Human Kinds in the Making: Race and the Census

This book review appeared in The Sciences, a very good magazine that, until its untimely death, was published by the New York Academy of Sciences.

The Sciences, September-October, 2000


Book Review by David Berreby

CHANGING RACE: LATINOS, THE CENSUS AND THE HISTORY OF ETHNICITY IN THE UNITED STATES by Clara E. Rodriguez New York University Press, 2000 283 pages; $55.00, cloth/​$19.00, softcover

They were waiting for the "walk" signal, two women in sundresses talking about their weekends, and one asked how it went on Saturday.

"Oh, well, you know," said the other. "She's Irish, so the food wasn't so good."

Her friend nodded, because she,did know, and so did I. We were in Brooklyn, and if you re seeking gastronomical pleasures among the hyphenated Americans of that borough, you think of Chinese food and Italian food and foods Arab and Greek, but you don't think Irish. And, generally speaking, you're right. As a matter of pure logic, of course, that makes no sense, for Irish identity is not a brake on cooking skills. And even if it were, a tendency among the Irish to avoid the kitchen is no predictor that any particular Mrs. Ryan will be a lousy cook. Groups don't cook. Individuals cook, and individuals vary.

No matter. Any reader who knows how Americans classify different kinds of white people can grasp why the first half of the woman's sentence was relevant to the second. Yet it's a guilty grasp, because everybody knows that group traits are probabilities, not certainties (somewhere, no doubt, there's a Colleen who's the best cook on her multi-ethnic block). One can't even talk about the racial, ethnic, cultural, geographic or religious groups to which a person belongs without carving out exceptions for that person--as in "Yes, their name is Peterson, but they're a Puerto Rican family," or "Yes, he's from New York, but he's really polite," or "Yes, she's Irish, but she's a fabulous cook." That inescapable yes-buttering means it is impossible to know for sure that any given Irish woman is a bad cook or a good Catholic or anything else. And if one can't be certain that a given Irish person has any particular trait, one can't be certain that she even possesses all the traits that are typically used to define people as Irish. Human groups are inherently slippery.

In certain narrowly defined circumstances--on a passport, for instance--one's racial or ethnic identity can be specified by the artificial clarity of the law. Yet even such paperwork has its limits. As Clara E. Rodriguez notes in her rich trove of lore on how Americans classify Americans, the United Nations has a hard time comparing census data from various countries because each country pigeonholes people in surprising different, even incompatible, ways. One study of the censuses that were conducted in fifty-one New World nations during the past forty years showed that they had no shared definition of ethnicity. Sixteen countries did not even ask about it; moreover, in many countries people skipped the ethnicity questions altogether or complained that they could not answer them. Even within the United States, definitions vary. Hispanics are considered a race by the federal agencies that track civil-rights enforcement, but they are not a race to the Bureau of the Census. Thus even in the realm of written, official definitions, the answers to questions about race depend on what is being asked, who is asking it, and why.

The upshot is that when one tries to examine these overwhelmingly important entities to which every person belongs, they dissolve into exceptions, qualifications, coded speech and navel gazing. Each category is like a portrait that, held up too close, turns into flecks and dabs of paint. And those of us who persist in applying the categories anyway are like the family in the old joke that doesn't want to cure the cousin who thinks he's a chicken. We need the eggs.

Globalization makes the problem urgent: we are all being forced to rub elbows with people who seem somehow all wrong. The questions of what makes them seem that way, and of what marks our common humanity, are more relevant to more people than ever before. We want to talk coherently about race, ethnicity, class, nation, religion and "culture." But we don't know how because, when we try, we realize that we don't know what we're talking about.

To admit confusion about such social categories may be fine for the ordinary citizen, whose folk sociology is probably no more coherent than his folk physics or folk biology. But in academia the fuzziness in thinking about group concepts is an embarrassment if not a crisis. The basic question social scientists seek to answer is straightforward: How do people categorize other people, and how do they make predictions on the basis of those categories? Put that way, the question sounds readily amenable to the scientific method. And, in fact, most every useful tool of science, from the statistical analysis of populations to genetics to evolutionary theory, has been applied, soon after its inception, to the question of human groups.

That application of sophisticated scientific tools to the problem of human "groupthink" will no doubt continue. Last summer, for instance, two teams of social psychologists, neurobiologists and medical-imaging specialists brought the magnetic resonance imaging scanner into the fray. They focused on the amygdala, a region deep in each brain hemisphere that helps spotlight emotionally important information: things that are dangerous, unpredictable or sexually arousing. One team, based at Massachusetts General Hospital in Boston, reported that the amygdala shows more activation in white people viewing black faces, and vice versa, than it does when people look at faces of their own color. The other team, which was made up of investigators from New York University and Yale University, found that among volunteers who showed race-based differences in amygdala activation, the amount of that difference could predict how the volunteers scored on two unrelated psychological tests that measure unconscious racial attitudes.

