"Mental" and "mental illness" are social concepts
Although Kandel (2) correctly observed that "all mental processes are biological," he failed to add that all higher mental processes are also social and contextually bound. Terms that we use for mental processes such as fear, anxiety, pleasure, attention, memory, mood, and so forth are not biological but social concepts. They are intrinsically tied to the way people relate to each other (8). Cognition, memory, perception, emotion, and anxiety vary considerably from one culture or historical period to another (9,10,11). Culture is not "out there" to be treated as an independent variable, but something that impinges on a person's thinking (12).
Thus, although it is typically assumed that there are distinct, objectively existing entities that correspond to the various categories of the mind, these categories have always depended on the use to which they were put in specific social contexts (13). Mental processes must always be social, just as they must always be biologically situated because they occur in biological beings. Both are necessary; neither is sufficient.
Diagnoses in psychiatry are starkly different from diagnoses in physical medicine in which asymptomatic or incidental disorders are often uncovered, such as an examination of the prostate that reveals a malignancy. To use a biological marker to define a mental illness would mean that we must be willing to accept a definition based on an incidental abnormality that might be found on neuroimaging, even if the subject never manifested any symptoms. Simply put, is it logical to define a contented person as "depressed" based on an abnormality in brain functioning?
Indeed, we can create new disease categories based on brain findings, but these categories would no longer rely on mental concepts and therefore cannot be mental disorders. Because mental symptoms involve verbal communications and behaviors—that is, cultural understandings—psychiatric illness can never be transmuted into an anatomical or physiological entity in the same way as physical illness. As Satel (14) wrote about addiction, "You can examine brains all day, but you'd never call anyone an addict unless he acted like one."
Social forces transform biological and psychological structures
Evidence is accumulating that social and environmental forces can directly influence neural functioning, synaptic connections, gene transcription, ventricular size, and neurotransmitters (15,16,17,18,19,20). Analogous changes occur at the psychological level. Many years ago, the Russian developmental psychologist Vygotsky concluded that a person's "psychological nature represents an aggregate of social relations that have been transformed internally and that have become functions of personality and forms of structure" (21). Luria (22) further recognized that not only the content but also the actual form of psychological processes, including "higher, specifically human forms of psychological activity such as voluntary attention, active memory, and (especially) abstract thought" are altered with changes in social structures.
Social forces can alter psychiatric disorders
Cochrane (23) observed that "the fluctuations in the incidence of disorders over time, place, and social category are too large to be explained away by the differential incidence of purely biomedical imbalance." A variety of social factors have been implicated in affecting the incidence and outcome of menta1 illness. For example, social class, gender, marital status, unemployment, housing conditions, social stressors, life events, social supports, and immigrant status have all been linked to psychological ill health, often in a complex, interactive way (24).
Considerable data support historical changes in the incidence of various psychiatric disorders such as anorexia and depression, both of which have increased during the 20th century (24,25,26). Although most psychiatric disorders are thought to be universal, the prevalence rates of disorders vary considerably (24). Thus the World Health Organization Collaborative Study on Psychological Problems found that among 15 centers worldwide, the prevalence of ICD-10 psychiatric disorders ranged from 8 percent to 53 percent (27). Rates of specific psychiatric disorders varied across sites; for example, rates of depression ranged from 2.6 percent to 15.9 percent. The content and form of disorders also vary cross-culturally. Rates of visual hallucinations are higher among persons with schizophrenia who live in developing countries (28). Susser and colleagues (29) observed that "the varieties of human affliction owe as much to the inventiveness of culture as they do to the vagaries of nature."
Finally, the outcome and course of a disorder, once defined, are influenced by social forces. For example, the outcome of depression has been found to be influenced by social support, employment, and coping styles (30). Persons with schizophrenia from lower socioeconomic classes have worse long-term outcomes in developed nations than in developing countries (31,32).
In the absence of a strong complementary social component, biomedical models in psychiatry have been girded by various assumptions, the validity of which remains an open question. In this section, I describe some of the principal assumptions of the biomedical models and show how a social perspective can modify, complement, or provide a useful alternative to prevailing models.
