For years a debate has raged within the various literatures of philosophy, psychiatry, and psychology over whether, and to what degree, the concepts that characterize psychopathology are social constructions that reflect cultural values. While the majority position among philosophers has been normativist, i.e., that the conception of a mental disorder is value-laden (Englehard, 1974; Reznik, 1987), a vocal and cogent minority have argued that psychopathology results from malfunctions that can be described by terminology that is objective and scientific (Boorse, 1997; Wakefield, 1992). Scientists and clinicians (Guze, 1992; Klein, 1978) have tended to endorse the objectivist position, with some notable exceptions (Szasz, 1983).

Objectivists assert that psychopathology is not determined exclusively by the values or norms of a particular culture or historical period and, therefore, is not simply a matter of behavioral fashion. At the core of all objectivist positions is the naturalist view that mental disorders involve a departure from the proper and normal functioning of the human organism. Psychopathology thus represents a breakdown or malfunction of some psychological mechanism that is part of the design of our species. Objectivists believe that such broken or malfunctioning psychological mechanisms cause some, but not all, instances of what is presently labeled mental disorder in the contemporary medical nosologies of DSM-IV and ICD-10. Objectivists contend that such underlying malfunctions can be discriminated from proper functioning by a value-neutral empirical analysis.

Unsurprisingly, normativists and objectivists differ notably in how they conceive the metamorphoses of such memorable psychopathologies of the past as drapetomania and masturbation. Presently, both camps regard the desires of slaves to escape captivity as appropriate and the satisfaction of sexual urges through self stimulation as clinically benign. Normativists, believing that illness categories are essentially social constructions, tend to see such bygone "mental illnesses" as evidence for the cultural relativity of psychopathology and to view developments like the depathologization of drapetomania as examples of moral, rather than scientific, progress. Objectivists, of course, do not deny that social forces determine the application of illness labels, as in the notorious case of Soviet psychiatrists diagnosing political dissidents with "sluggish schizophrenia" (Bloch & Reddaway, 1977). But objectivists attribute such retrospective psychiatric embarrassments to bad science, and conceive of the attendant nosological revisions as resulting from empirical advances, e. g., scientific progress demonstrating that there is no flawed psychological mechanism underlying the desire to escape slavery.

The conception of malfunction that is central to the objectivist position is intelligible only in relation to a conceptualization of proper or normal functioning, which is intelligible only in relation to the concept of a function. Functional explanation, with its teleological language of means and ends, is an integral aspect of the social and biological sciences. Griffiths has asserted that functional analysis has a broad purview, being applicable "whenever the aim of scientific investigation is to explain the overall capacity of a complex system"(1993, p. 410).

In the various philosophical literatures that have addressed the underpinnings of biology and of psychology, several concepts of function have been advanced. Each concept of function has its own distinctive character and implications for the way proper function and malfunction are conceived. Functional language has featured prominently in philosophical writings on psychopathology, albeit sometimes without adequate elaboration of the theoretical commitments the various understandings of function entail.

My approach in this paper will be to explore the crucial concepts of function, proper function, and malfunction. In the pages that follow I shall attempt to analyze these concepts, in order to explore the various ways that mental disorder is comprehended and to sharpen our understanding of how values figure in various concepts of mental illness.

An etiological function is defined in relation to its history of natural selection. The ascription of function is intended to explain the origin or maintenance of a trait or organ in a given population. Such functions are conceived as the effects of traits that enhanced fitness (the ability to survive and to reproduce) and, therefore, were favored by natural selection (Wright, 1976). When we describe the function of the heart as the circulation of blood, in an etiological sense, we imply that hearts exist, in the first place, because they perform that function. The related concepts "proper function" of the heart or "normal" heart derive from the evolutionary story. Normal hearts do what they were selected for and the proper function of the heart is to perform the role for which it was designed by natural selection (Millikan, 1989). An organ is functioning properly in the etiological sense when it is performing the function that caused it to exist in its present form:

The definition of malfunction that follows from the etiological concept of function is, simply, the failure or deficiency of a organ in doing that which caused it to become part of our equipment through natural selection. Etiological malfunctions are conceptually unproblematic, in that as good Darwinians, we can posit no mechanism other than natural selection that contributed to our species design and, therefore, assume that there is an objective, historically accurate answer to every question asked about human evolution. Most philosophers of biology assume that etiological functions and their related malfunctions can be determined objectively, at least in principle, by empirical evolutionary science. As a practical matter, however, determining the exact contribution of current traits to fitness in times past is, at best, an uncertain endeavor, given the rather primitive state of evolutionary science.

