Meehl, Paul Everett
MEEHL, PAUL EVERETT
(b. Minneapolis, Minnesota, 3 January 1920; d. Minneapolis, 14 February 2003),
psychology, philosophy, psychiatry.
Meehl was a theoretical psychologist and philosopher of science who contributed several important conceptual formulations to psychology, most notably a demonstration of the superiority of actuarial against intuitive prediction. A practicing psychotherapist, he played an important role in the medicalizing of American psychology.
Polyvalent Beginnings . Meehl came to psychology out of deep personal motives: His father’s suicide during the Great Depression and his mother’s death five years later from a misdiagnosed brain tumor led him to question fatal human motives and fallible judgment. Intellectually precocious and voracious, his first encounter with his future work was his reading, at age twelve, Karl Menninger’s The Human Mind. In his autobiography, Meehl said he found this blend of contemporary psychiatric nosology and therapeutic optimism a “healing Damascus experience” (1989, p. 339). Before Meehl was fifteen he had read extensively in psychology—popular behaviorisms as well as Sigmund Freud—and revealed a taste and talent for epistemology, reading Bertrand Russell. He was especially influenced by Alburey Castell’s lucid 1935 textbook A College Logic, and honed his formal logical skills debating with adolescent peers. Castell, a University of Minnesota faculty member, was limned as Frazier’s critical interlocutor in B. F. Skinner’s Walden Two; similarly, Meehl became a gadfly critic of psychology’s scientific enterprise.
Meehl entered the University of Minnesota in 1938, intending to follow the premedical course, but he migrated to psychology, graduating with an undergraduate degree (with a minor in biometry) in 1941. His main undergraduate influence in psychology was Donald G. Paterson, a practical psychometric specialist and an early proponent of vocational guidance, from whom he gained expertise in the measurement of human abilities and a predisposition to consider psychology as inseparable from its applications. It was fortunate that, in 1940, Herbert Feigl moved from Iowa to Minnesota, where he finished the rest of his career. Feigl, a highly cultured colleague of the Viennese logical positivists, found in Meehl a kindred spirit and a lifelong friend. Through Feigl, Meehl had access to a rich, well-articulated theoretical approach to the unity of science and liberalized versions of neopositivism. Meehl remained at Minnesota for graduate work—indeed, he remained at Minnesota for the remainder of his career, retiring as Regents’ Professor of Psychology in 1990—and was mentored by another exponent of the style of Minnesota psychology known as “Dustbowl Empiricism,” Starke Rosecrans Hathaway. Hathaway was one of the earliest medical clinical psychologists in the United States and had just finished assembling, in conjunction with the neuropsychiatrist J. Charnley McKinley of the medical faculty, a psychometric instrument designed specifically for identifying and classifying psychiatric disorders, the Minnesota Multiphasic Personality Inventory, or MMPI.
Meehl, rated unfit for service during World War II because of a heart defect, remained at Minnesota during the war years. While continuing to work with Feigl and Hathaway, Meehl became one of the group of graduate students who associated closely with Skinner in the midst of a most fertile conceptual period, when he was working simultaneously on shaping, language, and the social implications of behaviorism, leading to the writing of Walden Two. Skinner, after moving to Indiana from Minnesota at the end of the war, thought highly enough of Meehl that he offered him his first academic position. Meehl, however, declined it, having established himself as not only a psychometric specialist but also as a clinician, seeing his first clients in 1942. Meehl received a PhD in clinical psychology in 1945, entering that profession just as it came of age in economic and social terms. His thesis focused on people with psychopathology who are able to compensate for it and appear normal. The search for latency, psychopathic and otherwise, determined the direction of his career.
The Search for Latent Entities . From the outset, Meehl strove to create a theoretical framework that could contain both clinical and experimental psychology and that would allow the existence of latent mental entities. Meehl was impressed early on by an experience in which Hathaway gave him several psychiatric case studies and asked him, blind to the diagnosis and to the MMPI profile, to find similarities between them. Meehl identified a behavioral regularity (sexual nonaggressiveness) the group of cases, and was surprised when Hathaway revealed that they had been sorted not by psychiatric diagnosis but only by the similar simple pattern of their scores on the MMPI. Meehl likened this procedure of “looking at the person through the test” to pulling oneself up by one’s own bootstraps, and “bootstrapsing” became a familiar Meehl theme and technique.
