Note: scanned and partially corrected. MURRAY G. MURPHEY, "American Civilization as a Discipline," THE EMORY UNIVERSITY QUARTERLY, VOL. XXII:I; SPRING 1967 AMERICAN CIVILIZATION is a discipline concerned with the study of American society and culture, past and present, with particular emphasis upon the functioning of that socio-cultural system over an extended period of time. Let me emphasize that the focus is upon the system, not upon the unique event, and upon the processes of change within the system over long periods of time; by "long" I mean multi- generational periods. This means, if you like, that American Civilization is "historical" in orientation, since of course it is concerned with what has happened in this socio-cultural system in the past, but it is equally concerned with what is happening to it now and with what will happen to it in the future It also means that American Civilization emphasizes a single society, albeit over a long time period, but that society is viewed in cross-cultural perspective as but one among many societies, used as a microcosm through the study of which we hope to arrive at broader statements concerning the behavior of any similar society American Civilization is therefore interested in long-term human behavior, and although our field work is done in this particular society, as indeed the anthropologist usually does his in but one or. at the most, a very few societies, we are interested in that society as a special case of universal processes. We are not interested i,] what is "uniquely American"; we are not certain that anything is "uniquely American," but we are sure that nothing unique is intelligible until it is brought under some higher generalization, ill which case it is no longer unique. We regard our subject, therefore as the field of human behavior, past and present, and the case example with which we work is America. The basic theories of the nature of society, culture, and personality with which we work are in no sense specific to the discipline of American Civilization. They are indeed the common possession of all the social sciences and are as much the property of any one as of any other. Because anthropologists have most frequently worked with small primitive societies, there has sometimes been in the past a presupposition of homogeneity an assumption that there is but one pattern of behavior characterizing all members of the society. The fallacy of such an assumption, even respecting primitive societies, has been clearly pointed by Wallace (Culture and Personality, 1961. Ch. 1 ), and is even more glaringly apparent when one considers so diverse and heterogeneous a society as the United States. In speaking of the socio-cultural system, therefore the use of the singular descriptive article does not imply any such homogeneity it rather implies that one is dealing with a society which is divided into many socially significant positions, and that although behavioral patterning related to membership in such positions does occur and does involve sufficient complementarity to permit the society as a whole to achieve at least a minimal adjustment to its environment (if the society survives), the degree of uniformity or homogeneity, the amount and range of deviance, and the relation of conceptual, normative, material, motivational, and behavioral systems to the social structure are all questions to be empirically answered. These remarks are of course platitudinous as far as the social scientist is concerned, but they may serve to avoid some misconceptions here. In particular, there is no assumption involved in the discipline of American Civilization that there is any such thing as one American culture, or that one is at all times dealing with one homogeneous society. The merest consideration of the diversity of the colonies should make that point clear, and it remains ( at least to this writer) an open question as to just when the North and South may i~)c said to be one society. Moreover, all present indications point to the fact that, at least in the nineteenth century, religious and ethnic groupings were of such major importance as determinants of behavior that in numerous cases they overrode economic divisions and determined political behavior. Similarly, and most obviously, racial differences in the South before (and after) the Civil War have been determinants of such radically different patterns of behavior that, while these patterns have certainly possessed some degree of complementarity, they are so different that as Stanley Elkins (Slavery, 1963. Ch. 3) has suggested they may even have involved the production of fundamentally different modal personality types. The task of the student of American Civilization therefore, is the identification of the socially and culturally significant positions and relations among positions within the society which constitute the social structure, the delineation of the patterns of practice, value, motivation, and belief associated with these positions, and the study of the temporal covariation among these factors. Naturally, in the course of this work one is seeking ultimately! for explanatory theories, but I think it is by now obvious that there can be no such thing as a body of theory specific to the field of American Civilization. In so far as our endeavor is to study American Civilization as a particular example of a socio-cultural system over an extended time period, there can be no unique theory of this case by its very nature, theory must be general and must apply to all like cases. As our concern is with the full range of human behavior, personal and social, so all behavioral theory, personal an(i social, is relevant to our endeavor, just as it is relevant to the work of any other investigator seeking a rational account of human behavior. If this is the nature of the subject studied by the student of American Civilization, what constitutes his data? The answer to that question is almost self-confident: everything made or said b any member of a society can tell a sufficiently alert investigator something about the culture of that society. Since we are interested in the society as a whole, the prime data for us is data which is descriptive of the behavior of the whole society or of the significant subgroups within that society. The ideal (which is of course unattainable) would be precise information concerning the variables of interest for every individual in the population under studs This ideal is most nearly attained with respect to the data concerning the present, for despite certain exceptions e.g., current operations of the government or of some businesses such data can be increased to a degree limited only by the ingenuity, perseverance, time, and money