:
I conjecture (and am frank to admit, as practitioner, I believe—"everybody is of course betting on a horse, so I feel free to bet on mine!) that in the non-psychotic realm of psychological disturbance that we put under such arbitrary (and nontaxonomic?) rubrics as neurotic depressive reaction, anxiety neurosis, transient adult situational maladjustment, marital problem, etc., there is fre-quently—I do not say always— operative in the complex skein of causality whichbrings the patient to us, one or more (typically, I think one should assume more) hereditary variables of polygenic kind that represent not pathological entities such as Huntington's disease or schizophrenia but what may be called "normal individual difference variables" of temperament. Among these I would list such factorsas anxiety proneness, energy level, the heritable component of the so-called general intelligence factor (which I am old-fashioned enough to persist in believing), social dominance, some components of so-called introversion, sexual drive, frustration tolerance, orality, rage-readiness, and the like. I think it difficult to set any strong limits on which aspects of socially learned behavior may involve acquisition-parameters that are in some degree genetically determined, given the growing body of evidence that there is a heritable component in a number of characteristics which might not antecedently have been expected to have anything to do with one's genes (e.g., social dominance in the mouse, introversion in the human, and I think particularly of some surprising studies indicating a heritable component even for vocational interests!) Genetic influences are currently unpopular among American psychologists, but the evidence is accumulating very rapidly; and I detect, especially among recent crops of clinical psychology students, a greater readiness to allow for the possibility of heritable contributors to behavior disorder than was true even 10 years ago. Of course when human family or physiological and biochemical data are skimpy or lacking, one who feels impelled to arrive, however tentatively, at reasonable conjectures about these matters will use whatever "personalistic probabilities" he can lay hold of.
I’ll say up front that I think Meehl was wrong about this, for an interesting set of reasons. His appeal to an genetic core is a nice way to put what we now call genetic essentialism, the idea that there is an inner genetic structure of some kind that causes the phenotype we observe. Notice that essentialism is more than just believing that something is “heritable” in the quantitative genetic sense, ie, MZ twins are more similar than DZ twins. The whole drift of my theoretical work has been to argue that traits can be heritable without being essential.
In 1973, twin studies were in the early days of making it clear that everything that interested Meehl was turning out to be heritable, which he took to mean, “genetic.” Intelligence, personality, schizophrenia, depression, anxiety, you name it. It was a time of considerable excitement in behavioral genetics, because the big problem hadn’t yet arrived: everything was heritable in more or less the same sense. You can hear Meehl anticipating the interesting science that was going to follow from parsing out which aspects of normal and abnormal behavior were or weren’t going to turn out to be “genetically determined,” allowing the scientifically-oriented psychotherapist to pick and choose a route between the unchangeable genetic core and the socially-determined traits that were more amenable to treatment.
But that never happened. Fifty years later, there isn’t a single distinction to be made between a mental disorder that is largely genetic and a social adjustment that is largely environmental. Everything is genetic in some very limited sense— MZ and DZ twins, from which we can weakly conclude that genes have “something to do” with all human differences, behavioral or otherwise. Modern DNA-based human genetics hasn’t changed that conclusion: using GWAS, all human differences are (more modestly) heritable, and in a general way one can show that bits of DNA are correlated with behavioral differences. But there is nothing in modern GWAS that would be useful to a working psychotherapist.
This is precisely because there are no genetic essences that can be meaningfully separated from the observable phenotype. If we want, we can believe that some people are depressed because they have a strong genetic predisposition to depression, while other people are depressed because of the things that have happened to them, perhaps in spite of an otherwise non-depressive genotype. I personally don’t doubt that is true, but there is no empirical way to separate those things in real people. Why is Turkheimer introverted? You can say, because he has a genetic predisposition to being introverted like his Mom, but that is no more than working backwards from the phenotype to a tautological assumption about my genotype. You could do the same thing with the fact that I prefer baseball to football, like my Dad. (Take my word for it, it’s heritable.)
This is a very important point, because it violates an easy, scientifically gratifying assumption about the relationship between genes and phenotypes. Just one more example— think about the concept of innate (i.e., essential) intelligence. IQ is heritable. Does that mean that some people are “born smart” with genes that produce brains that make them smart, while other people become smart because of good parenting or hard work? Maybe, in theory, but on this planet there is no way to tell those two things apart. People talk as though people “have” an inner IQ that is then expressed to one extent or another in the real world, but that isn’t right. People have one IQ, in the here and now, as measured by a good intelligence test. Useful for many purposes, “heritable” in the twin sense, but not a reflection of anything essential.
Thanks for this entry! I was lucky enough to be a grad student at the Dustbowl of Empiricism in the late 80's and got to hear Meehl lecture in an unregistered seminar open to anybody to attend. I was not smart enough then to totally track everything he talked about but I was struck by how easily he could slip between statements extolling the need for objective tests of personality and performance (he thought physicists were the smartest of all people because of how they scored on the Millers Analogy Test) and anecdotes about patients from his psychoanalytic practice that relied on metaphors and unconscious drives. I remember him discussing one particular depressive patient, stating that this man wore a brown suit all the time, "shit-colored clothes" that matched how he viewed the world through "shit-covered glasses."
What do you mean by "heritable" in the twin sense, but not a reflection of anything essential? What is heritable in the twin sense? And do you still think that genetics has no use in psychotherapy (okay, maybe not in treatment, but not in theory?)?