One effect of stipulating evolutionarily defined malfunction as a necessary condition for mental disorder is to restrict the range of phenomena to which the category applies. The etiological concept of malfunction underlies Wakefield’s (1992) harmful dysfunction theory of mental disorder. Within Wakefield’s formulation, an etiological malfunction is a necessary (but not sufficient) condition for mental disorder. He avers, correctly I believe, that many disorders recognized by Western psychiatry do not comprise malfunctions so defined. Wakefield has argued that limiting disorders to cases of malfunction not only keeps psychiatry more closely allied to the biological and cognitive sciences but also serves to protect us from such abuses as clitorectomies of masturbators and the medicalizing of slaves’ aspirations for freedom as in drapetomania.

The difficulties connected with applying etiological functions to psychiatric nosologies are numerous. What, for example, are we to make of that erstwhile mental illness, homosexuality, from an evolutionary perspective. Clearly reducing, on average, the procreative potential of any individual person with the trait, homosexuality looks like a good candidate for a mental illness from an evolutionary perspective. Homosexuality, however, could have enhanced the inclusive fitness of kinship groups, as do the sterile castes within naked mole rat and some insect populations. The trait may result from what biologists call "genetic drift," random variation that escapes natural selection. It is also possible that homosexuals may suffer from maladaptive sexual preferences, albeit caused by underlying genes that are linked pleiotropically with other fitness-enhancing traits. And last, it is still plausible and certainly possible, that the trait may arise largely because of environmental influence. This last view is that of social learning theory, that sexual preferences are acquired through learning and conditioning, just as musical tastes are.

The parsimony produced by limiting psychopathology to cases of etiological malfunction, however, is both theoretically and practically problematic. If we choose to define malfunction etiologically, many human capacities and proclivities and some mental disorders will lie outside the function/malfunction dimension. To possess an etiological function, a trait or capacity must have a history of selection, must have enhanced fitness and been selected for. It also must have a function in the evolutionary sense in order to have the capacity to malfunction.

But some of our current psychological apparatus may have no function at all in the etiological sense. It is probable that some features of the mind are what Gould and Lewontin (1979) term "spandrels." Spandrels are features without adaptive value that arise as by-products of other correlated developments that are fitness-enhancing and, hence, favored by natural selection. The mammalian navel, the white color of bones, and the human chin frequently are classified as spandrels. If there are numerous mental spandrels, as Gould (1997) has suggested, then we are replete with mental mechanisms that are indirect products of evolution that have not been selected for and, therefore, possess no etiological function. Because spandrels have no etiological function, they cannot, in the etiological sense, malfunction.

Just as we can conceive of mental disorders that involve no malfunction because they never had any evolutionary function, we also can imagine that a mental disorder might result from a mechanism functioning exactly as evolution designed it (Murphy & Woolfolk, in press). The adaptive problems which caused the mind to evolve as it did are the ones we faced in long-vanished environments, of which our minds still bear the imprints. It is entirely likely that we have mental mechanisms ("modules") which evolved to meet challenges in those earlier environments. These modules may no longer be apt to bring about desirable behavior in our contemporary circumstances.

Mental disorders characterized by fear and avoidance may be cases in which internal mental mechanisms are functioning as designed. Identification and avoidance of dangerous situations is adaptive, and the "better safe than sorry" strategy is, relative to suffering false negatives, adaptively superior across the spectrum of species. Avoidant creatures from from guppies (Poecilia reticulata; Dugatkin, 1992) to homo sapiens live to procreate another day:

The evolutionary account would have it that the proclivity to develop a conditioned avoidance of snakes, strangers, or heights is adaptive. Some evidence exists that humans and closely related species are "prepared" by natural selection to develop certain fears very rapidly (Mineka, Davidson, Cook, & Keir, 1984; Seligman, 1970. Thus it may be that at least some phobias do not represent the dysfunction of a mental mechanism but rather its functioning as per design, albeit inconsistently with the current pattern of human interests.

