The illustrations spread out on my desk look like freeze frames from some 1940's melodrama. In one, a man who might be Fred MacMurray, his brow furrowed in a way that strongly suggests limited acting range, turns his back on an elderly woman wearing a martyred expression. In another, a young woman is shadowed by a crone swathed in black. And in a third, a man wearing a tie and pleated pants covers his face as he stands beside the bed of an alarmingly inert woman, whose bare breasts are rendered in Vargas-vintage pinup style. The women remind me of Judy Garland or Gene Tierney, their hair upswept and their lipstick dark and dramatic.

The pictures are fun to look at -- the way that a Douglas Sirk movie is fun to look at -- but that isn't why most people do it. Most people look at them because they have been asked to by a psychiatrist or a prospective employer, for these are among the 20 pictures that constitute the Thematic Apperception Test, or TAT, one of the most influential and widely used of personality tests. The fact that these anachronistic images are still widely used makes you wonder not only about the TAT but also about all personality tests and what it is, exactly, that they measure.

Modern personality testing -- which encompasses everything from artsy interpretative exercises like the Rorschach inkblot to exhaustive questionnaires like the Minnesota Multiphasic Personality Inventory (M.M.P.I.) -- is a child of 20th-century research psychology, born of the dream that we can crack the code of human behavior if only we can devise the right set of questions. These tests grew up with the help of modern bureaucracies, like corporations and the military, that needed an efficient means of categorizing people by temperament, the better to predict their on-the-job behavior.

Psychologists have been happy to oblige in this quest, for it has allowed them to indulge in a fantasy of their own: that personality assessment may someday attain the authority and respect of more objective medical tests, helping, in turn, to endow psychology with some of the status of the hard sciences. Indeed, from the 1920's on, the inventors of such tests have resorted to a favorite, telltale metaphor from the world of medicine. "As a rule," wrote Henry Murray, the Harvard psychologist who invented the TAT in the early 1940's, "the subject leaves the test happily unaware that he has presented the psychologist with what amounts to an X-ray of his inner self."

Consider the logic of the TAT, which was intended to measure unconscious preoccupations. A subject looks at the pictures and is told to make up stories about them, and these stories are assumed to reflect his own covert fears and fantasies. If the subject says the Vargas girl on the bed is dead, then he is hostile "toward his wife or women in general," as one manual puts it. If he describes the naked muscleman climbing a rope in another picture as both going "up and down," then he is preoccupied with masturbation. And so on. Yet surely what the TAT reveals just as well are the story-telling or imaginative capacities of those tested -- their ability to distance themselves (or not) from the B-movie cliches in the pictures.

It's not only that personality is a harder thing to measure than, say, liver function. It's that personality tests, perhaps even more than intelligence tests, mirror their cultural moment. We tend to think that the personality traits we value now have always been so valued. But the truth is that they have changed even over the 80 years or so that scientific personality testing has been in vogue. Whole personality types -- the self-abnegating mom, the perfectionist housewife, the celibate -- have fallen out of favor. Proclivities that once seemed like pathologies -- homosexuality, as late as the 1950's and 60's, for example -- no longer do. And traits that were once regarded as benign, like extreme shyness, have been newly redefined as epidemic illnesses, complete with Diagnostic and Statistical Manual labels and pharmacological remedies.

The history of personality testing neatly reflects these shifting notions of the optimal self. How could it not? For while the people who take them may regard such tests as a discreet way to unlock the code of their unique inner selves, the fact is that the whole enterprise has always been eminently social and practical, its origins tied to the needs of large organizations. And what those organizations needed, more often than not, were people who conformed to their model of the company man or company woman, the team player.

The first modern personality test -- the Woodworth Personal Data Sheet of 1919 -- was designed to help the Army screen out recruits who might be susceptible to shell shock. The TAT was commissioned by the Office of Strategic Services (O.S.S.) to identify personalities that might be susceptible to being turned by enemy intelligence. Many of the latest personality tests have been developed to help corporate employers decide whether Joe has what it takes to be district manager or whether Kathy ought to be hired in the first place. (That's one reason that so many of them seek to measure qualities like extroversion, leadership and self-motivation, as opposed to introspection, spirituality and creativity.)

The peculiar thing about some of the classical tests -- the TAT, for example, or the Minnesota Multiphasic Personality Inventory, which came out in 1943 -- is that since much of their usefulness depends on their having been around for a while, they must continue to be administered in more or less their original form, no matter how anachronistic. "If personality tests have been around a long time," says Robert Bornstein, a professor of psychology at Gettysburg College and an expert on personality testing, "even if they aren't necessarily the best, there's an incentive to keep using them because they have generated decades' worth of data -- norms -- that you can compare them to." Each personality test creates its own autocracy of data, in other words, and they are hard to overthrow.

The M.M.P.I. was updated in 1989, it's true, and some fusty language was removed -- the statement "I used to like to play drop-the-handkerchief," for example. Among the true-false statements that remain, though, are quite a few along the lines of "My neck spots with red often," "I am not bothered by a great deal of belching of gas from my stomach" and "I have to urinate no more often than others." Hypochondria, it turns out, was one of the test's prime targets. Yet today, hypochondria as a freestanding diagnosis is far less common.

