Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Knowledge from Me Behavioral and Social Sciences: Examples It is illuminating to compare our understanding of the nature of human behavior and social organization with beliefs held just a few decades or even a few years ago. In almost every area of the behavioral and social sciences, widely accepted understandings have been transformed by ongoing research. To illustrate the breadth and variety of basic research in the social and behavioral sciences, this chapter presents a set of vignettes briefly describing developments in a few areas of research. The vast range and variety of research in the behavioral and social sciences makes a comprehensive survey of such efforts impossible in a report of this kind. The committee has found it possible to proceed only by a process of heroic selection of a few examples in which there has been substantial advancement of understanding in the recent past. We imagine that different choices would have been made by another similarly constituted group, but the necessity for choice could not have been avoided. Even with regard to areas of research in the social and behavioral sciences in which rapid progress is occurring, the topics chosen for inclusion are but a small fraction. They are intended to provide a sense of the variety of activity in these disciplines, in sufficient detail to suggest the richness of discovery sometimes attending it. Some but not all of these topics are developed more fully in the papers that accompany this report. These vignettes serve also to illustrate three fundamental features of basic research. First, it is often interdisciplinary; second, substantive advances often depend on improvements in measurement procedures and research 33
34 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I strategies; and third, research intended purely to improve understanding often leads to practical applications. This last point is merely touched on here, since it is the main topic of the next chapter. As these vignettes illustrate, there is a great deal of cross-fertilization between fields, both within and beyond the behavioral and social sciences, with respect to theories, concepts, and methods. The study of primate evolution, for example, has profited enormously from the development of biochemical methods for assessing species similarity. Similarly, the study of social stratification was transformed through the use of statistical procedures first developed for use in population genetics and, in a somewhat independent line of development, elaborated by econometricians. The new field of cognitive science has drawn from the insights of linguists, computer scientists, and logicians as well as psychologists. And theorists in anthropology, sociology, and political science who characterize their respective domains as systems have, over the decades, borrowed organismic and other models from the biological sciences. Statistical procedures for the analysis of events occurring over time (time series), developed by statisticians working on problems in the social sciences, have been adapted for use in weather and climate prediction, statistical astronomy, and epidemiological studies. The propensity for disciplines to borrow from one another should serve to remind us of the arbitrary nature of the labels physical, biological, behavioral, and social science. Is the study of human origins a biological or a social science? Its methods are now largely drawn from the physical and biochemical sciences, while its explanations of human evolution depend heavily on ideas about the nature of social interaction. Similarly, interpretation of the fossil record depends in part on an analysis of the implements and cultural debris found near fossil remains, while the construction of a chronological framework would often be impossible without the dating methods derived from the physical sciences. The vignettes are arranged deliberately in no particular sequence to emphasize that they cannot be fully representative of the totality of work within particular disciplines; they should be taken in their entirety as an attempt to convey the variety of research activity encompassed by the behavioral and social sciences. VOTING The centrality that democratic theory assigns to elections and therefore to the act of voting has naturally promoted studies of voting in American political science and sociology. The essential questions of basic research on voting are easy to state, although difficult to answer: first, at the individual level, how citizens decide whether to vote and who to vote for; and, second,
Knowledge from the Behavioral and Social Sciences 35 how individual votes are aggregated to make collective decisions (a question that has close ties with the theory of social choice, discussed below). Modern voting research began about 50 years ago with the collection of the first survey data on how individual voters made their decisions. It was given a strong boost after World War II by the failure of commercial polling houses to predict accurately the outcome of the 1948 presidential election. Subse- quently it has become an extremely vigorous area, supported by an ongoing series of national election surveys conducted by the Survey Research Center of the University of Michigan. The history of this topic has been one of formulation and testing of successively more complex (and more realistic) models of voting behavior. Early research by Lazarsfeld and his colleagues (1944) revealed that actual voting behavior runs counter to widely held beliefs about how voters decide among candidates particularly to the notion that the typical voter makes a rational decision based on careful consideration of the candidates' records and stated positions on issues. These investigators started their research with the idea that voters behaved somewhat like consumers exposed to advertising campaigns: Over the course of the campaign, voters would be exposed to competing claims by the candidates, and they would weigh these claims to arrive at a final voting decision in much the same way as consumers were presumed to digest the competing claims of advertisers to arrive at a decision about a brand to purchase. To capture the process of decision making, they designed a panel survey, in which the same voters were interviewed on a monthly basis, beginning in May 1940, well before the presidential campaign began, and continuing through the election in November. As it turned out, their fundamental premise was erroneous: Almost all voters had made their decision in May, before they even knew who the candidates would be. That is, they voted a party preference, rather than for a particular candidate. Less than 10 percent of the panel changed their preference during the campaign, and these turned out to be the least sophisticated voters, who were least involved in the election; their willingness to change preference was due to indifference rather than a careful consideration of the alternatives. These unanticipated results fundamentally altered the conception of voting behavior and led to the investigation of important new questions: What does account for swings in the popular vote from one election to the next? Under what circumstances do such shifts lead to a fundamental realignment of the electorate? A series of biennial surveys of national elections initiated at the University of Michigan in the 1950s plus a series of long-term historical studies provided answers to these questions: While most voters simply follow a party preference, under circumstances of special stress or special attraction to the candidate of the opposing party, they may engage in short-term defection. Voters recognize
36 ' BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I their defection as such and ordinarily return to their original party in the next election. However, occasionally there are ''critical" or "realigning" elec- tions, in which major shifts take place in long-term party alignments. The last definitively critical election period was that of the 1932-1934 presidential and congressional elections; it is still too early to decide whether the 1980 presidential election signaled a new realignment. The Michigan studies and others also led to the solution of a number of other puzzles regarding voting behavior. Why does the party occupying the White House invariably lose congressional seats in off-year elections? It turns out to be a matter of who votes. In presidential election years the voter turnout is higher, drawing people with unstable voting preferences as well as party regulars, and the party that gets these swing votes captures the White House. But in off-year congressional elections only party regulars bother to vote, pulling the vote closer to the long-term equilibrium. Why do presidential and other candidates often seem to take very similar positions near the center of the ideological spectrum? It can be shown that by taking a position as close to the center as possible candidates will maximize the number of votes they get. This result also accounts for the nonideological character of much votin~there is not much room for decision making on the basis of the issues if very little distinguishes the candidates and has led to a much more sophisticated version of the rational voter model. In this model the emphasis is on explaining the behavior of the small fraction of voters in any election for whom real choices are available; it thereby focuses on explaining shifts in the vote from election to election rather than on explaining the total vote in each election. Such shifts turn out to be largely a matter of rational responses to candidates' positions on issues perceived to be important by voters. Subsequent research has expanded in three directions. First, there has been an increase in the range of influences implicated in the voting process, and current models encompass a wide range of influences: voter characteristics, such as personality-rooted motivations, learned social and political attitudes, role-dependent perceptions and expectations, interpersonal relations and pressures, and group-linked and institutional constraints; self-presentations and behaviors of candidates in campaigns; the behavior of the mass media; the influence of campaign finances and the activities of party organizations and interest groups; the impact of long-term social and economic conditions; and the facilities and impediments of the electoral mechanism itself. Second, there has been a very substantial export of the Michigan surveys to other countries, and data are now beginning to accumulate that will permit an assessment of whether the voting behavior observed in the United States is characteristic of industrialized democracies in general. The accumulation of data on past elections also permits us to see how stable voting behavior has been in the United States over a period of nearly 30 years. A preliminary