A social perspective and Cartesian modeling
Despite dramatic strides in molecular research, efforts to integrate biological elements into aspects of human psychology have been hampered by ignorance about the processes by which such integration can occur. On one hand, it is recognized that the complex chemical reactions induced by a gene, mediated through an intricate series of life circumstances, does not equal an identification between a complex human behavior and a gene (33). On the other hand, because of the inchoate state of our knowledge, researchers typically use simplified models based on the Cartesian notion that the whole is the sum of separate, irreducible, unchanging parts (34). For example, to explain a personality trait, biological and social variables may be added together in a large multivariate analysis.
In the absence of a strong, sophisticated social component in psychiatric research, no countervailing forces to various forms of biological overdeterminism have existed. In one type of analysis, several biological variables and a few basic social variables—typically race, gender, and education—are entered into an analytic model. Limiting or simplifying the social variables attenuates their effect on the outcome variable as well as their potential to detect interactions with the biological variables.
In a second type of analysis, social variables are discarded altogether. For instance, using the modeling described above, investigators of genetic influence on human behavior, or heritability, assume that genetic and environmental elements are both additive and uncorrelated (35). If these assumptions are correct, then it is permissible to look solely at heritability and assume that anything not explained is due to environment, even though no social variables are actually measured. One byproduct of this approach has been to implicate inherited particles within the individual—atoms of behavioral propensity—to explain a particular disorder or behavior (33). Statements that a behavior or disorder is 55 percent genetic and 45 percent environmental are increasingly common and represent reductionistic thinking based on an atomistic assumption of human behavior (33).
A broader and more sophisticated social research component is needed to provide complementary variables for these analyses as well as to provide detailed descriptions of the environments from which samples are recruited. With respect to the former, social variables may modify effects attributed to biomedical variables. To cite an illustration from my own research, I found that among older persons with schizophrenia, findings about the impact of positive symptoms on depression were modified by adding to the analysis several social network variables, including network member interconnectivity and proportion of members providing material assistance (36). Merely examining the relationships between clinical variables would not have adequately explained the depressive symptoms of older persons with schizophrenia nor have taken into account the interactions between depression, positive symptoms, and these social network variables.
Likewise, with respect to environmental effects, it has been customary in twin adoption studies to assume that different family environments are not correlated with each other. However, environments are often quite similar, especially within a single country or group of countries. For instance, height might be deemed nearly entirely genetic in the United States but considered socioeconomically based in Bangladesh.
Although it is a step forward to include social variables in analyses, it is also critical to move beyond global social variables such as race or income to develop more nuanced variables. Thus it is now recognized that race may be a proxy for a variety of social experiences that affect health such as access to goods and services; residential segregation, including quality of housing and exposure to environmental factors; access to labor markets; political power; and belief systems (37). Similarly, income may not be as important as accumulated wealth, levels of neighborhood wealth, the number of persons being supported by an income, the level of social deprivation, socioeconomic status in childhood, or the method used for demarcating poverty, such as deciles, quartiles, or poverty level (38).
The inclusion of social variables does not immediately obviate the limitations of Cartesian atomistic modeling, although a deepened understanding of the relationship between biological and social variables should further the development of process models. Clearly, without such social research, no substantial progress can be made.
A social perspective and the disease model of biology
Although the decision to render symptom clusters into psychopathological categories unquestionably has increased nosological conformity and improved communication among clinicians and researchers (39), it has created a variety of problems that can be adequately addressed only by the inclusion of a social perspective in research and practice.
Biology does not necessarily have priority over social perspectives in explaining mental disorders. Little empirical support has been found for strict DSM-IV boundaries (40). As Kleinman (41) observed, "The difference between 3 and 4 'vegetative' complaints is the difference between becoming a case of depression as a disease or being an instance where depression is a socially caused type of human misery. DSM-III was not created with such problems in mind, but it is applied in such settings."
Consequently, mental illness, which was developed as an analogue to physical illness, is increasingly viewed as an underlying biological (brain) entity rather than a clustering of symptoms. Andreasen (42) wrote, "Psychiatry now recognizes that serious mental illnesses are diseases in the same sense that cancer or high blood pressure are diseases. Mental illnesses are diseases that affect the brain, which is an organ of the body just as the heart or stomach is. People who suffer from mental illness suffer from a sick or broken brain, not from weak will, laziness, bad character, or bad upbringing."
