three broad groupings: the Mongoloid, the Negroid, and the Caucasoid. An influential paper by Masatoshi Nei and Arun K. Roychoudhury, entitled  Gene Differences between Caucasian, Negro, and Japanese Populations,  which appeared in Science, in 1972, found that the genetic variation among individuals from these racial groups was only slightly greater than the variation within the groups.
In 1965, the anthropologist Stanley Garn proposed hundreds, even thousands, of racial groups, which he saw as gene clusters separated by geography or culture, some with only minor variations between them. The paleontologist Stephen Jay Gould, for one, has proposed doing away with all racial classifications and identifying people by clines regional divisions that are used to account for the diversity of snails and of songbirds, among many other species. In this Gould follows the anthropologist Ashley Montagu, who waged a lifelong campaign to rid science of the term  race  altogether and never used it except in quotation marks. Montagu would have substituted the term  ethnic group,  which he believed carried less odious baggage.
Race, in the common understanding, draws upon differences not only of skin color and physical attributes but also of language, nationality, and religion. At times, we have counted as  races  different national groups, such as Mexicans and Filipinos. Some Asian Indians were
counted as members of a  Hindu  rac in the censuses from 1920 to 1940; the. they became white for three decades Racial categories are often used as eth nic intensifiers, with the aim of justify ing the exploitation of one group by an other. One can trace the ominou example of Jews in prewar Germany who were counted as  Israelites,  a re ligious group, until the Nazis came t< power and turned them into a race Mixtures of first- and second-degre, Jewishness were distinguished, much a quadroons and octoroons had been ii the United States. In fact, the Nazi ex perience ultimately caused a widespreac reexamination of the idea of race Canada dropped the race question fron its census in 1951 and has so far resiste< all attempts to reinstitute it. People whi were working in the United States Bu reau of the Census in the fifties ani early sixties remember that there wa speculation that the race question wouli soon be phased out in America as wel The American Civil Liberties Unio tried to get the race question droppe from the census in 1960, and the Stat of New Jersey stopped entering race in formation on birth and death certifi cates in 1962 and 1963. In 1964, how ever, the architecture of civil-rights law began to be erected, and many of th new laws particularly the Votin Rights Act of 1965 required high! detailed information about minoi ity participation which could be gathere only by the decennial census, the nation supreme instrument for gathering dem< graphic statistics. The expectation that tl race question would wither away surrei dered to the realization that race data we: fundamental to monitoring and enforcir desegregation. The census soon acquire a political importance that it had nev had in the past.
Unfortunately, the sloppiness ar multiplicity of certain racial and ethn categories rendered them practical meaningless for statistical purposes. . 1973, Caspar Weinberger, who w then Secretary of Health, Educatk and Welfare, asked the Federal Inte agency Committee on Education (nc to develop some standards for classif ing race and ethnicity. An ad-hoc cor mittee sprang into being and propos to create an intellectual grid that woi
9. UNDERSTANDING CULTURAL PLURALISM
sort all Americans into five racial and ethnic categories. The first category was American Indian or Alaskan Native. Some members of the committee wanted the category to be called Original Peoples of the Western Hemisphere, in order to include Indians of South American origin, but the distinction that this category was seeking was so-called  Federal Indians,  who were eligible for government benefits; to include Indians of any other origin, even though they might be genetically quite similar, would confuse the collecting of data. To accommodate the various, highly diverse peoples who originated in the Far East, Southeast Asia, and the Pacific Islands, the committee proposed a category called Asian or Pacific Islander, thus sweeping into one massive basket Chinese, Samoans, Cambodians, Filipinos, and others peoples who had little or nothing in common, and many of whom were, indeed, traditional enemies. The fact that American Indians and Alaskan Natives originated from the same Mongoloid stock as many of these peoples did not stop the committee from putting them in a separate racial category. Black was defined as  a person having origins in any of the black racial groups of Africa,  and White, initially, as  a person having origins in any of the original peoples of Europe, North Africa, the Middle East, or the Indian subcontinent  everybody else, in other words. Because the Black category contained anyone with any African heritage at all, the range of actual skin colors covered the entire spectrum, as did the White category, which included Arabs and Asian Indians and various other darker-skinned peoples.
The final classification, Hispanic, was the most problematic of all. In the 1960 census, people whose ancestry was Latin-American were counted as white. Then people of Spanish origin became a protected group, requiring the census to gather data in order to monitor their civil rights. But how to define them? People who spoke Spanish? Defining the population that way would have included millions of Americans who spoke the language but had no actual roots in Hispanic culture, and it excluded Brazilians and children of immigrants who were not taught Spanish in
their homes. One approach was to count persons with Spanish surnames, but that created a number of difficulties: marriage made some non-Hispanic women into instant minorities, while stripping other women of their Hispanic status. The 1970 census inquired about people from  Central or South America,  and more than a million people checked the box who were not Hispanic; they were from Kansas, Alabama, Mississippi the central and southern United States, in other words.