No matter how many new technologies they apply, however, all social scientists face a second, confounding question about human groupthink. To understand how people categorize people and make predictions, the scientists themselves must categorize and make predictions. But how can scientists be sure that "scientific" categories are different from--more objective than--folk categories based on emotion, self-interest and bias, whether unwitting or otherwise? Until that conundrum is resolved, the dangers of wishful thinking and pseudoscience will always be close at hand.

All one can be sure of so far is that what seems to be good people-science in one epoch looks tainted a century later. American medical science, for instance, once included a diagnosis of "drapetomania"--a regrettable condition that afflicted slaves and whose defining symptom was a desire to run away. Yet an awareness of that history of egregious error has led some scholars to throw the baby out with the bathwater. In 1997, for instance, a trio of anthropologists, writing in the journal Evolution and Human Behavior, declared tribes, clans, villages, societies and cultures to be the inventions of intellectuals, and unworthy of serious attention from evolutionary theorists. (The mass-market version of that idea is Margaret Thatcher's line: "There is no such thing as society. There are individual men and women, and there are families.")

Without a consensus on fundamentals, then, scholars who study groups are free to say whatever suits their purposes. Early in her book, Rodriguez observes that most scholarly writing on ethnicity is not particularly rigorous. In fact, she notes, a recent analysis of nearly 200 studies of ethnicity published by social scientists between 1974 and 1992 found that more than 80 percent of the studies did not adhere to a coherent theory of the subject. Slightly less than a quarter of them even acknowledged the central problem: people and the groups to which they belong are separate kinds of entities.

By definition. the group known as African-Americans is all African-American and nothing but. Any individual African-American, though, may be the holder of many other identities as well: man, father, veteran; lawyer, resident of the East Side and so on. Identity depends on who is asking, and for what purpose. And it is the process of asking the questions and getting the answers that is not understood: What is it about a trait that makes someone decide that all who possess it form a group? How do people determine that a stranger belongs to this group instead of that one? How much of that decision is even accessible to the conscious mind? How do people reconcile the exceptions with the rules? And how does the process of group making change with a person's--or a region's or a society's--circumstances?

Rodriguez writes that "many people have a core of identity, or a self, that is made up of multiple identities." She herself, Rodriguez explains, is "a light-skinned Latina with European features and hair texture ... born and raised in New York City," and her first language was Spanish. In her South Bronx neighborhood her "natural tan" was attractive, whereas "downtown ... it was `otherizing.'" "In the United States today," she points out, "a person may be Puerto Rican or Mexican on a personal level, Latino on an instrumental level, and Hispanic to the government. Some people might classify this person as black, white or Asian." ("Hispanic" is a rather vague term, encompassing new immigrants from Spanish-speaking countries, as well as North Americans who speak no Spanish. The U.S. government uses the term to refer to people whose ancestors came from Central or South America. Many people, however, prefer the term "Latino," which, among other things, avoids labeling Portuguese-speaking Brazilians as "Hispanic.")

It is not, of course, merely interesting that no one quite grasps the relation between group and member. The consequences in law enforcement have become well known, thanks to the current campaign against "police profiling"--the practice of viewing members of one ethnic group with extra suspicion. The effects of groupthink on the practice of medicine are also becoming better known. In a 1996 paper published in the Annals of Internal Medicine, for instance, Ritchie Witzig, a physician and epidemiologist, cited two relevant case histories: an eight-year-old boy who was scheduled for surgery because no one associated his stomach pains with their true cause, sickle-cell anemia (he wasn't black); and a twenty-four-year-old black man who was treated for sickle-cell anemia (which he didn't have), and then died of a bleeding peptic ulcer [see "Bred in the Bone?" by Alan H. Goodman, March/​April 1997].

Rodriguez's sprawling yet intriguing book makes the case that there is a third arena in which a flawed understanding of groupthink affects contemporary Americans: the censuses conducted in the United States every ten years. Much more than a head count, each census since the first one in 1790 has been a taxonomy of the American people, with serious, practical effects on their lives: their taxes, their freedom of movement, even their right to stay in the country. As she puts it, the census categories "represent public consensus on how populations are viewed and counted." The census is an explicit social construction--debated, decided and recorded--and the classifications it makes are just as arbitrary as the ones imposed by law enforcement or medicine. But because it was so early in the country's history that the government came to believe in an absolute division of human beings into white and black, Rodriguez points out, the arbitrariness of classifications has been most apparent in the case of people who are neither.

