The panel The interview panels differ in style and substance between schools but typically consist of three or four members of staff and often a student. The

panel is a mixture of basic scientists, hospital consultants, and general practitioners, one of whom, often the dean or admissions tutor, will take the chair. Members of panels attend in an individual capacity and not as representatives of particular specialties. They know that medicine offers a wide range of career opportunities, that most doctors will end up looking after patients but not all do, that more will work outside hospitals than in, and that both the training and the job itself are demanding physically and emotionally. They also know that whatever their final occupation doctors need to make decisions, deal with uncertainty, and communicate effectively and compassionately with patients and colleagues alike as well as maintaining moderately exacting academic standards. The aim is not to pick men and women for specific tasks but to train wise, bright, humane, rounded individuals who will find their niche somewhere in medicine. The format may be formal, with the interview conducted in traditional fashion across a large table, or more informal, sitting in comfortable chairs around a coffee table by the fireside. The tenor of the interview, however, depends much more on the style of questioning; no matter how soft the armchairs are, they can still feel decidedly uncomfortable if you are made to feel like you are being grilled and about to be eaten for breakfast.

H. Mental Illness Mental disease is a subcategory which is itself radial with core members which are generally accepted as pathological (although not universally so) and peripheral members whose status is disputed. The core members are the psychoses and the profound mood disorders. Most neuroses, phobias and panic disorder are intermediate, and entities like “social anxiety disorder,” “attention deficit hyperactivity disorder,” “oppositional and defiant disorder” and “adolescent adjustment reaction” are more marginal and controversial. With a “mental illness:” (1) Premorbid health is suspected to have been flawed. (2) Most cases are chronic and/or relapsing. (3) Often the causes are not thought of as single (except in cases such as mercury poisoning or acute drug psychosis), but are multi factorial or unknown.

(4) They may originate “inside” the person as with neurotransmitter imbalances or “outside” in the case of traumatic experience. The patient, under the influence of the disorder, typically does not view its causation as does inter subjective community consensus. In psychoses, the commonly accepted distinction between the “inside” and the “outside” of the person has deleteriously altered, reducing the ability to function socially and survive. (5) The cause is not stereotypically physical, but is usually a matter of vigorous contention among all the parties concerned. There is a longstanding schism in the health professions themselves about the role and interaction of “organic” versus “psychological” causation for most of these conditions.

There is not even any general consensus about the distinction between these terms.

I claim, subject to empirical confirmation, that the prototypical, central “diseases,” at least in Anglo American folk culture and scientific medicine, are such entities as “pneumonia,” “colds,” “bladder infections,” and “gastroenteritis” (”stomach flu”).

Close to, but slightly off center are “heart attacks,” “appendicitis,” “strokes,” and “cancer.” Fanning further out we encounter the “chronic diseases,” “mental illness,” and “dementias,” with instances like “latent” or “asymptomatic disease,” “learning disabilities,” “attention deficit disorder,” “sexual addiction,” “character disorders” and “genetic carrier states” at the very margin.

To confirm or reject this hypothesis it would be necessary to study both lay people and medical professionals to discover what names come to their minds most readily as representative examples of disease, how they think and reason about diseases, and how quickly and easily they assent to the inclusion of any givenentity in the “disease” category. Also, if category membership in “diseases” were presented to an experimental group as a matter of degree this should result in some informative rankings.

CENTRAL MEMBERS OF THE DISEASE CATEGORY “Pneumonia” and other central members share a number of features which give them their central place and vividness for identification as diseases. I have come up with a list of 13 features which characterize disease in contrast to health. These features may not be the only significant ones, but I think they are sufficient for picking out the most readily accepted examples of disease. The central, prototypical ensemble of these features makes the sharpest possible distinction of a disease from health. Clinical entities manifesting the features of core diseases stand out to be grasped readily and are blatant, as opposed to the less overt and more subtle features of less exemplary “diseases.” As noted above, these central diseases share little with the non disease categories which are adjacent to and overlapping somewhat with “disease.” They are relatively pure examples. Furthermore, they are readily suited to at least one or more of the important ideal cognitive models of disease, such as Being Under Attack. Parenthetically, note that the worst diseases, such as rabies, pancreatic cancer and AIDS are not necessarily the most prototypical.