Knowledge from the Behavioral and Social Sciences assessment is that party identification is lessening and voters are becoming more issue-oriented but that these trends are relatively weak. Third, there has been increased attention to other parts of the voting system, apart from decision making by individual voters, with particular emphasis on strategies adopted by candidates and the legislative behavior of elected representatives. These and other topics are reviewed in more detail in the paper on voting by Converse et al. (in Part II). 37 HISTORY OF THE FAMILY As the result of efforts by social scientists in many disciplines who have vigorously explored data from numerous sources with the aid of multivariate statistical procedures, it is now possible to describe modern family life with a fair integration of the particulars of internal family practices, ethnic, class, and national cultural settings, and the location of these styles of life within vast historical and world trends. As yet no theories predict or explain such wide ranges of human behavior, but we can now observe and judge family and population problems with a much higher degree of precision than was possible 30 years ago. The paper by Featherman (in Part II) illustrates the approach of life-span research to this topic. It is fitting that American researchers should date the origins of their studies of the family with the publication in 1798 of Malthus's Essay on the Principle of Population, since the Reverend Malthus based his reasoning on the first population census of the United States. To Malthus and those who followed him for the next century and a half, the family was the product of historical events, not the maker of history in its own right. Such an outlook had almost been forced on them. Those were the years of the population explosion of the Atlantic world, of vast international and transoceanic migrations, of crop failures and famines, of sudden epidemics of cholera, typhus, and yellow fever, of rapid industrialization, and of the sudden appearance of new cities and towns and giant new metropolises. In such a climate of social transformations, it is no wonder that demographers and historians viewed modern families as the product of massive environmental shifts, and no wonder, either, that they focused particularly on the most common new type the nuclear family of husband, wife, and a few children. By the early decades of the 20th century, through careful statistical work, social scientists had charted the major paths of European and Atlantic migration, mapped the spread of urbanization and industrialization (Weber, 1899), and established an overall historical sequence that seemed to explain the flourishing of the small nuclear family in urban industrialized societies. At this point in the progress of family scholarship, the anomalies grew in importance. The French population did not migrate overseas as readily as
38 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I other Europeans, and the French rural population had been sharply limiting its family size for decades before any mechanized factories or industrial towns appeared. The fertility of the Irish population never recovered from the potato famines of the 1840s. In the United States the birth rate had fallen steadily with every census from 1800 on, despite ample harvests and abundant land (it did not begin to rise until 19501. After World War I European populations did not replace their losses either as fast as previously or as fast as their governments urged them to. Clearly there was a latitude to family decision making about marriage, births, and employment that far exceeded the Malthusian cycles of food and famine or the later model of demographic transition from premodern culture to urban industrialized culture. The desire to improve the quality of family and demographic history gained urgency after World War II, when populations of non-European nations exploded and the question arose as to whether the European experience with the flourishing of the small nuclear family would repeat itself throughout the modern world. Today's answer to this postwar question is a complicated one, because we have learned that the European experience compounded many more ingredients than formerly were recognized. The first step in recognizing this complexity began with a careful reexamination of time series data on American and European families, a process facilitated by the introduction of the computer and the improvement of demographic estimation techniques (which are described in the paper by Menken and Trussell in Part II). At the same time the question of non-European economic development spurred another reexamination of American history. In this case economists concerned with the relationships among capital formation, migration, and population growth began to tabulate these variables on a regional basis, comparing one aspect of the nation's experience with others. In this research the family and its aggregated decisions to move, to marry, and to have children appeared as an active force in history, not merely as an outcome of other forces (Kuznets and Thomas, 1957-19641. From quite another direction sociologists began to question the proposition that the modern nuclear family had taken its form to meet the demands of the modern factory. Perhaps, they reasoned, the factory itself, in some of its history and in some of its current particulars of hours, work rules, and employment policies, reflects the preferences of families. And so it turned out. A core institution of industrialization thus proved to be part of a complex interaction among large and small social forces (Smeller, 1959; Hareven, 1975; Hirsch, 1978; Hershberg, 19811. Freudian psychology provided yet another impetus for family and demo- graphic study. The Freudians' stress on the importance of childrearing practices (E. Erikson, 1950) encouraged the interests of anthropologists, who
Knowledge from the Behavioral and Social Sciences 39 in turn speculated that perhaps alternative family practices might have major consequences for modern ethnic, class, and even national cultures. This anthropological work in turn stimulated historians, who began careful studies of villages in which the family histories of entire settlements were undertaken. The anthropological example was further imitated in that attention was focused on the interrelations among all the local institutions the religion, economy, and technology. As might be expected, villages proved to be far from uniform, and, more surprising, they revealed many modern character- istics, even in the Middle Ages for example, nuclear families and high rates of geographic mobility. Centered in a group at Cambridge University in England, this historical work on villages and families soon generated fresh interpretations of European and American social and family history (Laslett, 1965, 1969; Wrigley, 1969; Grevern, 19701. As a result of the rapid accumulation of such a variety of studies, the conception of the family shifted from that of a mere mechanical respondent to that of an interacting human institution that both molds and is molded by society and history. No overarching theoretical synthesis has yet resulted. Social scientists cannot yet link in any causal statement all the important variables of family life, which range from courtship practices and child care to industrialization and urbanization. But the cumulation of recent research does permit us to be much more specific about what must be attended to in order to describe or discuss the present or past circumstances of a nation's population. We have learned that family practices with respect to marriage, children, employment, property, and extended kin are culturally specific. Whether we are observing the impact of a factory's coming to a village or of a countryman's moving to the city, we know to watch for a range of possible responses. Sometimes the responses of families are communitarian and involve extensive connections with churches, clubs, and labor unions; other families at other times respond to change by elaborating their kin networks; others pursue highly individualistic education and property-oriented responses; still others try to carry on as if no changes had taken place around them (Barton, 19751. Furthermore, we have learned that modern urbanized populations conduct their family life according to class subcultures, and, to make matters more complicated, each of these class subcultures is subject to trends of fashion in family goals and family behavior. For example, the current middle-class American pattern of attentive and permissive childrearing became a fashion among that class in the late 19th century. How long the fashion will flourish is unpredictable, but it is useful to notice that a similar fashion, among a similar class, prevailed in England from the late 17th to the early 19th century, when it was replaced by a more repressive mode. All the latest
40 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I summary work in family history stresses the class structure and class waves of family fashions, and this work contends that the waves of fashion do not correspond to economic or political cycles (Stone, 19771. BEHAVIOR AND HEALTH As the effects of psychosocial and behavioral variables on biological processes have come to be recognized, the social and behavioral sciences have made increasingly important contributions to the understanding of problems of physical health. Many medical problems, including heart disease and cancer, appear to be influenced by behavioral and social variables, such as habits of living (e.g., smoking, diet, alcohol, exercise), and by what has been termed psychosocial stress. The processes linking behavior to physical illness are of three types: (1) direct alterations in tissue function through the brain's influence on hormone production and other physiological responses to psychosocial stimuli, particularly stress; (2) health-impairing habits and life- styles, such as smoking, heavy drinking, lack of exercise, poor diet, and poor hygienic practices; and (3) reactions to illness, including minimization of the significance of symptoms, delay in seeking medical care, and failure to comply with treatment and rehabilitation regimens. These processes are reviewed in greater detail in the paper by Krantz et al. (in Part II). DIRECT PSYCHOPHYSIOLOGICAL EFFECTS In the first category are bodily changes that occur under conditions of stress in the absence of external agents such as cigarette smoke or high-cholesterol food. Stress has been defined as a nonspecific response of the body to external demands that are placed on it (Selye, 19561; in a more psychological sense, stress refers to a perceived imbalance between the demands imposed on an individual and his or her felt ability to cope with them (Cox, 19781. Examples of stress-inducing stimuli are work pressures, marital disruption, and geo- graphical mobility. Physiological responses to stress include neural and endocrine activity, which in turn can influence a wide range of bodily processes, including metabolic rate, cardiovascular and autonomic nervous system functioning, and immune reactions (Mason, 1971; Levi, 19794. Short-term stress responses include hormonal and cardiovascular reactions (e.g., increased heart rate, higher blood pressure), which may precipitate clinical disorders (e.g., stroke, cardiac instabilities, and pain symptoms) in predisposed individuals. If the stressful stimuli are pronounced, prolonged, or repetitive, the result may be chronic dysfunction in one or more systems (e.g. the gastrointestinal or cardiovascular systems).