While continuing to work on the development of the MMPI, Meehl, by now a faculty member in psychology at Minnesota, entered into a long-term collaborative relationship with his behaviorist colleague Kenneth Mac-Corquodale, researching aspects of Edward C. Tolman’s theory of latent learning. Tolman and his coworkers had contended that, in rats, learning could take place implicitly without direct reinforcement. By the mid-1940s Tolman had evolved several versions of a theory incorporating not only empirical behavioral data but also theoretical terms that he eventually came to call cognitive structures. In 1948 MacCorquodale and Meehl published “On a Distinction between Hypothetical Constructs and Intervening Variables” in Psychological Review, streamlining the vocabulary used to define the relations between theory and data across several contemporary learning theories. They conceived a distinction between terms that functioned to summarize or abstract empirical data and that could, if necessary, be reduced to the data level (intervening variables) and terms at a high level carrying “surplus meaning,” which could not be completely reduced to data (hypothetical constructs). This was a move away from a narrow operationalism and toward a modern neurocognitive view of learning, the brain, and the mind. Meehl and MacCorquodale made an explicit connection between hypothetical constructs and underlying neurology, and D. O. Hebb utilized this distinction as a basic element of his seminal 1955 paper “Drives and the Conceptual Nervous System.” The hypothetical construct-intervening variable conception has proved durable and useful not only in psychology but in fields as diverse as computer science and molecular biology.
Meehl’s next significant contribution to theoretical psychology was again collaborative, the product of a working group investigating test validity for the American Psychological Association in 1954. This paper, “Construct Validity in Psychological Tests,” coauthored with Lee Cronbach and published in Psychological Bulletin in 1955, introduced the idea of the nomological net, an idea drawn from several sources in empirical-realist philosophy. The net provides coherence and explanation through individual theories that exist at a conceptual level as constructs, ligated to each other through logical coherence and to physical reality (through a level of intervening variables) by connections to empirical data. Between 1955 and the end of his career, Meehl, in several papers, evolved the following formal logical schematization of the network of relations between theory and data (simplified here to its basic elements):
(T • A • C ) → O
with T and A representing primary and auxiliary theories respectively, C representing conditions of observation, experimentation, or individuals, conjunctively leading to O, empirical observations. Strongly influenced by Hans Reichenbach’s 1938 text Experience and Prediction—he referred to it in both his first and last (posthumously) published papers—as well as by his own psychometric background, he conceived this conceptual network as essentially probabilistic. Theories, maintained Meehl, borrowing from his philosophical contemporary Arthur Pap, are “open concepts,” provisional and corrigible. Meaning emerges in reciprocal relations between all elements of the system, “seeping upward” from data, and spreading downward and laterally from theories and auxiliaries. The nomological net, in Meehl’s view, is continually being mended by the addition of new postulates, new observations, and new interconnections at all levels.
During the 1950s Meehl’s search for latent entities took a more explicit clinical turn, as he shifted focus to psychiatric syndromes, in particular schizophrenia. Meehl, now established in clinical practice, had already garnered extensive experience with the phenomenology of schizophrenia, contributing a case study on catatonia to Arthur Burton and Robert E. Harris’s Case Histories in Clinical and Abnormal Psychology in 1947 and editing, along with Hathaway, an extensive collection of case studies aligned with MMPI profiles. Sometime during this period Meehl encountered the description of “pseudoneurotic schizophrenia” published in 1949 by Paul Hoch and Philip Polatin. This syndrome, in which underlying schizophrenic pathology is overlaid and masked by virtually all other psychiatric syndromes including depression, anxiety, and hypochondria, was at once a supreme diagnostic challenge and an apt exemplar of latency. Meehl became more explicitly psychoanalytic in orientation, undergoing training analysis with Bernard Glueck, a student of Sandor Rado, an independent-minded Freudian. Rado theorized that a collection of personality traits lay behind schizophrenia, terming this the schizotype.
In 1962 Meehl presented, as his presidential address to the American Psychological Association, a theory of schizophrenia that postulated an underlying, genetically determined neural deficit (schizotaxia) leading to, through learning, the formation of a set of personality characteristics (the schizotype, or schizotypy, in Meehl’s terms), which could, under certain environmental and social conditions, lead to the expression of phenotypic schizophrenia. This theory was one of the earliest modern postulates of a behavioral-genetic theory in psychology and has proved a durable source of research conjecture in psychiatry. Meehl’s work on schizophrenia and diagnosis led to two significant related contributions: the revival of the concept of anhedonia and a statistical taxonomic approach to mental entities.
