Defining who is normal and who is mad is fraught with peril. The modern history of insanity begins in the thirteenth century when “leprosy” became a popular catch-all diagnosis. At one time there were as many as 19,000 leprosariums in Europe alone. By the middle of the fifteenth century however leprosy had spontaneously regressed. It was replaced in the Renaissance by the “madman” who typically was a poor vagabond, criminal or other miscreant. They were expelled by villages and took to wandering the countryside where they became a public nuisance (much like today’s homeless population). In fact the phrase “ship of fools” is not metaphorical. Etymologically it refers to an actual boat onto which communities loaded undesirable persons and then set them loose to glide along Europe’s rivers and canals. The institution of the “madhouse” first appeared in the seventeenth century as a way of restraining free-roaming vagrants, in response to population growth and changing land-use pressures. It was deployed primarily as a means of law enforcement. At one point one out of every hundred residents in Paris was confined to a madhouse. They made no pretense of being medical institutions but rather simply were a way to warehouse the indigent and the unemployed and remove them from public view. The madhouse then disappeared at the end of the eighteenth century to be replaced by the “workhouse.” The labor of the inmates was not particularly valuable from an economic standpoint. Rather, reflecting the mores of the time, it was thought to be “good for them.” It was not until the nineteenth century that a sharper distinction was made between those who should be in insane asylums (which acquired an ostensible medical purpose and thus a veneer of respectability) and those who simply were down on their luck. Most of the latter were absorbed by the European wars of the nineteenth century (with an increased demand for soldiers of whatever mental status to act as cannon fodder). Source for this brief history: Michel Foucault, Madness and Civilization (1965).
Foucault’s point is that insanity is a complex phenomenon rooted in art, religion and tradition. Phenomena that qualify as socially inappropriate or socially deviant under one set of cultural conditions or constraints may be acceptable under others. The main factor affecting the concept of what counts as deranged is the community’s altering perceptions of just what madness is. Foucault highlights the first main issue with understanding aberrant behavior and determining the presence of psychopathology, which is that it is culturally relative. The literature abounds with illustrations of culture-bound mental illnesses such as amok, dhat, grisi siknis, koro, kundalini, latah, piblokto and windigo. More recently there are new unexplained illnesses such as gambling, obesity, sex addiction and “internet addiction,” which per se could not exist in environments without the internet. Is there enough empirical evidence for them to meet the threshold for what counts as mental illness, or do they just reflect modern times?
The second main issue is the problem of “fit.” All theories or sets of rules have four possible orientations towards facts: good positive fit, good negative fit, bad positive fit or bad negative fit. Thus for example one might say a set of diagnostic criteria supplies good positive fit to the extent it correctly identifies patients who actually have the underlying pathology and bad positive fit to the extent it fails to do so. It has good negative fit to the extent it excludes inappropriate candidates and bad negative fit to the extent it accommodates them.
Fit problems typically occur with principles or standards that are operationally non-specific or insufficiently specific. They either are vague or allow too much interpretive discretion in their application by those implementing them. Fit problems abound with psychopathological definitions where the criteria for specific mental illnesses are loosely grouped around imprecise symptom clusters. This allows for conflicting diagnoses, overlapping diagnoses (comorbidity) and multiple diagnoses (different disorders or categories share the same symptoms), none of which are desirable from a scientific perspective. It also creates the risk of medicalizing the normal range of human behaviors.
Closely related to fit is the problem of empirical validation. Many psychopathological interpretations of human behavior originated with vague psychoanalytic doctrines, which have no scientific or empirical basis. (1) They do not meet the criteria for what counts as a scientific theory such as potential falsifiability and internal consistency. At best they are “symptom-based,” not premised on neurophysiological or even behavioral outcomes. Others were developed from limited population samples and then wildly generalized into overly broad theories, which then gradually recruited adherents and constituents until they became generally accepted. It is a logical error though to extrapolate from individuals or small groups of individuals to communities at large. To the extent it works at all, quantification works the other way around (that is, from group to individual, not vice versa). In statistics such an error of interpretation is known as the “ecological fallacy” – that all members of a group exhibit or display the characteristics of the group at large. It also might be characterized as a form of “fundamental attribution error,” that is, analyzing psychological phenomena in terms of individual or cognitive explanations as opposed to situational ones.