Knowledge from the Behavioral and Social Sciences 41 Research on psychosocial variables related to the pathogenesis of cardio- vascular disorders provides a good example of direct psychophysiological effects in the etiology of disease. Perhaps the most thoroughly investigated psychosocial risk factor for coronary heart disease is what is called the type A behavior pattern (Rosenman and Friedman, 19741. Type A personalities are characterized by extreme competitiveness and striving for achievement, a strong sense of time urgency and impatience, hostility, and aggressiveness. Personalities that do not exhibit this syndrome of traits are designated type B. Although several studies have documented an association between type A behavior and coronary heart disease, the best evidence comes from the Western Collaborative Group Study (Rosenman et al., 1975), in which more than 3,000 initially healthy men, ages 39 to 59, were assessed on a comprehensive array of social, dietary, biochemical, clinical, and behavioral variables. A follow-up study after eight and a half years showed that subjects exhibiting type A behavior at the study's inception were about twice as likely as type B individuals to develop coronary heart disease. This differential in risk persisted when statistical procedures were used to control for the influence of other risk factors, such as cigarette smoking, serum cholesterol, and high blood pressure. In other words, these other factors do not account for the difference between type A and type B individuals in the probability of developing heart disease. HEALTH-IMPAIRING HABITS AND LIFE-STYLES Another way in which behavior leads to physical illness is the pursuit by some individuals of habits and styles of life that are damaging to health. Personal habits play a critical role in the development of many serious diseases. Cigarette smoking is probably the most salient behavior in this category, for it has been implicated as a risk factor in the three leading causes of death in the United States: coronary heart disease, cancer, and stroke. Poor diet, lack of exercise, excessive consumption of alcohol, and poor hygienic practices have also been linked to disease outcomes. These habits may be deeply rooted in cultural practices or initiated by social influences (e.g., smoking to obtain peer group approval). They may be maintained as part of an achievement-oriented life-style or by the interaction of biological and behavioral mechanisms of addiction. Therefore a major focus of research in behavioral medicine has been on the role of sociocultural systems, life-styles, and psychophysiological processes in the etiology and pathogenesis of chronic diseases. Considerable attention has also been directed toward the development of techniques to modify those behaviors that constitute risk factors for illnesses. A particularly interesting example of the complex interplay between
42 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I psychological and biological processes is the demonstration by Schacter et al. (1977) that heavy smokers adjust their smoking rate to keep the nicotine level in their body at a roughly constant level, so that the rate of smoking depends on the rate of nicotine excretion and breakdown by the body. The rate of nicotine excretion depends in part on the acid-base balance (pH) of the urine, which in turn can be altered by psychologial stress or anxiety. A physiologically mediated craving for cigarettes thus depends directly on the smoker's psychological state, explaining why smokers tend to light up a cigarette when they are nervous. REACTIONS TO ILLNESS AND THE SICK ROLE A third process through which behavior leads to physical illness is the practice of some individuals of minimizing the significance of symptoms, delaying in seeking medical care, or failing to comply with treatment and rehabilitation regimens. One prominent example is the sizable number of heart attack patients who procrastinate in seeking help, thereby endangering their chances of survival. Many other examples exist of circumstances in which people refuse to acknowledge that they are sick and hence do not do what is necessary to get well. These actions are representative of a larger area of study concerned with the way people react to the experience of organ dysfunction as well as to the experience of being in the role of a sick person. To succeed, medical therapies require that a patient follow his or her physician's advice, yet an extensive literature reports low rates of compliance with health and medical care regimens (Sackett and Haynes, 19761. Accord- ingly, there has been considerable research on social and psychological processes involved in patients' reactions to pain and illness, the decision to seek medical care, and compliance with medical treatments. A good deal of attention has been given to isolating factors that influence or predict compliance (Sackett and Haynes, 1976; Becker, 19791. Contrary to the expectations of the medical profession, social and demographic characteristics such as age, sex, marital and socioeconomic status, and educational level turn out to have little relation to compliance; the most important influence on compliance is the quality of the doctor-patient interaction. Physicians who engender trust on the part of their patients and ~. . . c;7 who explain the treatment regimen in comprehensible ways achieve much greater compliance than do others. ~c - - PRIMARY GROUPS IN LARGE SCALE SOCIETY One of the laments of modern life, promoted heavily in the rhetoric of the popular press, is that the scale of society is too large. Big cities, big government, big business, big schools have led to a depersonalization of
Knowledge from the Behavioral and Social Sciences 43 relationships and the decline of the primary group-composed of family, friends, and neighbors- as the organizing basis of social life. As with all such generalizations, there is a grain of truth to this one. There no doubt has been a shift in the character of modern life, epitomized in such phrases as "mass society" and "the eclipse of community." Yet, to paraphrase Mark Twain, rumors of the death of the primary group have proved greatly exaggerated. Social scientists working in different corners of society have repeatedly confirmed the hardiness and tenacity of the primary group, often in unlikely places. Consider the following: In the Chicago school of urban sociology, which developed mainly during the period of very rapid industrialization and urbanization in the first third of this century and which dominated sociology at that time, cities had come to be regarded as a kind of epitome of depersonalization if not outright disorganization of modern social life (Zorbaugh, 1929; Wirth, 19381. Careful empirical research has, however, corrected this view. One of the classic studies in sociology employing field observation, Whyte's Street Corner Society (1955), revealed the pervasiveness of men's friendship groups, which constituted the main source of social cohesiveness for men in the Italian-American slums of Boston. Years later, Elliot Liebow (1967) observed the same phenomenon among lower-class blacks in Washington, D.C., who, while experiencing the greatest fragility and instability of their work and family lives, fell back into neighborhood friendship groups for support, intimacy, and material help. Still more recent research by Carol Stack (1974) has refuted the stereotype of family disintegration among lower-class blacks, documenting the stability of kin-based networks of social support. Recent survey research in urban areas has shown a similar persistence of familial, friendship, and ethnic bases of integration for most people, although the "skid row" population and the urban elderly are major exceptions (Fischer, 19761. And even among these people, a considerable degree of social cohesion and mutual aid exists (Hochschild, 1973~. The myth of the factory as a stronghold of depersonalization has also been debunked. The classic work on this topic was earned out by a research team studying the Hawthorne plant of the Western Electric Company in Chicago in the 1930s. This research revealed that work groups tended to develop into cohesive social systems that had a profound influence on workers' morale and even productivity. In particular, members of these informal groups were shown to influence the rate of production on the line by sanctioning fellow workers who deviated too much from an informally set rate (Roethlisberger and Dickson, 19471. Subsequent research has confirmed the power of the informal group to either facilitate or disrupt the workings of formal bureau- cracies and to either reinforce or subvert higher authorities (Bensman and Gerver, 1963; Roy, 1972; Stoddard, 1972~. The market, also a prime symbol of depersonalization in contemporary