An unfortunate confusion has been created by the legitimate need to avoid the Cartesian dualism of mind and spirit in which diseases of the mind are seen as diseases of the brain (43). However, what is not always recognized in biomedical models is that psychological and social aspects of the mind are not spiritual but are likewise material states. They represent a different mode of discourse for addressing how the mind is organized. Just as physical reality can be explained only in part by physics and requires chemistry for further clarification, biology cannot fully explain the mind and requires social and psychological discourses, although the latter should be consistent with biology.
Dichotomizing mental symptoms into "disease" and "nondisease" does not mean that nonbiological factors are less important than biological ones or that biology has explanatory superiority over other ways of understanding the mind and its problems (43). In some instances, biological factors may account for proportionately more of the etiology; however, this is a researchable question that will not be satisfactorily addressed by prematurely excluding social variables from analyses.
Psychopathology cannot be divorced from social context. The psychopathology model is also prone toward decontextualizing mental disorders, and a social perspective can correct this tendency. One type of decontextualization occurs when social concomitants of a disorder are attributed to the underlying features of the illness. Thus among persons with schizophrenia who are homeless, homelessness is attributed primarily to their mental disorganization rather than to the lack of affordable housing or inadequate incomes (44).
Another type of decontextualization occurs when social concomitants of an illness are not considered relevant to the fundamental structure of the disorder. Thus although many persons with schizophrenia live in poverty or have near-poverty levels of income, few studies have examined the role that poverty plays in affecting symptomatology, course, and outcome of the disorder (45).
A third type of decontextualization involves the propensity to focus on individual risk factors and proximal causation—that is, factors for which a "biological plausibility" can be argued—and to neglect or minimize the social context of such risk factors (46). For example, investigators may examine the neurodevelopmental changes predisposing to schizophrenia, the structural brain abnormalities associated with various mental disorders, or the individual stressful circumstances that precipitate a disease episode. However, they fail to examine why people come to be exposed to risk or protective factors and fail to explore the social conditions under which individual risk factors are related to the disorder.
In other words, we must ask what it is about persons' lives that might shape their exposure to neurodevelopmental, brain, or stressful risk factors. Without comprehending the context that leads to a risk, the responsibility for alleviating this risk falls on the individual, and nothing is done to alter the more fundamental factors that put people at risk of risks (46).
The structure of a disorder is not necessarily more important than its content. Kleinman (41) observed that the model of pathogeneity-pathoplasticity "comes fairly close to being a professional orthodoxy." In this model, biology is presumed to "determine" the cause and structure of the disorder, while cultural and social factors, at most, "shape" or "influence" the content of disorder. Thus paranoid schizophrenia is a biological disorder in which the content of delusions varies between cultures. Similarly, the extensive somatic symptoms accompanying depression in some cultures are merely viewed as epiphenomena of the underlying biological disorder. However, no evidence exists that content is any more or less biological than structure. Certainly, strong evidence suggests that the "epiphenomenal" expressions of physical disorders, such as physical pain in arthritis, may be more potent determinants of disability and outcome than the so-called underlying disorder (41).
Moreover, even if there were an underlying biological disorder, it is likely to be transformed at least in part by social content because the latter can transform the circuitry of the brain. An example that tends to bolster this argument is that both the form and content of schizophrenia have changed historically—there are now fewer cases of hebephrenia and catatonia and more paranoid and undifferentiated cases (32). It is here that social research can elaborate on the precise role that content plays on the outcome and course of disorders.
A disease model should not preclude prevention and public health approaches. During the 1960s, a combination of economic and sociopolitical conditions favored the development of the community psychiatry movement, which made prevention of mental illness one of its goals. However, by the late 1970s two prominent psychiatrists, Lamb and Zusman (47), baldly stated that "there is no evidence that it is possible to strengthen 'mental health' and thereby increase resistance to mental illness by general preventive activities." They maintained that "the cause and effect relationship between social conditions and mental illness is extremely questionable'' and that "mental illness is probably in large part genetically determined and it is probably not preventable, at most only modifiable. Even that it can be modified is questioned by many and there is little hard evidence one way or the other."