Various other forms of psychopathology also may represent adapive strategies that have been favored by natural selection. The social competition theory of depression (Price, Sloman, Gardner, Gilbert, & Rohde, 1994) hypothesizes that depression is an adaptive response to a fall in status or a loss in resource holding power. The fitness enhancing effects of a depression in such circumstances would include the conservation of energy, the reduction of costly behavior, the elicitation of aid from others, and the opportunity to develop new and potentially more effective tactics for gaining status. Individuals with Antisocial Personality Disorder and those with Histrionic Personality Disorder have been concpetualized as social "free riders" who seek to acquire group resources without adhereing to the norms governing the acquisition of social status (Harpending and Sobus, 1987; Stevens and Price, 1996). Histrionics and sociopaths are conceived as nonreciprocators (cheaters) who employ disingenuous but successful strategies to gain resourses and status. Finally, McGuire and Triosi (1998) argue that Attention-Deficit/

Hyperactivity Disorder and its associated aggressiveness and high activity levels may lead to reproductive success in some environmental circumstances.

Thus behavior that qualifies for a DSM diagnosis could occur in the absence of any mechanism that is, in the etiological sense, malfunctioning. In one case, spandrels, there may be no adaptive function involved at all, hence, no malfunction. In the other case, mechanisms selected by evolution may, when functioning as designed, bring about patterns of behavior that for various reasons we would choose to classify as pathological.

Many species, including our own, no longer occupy environments similar to those in which they evolved, resulting in mismatches between design and current environmental demands. Propensity functions take into account that a previously adaptive trait may no longer function to enhance fitness in its current environment, e.g., the predatory equipment of domesticated animals that no longer hunt for food. Or traits that, themselves, originally conferred either no adaptive advantage or disadvantage subsequently could become fitness enhancing in a novel environment.

A propensity function (Bigelow and Pargetter, 1987) is defined in relation to its propensity to enhance fitness in any possible environment: historical, current, or future. When contrasted with the historical view of etiological functions, propensity functions are said to be "forward-looking" in their import, in that they refer to probable future selective success rather than to a history of selection. Walsh (1996) has elaborated the propensity concept of function within the rubric of what he terms a relational function. This view is that a function is the contribution that a trait makes to fitness, but that the contribution is always contextualized within an actual or potential environment and can be specified only relative to a given environment.

If our concept of malfunction is predicated upon propensity functions rather than etiological functions, a somewhat different picture of malfunction emerges. The reason a heart is malfunctioning today is not because it fails to do what it was selected for, but because it performs in a fashion that diminishes our current capacities to survive and procreate. Because propensity functions are also environment-relative, unlike etiologically-defined malfunctions, their malfunction can be a direct result of a mismatch between environment and evolutionary design, rather than a simple failure of a mechanism to function according to its design. Propensity functions allow us to state that our antediluvian limbic systems, albeit functional and salutary in earlier ages, now malfunction and produce pathologies even though they operate as designed originally.

Because propensity functions involve the conferring of advantage relative to a specified environment, either actual or potential, categories predicated upon propensity concepts are rather malleable. Almost any trait, in some environment, might enhance the chances of survival or procreation either of individuals or the kinship groups to which they belong. The trait would be subject to malfunctioning relative to its fitness enhancing properties in that environment. For example, dark pigment would advantage individuals in tropical environments but not those who live close to the poles. In the tropics albinism might be properly regarded as a genetic malfunction, but not so in locales where pigmentation conferred no advantage.

One consequence of employing propensity functions and the standards for proper functioning that derive from them is that, to the degree that adherence to cultural values enhances fitness, proper functioning becomes intertwined with values. It is indeed likely that what would enhance fitness in current and future human environments is not independent of culture and its values. We can imagine a range of possibilities in this regard. For example, human aggressive proclivities, when they are high, might be regarded as functioning properly (in a propensity sense) within a warrior society. Pacific societies that abhor violence, however, may regard bellicose individuals as pathological.

What it takes to "get ahead" in different social networks is highly variable, frequently depending on norms and values. The propensity analysis of function introduces a degree of cultural and historical relativism into the determination of malfunction. Being labeled "pathological" may in itself influence survival chances and the ability to procreate; thus behavior might become pathological simply by being labeled "pathological." Malfunctions predicated upon propensity functions create the possibility of allowing back into psychiatric nosologies such infamous cases as drapetomania, simply on the assumption that attempting to escape slavery may have adversely affected slaves’ chances of survival.