The concept of a lack of pleasure had fallen into disuse as interest in psychological hedonism waned during the 1920s. Meehl’s approach to pleasure is a good example of his approach, his eclecticism, and his range. In 1950 Meehl established the noncircularity of the law of effect by showing that reinforcers were transitive across situations. Then, proceeding to clinical observations, he focused specifically on the lack of pleasure observed in schizophrenics. To describe this lack of pleasurable feeling he borrowed Rado’s term anhedonia, which had come into psychoanalysis from older psychologies such as those of William James and Théodule Ribot, to name an internal entity, a measurable individual-difference variable for which he proposed measurement scales. Finally, late in his career, he considered this latent hedonic dimension in the context of economics and proposed a theory of measurable relative pleasure utilities in contrast to Pareto optimality.
Meehl became interested in problems of classification while developing methods to test competing genetic hypotheses of schizophrenia. Eventually he came to the conclusion that many psychological types and traits— pathological and nonpathololgical—form taxonic classes. “There are gophers,” said Meehl, “there are chipmunks, but there are no gophmunks” (1995, p. 268). Meehl came to see “carving nature at its joints” as a prime focus of scientific interest, and he invented several novel statistical techniques to reveal patterns indicating taxonic class membership from empirical data, which he grouped under the heading “bootstrapsing taxometry.” Development of these techniques with several colleagues, notably David Faust, Robert Golden, and Niels Waller, occupied much of Meehl’s attention after 1980.
The search for mental entities forms the contextual background for Meehl’s philosophy of mind. Though Meehl thought of himself as an avocational philosopher in contrast to the “card-carrying” members of that discipline, he had substantial credentials in the field and was one of the founding members of the Minnesota Center for the Philosophy of Science in 1953. During the next forty years he produced, singly and in collaboration with others, including Wilfrid Sellars and Herbert Feigl, substantial contributions on emergence, the mind-body problem, and freedom and determinism. In conjunction with a defense of Feigl’s mind-brain identity thesis, Meehl utilized thought experiments involving the autocerebroscope, by which one’s brain activity could be read off the cortex and projected onto a screen, allowing comparison between reports of perceptions and brain activity. This conceptual device anticipated, as Meehl noted late in his career, current progress in neuroimaging and its correlation with cognitive states.
While Meehl maintained that, from a physical perspective, identity held its own even when conflicts were introduced between cerebroscope readings and cognized self-reports of states—a foreshadowing of current modular theories of mind—he also demonstrated considerable weakness in this monistic view at the semantic level of qualia. Likewise, regarding freedom and determinism, Meehl opted for a middle-ground view, locating decisions in a probabilistic space within a deterministic system, a “tertium quid.” Meehl adopted a “cafeteria” approach to philosophy, assembling what he felt was needed for the tasks at hand. He also kept an open mind about the subjects that could be admissible in a scientific, realist psychology, writing extensively on religion and parapsychology (he contributed the article on the subject to the Encyclopaedia Britannica in 1962) as well as on Freudian theory—dream analysis in particular.
Critique of Practice . Complementary to Meehl’s focus on entities is his extensive critique of how theory guides practice both in science and in psychotherapy and the ways in which this might miscarry. Meehl’s first and most well-known contribution in this area was the result of ten years’ effort at the beginning of his career. Published in 1954 after it had been turned down by several publishers, Clinical vs. Statistical Prediction: A Theoretical Analysis and a Review of the Evidence became an immediate classic that was reprinted forty-two years later after more than thirteen thousand copies had sold. This work is the Rosetta Stone of Meehl’s critical approach, containing the essence of his critiques of method and clinical practice. His starting point was Theodore Sarbin’s contention that clinicians’ predictions of outcomes are always inferior to actuarial predictions in principle.
In contrast to this view, Meehl invoked Hans Reichenbach’s distinction between the “context of discovery” and the “context of justification.” Clinicians, maintained Meehl, function both as theory constructors and theory verifiers. As constructors, they often make predictive statements emerging from novel combinations of unique clinical data—Meehl used a dream interpretation from Theodor Reik as an example—for which no actuarial track record exists. The unique functions of clinicians, Meehl stressed, were in pattern finding and theory building, rational but at a level different from that at which precise actuarial verification could obtain. However, when actuarial data existed that could be combined to make predictions via regression equations, the situation was different. Meehl conducted a pioneering meta-analysis, tabulating the results of twenty studies in which clinicians and actuarial clerks worked with the same data. The results showed that, with one exception that Meehl admitted with many qualifications, clinicians were not superior to, and frequently inferior to, the actuarial calculation, a result verified several times over in subsequent studies by Meehl and many collaborators and independent researchers. Clinical vs. Statistical Prediction was a stimulus both for development of automated systems for test interpretation and diagnosis, as well as an important step in the development of a psychology of cognitive biases. Meehl maintained that it was unethical and socially irresponsible to fail to use actuarial prediction systems when they were available, and he campaigned vigorously for their use through the rest of his career.