Lamb and Zusman's argument (47) was based on the assumption that mental illness is analogous to physical disease. In their paper they sharply distinguished between "real" mental illness and emotional distress, and they likewise assumed that psychopathology cannot be prevented unless the specific organic cause of the mental disorder is identified. However, as previously noted, the term "mental illness" has been used analogously to "physical illness," but it is not identical to physical illness. Further, even with physical illness, the role of the social environment has been increasingly recognized as enhancing susceptibility to disease (48,49). Cassel (48) showed that psychosocial stress heightens general susceptibility to illness and concluded that "a remarkably similar set of social circumstances characterize people who develop tuberculosis and schizophrenia, become alcoholic, are victims of multiple accidents, or commit suicide. Common to all these people is a marginal status in society."
It is now appreciated that many social factors have general effects on disease and that singular, one-to-one correlations are unlikely to be identified. For example, Link and Phelan (50) have described "fundamental social causes" of disease, which involve resources like knowledge, money, power, prestige, and social connections that strongly influence people's ability to avoid risks and minimize the consequences of disease once it occurs. Because of the very general utility of these social and economic resources, fundamental causes are linked to multiple disease outcomes through multiple risk-factor mechanisms.
Regardless of what the profile of diseases and known risks happen to be at any given time, those who have greater access to critical economic and social resources will be less affected by disease. On a biological level this situation may mean that those with fewer resources may experience greater social-environmental stressors, which in turn diminish host resistance, perhaps through effects on hormonal response. In any case, such effects are likely to be general across various disorders, both mental and physical.
A focus on psychopathology should not exclude positive mental health. Despite some resistance, much of the health field has moved away from treating only disability and from defining health as the absence of disease and has moved toward improving robustness. For example, Rowe and Kahn (51) described "usual" versus "successful aging." That is, persons who exercise regularly, have strong social supports, and have more personal autonomy seem to exceed age norms on various physiological and psychological measures. The emergence of a new journal, the Journal of Happiness Studies, suggests that there is increasing interest and a body of research that looks at positive mental health rather than disease.
Psychiatry's focus on the disease model favors theories of pathogenesis based on biological deficits rather than positive attributes or strengths. For example, the proportionately greater number of women who develop schizophrenia in later life is attributed to a reduction in estrogen production or a neurodegenerative process (52,53). Using a socially oriented paradigm, one might hypothesize that psychosocial factors exist that protect older men from schizophrenia, such as their greater economic resources, their higher levels of autonomy and social power, or their greater likelihood of being married (nonwidowed).
Psychopathology does not require an underlying biological abnormality, and "toxic" environments must also be examined. The psychopathology model, with its emphasis on an underlying biological deficit, has also influenced our way of looking at how genes determine mental illness. The predominant paradigm is that abnormal genes produce mental disorders. Franche (43) observed that once we recognize that the heritability of psychopathology may be due to nonpathological traits and "normal'' wiring, the exploration of social factors becomes more critical.
For example, while fair-skinned people are more susceptible to sunburn, there is nothing pathological about having fair skin, and there is no gene for sunburn. Similarly, we do not know whether a genetic relationship with psychopathology represents a biological abnormality or a biologically well-integrated set of systems that under specifiable conditions results in distress. Therefore, genetic correlations do not tell us much unless we examine the role of environmental conditions in which genes are operating.
In a similar fashion, it is also possible that a "toxic" environment produces mental distress, and that genes and other biological factors enhance well-being by modifying the toxic effects of the environment on persons. Thus, while the disease model assumes an abnormal gene for abnormal behavior, this assumption is surely not the only way to approach this issue. Here again, more fully engaging the social world opens up alternative paradigms and new points of departure.
A social perspective and the technical approach to mental illness
As psychiatry has tried to replicate the natural sciences by turning toward what Habermas (54) described as "technical control," the cultural and moral dimensions of life have been collapsed into merely technical and instrumental considerations, or "value-neutral" knowledge. Although this approach may enhance our control or mastery over events and things, it also potentially undermines our ability to adequately assess the value of our ends and the social ideology underlying these values.