Propensity functions have in common with historical functions that they make fitness (survival and procreation) the test. The fitness criterion is appropriate in the case of etiological functions since selection due to fitness caused the feature to be there in the first place. But how do we export the propensity model of malfunction to societies where people’s chances to survive and procreate are not systematically biased by their psychological traits? The propensity perspective suggests that no psychopathology (at least that requiring malfunction) would be present there. We, of course, easily can imagine a society where the brightest and the best (and the "healthiest") choose not to procreate or to procreate less than other groups. Birth rates in the United States and Europe already reflect this trend. Fitness seems an odd criterion to apply in these cases, but any alternative standard would surely derive from cultural values and related concepts of human flourishing.

An ahistorical function, sometimes referred to as a Cummins (1975) function, after the philosopher who proposed it, makes no reference to the history or origin of the trait or organ involved. Functions simply describe the causal relations among systems and their component parts, such that "the function of a part of a system is its causal contribution to some specified activity of the system" (Walsh and Ariew, 1996, p. 493). Within this kind of framework, it is the function of heart to pump blood, not because of its history of natural selection, but because the heart is a component of a larger system, the circulatory system, within which it plays a vital causal role.

Cummins functions are termed "interest relative," meaning that the function of a component is always relative to a given analysis of a system that comprises the component. The designation of the system may be the result of an arbitrary explanatory interest. Many different systems can be posited that concurrently contain the component. The Cummins function of the heart is to pump blood, but only within the context of explaining human physiology. Alternatively, if we wish to give an account of an electrocardiogram, we can describe a electrophysiological system wherein it can turn out to be the case that the function of the heart is to produce the electrical signals that result in EKG tracings. The heart can also function as a lethal target, one that is penetrated by bullets and arrows and thereafter exudes blood, this, if our explanation is of assassination.

The Cummins function refers solely to cause and effect within a containing system specified by an analytic inquiry. No background context of inquiry is privileged over any other, as is the case with the privileging of an evolutionary account by historical functional analysis.

Cummins functions potentially can provide great latitude with regard to the designation of function and malfunction. For example a kidney malfunctions when it fails to remove water from the blood, but its failure in a Cummins sense has to do with a preestablished systemic goal, i. e., maintaining salutary blood hydration. Its malfunction is only relative to its causal role in effecting a specified outcome within a system.

But many different kinds of outcomes are associated with living systems. Such systems subsist and procreate, but they also become diseased and die. And if our interest is in the explanation of the processes of disease and death, we may construct functional analyses that illustrate the contributions of our organs to such baneful outcomes. In this kind of endeavor various containing systems with interacting components can be described. In such analyses we can speak properly, in the Cummins sense, of the coronary arteries functioning to accumulate plaque, the lungs to host pneumococci, and of cancer cells functioning to interfere with cellular nutrition.

Because the causal con tribution of a component of a system is relative to an explanatory interest and to a context of inquiry, virtually any effect of any organ can be classified as a Cummins function. This situation has led some writers to claim that the concepts proper function and malfunction do not apply to Cummins functions. Other writers have stated that the malfunction of a Cummins function is intelligible only in relation to goals provided by background theories. Wachbroit (1994) argues that malfunction presupposes a concept of "biological normality," (as distinct from mere statistical or normality in relation to social norms). This is the conception of biological normality presumably employed in the notions of a normal heart, a normal response, or a normal environment.

Cummins functions are employed, unproblematically, in medicine and the biological sciences to describe cause and effect relationships involving organisms. There is universal acceptance of the proposition that a heart that cannot not pump blood is a malfunctioning heart and is, therefore, pathological. The consensus on such judgments of malfunction does not depend upon an evolutionary account of why hearts exist, but rather unanimity of opinion on the question of what hearts are for. What is the source of this consensus?