44 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I life, is infused with primary groups and their influence. Pioneering work by Lazarsfeld and Katz uncovered what they described as a "two-step flow of influence" in decisions to buy products, attend movies, and the like. Any given community includes certain "influentials," who maintain close touch with national and international advertising. The remainder of the buyers, normally out of touch with or uninfluenced by advertising, are nonetheless influenced personally, largely through informal contact with the influentials in the community (Katz and Lazarsfeld, 19551. A subsequent study of the spread of the use of an antibiotic drug also indicated that one of the most important determinants of whether a physician would adopt the drug was the degree to which he or she interacted with and was influenced by other physicians (Coleman et al., 1966~. Even under circumstances of outright and extreme disorganization, the influence of the primary group persists. Studies of disasters (tornadoes, floods, explosions) show that much of what appears superficially to be panicky or random behavior on the part of community residents is actually interpretable as evidence of their efforts to join or help family and friends elsewhere in the community (Killian, 19521. And in a recent study of a flood disaster in a mining community in West Virginia, whose informal neighbor- hood and community life had been decimated over the decades by economic hardship and out-migration, Kai Erikson (1976) observed how the absence of primary group ties contributed greatly to the depth and prolongation of the community's subsequent despair and depression. Studies of drug use have revealed the importance of primary groups in understanding a form of deviant behavior that previously had been thought to be either a consequence of biological vulnerability or the product of forces affecting society at large. Drug use in the contemporary United States is now seen to begin as casual partying behavior among pairs or small groups of close (but generally short-term) high school friends, to continue in a series of casual, mainly adolescent relations, remaining most strongly entrenched when family members, especially siblings or spouses, regularly share drugs, and to decline or end when family or coworker ties not involving drugs are strong. This picture seems to be broadly true of illicit drugs ranging from marijuana (Kandel, 1980) to PCP (Feldman et al., 1979) to heroin (Gould et al., 1974; Hughes, 19771. THE ANALYSIS OF STATUS ATTAINMENT The study of social stratification the unequal distribution of power, privilege, and prestige in society and social mobility the stability and change in these attributes from generation to generation and during the life course- was radically transformed with the publication in 1967 of Blau and Duncan's The American Occupational Structure. Prior to the work of Duncan and his
Knowledge from the Behavioral and Social Sciences 45 colleagues, stratification and mobility generally had been conceived as separate phenomena, and the study of social mobility had been concerned mainly with assessing whether American society was becoming less open, that is, whether equality of opportunity was declining. This question was pursued through the analysis of cross-tabulations of father's occupation, divided into roughly hierarchically ordered categories, by son's occupation. Blau and Duncan reconceptualized mobility as a process of status attain- ment, in which a father's occupation was but one of many factors~ducation, occupation, income, and similar attributes influencing his son's accomplish- ments. (The initial restriction of analysis to men was strictly a convenience; subsequent research extended the analysis to women and to characteristics of both parents.) Moreover, borrowing statistical procedures developed by the geneticist Sewall Wright (and independently by several econometricians), Blau and Duncan were able to represent and quantify complex processes of status attainment through what are called path models and corresponding systems of structural equations. In this way they moved beyond mere static description of people's status characteristics to a dynamic characterization of the American stratification system (for useful reviews of these developments see Bielby, 1981; Featherman, 1981; and Mare, 19801. For an interesting approach to this work from a life-span perspective, see the paper by Featherman (in Part II). The creation of scales to measure such concepts as educational attainment and occupational status has been an important part of research in this area. The measurement of occupational status, in particular, has been the subject of an extensive and cumulative research tradition of great theoretical importance. One tradition has focused on the relative prestige of occupations, studying how and why occupations differ in their prestige, demonstrating the essential invariance of occupational prestige hierarchies across time and space (Treiman, 1977), and developing prestige scales to measure the status of occupations (Siegel, 1971; Treiman, 19771. Another tradition has focused on the socioeconomic aspects of occupations as the basis for creating an occupational status scale that is widely used in status attainment research (Duncan, 19611. The model of status attainment shown in Figure 1 relates father's educational attainment, father's occupational status, son's educational attainment, the status of the son's first job, and the status of the son's job in 1962, the time of the data collection. This model states that a son's educational attainment depends on his father's educational and occupational status (plus a set of unmeasured factors uncorrelated with father's education or occupational status); that the status of the son's first job depends on his own education and his father's occupational status; and that the status of his job in 1962 depends on his education, the status of his first job, and the status of his father's job. The coefficients in the model (which are in fact standardized
46 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I :86 father's ~ educational / attainment .52 \~ father's ' occupational status .31 .28/ son's educational attainment \ 44 .22 status of son 's f i rst ~ Job .82/ ~ .75 .39 /.28 son's 1962 occupational status FIGURE 1 Path model of the process of status attainment in the United States. for males from nonfarm origins aged 20 to 64 in 1962. Educational attainment is measured by years of school completed; occupational status is measured by Duncan's Socioeconomic Index (Source: Blau and Duncan, 1967:170). Reprinted by permission of the Free Press, a division of Macmillan Publishing Co.. Inc. regression coefficients estimated from data) indicate the relative importance of these causal connections. From models such as these it is possible to draw important substantive inferences. Consider the role of education in social mobility. A good deal of both sociological and popular writing in the 1960s portrayed American society as one in which individuals largely inherit their status from their parents and in which the educational system serves mainly as an instrument of status transmission rather than as a means of freeing people from their social origins. Applying the algebra of path analysis to the model shown in Figure 1, Blau and Duncan (1967:200-201) were led to conclude the contrary, that a "very considerable amount of 'status modification' or occupational mobility does occur," and that "far from serving in the main as a factor perpetuating initial status, education operates primarily to induce variation in occupational status that is independent of initial status." This is an important conclusion in terms of what it tells us about the structure of American society. And it is by no means intuitively obvious, but is rather a derivation from the model. This simple five-variable model can hardly be taken as a fully adequate representation of the process of occupational attainment in American society. A great deal of subsequent work has been done to elaborate the model by including additional social origin variables, personality and attitude variables, indicators of the character of the interpersonal environment, and multiple measurements of occupational experience through the course of the career
Knowledge from the Behavioral and Social Sciences 47 (see in particular Duncan et al., 1972; Jencks et al., 1972; Sewell and Hauser' 1975; Featherman and Hauser, 1978; and Jencks et al., 1979~. It is of interest to note that even the elaborated models leave much of the variation in occupational status unexplained. Rather than regarding this as evidence of the inadequacy of the model, however, sociologists regard it as a consequence of the openness of the stratification system of the United States. Would many wish to live in a society in which the occupational attainments of individuals are entirely or even largely determined by their social origins and a handful of additional variables? Similar models have also been estimated for women and for members of minority groups, and these models have been elaborated in ways that permit systematic comparison of sources of inequality of opportunity and achievement on the basis of race, sex, and ethnic status. Similar procedures have been used to assess trends in opportunity structures and achievement patterns, especially through a careful replication of the 1962 Blau and Duncan study by Featherman and Hauser a decade later (Featherman and Haus_r, 19781. Finally, the Blau and Duncan research has served as a model for a number of similar studies conducted in foreign countries. This work will permit systematic comparisons of the process of status attainment in different types of societies, although substantial measurement problems will need to be solved before this becomes a reality. Work currently under way is beginning to address the relationship between the achievements of individuals and the way educational systems and labor markets are structured to create differential opportunities and incentives for schooling, jobs, and promotions. This work shows promise of providing a much richer description of the processes leading to variations in success in the United States and in other countries. Much of the power of the modeling procedure used in the study of status attainment path analysis and structural equation models-derives from its requirement for precise specification of the hypothesized model and its capacity to provide statistical estimates of the adequacy of the model. These requirements make it much easier to understand and to criticize the work of others, even that in other disciplines (e.g., economics), and so have increased the rigor with which theoretical disagreements can be pursued. Perhaps for this reason, these causal modeling procedures have been widely adopted in sociology and other social sciences (see Bielby and Hauser, 1977, for a review of this development). INFORMATION PROCESSING PSYCHOLOGY The past quarter century has seen the birth of a new discipline cognitive science which encompasses parts of psychology, linguistics, computer