Mexicans once constituted a race of their own, in the 1930 census. Armenians were "Asiatics" until a court decision in 1909 promoted them to "white." And the explicit construction of categories goes on: Before the 2000 census, Congress held hearings on a proposal to create a Middle Eastern/​Arab group. The government decided against the proposal, because that rather remarkable category (it would lump Arabs, Turks, Iranians, Afghans and others into one unit) is still under construction.

But it is among Hispanics, Rodriguez argues, that the fluidity of ethnic identity in the United States is easiest to observe. For example, Hispanics were defined by language in the 1940 census, surname in the censuses of 1950 and 1960, and "origin" in the 1970 census.

To say race is a social construction is not to say it doesn't matter, any more than telling people that their chronic pain is psychosomatic makes it go away. Race and ethnicity are facts of people's lives. Most Hispanics feel discriminated against in American society, Rodriguez reports. Some of her informants recounted stingingly clear reminders of their status, having been treated as "white" until the sound of a name or an accent led someone to reclassify them. Race is a fact; it is just not the kind of fact--biological, inevitable, unchanging--that most Americans think it is. What kind of a fact it actually is is the unanswered question.

Rodriguez, a professor of sociology at Fordham University in New York City, brings several disparate methods to bear on the problem: surveys of Hispanics, in-depth case studies of individual Latinos, a historical analysis of the U.S. census, and an analysis of contemporary racial politics as reflected in government hearings on the 2000 census. Her argument can feel a bit scattered, but her accumulation of relevant evidence from so many different perspectives is, ultimately, a strength. That variety of approaches underscores an often-ignored fact that any successful theory of race and ethnicity must account for: On most any timescale you care to choose, be it the day, the week, the decade or the century, ethnic identity varies with the circumstances.

Americans are by now so used to defining themselves--in conversation, on official forms, in Web-based "user profiles"--that it may come as a shock to learn that until the 1980 census, respondents were not allowed to check their own racial or ethnic category from a list on the census form. Instead, race was reported--and, in all cases before 1970, decided--by the census taker who counted each household. The "race" question (though it was not explicitly labeled as such) on the 1980 U.S. census form asked: "Is this person--?" followed by a list. The list included such categories as "Black or Negro," "White," "Hawaiian," "Korean," "Samoan" and so forth, and ended with the option "Other--specify." It did not include "Hispanic." Yet just three questions later the form asked, "Is this person of Spanish/​Hispanic origin or descent?" The result was that some 40 percent of Hispanics--7.5 million people--checked "other" on their census forms, causing deep consternation among the tabulators.

In 1990 the Census Bureau changed its language, dropping the open-ended question used in 1980 in favor of one that explicitly mentioned the word "race." Yet in that census even more of the country's Hispanic population--43 percent, by then 9.7 million people--chose the "other race" box. The jump in the nonwhite numbers caught the imagination of the media. Time magazine published a story titled, "The Browning of America," and American capitalism adjusted to the new kind of consumer it was told to expect. Companies changed their marketing plans, and an era of corporate-sponsored celebrations of "diversity" began.

All of that, Rodriguez shows, was overblown. A good part of the "browning of America" was the result of the shift, after the 1970 census, to self-reporting. Under the old system, census takers in 1970 classified 93.3 percent of Hispanics as white. In 1980, when people first categorized themselves, only 57.7 percent of Hispanics were white. Self-classification had revealed their reluctance to adopt the American racial taxonomy. They flocked to the "other" designation. And often they added explanatory notes, writing "Dominican" or "Honduran" or "Puerto Rican" in the box.

The trouble, Rodriguez argues, is that in the Latin tradition, culture, national origin and upbringing are also important factors in determining a person's race. North Americans tend to think instead primarily in terms of biology. That is not to say that Latin countries lack racial hierarchies--with whites at the top. Latinos who classify themselves as white fare better in the United States, and often they come from countries where white skin and traits are valued more highly than nonwhite ones. One of the case studies recounted in the book is that of a Mexican-American woman who recalled that the lightest-skinned of her sisters was always given the easy chores. At family gatherings even today, the darker sisters cook for hours, while la favorita brings the paper goods.

But the "pigmentocracies" of Latin America and the Caribbean are societies that recognize many gradations of racial category. Terms such as moreno, indio, jabao and trigueno signify people between the two ends of a black-white racial continuum. Rodriguez cites a study in Latin America that found eighty-two racial terms in use. Many Hispanics, she reports, react with puzzlement to America's few, immutable categories: black or white, for the most part, with a few Indians and Asian/​Pacific islanders in the margins. As one of the people interviewed in Rodriguez's surveys said resignedly, "I do not consider myself white, but this is what the government says I am."