Cummins functions appear to be scientifically unproblematic and objective to the extent that there is agreement on the proper characterization of a system. We all want the heart to circulate the blood, whether it evolved to do so or not. What we wish to explain and why those explananda are important to us, the relation of knowledge and human interests, seem crucial here in making the functional analyses of somatic medicine appear to be evaluatively neutral. The "proper functioning" or "normality" of the body is uncontroversial to the extent that there is consensus on what constitutes health and infirmity. Human beings universally disvalue death, discomfort, and incapacity. Therefore, to the extent that somatic structures and processes are relatively invariant and reliably related to universally desired health outcomes, Cummins functional systems can be constructed within a common background of explanatory and valuational interests.

A number of writers have employed functional language to sustain what amounts to a distinction between theoretical and practical concepts of mental illness. In his early work, Boorse distinguished between disease and illness. The former concept, he argued, could be purely theoretical, given a value-free definition based on the concept of malfunction and grounded in the science of pathology. Illness, for Boorse, was value-laden a practical concept intertwined with cultural criteria for administering health care. The disease/illness distinction has also been employed in medical anthropology, notably by Arthur Kleinman (1986). Wakefield distinguishes between malfunction and disorder, where disorder is malfunction plus a negative value judgment. Each of these views assumes a value-free realm of function and malfunction that is independent of human interests.

Psychological theories of mental illness that employ functional analysis require us to conceive of mental modules (since an undifferentiated mind is too complex for a functional analysis) and thereby may contribute a successful cognitive science. Recent studies of autism (Leslie, 1987) and psychopathy (Blair, 1995) point to the empirical utility of an evolutionary/modularity framework. But to the degree that we attempt to make etiological malfunction a necessary condition for psychopathology, we negate a postulate of writers as diverse as Freud, Cannon, and Selye. This view, that has guided psychiatry and clinical psychology for a century, is that our species is in many respects not cut out for the contemporary world and that many of our emotional difficulties stem from the stresses placed by modern society upon a stone age psyche reacting as it was designed to.

When malfunction is defined in relation to either propensity functions or to Cummins functions, the distinction between theoretical and practical concepts of illness becomes difficult to sustain. Since propensity malfunctions imply lack of success within a given environment, they seem to resuscitate for psychopathology that shibboleth, "adjustment to society." In the case of Cummins functions, we require background assumptions concerning biological or psychological normality or desirability to impute malfunction to a component within a system. These assumptions likely will emanate from the practical arena of clinical medicine or from the culture at large.

A perennial issue in medicine has been the relationship of mental illness to somatic illness. Szasz (1983), it will be recalled, wanted to draw a sharp distinction between most psychiatric conditions and physical diseases, a distinction that corresponds to mind versus body and responsible agency versus involuntary mechanism. Inveighing against the original, psychoanalytically-based medical model of mental illness that analogized between dysfunctional psychodynamics and dysfunctional somatic processes, Szasz contended that most psychopathology does not belong in the clinic since it is, along with criminality, nothing more than social deviancy. Exceptions to this formulation were made for disorders such as general paresis, to which Szasz was happy to give the disease label, since it resulted from a physical process, or in Szaszian parlance, a lesion.

Ironically, the contemporary biopsychiatric medical model is entirely consistent with Szasz’s views, in that it substitutes for the figurative relation (i.e., mental illness is like physical illness) a literal identity relation (i.e., mental illness is physical illness). For biopsychiatrists mental diseases are bona fide diseases. In the world view of biological psychiatry, conditions such as schizophrenia are diseases, in the way that tuberculosis and cancer are, in that each involves one or more malfunctions of physical processes that eventuate in disease symptomatology (Woolfolk, 1998). Depression and diabetes are categorical birds of feather, both straightforward medical disorders resulting from metabolic imbalances that readily are remedied through biomedical technology.

A malfunction/modularity-based model, no matter what definition of function we employ, allows us to place mental and somatic disorders on the same conceptual footing. This is achieved without the additional premise of biological psychiatry, that a materialistic somatic etiology must be specified. A mental organ, module, or device can malfunction, just as a bodily organ can; we only need be able to give it a scientifically acceptable characterization and attribute a function to it. Contemporary cognitive science approaches psychopathology in just this fashion (Murphy and Stich, in press; Pinker, 1997).

References

Blair, R. (1995) A Cognitive Developmental Approach to Morality: Investigating the Psychopath. Cognition, 57:1-29.