48 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I science, anthropology, and philosophy and bears a close connection to neuroscience. At the heart of this development was a shift from the behavioristic approach of experimental psychology and its preoccupation with simple memory and learning to an information processing approach and its preoccupation with human thinking and problem solving. For more comprehensive reviews of this topic see National Research Council (1981) and Simon (1980, 1981~. A central aspect of the information processing approach is the use of detailed computer programs to make theories about human problem solving explicit. Two important contributions of this work are (1) the development of computer programs that can make decisions, solve problems, and carry out other processes that in humans we call thinking and (2) the discovery that human thought can be described in terms of programs consisting of the storage, comparison, and manipulation of symbols. Computer simulation of human thinking has done much to release the study of thinking from the hold of vitalism and make it instead a systematic domain of scientific exploration. One outgrowth of this work is the field known as artificial intelligence, which is concerned with programming computers to "think." Computers have been successfully programmed to play expert or master-level chess, to diagnose disease, to discover reaction paths for the production of organic compounds, and to recognize spoken language. Many of these programs, however, rely on a brute force approach that would be impossible for humans, taking advantage of the size and speed of modern computers to engage in massive searches. In this sense they simulate the function but not the process of thinking, much as airplanes simulate the function but not the process of birds in flight. Some artificial intelligence programs, however, incorporate what is known about the corresponding human processes, especially strategies that enable humans to be highly selective in their search for solutions. These strategies have been revealed by research analyzing the detailed character of human thinking and problem-solving processes. Much of this research has been conducted by creating computer programs that simulate the process by which humans think, process information, and solve problems. This is done by having human problem solvers "think out loud" about what they do, then programming computers to do similar things and comparing the program in detail with the behavior of the human problem solvers to ensure that the simulation of human behavior is accurate. An underlying assumption of the information processing approach is that the programs of thought consist of a set of elementary mental operations. Since many of these mental operations are not available to consciousness, the empirical analysis has been an important branch of the field and a major
Knowledge from the Behavioral and Social Sciences 49 way in which its techniques have received applications in educational, industrial, and medical settings. To take one example, it is well known that a small amount of information can be stored in a short-term, readily accessible buffer memory. It has been shown that such memory consists of a small number of chunks of information (Simon, 19741. These chunks can be searched at a rate of about 30 milliseconds per chunk (Sternberg, 19661. Individuals differ greatly in the capacity and search rate of this system, but the overall structure appears similar among individuals, over the life-span, and in greatly differing tasks such as mental arithmetic, reading, and chess playing. Much of the increased development of such capacity in early life is due to improved familiarity with the items to be stored (Chi, 1976~. It is possible to increase the size of the buffer memory through altering the formation of chunks. One study showed that training could increase the number of digits stored from 7 to about 80 by the development of proper chunking strategies (Chase et al., 19811. Teaching these strategies has been shown to have important effects on the functional memory capacity of both normal and retarded persons (Belmont and Butter- field, 1971~. Such chunking strategies greatly improve short-term memory capacity; since they do so uniformly across individuals, however, they are not the source of the differences among people in memory span (Lyon, 1977~. Stored information may be viewed in terms of distinct formats. Thus a written word can activate a visual code, a phonological code, and a complex network of associations, a meaning code (Posner, 19781. In reading it is possible to access the semantic network either from the visual or phonological code. The visual form of the word serves as the major means of unlocking the meaning of the word for most readers, but the phonological code appears necessary for storing information in a way that will lead to understanding the meaning of the passage (Kleiman, 19751. Deaf persons have difficulty in reading largely because they have deficits in representing the phonological structure of the language, and therefore have problems with storage and retrieval. These findings have obvious importance in diagnosing reading difficulty among those who suffer from what is called dyslexia. In striking cases of dyslexia induced by brain injury, the patient may be able to access the meaning and be unable to read the name of a word whose associations he or she knows quite well (Coltheart et al., 1980; see also the paper by Carpenter and Just in Part II on reading as a cognitive process). While the information processing approach has been productive in linking human cognition to artificial intelligence through the concept of mental programs, it also provides creative opportunities to link cognition to the brain's underlying neural structure. The ability to time-lock elementary mental
50 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I operations (that is, to identify precisely when one occurs in relation to when others occur) provides a number of experimental techniques for linking cognition and brain processes. It is known that the subjective experience of concentration on a stimulus event increases brain wave activity, as can be indexed in an electroencephalogram in terms of an amplification of components that occur in the first few hundred milliseconds after the stimulus event (Donchin et al., 19781. In the case of spatial attention (that is, recognition of spatial relations), a cognitive capacity that humans share with other organisms, it is possible to examine changes in cellular activity that correspond to the mental operation of shifting attention from one part of the visual field to another. When people shift their eyes to a peripheral visual event, it has been shown that the processing of a target occurring in the neighborhood of that event begins to improve about 200 milliseconds before the eyes start to move (Posner, 1980~. The information processing approach has already led to enhanced under- standing of the complex mental performance involved in creativity, memory, language, and the acquisition of other skills (Newell and Simon, 1972; Anderson, 19801. It has also provided a framework that shows promise of guiding explorations in the neural substrate, the functioning of the brain as an information processing system, which has to date proved nearly intractable. As Herbert Simon notes (1981:71: We have a situation exactly like that in Nineteenth Century chemistry, where a quite elaborate and powerful theory of chemical reactions was built up long before physical theories of atomic structure were available to explain why certain compounds were stable and others unstable; why certain reactions occurred and others did not. As the present situation in the theory of elementary particles also illustrates, this kind of top-down, skyhook support for natural science theories is not at all unusual. If it were not so, the whole structure of scientific knowledge would tremble at each perturbation of the theory of quarks. Thus, cognitive science is concerned with a kind of "mental chemistry," and has made few connections, up to the present moment, with the physiological substratum, which is surely there but whose relations with the symbolic systems it supports are not yet understood. The hope is that understanding of the symbol system will lead to better understanding of the neural substrate, an instance of behavioral research paving the way for physiological research. ORIGINS OF AGRICULTURE Studies of the origins of agriculture essentially began in the years following World War II. Earlier archaeological interest in the Near East, in Mesoamerica, and in the central Andes, the regions where these origins are now thought to have occurred, had largely centered on the fully developed civilizations that followed some millennia later. The monumental art, architecture, and
