The first, and perhaps the most important, thing to be said about this book is that its title is misleading. Social Causes of Illness is not about whether social habits, customs, conditions, or structures cause illnesses—or whether social changes are reflected in changes in the incidence of illnesses. It is about the theoretical and epistemological implications of the fact that many physical illnesses seem to occur shortly after a significant event in the patient’s life, particularly after one which has changed his style of life. “Recently,” Dr. Totman writes,

a considerable body of evidence has accumulated which suggests that many serious illnesses, as well as being related to acknowledged physical causes, occur with some regularity following particular types of life experience. The significance of this evidence is not to be under-estimated. For if it is true that specific circumstances can render a person susceptible to illness, we are forced to recognize a degree of control over health prior to the point at which the accredited physical “causes” can exert any influence.

The physical aspects of a disease—the organic manifestations we have come to regard as the disease itself—may, in fact, represent a comparatively late stage in an “underlying” dysfunction. If this is the case, then treating illness by conventional methods (with drugs, surgery, radiation, etc.) is merely scratching the surface of the problem. Physical intervention may be attending to the symptoms of the basic affliction rather than to the real cause.

Dr. Totman’s book is, in fact, an essay on the theory of psychosomatic disease—not a contribution to the sociology or politics of medicine. It is an attempt to formulate a structuralist theory of mind which would make it possible to conceive how a social, psychological event, e.g., marriage, divorce, bereavement, retirement, could act as the cause of a physical illness, e.g., cancer, heart disease, or hypertension, and how there can be such a thing as “a degree of control over health prior to” the moment at which physical causes begin to operate.

At this point it must be stated that Dr. Totman is a clinical psychologist not a physician, a PhD not an MD, and that, therefore, the experience which has persuaded him of the reality of psychosomatic illness must be very different from that of those physicians and medically qualified psychoanalysts who introduced and popularized the concept during the early years of this century. They would, I imagine, have based their conviction of the reality of psychosomatic illness on their clinical experience, or perhaps on their philosophical position. Totman bases his on his familiarity with the accumulating statistical evidence provided by papers written, mostly, by psychologists, notably by R.H. Rahe and G.W. Brown, both of whom have developed sophisticated interviewing techniques for assessing the incidence and quantifying the impact of disturbing life events.

This evidence is, indeed, most impressive and it has, I think, to be taken as fact that human beings are at greater risk of developing an illness, physical or mental, shortly after an upheaval in their lives than they are when living on an even keel or in a rut. Previous researchers seem, on the whole, to have assumed that this period of greater risk was an effect of loss and bereavement, but Totman follows Rahe in insisting that the crucial aspect of the risk-increasing “life experience” is its tendency to cause “psychological upheaval” following “social change and reorganization.” According to their view, getting married, getting a new job, being released from prison are potentially risky upheavals, though not, it appears, as risky as being divorced, bereaved, or retired.

This evidence is, however, all statistical and a matter of probabilities, and Totman’s structural theory has therefore to take account of the fact that not everyone who undergoes an upheaval such as divorce or retirement falls ill. It even has to contend with the fact, though Totman does not even mention the possibility, that upheavals, even losses and bereavements, may sometimes lead to an improvement in health—I imagine we all know people who have taken on a new lease of life after losing, and mourning, a spouse or being sacked from a job. As a result his theory has to postulate some factor or attribute in people which, if present, may counteract the increased susceptibility to illness that follows an upheaval.

In Totman’s view only a structural theory is capable of explaining why some people fall ill after an upheaval while others do not. A structural theory is “based on the premise that a person’s understanding of the social world derives from an underlying structure of relations, and that his own behavior is the product of a system of social rules.” Claiming to follow the same line of reasoning that led Chomsky to propose that children learn, not sentences, but sets of rules by which sentences can be generated, Totman argues that human behavior is governed by a “deep structure” of standard rules by which all the specific actions that constitute behavior are generated. This way of viewing behavior “assumes the existence of some unconscious organizing system [which] functions rather like a computer programme, determining and overseeing an end-product. In computers, the end-product is printout, or a visual display. In people it is actions.”