Bloch, S. and Reddaway, P. (1977) Psychiatric Terror. New York: Basic Books

Boorse, C. A Rebuttal on Health. In J Humber and R Almeder (Eds). What Is Disease? Totowa, NJ: Humana Press.

Cummins, R. (1975) Functional Analysis. The Journal of Philosophy, 72:741-765.

Dugatkin, L. (1992) Tendency to inspect predators predicts mortality risk in the guppy (Poecilia reticulata). Behavioral Ecology, 3:124-127.

Engelhardt, H. (1974) The Concepts of Health and Disease. In T Engelhardt and S Spicker (Eds). Evaluation and Explanation in the Biomedical Sciences . (pp. 125-141) Reidel: Dordrecht.

Gould, S. (1997) The exaptive excellence of spandrels as a term and prototype. Proceedings of the National Academy of Sciences of the United States of America, 94:10750-10755.

Gould, S. and Lewontin, R. (1979) The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme. Proceedings of the Royal Society, London, 205:581-598.

Griffiths, P. (1993) Functional Analysis and Proper Functions. British Journal for the Philosophy of Science, 44:409-422.

Guze, S. (1992) Why psychiatry is a branch of medicine. New York: Oxford University Press.

Harpending, H. and Sobus, J. (1987) Sociopathy as an Adaptation. Ethology and Sociobiology, 8: 63S-72S.

Klein, D. (1978) A proposed definition of mental disorder. In D Klein and R Spitzer (Eds). Critical Issues in Psychiatric Diagnosis. (pp15-40) New York: Raven Press.

Kleinman, A. (1986) Social Origins of Distress and Disease: Depression, Neuroasthenia and Pain in Modern China. New Haven CT: Yale Univ Press.

Leslie, A. (1987) Pretence and Representation: The Origins of "Theory of Mind". Psychological Review, 94:412-426.

Marks, I. and Nesse R. (1994) Fear and Fitness: An Evolutionary Analysis of Anxiety Disorders. Ethology and Sociobiology, 15:247-261.

McGuire, M. and Triosi, A. (1998) Darwinian Psychiatry. New York: Oxford University Press.

Millikan, R. (1989) In Defense of Proper Functions. Philosophy of Science, 56:288-302.

Mineka, S., Davidson, M., Cook, M., & Keir, R. (1984) Observational Conditioning of Snake Fear in Rhesus Monkeys. Journal of Abnormal Psychology, 93:355-372.

Murphy, D. and Stich, S. (in press) Darwin in the Madhouse: Evolutionary Psychology and the Classification of Mental Disorders (to appear in Evolution and Cognition (Caruthers Ed)

Murphy, D. and Woolfolk, R. (in press). The Harmful Dysfunction Analysis of Mental Disorder.

Neander, K. (1991) Functions as Selected Effects: The Conceptual Analyst’s Defense. Philosophy of Science, 58:168-184.

Pinker, S. (1997) How the Mind Works. New York: Norton.

Price, J., Sloman, L., Gardner, R., Gilbert, P., & Rohde, P. (1994). The Social Competition Hypothesis of Depression. British Journal of Psychiatry, 164: 309-315.

Reznik, L. (1987) The Nature of Disease. London: Routledge and Kegan Paul.

Seligman, M. (1970) On the Generality of the Laws of Learning. Psychological Review, 77:406-418.

Stevens, A. and Price J. (1996) Evolutionary Psychiatry. New York: Routledge.

Szasz, T. (1983) Thomas Szasz: Primary Values and Major Contentions. R. Vatz and L. Weinberg (Eds.). Buffalo: Prometheus Books.

Wachbroit, R. (1994) Normality as a Biological Concept. Philosophy of Science, 61:579-591.

Walsh D. (1996) Fitness and Function. British Journal for the Philosophy of Science, 47:553-574.

Walsh, D. and Ariew, A. (1996) A Taxonomy of Functions. Canadian Journal of Philosophy, 26:493-514.

Wakefield, J. (1992) The Concept of Mental Disorder: On the Boundary Between Biological and Social Values. American Psychologist, 47:373-388.

Woolfolk, R. (1998) The Cure of Souls: Science, Values, and Psychotherapy. San Francisco: Jossey-Bass.

Wright, L. (1976) Teleological Explanations: An Etiological Analysis of Goals and Functions. Berkeley: University of California Press.