Knowledge from the Behavioral and Social Sciences 51 (in some cases) historical records left behind by these civilizations were important acquisitions for museums, which sponsored a large part of the archaeological research. By contrast, the transition from food gathering to food production generally left few vestiges suitable for popular exhibit. Yet as a milestone of human achievement it was fully comparable to the rise of urban civilization. Following the postwar shift of the bulk of archaeological activity from museums to universities, conditions existed for a rapid acceleration of effort on the question of the rise of agriculture. Economic and social development had become a theme of intense comparative study within the social sciences, and this interest extended to archaeological research as well. In a fundamental alteration of viewpoint, archaeologists directed attention away from the study of individual archaeological relics and their stylistic associations and toward the contextual reconstruction of broadly changing patterns of subsistence, technology, and social organization, which lay at the core of the introduction of agriculture. This alteration of viewpoint was facilitated by the favorable opportunities for interdisciplinary work existing in universities and by the support consistently given to such studies by the National Science Foundation. Most of the major advances in our understanding of the origins of agriculture have depended on the interdisciplinary cooperation of social, biological, and physical scientists. Great improvements in dating techniques (e.g., radiocar- bon dating) are probably the most widely known example, but there were developments of comparable importance in other areas as well the analysis of the plant and animal remains of species that became domesticated and analysis of soils, pollen, and other sources of information for environmental reconstruction. Nor were artifacts neglected in these cooperative processes of technical advance. Emphasis has shifted, however, from stylistic analyses of relatively rare complete specimens to analyses of microwear patterns (to understand how tools were used), of the debris of manufacture, and of the behavioral patterns represented by thousands of broken fragments of tools and pottery. The primary objective of study is no longer the recovery of physical remains, nor even the description of the newly developing technologies associated with extraction, processing, storage, and consumption that were immediately responsible for the successful introduction of agriculture. What is now of greatest interest is the reconstruction of interrelated social and subsistence systems as they evolved over time. Just as the cultural equipment was subject to continuous modification and improvement, so also were the plant and animal species more and more sharply distinguished from their wild prototypes by selective harvesting or breeding. Subsistence strategies and perceptions-of the socially constituted groups adopting them, of the natural world, of risk and opportunity and how to meet them are and have
52 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I always been culturally influenced and passed on. For this reason the study of agricultural origins has been primarily organized and conducted by anthropologically oriented archaeologists. The adoption of food production as an alternative to hunting and gathering is now known to have begun independently in the Old and New Worlds not long after the Pleistocene or Ice Age drew to a close about 12,000 years ago. It appears to have occurred repeatedly and more or less independently in many different settings in both hemispheres, based on locally differentiated complexes of potentially domesticable food resources. There are indications that experimentation along these lines may even have begun some thousands of years earlier in especially favorable settings (e.g., the Nile Valley). Trade- offs were involved in accepting a more sedentary way of life through much of the seasonal cycle in order to promote husbandry, and it seems clear that the productivity and security we associate with agriculture became apparent only very slowly and gradually. Sometimes there is reference to a food- producing revolution, and the term correctly calls attention to the cumulatively decisive change that took place and to its relatively accelerated pace. As an advance that has proved to be in a broad sense irreversible, it certainly bears comparison with the industrial revolution of the late 18th and 19th centuries. But the earlier change is better visualized as a process linking numerous, increasingly complex and interdependent adaptations, rather than as an event proceeding swiftly and self-consciously from a limited series of identifiable . . discoveries. The two hearths where the process of conversion to agriculture is currently best understood are the Near East and Mesoamerica. Others, not yet so well documented but quite possibly of comparable antiquity, are in Southeast Asia and probably in China, East Africa, and tropical South America. The Near East was, on the basis of current evidence, the earliest locus. Coinciding with a period of climatic and vegetational shifts, settled village farming communities had made their appearance there by about 9,000 years ago. The region of earliest cultivation of the wild grasses that were the progenitors of modern wheat and barley lay immediately east-of the Mediterranean, while at about the same time sheep and goats may have been first domesticated along the flanks of the Zagros and Taurus mountains somewhat farther east. Additional crops and domestic animals were added to these over several succeeding millennia. The Mesoamerican equivalent of this process is best known from south central Mexico, where morphological changes in the remains of 9,000-year- old squash suggest the beginnings of domestication. Domestic maize and beans appeared two millennia or so later, but maize, in particular, required several millennia of further selective breeding before great enlargement of
Knowledge from the Behavioral and Social Sciences 53 the ears and kernels permitted it to serve as the foundation of subsistence for American Indians. Numerous fruits and vegetables were gradually added to the food crop complex. Domesticated animals were of minor importance in most New World areas until the coming of the Spaniards in the 16th century. Most social scientists today have grown increasingly dubious about the adequacy of simple, single-cause explanations for anything, and the archae- ologists among them are no exception. Few of those specializing in this subject believe that there is any single factor, or even closely related set of factors, that was repeatedly responsible for the initial introduction and subsequent spread of an agricultural way of life. One of the earliest plausible suggestions was that the enforced propinquity of humans and potential domesticates along the fringes of shrinking oases initiated a successful process of experimentation leading to husbandry. This hypothesis assumed that climatic change was a factor, an assumption that finds at least partial support in a number of converging lines of scientific evidence, although shrinking oases are no longer seen as a likely locus. The propinquity of human groups and potential domesticates is an obvious precondition (thus helping to locate some of the hearths) but not an explanation. More recent work has tended to focus on zones in which wild subsistence resources were subject to periodic, critical interruption, intensifying pressures for innovations that would provide greater security regarding food supplies. According to another hypothesis, relatively more permanent, year-round settlements were first dictated by other components of the food quest, such as a dependence on fishing or shellfishing. This led in turn to attempts to develop a new subsistence base with a greatly amplified range and quantity of localized food resources. Population pressure plays a part in a number of the causal explanations that are being tentatively advanced. It is an exceedingly difficult concept to test within the constraints of the fragmentary archaeological record, but it does direct attention to the special incentives for technological and other innovation that must have existed under changing, uncertain, or otherwise marginal conditions. Also evident in a number of current theoretical ap- proaches are applications of positive and negative feedback principles derived from modern systems theory. As this overview suggests, the field is in a state of active advance (for reviews of these developments see Flannery, 1973; Bender, 1975; Megaw, 1977; and Reed, 1977~. There is widespread debate over divergent explanations and unassimilated data, and it is premature to speak of an emerging consensus around a single synthesis. But it can also be said that a major, previously unrevealed chapter of human existence has been at least roughly blocked out, within the space of a single generation of scientific effort.