Before we consider in any detail how this assumption enables Totman to generate a theory of psychosomatic illness, it should be mentioned that structuralist psychology appears to be a reaction against both psychoanalysis and behaviorism. Although it assumes unconscious mental activity, this activity is conceived in cognitive not conative terms—i.e., it is concerned with categories and rules, not with drives. Although structuralist psychology regards behavior as an interaction with the environment, it conceives this interaction in terms of “input” and programmed “output,” not in terms of stimulus and response. Totman’s unconscious, however, resembles Freud’s in being regarded as the mediator between the mental and the physical. He asserts that his structural model is “potentially, a scheme of physiological, or brain, activity,” since it is couched in terms which could “lead on to physiological hypotheses should it prove successful in accounting for the facts at a physiological level.”

The key concept in Totman’s account of how a structuralist theory can explain psychosomatic illness is that of cognitive consistency. If I understand him rightly, he postulates that the “unconscious organizing system” monitors all actions with an eye on their consistency with the individual’s established set of rules. If it observes consistency, the rules are reinforced and “clarified”; if it observes inconsistency, it continues to compare the inconsistent action with the rules until it has found a way of reconciling them or of altering the rules to make them consistent with the action. In either case the individual will eventually, unless illness supervenes, reach that peace of mind known as cognitive consistency.

A curious feature of Totman’s argument is that it implies that normally all actions are sooner or later assimilated (or justified) and rendered consistent with the rules, even if these have to be altered, and that they are never regretted and a source of guilt or remorse. In view of the fact that both the religious and the psychoanalytical traditions have taken seriously the idea that physical illness can be the result of denied guilt, it is significant that Totman’s deep structure is of a kind that does not and, for reasons that will become clearer as I proceed, never could generate regret, contrition, guilt, or forgiveness.

Normally, then, people aspire to and eventually attain cognitive consistency, though circumstances (but not, I think, inherent or deep-seated defects in the organizing system itself) may render it more or less difficult to do so.

On the one hand, there is the “ideal” situation, where an individual is habitually “outputting” actions (either in performance or by public endorsement) which, on testing against his system of rules result in a high proportion of good matches, and the minimum of modification. And, on the other hand, there is the situation in which radical modification of rules is often necessary before consistency is achieved. The former will be associated with constant, solid, habitual social behaviour, and probably a sense of personal identity and worth (with each occasion of consistency further clarifying and consolidating rules). The latter, on the other hand, is likely to be associated with inconstant, fitful, and uncertain social behaviour, and a sense of personal ineffectiveness, alienation, and “loss of identity.”

As readers who have persevered with Dr. Totman and me will already have appreciated, the next and crucial step in his argument is that “susceptibility to disease is increased when the ‘ideal’ situation fails to hold. More precisely, the likelihood of symptoms appearing is increased in the absence of frequent registrations of consistency.” Such an absence may, it seems, occur under three conditions.

First, the rules may be biologically unrealistic, in the sense of demanding either more inhibition or more activity than the organism can maintain, in which case the individual will have to break his own rules in order to survive. Dr. Totman does not give examples of this category, which is perhaps not surprising, since to have done so would have confronted him with awkward questions about the origins in the “unconscious organizing system” of the capacity to generate self-destructive, biologically unsound rules—and in any case this category seems to run counter to his general tendency to assume that rules are altered to justify actions and not the other way round. But I presume him to mean ascetics, both sexual and dietary, and people who drive themselves too hard. This category, incidentally, enables Totman to subsume within his own theory some of the insights of an earlier generation of researchers who found correlations between susceptibility to illness and various personality traits, notably those of rigidity, perfectionism, and the tendency to self-sacrifice. He is, however, dismissive of all theories of personality.


Secondly, “frequent registrations of consistency” between one’s output of actions and one’s inner rules may cease if there is a change in a person’s social environment such that the familiar, well-practiced rules become inapplicable and the adoption of new, unfamiliar rules becomes necessary. This is the state of affairs to which Totman’s theory applies most neatly, since the kinds of life event which, the statistical evidence suggests, do cause physical illnesses, e.g., bereavements, are generally accompanied by a need to abandon old, familiar rules and to acquire new and unfamiliar ones. But Dr. Totman’s thesis really implies, as he himself says, that upheavals people wish for also increase the risk of falling ill. In fact, however, he adduces no evidence suggesting that getting married, coming out of prison, being promoted at work—to cite examples he himself mentions as entailing “major and abrupt revisions of well-practiced rules”—cause illnesses.