How did americans get the way they are about black and white? Rodriguez's approach to the question through the history of the census is fascinating. Before the great waves of immigration in the early twentieth century, the story told by the census is not one of a melting pot but, rather, of a social group worded about keeping its advantage over people who were not part of the "governing race." The census results of 1870, 1880 and 1890 included maps showing the density of the "colored" populations of the states. A report on blacks in Maryland from the 1850 census stated that, given the growth rates of the black population in the first two decades after the first census was taken in 1790, "there was in 1810, reason for apprehension that, in another half century, the blacks would become the preponderating race."

One intriguing table in Rodriguez's book charts the categories for nonwhite people. "All other free persons" was the term of choice in 1790, 1800 and 1810; "free colored," between 1820 and 1840; "color," between 1850 and 1880; "color or race, whether white, black, mulatto, quadroon, octoroon, Chinese, Japanese, or Indian" was used in 1890; and "race," between 1900 and 1950. After 1950 those relative constancies gave way to irresolution. Census takers in 1960, running down a checklist, asked "Is this person--?" In 1970 they asked for "color or race." The 1960 checklist reappeared in 1980, though, as I noted earlier, this time it was to be filled out by the respondents. The 1990 census simply asked for "race." As the questions change, so do the answers, both scientific and political.

So what kinds of questions is the census asking nowadays? Rodriguez relates how proposals to count Hispanics as a race in the 2000 census were dropped after detailed studies suggested that such a question would lower the numbers of both "Hispanics" and "whites." Instead, the race question was changed in a different way. This year's question is: "What is this person's race?"--but, for the first time, people are allowed to give more than one answer. The Office of Management and Budget estimates that fewer than 2 percent of the respondents will pick more than one racial category. Outside investigators think the number could be much higher. The consequences--for federal statistics, civil-rights programs and politics--could be far-reaching.

Changing Race presents the central problem of groupthink in all its complexities and contradictions, and relates it to a political concern that is just as important, in the long run, as policing and medicine. Rodriguez makes an excellent case that categories for people depend on context, and that context includes the category maker. Such a conclusion suggests that much of the accumulated so-called wisdom about group identity presents only half the picture: it provides categories, such as "African-American" or "gay person" or "soccer mom," without asking, Who made this category? What purpose of theirs does it serve? As Rodriguez shows, the decennial deliberations about the census are unusual in that they reveal both sides of the picture. In the records of those discussions one sees not only the categorized, but also the categorizers, doing in public what most of us usually do unconsciously, or at least in private.

In reaction to the terrible history of such categories as "the proletariat" and "the Aryan race," in whose name millions of innocent people have been killed, intellectual life is now, properly, oriented against the making of vast, all-explaining categories for people. The problem, though, is that such a taboo leaves people without any language for the generalizations they need to say anything about any aspect of human behavior. Rodriguez's book hints at a way around that impasse. Hers is an example of a study of group-think that does not resort to simplifying assumptions. She does not claim that census categories are only of interest for what they say about the people who made them up. Neither does she say they are objectively real distinctions among people. After all, it is obvious that concepts about human groups, both folk and scientific, are useful in their context. For example, because the vast majority of the people who checked the "other" or "other race" boxes in the past two censuses were Hispanic, it is logical to look for a cause for that behavior that is rooted in that shared trait.

If, as Rodriguez makes clear, categories of people are not eternal, a science of human categories must explain how those categories come to be made and unmade--including how they are made and unmade within science. In place of the great master category, something else is needed: the self-aware, locally useful, answers-one-well-defined-question category. But the language hasn't been worked out.

What is clear, though, is this: The value of contemplating the abandoned pigeonholes of the past is not in the cheap pleasure of tsk-tsk-ing the folly of long-dead people. Rather than shoring up contemporary smugness, Rodriguez's book teaches two useful lessons. First, what is important to understand is not the "objective" value of this or that category of person, but rather how the process of category formation works. Social science needs to understand how pigeonholes are made, not any particular pigeonhole.

Second, as citizens, we ought to be humble about our own certainties, which are as much a product of ill-understood processes as were those of 1850 or 1910. Our descendants will find our categories for people quaint and silly. But if our social science can decipher how we--and they--exercise the category-making system, then that science will have earned the kind of respect that is now accorded to physics and biology. The great social problem of the twenty-first century is waiting, in plain sight, to be solved.

Recent Writing

The mind is biased to see secret information as better than open knowledge.
How Not to Think About Obesity
Ghastly gastronomy, as published in Nature
A genomic nightmare
Us and Them 2008 Edition
Preface to the Paperback Edition
Are we really running out of attention?
Is it always the lone maverick who's right, when s/he refuses to play along? That's how American psychologists have taught it, but recently they have been taking another look at conformism.
Other Species