The psychiatric literature has virtually ignored the postmodernist debates that have engaged the academy, the media, and, at least to some extent, psychology and psychoanalysis (55,56,57). The postmodernist challenge has been directed at mainstream science and empiricist foundationalism—that there must be a guarantee of universal reason and truth—by emphasizing that language and knowledge do not copy, reflect, or reproduce reality, but rather, that language constitutes reality (58). Consequently, each language constructs specific aspects of reality in its own way, which is rooted in specific culture and community. Thus knowledge is saturated with perspectives that reflect aspects of culture and society (11). It further follows that because the object of study is not directly reproduced, a culture of science must also exist, with a distinct body of language, concepts, and values that interprets its objects of study.
Howard (59) observed that "although philosophers of science still debate the role of values in scientific research, the controversy is no longer about whether values influence scientific practice, but rather how values are embedded in and shape scientific practice." For example, Prilleltensky (60) described a variety of values and assumptions that underlie psychological theory and practice. Values may include how much a particular clinical approach or research agenda promotes caring and compassion, human diversity, collaboration and participation, and distributive justice. Assumptions implicit in various frameworks include notions of the good life, the good society, relationships regarding power, professional ethics, and to what end knowledge may be used.
Several writers have described how mental health professionals, in part due to their socialization and social status in society, may be prone to facilitate the societal status quo by endorsing dominant social values, emphasizing individualism that minimizes sociohistorical context, and reflecting these values in the form of "value-free" scientific statements in their research and practice (61,62). For example, biological psychiatry may claim that medications correct for deficits and restore persons to what they are supposed to be (43). However, we have no clear sense of what anyone's true nature is. Answers to this question cannot be purely technical, but presumably involve values of what it means to be human in a particular society.
Social values and relationships are also embedded in our technology (63), and a social perspective allows us to recognize and assess their impact on our theories and practice. For example, if we look at neuroimaging, we see that rather than being an objective measure used in psychiatric research, magnetic resonance imaging (MRI) or the computed tomography (CT) scan must be understood as a socially constructed technique organized in the context of the academic community's decisions about what form data can take (quantitative measures) and what technology or instruments are considered acceptable (64,65). Other factors that play a role are the idea that the knowledge product can be based on interindividual differences and pooled data (so that average data may not correspond with data of any single individual in the group) and a theoretical model that assumes that abnormalities in brain structure may produce symptoms of schizophrenia or depression (64,65). It also requires an economic system that can support the cost of MRIs and CT scans.
A social perspective and quantitative research methods
An attraction of the biomedical model has been that it lends itself to more empirical (quantitative) measurements and thus mimics the physical sciences. Quantification also suggests a greater degree of predictability and control over events. Psychiatry and psychology enhance their prestige and worth by predicting and controlling aberrant human behaviors (65). When statistical significance testing is used as a basis for hypothesis testing, theories are typically "confirmed" by rejecting the null hypothesis, no matter how small or trivial the explained variance may be (65). Such computations give the impression that lawful, universal mechanisms are being uncovered. Conversely, incorporating social elements into the analysis in any substantive way begins to increase unpredictability and complexity (66,67), and, consequently, the social sphere may be cast aside as "noise" or "unexplained variance" that requires nonquantitative, or "less precise," methods for its elucidation (64).
Moreover, in using quantitative (statistical) aggregate or group data, characteristics of the "average individual" take precedence over data about distinct individuals (67). At a practical level this approach creates problems for researchers and practitioners because the fully embodied person never resembles a list of abstract traits. As Cassel (68) asserted, doctors must "ask what it is that makes a particular individual this explicit individual rather than an abstraction." Such a model omits the concrete individual, which makes it difficult to explain how psychological, biological, or social processes account for such an individual. With respect to social processes, clinicians and researchers must attend to patients' life histories, personalities, relationships, cultural backgrounds, hopes, fears, interests, and values (68). Each of these areas contributes to the course of illness and suffering because each contributes to the meaning and experience of illness (69). Thus we cannot begin to fully develop knowledge of real people's lives without their accounts of concrete, everyday experience, although it is important to avoid biographies that are ahistorical and decontexual (70).
Rather than relying on outside researchers' assumptions, researchers must consider qualitative methods such as ethnography that are oriented toward "discovering" the social world through observation and participation in natural settings. Such research can capture social processes as they develop and change and also the social meanings that generate them (71).