54 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I SOCIAL CHOICE Social choice theory focuses on the identification and analysis of various principles by which groups or entire societies do, might, or should decide what policies to adopt. Can we speak meaningfully about the welfare of society as a whole, as distinct from that of the individuals who compose a society? This question has occupied philosophers since antiquity and is the principal subject matter of two areas of contemporary social science welfare economics and axiomatic social choice theory. In economics the issue is to analyze economic arrangements in terms of their relative benefit to the totality of people in society; in political science the main interest is to analyze different voting systems, ranging from committee decision-making processes to national elections, to understand how individual preferences are translated into a collective or social choice. Reviews of this area are available in Mueller (1976), Plott (1976), and Russell (1979~. A central problem of social choice theory is how to reconcile the desires, values, and interests of different individuals and groups when they are not in agreement, which is taken by social choice theorists as synonymous with the question of how to decide what is of greatest benefit to society as a whole. The problem is most easily seen with respect to the analysis of voting systems. In the United States we are used to thinking of majority rule as an attractive voting system, partly because it appears to treat both candidates and voters in a neutral and unbiased fashion. In a majority rule system the social decision should coincide with whatever preference is held by a majority of voters. Yet it can be shown that if there are more than two alternatives available, it may not be possible to determine the majority choice unambig- uously. Consider the following situation, with three voters (1, 2, 3) and three choices (candidates Reagan, Anderson, Carter). Suppose voter 1 thinks Reagan is best, Anderson is second best, and Carter is third best; whereas the preference order of voter 2 is Anderson, Carter, Reagan; and the preference order of voter 3 is Carter, Reagan, Anderson. Under these circumstances we have a paradoxical social choice: Reagan is preferred to Anderson (since a majority, voters 1 and 3, prefer Reagan to Anderson); at the same time Anderson is preferred to Carter, and, also, Carter is preferred to Reagan. In other words, no alternative is best in the sense that it is unambiguously preferred to all other alternatives. This violates a common assumption, transitivity, which requires that if a is preferred to b and b is preferred to c, then a will be preferred to c. It turns out that this problem is endemic to virtually all voting systems that treat voters and alternatives neutrally. One way out of the dilemma would seem to be to vote on each
Knowledge from the Behavioral and Social Sciences Reagan Anderson Reagan Carter Anderson Carter Carter Reagan Anderson . . . _ 1 FIGURE 2 Three different orders of voting lead to three inconsistent choices for winner. Reagan 55 Carter Carter ~ Anderson Anderson Reagan pair of options successively. But this would not resolve the dilemma, since the outcome will be seen to depend entirely on the order of voting: 1. If voters first choose between Reagan and Anderson, Reagan's two favorable votes will win out over Anderson's one favorable vote. Then, when the winner Reagan is pitted against Carter, Carter's two favorable votes against Reagan's one will render Carter the final winner. (If this example is
56 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I unpersuasive because of the actual election outcome, think back to 1964 and substitute Goldwater, Rockefeller, and Johnson. But mostly keep in mind that this is an analytic exercise.) 2. Suppose, however, that Reagan was first pitted against Carter. Again, he would win. But when he is in turn pitted against Anderson, Anderson is seen to win. 3. Finally, had we first pitted Anderson against Carter, Reagan would be the ultimate victor. In short, as Figure 2 demonstrates, we see that the alternative that is omitted from the first of the two votes always wins. So the paradox is not resolved, since there is no way of deciding which of the three possible orderings of pairs of votes is to be preferred. The dilemma illustrated by the three-person, three-choice situation has obvious implications for the American presidential electoral system. Although our electoral system is somewhat more complex, with choices among candidates in primary elections followed by a choice between primary winners in the general election, the dilemma is much more general as was shown by Kenneth Arrow in his impossibility theorem (19631. What Arrow showed is that under a set of plausible assumptions that most of us would readily agree to and that most designers of voting systems have taken for granted (e.g., the assumption of preference transitivity, mentioned above; the assumption that preferences between any two choices do not depend on what other options are available; and the assumption that individuals are free to select any choice of ordering), and if society must make choices in a wide variety of circumstances, then the society's choice must be that of a single one of the voters or it must be imposed from the outside; in other words, the only way to get a definitive social choice is through dictatorship. Were Arrow's theorem simply a mathematical curiosity, it could easily be dismissed as of interest only to scholars. The difficulty is that the assumptions that lead necessarily to Arrow's impossibility theorem are shared, implicitly or explicitly, by virtually all believers in democracy and are encountered in most actual voting systems. In consequence, Arrow's work, for which he won the 1972 Nobel prize in economics, has stimulated a great deal of subsequent analysis and research. This work is divided into three main categories: attempts to discredit the main theorem by challenging some of its assumptions; mathematical analysis of choice systems that operate under different assumptions; and empirical research on the behavior of actual voting and choice systems, although much of this latter research has had quite independent origins. (For a review of some of the latter material, see the paper by Converse et al. in Part II.) One major contribution of social choice theory thus far has been to challenge older understandings derived from
Knowledge from the Behavioral and Social Sciences 57 political theory about the operation of voting and other choice systems and to clarify the assumptions underlying such systems. The field is moving very rapidly and it is difficult to predict future developments. As is common with landmark breakthroughs in understanding, social choice theory is being shown by current researchers to apply to important areas of life that would seem remote from voting. For example, how can society devise cheat-proof methods of auctioning to determine the scope of public-goods programs (lighthouses, fire stations, defense and security systems)? These important questions turn out to be logically identical to the social choice calculus of Arrow, Vickery, and other social scientists. HUMAN ORIGINS Theories of human origins have been vigorously debated for more than 100 years. Although most scientists agree that humans evolved from some earlier form of primate, until very recently there was little consensus on when the hominid lineage became separate from other animals. The differences of opinion were substantial. Some believed that our ancestors were apes, similar to the chimpanzee or gorilla; others maintained that there never was an apelike creature in our ancestry. Some thought the human-nonhuman sepa- ration was recent (4 or 5 million years ago), while others thought it was ancient (more than 40 million years ago). The radically different opinions of competent scientists made it evident that neither comparative anatomy nor the fossil record could settle the issues. Since the 1960s a variety of biochemical methods have been developed that made it possible to compare humans to other contemporary mammals by quantitative methods immunology, electrophoresis, sequences of amino acids in proteins, and direct comparisons of the genetic substance, DNA (Goodman and Tashian, 1976~. For a general review of this topic, see the September 1978 issue of Scientific American, subsequently published by W. H. Freeman (Isaac and Leakey, 1979~. The main outlines of primate evolution revealed by these methods are shown in Figure 3. In order to simplify and summarize a large amount of information, the genetically measured distance between humans and chim- panzees is taken as 1. The numbers on the chart refer to this distance. The time of the separation of the various lines is calculated from molecular divergences. For example, the human-chimpanzee separation is estimated as 5-8 million years ago, the human-orangutan separation as 10-16 million years ago, and the monkey-human separation as 20-25 million years ago. Although there is not yet complete consensus on these dates, they are considered reasonable approximations and show clearly the close relationship of hominids and African apes (chimpanzees and gorillas).