The next element in Meehl’s critique concerns the actual practice of clinical psychology. Meehl drew on both his statistical expertise and his penchant for formal logical dissection to reveal, in a chapter in his 1973 compilation Psychodiagnosis titled “Why I Do Not Attend Case Conferences,” a bestiary of diagnostic gaffes and fallacies, ranging from statistical mistakes such as ignoring population base rates to logical ones such as the volunteering of entirely irrelevant information or overgeneralization from personal experience, for example the Uncle George’s Pancakes Fallacy (a client storing uneaten pancakes in the attic is not exhibiting pathology because “good ole Uncle George used to do that”). Both Meehl’s conception of the clinician as theorist as well as his logical analysis of practice are at the forefront here. Clinicians would produce more interesting insights were they to function at a higher logical level, and they would be more accurate if they would aspire to the logic embodied in neurology grand rounds. Beginning in the mid-1960s and continuing through the rest of his career, Meehl expanded his critique to the collective scientific progress of psychology, which he saw as fitful at best, especially in its “softer” areas (for example, personality and social psychology). Adopting much of Karl Popper’s approach, Meehl criticized both the statistical basis and the logic of psychology’s scientific practices. Psychology errs, he claimed, in trying to validate theories of dubious verisimilitude with weak statistical methods.
Adding to the contemporary critique of null hypothesis significance testing (NHST), Meehl claimed that tests of directional hypotheses showing statistical significance corroborated substantive theories weakly, failing not only to conclusively reject false theories but also producing ambiguous records of confirmatory evidence for true ones. Meehl also compared the specificity of prediction from theory in physics and psychology and found psychology wanting. Borrowing a meteorological metaphor, he said that psychology appears content to predict that it will rain in April rather than to predict precise amounts of rainfall on specific April days. From the viewpoint of formal logic, taking a positive significance test result as evidence for the truth of the theory under test was an elementary logical mistake—affirming the consequent. Until this faulty logic could be replaced with a valid modus tollens procedure, and theories were proposed that would generate point predictions that could be conclusively rejected, psychological theories would tend to generate high rates of spurious confirmation under NHST and would linger until they, like Douglas MacArthur’s old soldiers, “faded away.”
Later, reacting to criticisms by Imre Lakatos, Paul Feyerabend, and others, Meehl would adopt a more nuanced view of how theory might be amended even after failing strong falsification tests, and still later Meehl devised a taxonomic method, cliometrics, which utilized criteria of successful science to classify research programs, historical and current, as either productive or “degenerating.” As an immediate methodological corrective, Meehl advocated reporting confidence intervals rather than statistical significance in summaries of research results. But ultimately psychology will progress faster when its theories are sounder, and Meehl urged psychologists to look to less statistical but equally rational and empirical disciplines for models for theory building, especially history, law, and psychoanalysis. Interestingly, Meehl saw psychoanalysis, one of his essential sources of theory as a clinician (though he could not devise formal tests for its theoretical claims), as drifting into degeneration, describing himself as a “60% Freudian” in 1973 but a “33% Freudian” in 1992.
Principled Practical Eclecticism . Along with his principled eclecticism in theory and philosophy, Meehl had a plainspoken practical attitude toward applied psychology. He was, by his own account, temperamentally suited to the practice of law and eventually taught in Minnesota’s Law School. From that vantage point he addressed practical problems such as the conceptual foundations of the insanity defense and defended lawyers’ use of the “fireside inductions”—commonsense generalizations about human behavior—against psychologists’ insistence on experimental results. Sometimes, Meehl argued, it is practitioners outside psychology who are ahead in terms of knowledge. A practical behaviorist, he was not averse to quoting, in many situations, what he called “Meehl’s Malignant Maxim”: The best predictor of future behavior is past behavior. As a therapist, he practiced both classical psychoanalysis (he had a couch in his office) and, concurrently, Albert Ellis’s rational-emotive therapy, in which clients are encouraged to confront the reality of failures and rejection directly, admit mistakes, and dispute their bad logical choices.