"Now we recognize that most everyone has some physical complaints, inexplicable kinds of sensations and so on," says Bornstein. "And we have in effect said, 'That doesn't matter.' Only people who have, if you will, imaginary physical problems that seriously interfere with their functioning would receive such a diagnosis today." Or to put it another way, we're all obsessed with our health these days -- just look at the Internet, where health sites are about the only category that can compete with sex sites in popularity. Distinguishing between individual pathology and the socially acceptable preoccupation with tainted food, insidious fat content, environmental illness and the like seems far more daunting these days.

Some personality tests bear the unmistakable stamp of one particular psychological theory. "The Adventures of Blacky" -- like the TAT, a so-called projective test in which subjects are asked to respond to a series of pictures -- is my personal favorite of this type.

Blacky is a puppy of indeterminate breed and sex, though almost everyone seems to assume he's a boy. His "adventures" are of the most transparently and obligingly Freudian variety. Oedipal Blacky comes upon his father and mother holding paws and gazing amorously at each other (mama dog is troublingly drawn with humanlike bow lips). He bares his teeth and snarls. Oral-stage Blacky suckles happily at mama dog's teats. Sibling-rivalrous Blacky fantasizes a butcher knife winging straight for his blindfolded sister, Tippy. The Blacky pictures were first published in 1950, and like certain cinematic images of the couch and the sagacious, bearded analyst, they are emblems of the golden age of American psychoanalysis so pure as to evoke nostalgia in all but the most hardened Freud bashers.

But some of the biases in personality tests are of the type more familiar to us from debates about intelligence testing -- they have less to do with schools of psychological thought than with the values and experiences of the cultural elite that designed them. The M.M.P.I. is full of questions about religious belief -- so full that one begins to wonder whether its inventors equated faith with mental illness. (And The M.M.P.I., unlike later personality tests, was actually designed to separate the normal from the abnormal, not to parse more subtle distinctions in a normal population.) "Until relatively recently, personality tests were not very sympathetic to cultural and subcultural differences," says Bornstein. "And it has turned out that on the M.M.P.I., for example, members of minority and ethnic religious groups have tended to look disproportionately pathological."

Even tests in which subjects are asked to draw their own pictures have been liable to interpretation on culturally blinkered criteria. The House-Tree-Person test, which is often used to evaluate intelligence as well as personality, is a case in point. The assessments of sample pictures in a 1970 manual are certainly confounded by drawing ability: the better artists are rewarded with labels like "Adult Superior" while those who never graduated from the angular scrawl of Etch-a-Sketch are rudely dismissed as "Adult Moron" or "Adult Dull Average." And the bigger, more comfortably middle-class the house, the better the rating.

In the hands of experienced clinicians who rely on them as aids to other diagnostic tools, all of these tests can surely be helpful. Think, for instance, of the child psychiatrist who uses a balky 6-year-old's drawing of his house in flames to start a conversation with him about his fears. But where personality testing has really taken off lately is not in the doctor's office but in the personnel office. "Personality tests are used very frequently in pre-employment screening, even more now than 20 years ago," says Bornstein. "People see more liability issues in hiring and firing, so they want to be as sure as they can before they sign on the dotted line that somebody is not a difficult employee."

Employers can now choose from at least 2,500 new tests on the market, all promoted with varying degrees of hucksterishness and designed, with varying degrees of sophistication, to predict everything from who is most likely to steal office supplies or slug a co-worker to who will stay in a dispiriting job like telemarketing for at least a year. And with all those tests out there -- personality assessment is now a $400-million-a-year industry -- you can bet that some of them are being administered or interpreted by amateurs. At any rate, the fact that personality testing in the workplace is at an all-time high doesn't seem to have stopped the murderously disgruntled from wreaking havoc.

For all that, though, personality testing is here to stay -- not only because of the many institutions now hooked on it but also because of the deep human curiosity it promises, however teasingly, to satisfy. Long before scientific testing, we tried to classify temperament and character based on the shapes of people's skulls or the color of their humors. Before the Rorschach, a popular 19th-century parlor game called Blotto invited players to assess one another's creativity on the basis of their interpretations of inkblots. Leonardo da Vinci used a similar method to judge the imaginative potential of his students.

But in the end personality testing belongs as much to cultural history as to science. Tests fade from prominence when the constructs they were designed to measure no longer trouble or interest us, and new ones take their place. "A hundred years ago," says Bornstein, "Freud's contemporaries might have been interested in developing tests for things like hysteria. Fifty years ago, they wanted to develop tests for hypochondriasis. Today, you might find more people interested in developing tests for self-esteem." Remember all that the next time some test administrator tells you that what you're about to submit to is just like an X-ray -- or, as they're starting to say now, like an emotional M.R.I. Remember that human personality can't be subjugated to the tyranny of types, the logic of a questionnaire, the promise of instant self-knowledge.