A fourth problem with diagnostic criteria lies not so much with the criteria themselves but rather with the interpretive stance of the person applying them. Some patients present with overwhelming evidence of a particular type of cognitive, emotional or social dysfunction or distress. They may be experiencing subjective mental trauma or severe socio-pathological adjustment issues. Their behavior may be maladaptive resulting in learning disabilities, unemployment or even criminal behavior. Because they comprise the core statistical population these patients do not present a diagnostic issue.
One or two standard deviations away however the application of diagnostic criteria becomes more imprecise and elusive. The difficulty in administering them consistently introduces (among other factors) etiological bias, which is connecting specific behaviors with race, gender, age, physical health or SES. It allows for observer bias, which is associating them with particular features of the clinician’s background, training and experience. (2) It creates assessment bias, which is the dynamics of the clinical situation in which the data is extracted, including extent of trust, the existence of a power (authority figure-subordinate) relationship, interviewer style and technique. Selection bias (subjects in the sample population may be unrepresentative of the population at interest or the population at large) also contributes to the indeterminacy of results. It is very difficult to design a psychological experiment along the scientific model of a double-blind placebo study, which is the standard for medical research. Virtually all psychological studies rely either on self-reports, observer-subject interaction or experimenter interpretation (e.g. Rorschach tests or thematic apperception tests, which elicit interpretations of evocative images). Because they are inherently subjective the results of these studies often are non-replicable and therefore non-generalizable.
All of these problems affect the DSM. The DSM did not even exist until 1952. It has been through five revisions since. Some revisions have added mental disorders (such as BPD); some disorders have been dropped (such as homosexuality). The DSM attempts to mitigate against the problems outlined by adopting a five-axis approach and more explicit time-based and quantitative criteria (e.g. a symptom must be present for at least a year, or five out of eight symptoms must be present for a diagnosis). Often however these import just as many difficulties as they purport to resolve (e.g. what happens if a symptom lasts for a short period under the minimum length or if four out of eight symptoms are present instead of five). The presence or absence of a particular symptom is neither necessary (must be there) nor sufficient (in and of itself constitutes a diagnosis) and a simplistic taxonomy of symptoms (count up how many there are) does not indicate their severity or degree of mutual interdependence or substitutability. Different patients may have the same disorder but present with a different symptom profile or different disorders but with the same symptom profile. There is the disturbing Axis V for “global assessment of functioning,” which basically is the therapist’s subjective take on the patient’s condition.
Since it is symptom-based the DSM does not offer a scientific, neurophysiological or empirical explanation as to the underlying causes of pathology. It cannot predict behavior and therefore lacks explanatory power. Just as seriously any psychiatric label carries with it the possibility of stigma and prejudice. Once diagnosed a patient may find it hard to get a job and obtain insurance or existing insurances might be canceled. Some diagnoses (such as mood disorders) are permanent. When a patient no longer exhibits the relevant symptoms then the condition is said to be in remission, but it never goes away.
Undeniably the DSM has considerable utility. It facilitates communication between clinicians; schematizes access to scientific literature; defines research populations for scientific study; and is useful for insurance and actuarial purposes. A DSM diagnosis is the key to unlocking the availability of governmental assistance that will enable (say) an autistic child to get resources from the public school system. It is an imperfect tool at best, however, given our current state of knowledge, it may be the only one available.
(1) Psychoanalysis (and psychodynamic therapy) actually has the potential to be risky because a psychoanalyst may view behavior as some kind of a personality disorder when in fact it has an organic cause.
(2) Cosgrove et. al. (2009) found that DSM authors writing guidelines on depression and bipolar disease had significant financial ties to the pharmaceutical industry.
Cosgrove, L., Bursztajn, H., Krimsky, S., Anaya, M. & Walker, J. (2009). “Conflicts of Interest and Disclosure in the American Psychiatric Association’s Clinical Practice Guidelines.” Psychother. Psychosom, 78, 228 – 232.
Foucault, M. (1965). Madness and Civilization. New York, NY: Mentor.