The greatest dilemma was that there was no conceivable justification for calling Hispanics a race. There were black Hispanics from the Dominican Republic, Argentines who were almost entirely European whites, Mexicans who would have been counted as American Indians if they had been bom north of the Rio Grande. The great preponderance of Hispanics are mestizos a continuum of many different genetic backgrounds. Moreover, the fluid Latin-American concept of race differs from the rigid United States idea of biologically determined and highly distinct human divisions. In most Latin cultures, skin color is an individual variable not a group marker  so that within the same family one sibling might be considered white and another black. By 1960, the United States census, which counts the population of Puerto Rico, gave up asking the race question on the island, because race did not carry the same distinction there that it did on the mainland. The ad-hoc committee decided to dodge riddles like these by calling Hispanics an ethnic group, not a race.
In 1977, O.M.B. Statistical Directive 15 adopted the FICE suggestions practically verbatim, with one principal exception: Asian Indians were moved to the Asian or Pacific Islander category. Thus, with little political discussion, the identities of Americans were fixed in five broad groupings. Those racial and ethnic categories that were dreamed up almost twenty years ago were not neutral in their effect. By attempting to provide a way for Americans to describe themselves, the categories actually began to shape those identities. The categories became political entities, with their own constituencies, lobbies, and vested interests. What was even more
significant, they caused people t of themselves in new ways as bers of  races  that were little me statistical devices. In 1974, the j ad-hoc committee set to woi people referred to themselves : panic; rather, people who fell ir grouping tended to identify the by nationality Mexican or Don for instance. Such small cate however, are inconvenient for s and politics, and the creation meta-concept  Hispanic  has res the formation of a peculiarly A group.  It is a mixture of ethnic ture, history, birth, and a prest of language,  Sawyer contends, because of immigration, the / Pacific Islander group is considi fastest-growing racial group United States, but it is a  raci egory that in all likelihood ex where else in the world. The third growing category is Other mai the nearly ten million people, them Hispanics, who refused t any of the prescribed racial American Indian groups are als ing at a rate that far exceeds the of the population as a whol about half a million people in nearly two million in 1990  hundred-and-fifty-nine-per-< crease, which was demographic possible. It seemed to be accou by improvements in the censu procedure and also by the fact 1 five Americans had become fas! and people now wished to iden them. To make matters even m founding, only seventy-four pe those who identified thems American Indian by race repot ing Indian ancestry.
Whatever the word  race  rr elsewhere in the world, or to t of science, it is clear that in Am categories are arbitrary, confu hopelessly intermingled. In ma Americans don t know who 1 racially speaking. A National C Health Statistics study found per cent of the people who call' selves Black were seen as W1 census interviewer. Nearly a thi people identifying themselves were classified as White or Bia dependent observers. That was
264
of seventy per cent of people who identified themselves as American Indians. Robert A. Hahn, an epidemiologist at the Centers for Disease Control and Prevention, analyzed deaths of infants born from 1983 through 1985. In an astounding number of cases, the infant had a different race on its death certificate from the one on its birth certificate, and this finding led to staggering increases in the infant-mortality rate for minority populations 46.9 per cent greater for American Indians, 48.8 per cent greater for Japanese-Americans, 78.7 per cent greater for Filipinos  over what had been previously recorded. Such disparities cast doubt on the dependability of race as a criterion for any statistical survey.  It seems to me that we have to go back and reevaluate the whole system,  Hahn says.  We have to ask, What do these categories mean?  We are not talking about race in the way that geneticists might use the term, because we re not making any kind of biological assessment. It s closer to self-perceived membership in a population which is essentially what ethnicity is.  There are genetic variations in disease patterns, Hahn points out, and he goes on to say,  But these variations don t always correspond to so-called races. What s really important is, essentially, two things. One, people from different ancestral backgrounds have different behaviors  diets, ideas about what to do when you re sick that lead them to different health statuses. Two, people are discriminated against because of other people s perception of who they are and how they should be treated. There s still a lot of discrimination in the health-care system. 
Racial statistics do serve an important purpose in the monitoring and enforcement of civil-rights laws; indeed, that has become the main justification for such data. A routine example is the Home Mortgage Disclosure Act. Because of race questions on loan applications, the federal government has been able to document the continued practice of redlining by financial institutions. The Federal Reserve found that, for conventional mortgages, in 1992 the denial rate for blacks and Hispanics was roughly double the