As we have already seen, there is no classical criterion, no univocal set of necessary and sufficient features to define disease literally. On every level there is ambiguity, overlap and uncertainty. Depending on the vagaries of ongoing research, academic fashion and the mutually contradictory pronouncements of authorities at different times and in different places, category assignments shift, drift and are often in dispute. There is very little about this whole system which accords well with classical category structure.

The overall “disease” category is radial, not classical. Central members of this category are extended by cognitive proximity, analogy and metaphor to increasingly peripheral examples. If a history of disease identifications were undertaken, I suspect that the central prototypes would be found to have been the first ones labeled as “diseases.” The most central and exemplary diseases are those best exemplifying the main idealized cognitive models. Analogies and metaphors act cognitively like forces (such as gravity) or links in that the easily identified, clear cut central members present a cognitive pull on marginal examples, drawing them into association. At the very margins of the general “disease” category the most peripheral examples wobble in their orbits, so to speak, partially gravitating toward other large categories in the lexical neighborhood of disease: “old age,” “weakness,” “crime,” “harm,” “suffering,” “eccentricity” and “infertility.” The best examples of disease are the ones farthest from these adjacent categories (although they may be excellent examples of “suffering,” which is overarching, not just overlapping).

This work of Rosch has been amply confirmed and extended to many classes of categories. It has also been greatly refined and elaborated by Rosch herself, going far beyond the simple summary of her findings which is most pertinent here.

Representative members of categories are metaphorically placed in the center of a two dimensional category space, although three dimensional spaces representing categories and their neighbors seem possible. Less and less representative members are imaginatively farther and farther away from the center, giving the categories a radial structure. However, typicality is not the only feature of category members which accords them differential significance in reasoning.

There is also the ideal prototype. Consider your own concept of a typical doctor and then your concept of an ideal one. The ideal doctor is selfless, always available, calm, caring, intelligent and well informed. The stereotypical one ismore likely thought of as rich, intelligent but arrogant, intemperate, ambitious and emotionally distant. And then there are salient members of a class: particular ones coming to mind because of recency (you heard of them lately) or primacy (you heard of them first) effects, or something else causing them to be especially vivid in the imagination: Hippocrates, Everett Koop, Jocelyn Elders, Michael Debakey, Jack Kevorkian, your childhood doctor. In these and many other ways categories have texture which affects reasoning about them and about individuals as members.

• There are levels of categories. The “basic level” consists of middle sized

enduring objects and vivid, relatively discrete actions or states of being with which we are intimate early and throughout life, with which we deal more facilely, and which are the most accessible and recurrent entities in bodily experience. Ask someone under no particular mandate to describe objects in a waiting room and she or he will usually respond on the basic level, viz. chairs, tables, a desk, the counter, lamps, people and magazines. These are default, path of least resistance answers.

Fuzzy set theory, developed by Lofti Zadeh, has enabled the assignment of numerical values to partial category membership, expanding set theory applications. Additionally, probability numbers could be assigned to set membership. The flexible adaptation of amendments to a fundamentally mechanical theory, however, requires the use of judgment, which is none other than informal reasoning from experience.

• Categories are textured; they have an internal terrain. We manifestly do not

treat all their members alike and there are good reasons why. Eleanor Rosch discovered prototype effects. In a graded and indistinctly bounded category like “tall men” taller ones (unless exhibiting clear cut pathological features) are the best examples. But even categories usually taken to be well defined (they are not, really) like “bird, a feathered biped” have more or less representative, salient and ideal members as identified in studies of people using and dealing with the categories. In Western culture robins and sparrows are more representative of birds than emus and penguins. Desk chairs are more representative, prototypical chairs and come to mind more easily as examples than do dentist’s chairs and bean bag chairs.

Rosch found evidence that people rate certain members of categories as the better examples of those categories. Experimental subjects identified such prime examples as category members more rapidly than they did the poorer examples.

For instance, subjects would more quickly identify a chicken as a true bird than an emu. Also, when asked to come up with an example of a bird, robins were given much more readily than, for instance, penguins. And she found that when judging similarity, there were asymmetries: penguins were thought of as more similar to robins than robins to penguins. Furthermore, when new information was introduced about a prototypical category member, this information was more likely to be thought of as applying to all the members than when it was first revealed about a less representative member. Thus prototypical category members carry more weight in determining our general sense of the category than do less typical ones.