A social meliorist and libertarian, Meehl was, however, under no illusions about either human goodwill or the political process: correspondence from him bore the legend “Caligula for Proconsul” on the envelope, and he made a very good case for the idea that one might as likely be killed driving to the polls as that one’s vote would decide a national election. Yet Meehl saw participation in a democracy as an ethical imperative, and he considered the taxpaying citizen who underwrote his salary his primary social responsibility. Meehl reserved his strongest opprobrium, however, for intrusions of political views into science, and wrote, toward the end of his life, a lengthy defense of the necessity for the independence of science in the face of political objections to its content.
Meehl was extensively honored for his work, winning many of the main achievement awards offered within organized psychology, including the American Psychological Association’s Distinguished Scientific Contributor Award in 1958 and its Award for Outstanding Lifetime Contribution to Psychology in 1996. He was elected a member of the National Academy of Sciences in 1987. His ultimate goal for psychology was unity via what he called the “five noble traditions”: psychometrics, behavior theory, behavior genetics, descriptive psychiatry, and psychoanalysis. Meehl once said, referring to Clinical vs. Statistical Prediction, that he had been “socially reinforced for writing something which hardly anyone believes” (1996, p. v), but his formulations of psychology’s scientific vocabulary are ubiquitous and ineluctable. His challenge to psychologists to improve theory and practice remains open.
Meehl’s papers are cataloged on his Web site, “Paul E. Meehl, 1920–2003,” available from http://www.tc.umn.edu/~pemeehl/ Regularly updated by his wife, Leslie J. Yonce, this site contains a complete bibliography and links to more than half of Meehl’s papers.
WORKS BY MEEHL
Psychodiagnosis: Selected Papers. Minneapolis: University of Minnesota Press, 1973. Includes Meehl’s American Psychological Association address on schizophrenia, Cronbach and Meehl, Meehl and MacCorquodale, and the “Case Conferences” chapter.
“Hedonic Capacity: Some Conjectures.” Bulletin of the Menninger Clinic 39 (1975): 295–307.
“The Selfish Voter Paradox and the Thrown-Away Vote Argument.” American Political Science Review 71 (1977): 11–30. Meehl at the peak of his logical game.
“Autobiography.” In A History of Psychology in Autobiography, vol. 8, edited by Gardner Lindzey. Stanford, CA: Stanford University Press, 1989.
Selected Philosophical and Methodological Papers. Minneapolis: University of Minnesota Press, 1991. Collection of essential papers through 1989; contains Meehl’s main contributions to the philosophy of mind, to the critique of scientific method, and to law.
“Bootstraps Taxometrics: Solving the Classification Problem in Psychopathology.” American Psychologist 50 (1995): 266–275. Good entry point for taxometrics.
Clinical vs. Statistical Prediction: A Theoretical Analysis and a Review of the Evidence. Northvale, NJ: Jason Aronson, 1996. Reprint of the 1954 edition with a new introduction.
“Relevance of a Scientist’s Ideology in Communal Recognition of Scientific Merit.” Psychological Reports 83 (1998): 1124–1144. Discussion of the relation of politics and science.
“Cliometric Metatheory III: Piercian Consensus, Verisimilitude, and Asymptotic Method.” British Journal for the Philosophy of Science 55 (2004): 615–643. Complete discussion of cliometrics.
Cichetti, Dante, and William M. Grove, eds. Thinking Clearly about Psychology. Vol. 1: Matters of Public Interest; Vol. 2: Personality and Psychopathology. Minneapolis: University of Minnesota Press, 1991. These two volumes, Meehl’s Festschrift, survey the range of applications of his thought.
Peterson, Donald R. Tough Notes from a Gentle Genius: Twelve Years of Correspondence with Paul Meehl. Hillsdale, NJ: Erlbaum, 2005. Revealing exchange of letters between Meehl and one of his students, covering the last twelve years of Meehl’s life.
Peterson, Gail. “A Man of Many Gardens.” Behavior and Philosophy 33 (2005): 85–89. A sympathetic and personal life sketch.
Rozeboom, William W. “Meehl on Metatheory.” Journal of Clinical Psychology 61 (2005): 1317–1354. Thorough survey of Meehl’s methodological contributions.
Waller, Niels G., and Scott O. Lilienfeld. “Paul Everett Meehl: The Cumulative Record.” Journal of Clinical Psychology 61 (2005): 1209–1229. Part of a Meehl memorial issue of this journal, this biographical account also contains an analysis of the rate and content of Meehl’s reading.
David C. Devonis