58 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I ~ o ~ O ~a' ~ \ 5-8 million years: /~2 / \10-16 million years \/ \ / 20-25 million years \,/ 35 million years FIGURE 3 Diagram of human evolution. The numbers refer to the evolutionary distance between chimpanzees and human beings. The rates of evolution for some proteins are remarkably constant, and some portions of DNA also appear to evolve at constant rates. Genetically, small prosimians, tree shrews, monkeys, apes, and human beings have all changed approximately the same amount since the primates became a separate order of mammals. The amount of certain kinds of genetic change correlates with time not with number of young, length of generation, or morphological
Knowledge from the Behavioral and Social Sciences 59 change. These findings suggest that a significant portion of the genome (the body of genetic information that determines heredity) is not subject to natural selection but rather changes in a random way at a steady rate. Much change at the molecular level may be explained by this neutral theory of evolution; the relative importance of selection and neutral factors in evolution is a field of major controversy at the present time (Kimura, 19791. All the biochemical findings of the last few years support the beliefs that humans and the African apes are very closely related, that humans and the apes had a long period of common ancestry, and that apes and monkeys shared a long common ancestry before that. The main outline is not new; what is new is being able to base the theory on molecular information and eliminate theories that demand a very early separation of the human line. This is a major clarification almost no scientists believed that apes and humans are as close as they have proved to be. The fossil record of the early stages of hominid and nonhominid primate evolution is growing steadily. New techniques of anatomical analysis and new knowledge about living biological systems (from complex social systems to simple anatomical ones) are fleshing out the ancient bones, enabling us to see the extinct as the once alive. The earliest definite fossil evidence for hominids comes from the footsteps found by Mary Leakey and coworkers at Laetoli, Tanzania. The footsteps, more than 3.5 million years old, document a typically human bipedality. Fossils of a comparable age have been found by Donald Johanson and coworkers in Ethiopia. These hominids had small brains, no larger than those of contemporary apes, walked upright, and probably did not make stone tools. These finds support the evidence of the fossil finds in South Africa many years ago, which showed that human locomotion evolved long before human brains enlarged significantly. Judging by the behavior of chimpanzees and the hands and teeth of the fossils, the small-brained bipeds (Australo- pithecus) were probably using tools, although perhaps not making patterned tools from durable materials. Stone tools appear in the fossil record by 2.5 million years ago, but the time is uncertain. In the period 2 million to 1 million years ago, stone tools were present, at first in simplest form, but becoming quite advanced. During this million-year period the brain doubled in size and Homo erectus replaced the earlier forms. Homo erectus, who occupied much of the Old World, made complex stone tools, killed large animals, and controlled fire. But technical progress was exceedingly slow- the same kinds of tools were made for hundreds of thousands of years (Washburn, 1979~. By about 40,000 years ago, people anatomically like us had appeared, possibly by rapid spread from a quite restricted area. Then progress accelerated;
60 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I these modern humans spread all over the world. They crossed large bodies of water, reached Australia, adapted to life in the Arctic, and finally entered the New World. Art, housing, compound tools, boats, fishing, and tailored clothing all appear in the record of technological acceleration. It is tempting to think that this great acceleration was the result of communication, of language as we know it. There must have been earlier and simpler forms, but these have left no trace in the archaeological record. This latest phase of human evolution, in which we know that human nature was basically the same as it is today, represents less than 1 percent of uniquely human evolution (measuring from the time after our ancestors separated from the apes). Five million years is a small part of the total history of life on earth. In these 5 million years humans became bipedal, became skillful with their hands, became hunters and warriors, and began to control nature. Evolution produced a brain capable of learning those behaviors important for survive in the distant past. It is this biological ability to learn that is the most important aspect of human nature. The ability is old, a product of evolution, but what we learn is new, a product of the last few hundred years, particularly of sciences; and how we learn it is also new. When considering human evolution, it is very important to remember that the evidence is changing. Science is not static, and the study of human evolution is no exception. Progress is under way and can be expected to accelerate in many areas. The fossil evidence accumulates at an increasing rate, and our ability to make sense of it, to produce descriptions and explanations of past adaptations, expands as our application of what we learn from studies of living animals becomes more sophisticated. Twenty years ago the dates given for fossils were informed guesses. Today there is a series of methods (potassium-argon, magnetic reversals, fission track) that yields accurately determined dates not dependent on any single individual's opinion. Molecular biology has revolutionized methods of comparing animals and has greatly changed evolutionary theory. Jerold Lowenstein (1980) has shown that there may be enough unaltered protein left in fossil bones so that these methods can be directly applied to some fossils, which holds promise of definitively settling ongoing arguments about relationships between various fossils and leading to much firmer conclusions from the fossil record. SOCIAL BEHAVIOR OF MONKEYS AND APES Despite centuries of fascination with monkeys and apes, reflecting intuitive recognition of their close relationship to humans, until very recently almost nothing was known about the behavior of nonhuman primates in their natural habitats. In the late 1950s a series of field studies was initiated by zoologists, psychologists, and anthropologists motivated by concerns ranging from an
Knowledge from the Behavioral and Social Sciences 61 attempt to understand the origins of human behavior to an interest in the functioning of small groups. In the past 30 years the study of the behavior of nonhuman primates has rapidly developed into one of the most sophisticated and productive areas of research at the interface between the biological and . . socla. . sciences. The study of nonhuman primate behavior falls into four interrelated areas: social behavior and development, social organization, ecology and life history patterns, and intelligence. Some examples illustrate the sorts of problems that primatologists address. In his famous experiments on monkeys raised in isolation, the psychologist Harry F. Harlow showed that young monkeys who are denied contact with other monkeys develop into listless, hollow parodies of their group-raised, socially adept peers (Harrow and Harlow, 1962, 19651. Long-term field studies of individually recognized animals make clear why this is so: Primate infants, helpless at birth, depend for months or even years on their mothers for survival (Altmann, 19801. This dependence is reflected in a prolonged emotional bond between mother and infant. In the wild the intensity of this bond is most clearly revealed after death: Monkey mothers whose infants die often carry the body for days, repeatedly staring at the lifeless form with a dazed expression. Chimpanzees orphaned before the age of five sink rapidly into a severe depressive response characterized by loss of appetite, absence of response to environmental stimuli, and stereotypic body movements. Most orphans recover, but some die soon after their mothers, due in part at least to their depressed condition. Many aspects of nonhuman primate social organization were not understood until groups of recognized individuals had been observed for a number of years (Hamburg and McCown, 19791. Perhaps the most important finding to emerge from long-term field studies is the fact that all monkeys and apes live with close kin and that group social organization is based on kinship ties. In most species, females remain in the group of their birth, whereas males leave at adolescence. In these species, lifelong relationships among females linked through a common female ancestor are the key to group structure (Kurland, 1977; Wrangham, 1980~. In a few species, including our closest relative, the chimpanzee, males remain in their natal groups and females leave; in this case the kinship group is based on father-son and brother-brother ties. In all nonhuman primates that have been well studied, it is clear that individuals form their strongest and most persistent bonds with close relatives and that they cooperate with these relatives against unrelated individuals from other groups. Blood ties form the fabric of nonhuman primate society. Detailed studies of the activity patterns of nonhuman primates in natural habitats have shown that they approximate an optimal balance between energy
62 BEHAVIORAL AND SOCIAL SCIENCE RESEARCH: PART I expended (in survival and reproduction) and energy obtained through food-. All aspects of the life of a nonhuman primate its movements, its interactions with others, the kind of group it lives in, the timing of sexual maturity, reproduction and old age (life history patterns)- are reflections of the nature and distribution of its food resources (Richard, 1981~. Several long-term studies that covered periods of naturally occurring food scarcity show that monkeys reacted immediately to changed conditions: Average weights decreased, age at first reproduction went up, morbidity and mortality increased dramatically-the very old, the very young, and those low in dominance and therefore least able to compete for food were hit the hardest (Grander, 1977~. Parallel studies in the laboratory on the effects of different nutritional regimens on maternal milk production, infant growth and development, infant mortality, and maternal mortality have expanded our understanding of the fine line that exists between a well-fed and an undernourished mother-infant pair. Because the diets, digestive physiology, social organization, and life history patterns of many nonhuman primates are similar to our own, they provide ideal subjects for studying the effects of a changed environment on behaviors associated with survival and reproduction. The earliest research on nonhuman primates in captivity focused on ape intelligence, and recently primatologists have returned to this subject. American sign language, the gestural language of the deaf, has been taught to over a dozen chimpanzees, gorillas, and orangutans, and several have learned over 500 of its "words." Although scientists disagree as to whether the apes' usage of sign language constitutes true language, these animals clearly have a capacity for symbolic communication that has amazed (and delighted) everyone who comes into contact with them. These studies are of interest for two reasons. First, they allow us to conduct experiments on the process of language acquisition that for ethical reasons cannot be done using human children. Second, American sign language provides a medium of communication between human and ape that allows us to see into the minds of another species an accomplishment that, 15 years ago, would have seemed credible only in a science fiction novel. Human paleontologists now believe that the human ancestor who first walked upright, made tools, and invented spoken language had a brain very much like that of the living apes (see the discussion of human origins above and also Lovejoy, 1981~. Thus a better understanding of their minds is one of the most promising avenues of research for shedding light on the origins of the most awesome human attribute, the brain.