Furthermore, his objection to the idea that loss can be a cause of illness is ultimately a logical one: loss produces grief, grief is an emotion, emotions are subjective states, and subjective states cannot logically be adduced as causes of organic states. How Dr. Totman has survived long enough to write a book without discovering that emotions are simultaneously subjective and organic states—and could, therefore, be used as the keystone of a theory of psychosomatic interaction—strikes me as entirely mysterious. So also does his idea that cognitive concepts such as rules and categories can be included in the “same causal account” as organic states, when affective and conative concepts cannot. Has he never seen guilt, or despair, or grief, or even happiness written on someone’s face or sculpted into his posture?

Thirdly, Dr. Totman argues that the ideal situation of health-maintaining cognitive consistency may be undermined by lack of opportunity to perform “outputting” actions which consolidate and clarify the established rules or to engage in rule-affirming conversation with at least one other like-minded person, as will happen if a person becomes socially isolated or moves into a setting in which his familiar rules come under attack. The reason for this, Dr. Totman says, is that structural rules are ultimately social rules and require regular endorsement from the group of which the individual is a member and within which his rules have been acquired. In other words, each person’s “unconscious organizing system” is part of a “social communication network” and is incapable of functioning smoothly and consistently in the absence of at least one other unconscious organizing system with which it interacts. Human beings differ, it seems, from the computers to which Totman’s model likens them in requiring input from like-minded, or rather like-programmed, others, without which they are at risk of falling ill. “Most resistant to illness is the socially involved individual: the person who is well adjusted to a stable role within a supportive community.”

This is a view of human nature which precludes the possibility that individuals may generate their own inconsistencies or that social groups may generate what Marxists call internal contradictions. Totman’s assumption seems to be that there are such things as stable roles and static communities, and that problems only arise when people are compelled by circumstances to move from one stable setting to another, when they have to learn new sets of rules. There is, in fact, something undynamic, undialectical, and mechanical about his conception of both individuals and groups. They function smoothly so long as they are not interrupted, disturbed, or transplanted.

As a result “the preventive and remedial measures” Totman recommends in his last chapter are banal and superficial. They lack the bite of radical criticism of society that would be indicated if psychosomatic disease really is as common as he believes it to be and has the cause he believes it to have. He confines himself to such speculations as the suggestion that meditation may “induce temporary plasticity in the neural structures responsible for rule-following, thereby facilitating adaptive modification of rules” and such platitudes as suggesting that people should prepare themselves psychologically for predictable changes in their life styles, and that hospital regimens should not, as indeed they all too often do, teach patients the hospital’s set of rules all too well and undermine the set of rules they will need again after recovery and discharge.

Another result of Totman’s static, undynamic approach is that he entertains seriously the idea that psychosomatic illness has evolutionary value by eliminating those who are “socially uninvolved,” who “contribute nothing to the community’s cohesion, and in all probability detract from it in that they draw on its resources.” “It could therefore be conjectured that psychosomatic disease represents an automatic ‘self-destruct’ mechanism the purpose of which is to protect the species by selecting against individuals who, as a result of social change, become socially redundant.”

Totman himself admits that he finds this idea “morally and emotionally repugnant,” but he must, I think, be convicted of intellectual naïveté in supposing that it is possible to construct a purely cognitive model of mind based on mechanical analogies, excluding all conative and affective considerations, without coming to conclusions that are morally and emotionally repugnant. His sin is one of omission. If will and desire, guilt and grief, joy and sorrow are excluded by fiat at the onset, the resulting structural model cannot but be mechanical, lifeless, and inhuman. If Dr. Totman’s book is a representative example of structuralist thinking applied to psychology, God help us—the psychologists won’t.

This Issue

April 3, 1980