of the investigator. By far the more difficult question concerns data respecting the past. In a limited number of cases sc have statistical data concerning a past population, or subgroups thereof, collected at the time by federal or state censuses or by some other collection agency e.g., election statistics. But in the general case, the documentary evidence consists of reports or records, official or unofficial, letters, diaries, autobiographies, reminiscences in short, of the vast melange of published and unpublished documentary material ranging from state papers to unpaid bills which is traditionally called "historical data." So far as the period since 1600 is concerned, the problem is not so much the lack of such material (although that is serious enough) as the development of methods which will permit their effective use to answer questions about the society and culture. What must be insisted upon is that while the amount of such data decreases reciprocally as one goes back in time and while it is still possible, by rigorous search, to increase the amount of such material known, one cannot expand the domain at will as one can with contemporary .data, nor can one dictate the form in which the data will be created, as to a very large extent is possible for students of current behavior. The crucial problem here then is methodological: how are ~c to utilize this variegated mass of data to answer our questions? Because of the limitations with respect to availability of data from the past, it is particularly important to stress the significance of the use of material objects as evidence. Archaeologists have long since demonstrated how fruitful the detailed analysis of material artifacts can be in the process of the reconstruction of the culture of antique peoples, and anthropologists have demonstrated the centrality of material artifacts to the understanding of the way of life of any people, extinct or contemporary. In fact, it is quite astounding that historians have shown virtually no interest in this domain of evidence a domain so obviously vital to the understanding of one of the most technologically oriented societies in all history and one which offers a broad range of data for immediate use. In the University of Pennsylvania Department of American Civilization, of which the author is a member, courses in historical site archaeology and others in the analysis of material culture form one of the major subdivisions of the program, and it may be expected that as the significance of this form of data, and sophistication in handling it, become more widespread, other American Civilization departments will give more serious attention to it. It is partly a historical accident that the use of literary and artistic evidence has been so closely associated with the field of American Civilization. Many programs first came into existence as the offspring of affairs between history and literature departments, and the literary men have retained an often disproportionate influence in the conduct of such programs. Yet it is obvious that aesthetic expression forms an important category of human behavior, an(i when the realm of the arts is conceived as including not only belles- lettres but also the entire gamut from painting and sculpture to the decorative arts and technological design, one is dealing with an area without which no ethnography could be complete. Great progress has been made, I believe, in relating such areas as city planning, furniture, and the decorative arts to socio-cultural variables of interest, and some very promising starts have been made with the relation between music and certain cultural variables.l Unfortunately, despite the enormous wealth of literary material! the greatest difficulties in the field have arisen in the attempt to subject this material to uses which can tell us much about the society and culture. The fault here is dual: literary scholars have generally been concerned to prove the importance of their material or racial evidence and therefore have made rather grand and naive claims for it, while more scientifically oriented students, having found that literary material is a poor description of actual behavior in the society, have too quickly depreciated its importance. This sort of impasse is most apt to arise where American Civilization is viewed as an interdisciplinary field, since the representatives ,)f the several disciplines then feel honor bound to proclaim the significance of their respective fields. On the other hand, when American Civilization is conceived as itself a discipline, one can forget such political problems and turn to the far more rewarding question of the role of certain types of aesthetic productions in particular socio-cultural systems at particular times. It is strongly to be hoped that further work along such lines will rid us of some of the tedious and utterly futile controversies which have beset the field in the past. The basic questions with which the discipline of American Civilization is concerned are those of the structure and change of socio-cultural systems. It is therefore obvious that the methods adopted must be such as will permit answers to those questions. Since we are always concerned with how individuals behave as members of groups, we may first turn to means of characterizing the behavior .)f many individuals, or of aggregates of individuals. There are of course a variety of standard statistical methods which are applicable to this type of problem; indeed, so far as contemporary data is concerned the only limitation upon the exploitation of current techniques of measurement and statistics is the expertise of the investigator and the problem of the comparability of the results so obtained from past data. The problem of the application of statistical techniques to materials from the past is a far more difficult and as !et largely unexplored area, but one which will certainly undergo very rapid development in the future. Many standard statistical techniques are applicable now to limited sets of historical data, but general applicability has yet to be achieved. The fundamental problem here is that the set of individuals from a past population for whom information relevant to a question of interest now remains is not only not a random sample of the original population, but there exists no clear way in which to estimate either the magnitude or, in many cases, the direction of the bias. This is a problem in theoretical statistics for which mathematical solutions will have to be achieved, if indeed solutions are possible. Meanwhile, there remain innumerable special questions in which a sophisticated use of present techniques can greatly facilitate research. To mention but a few examples, if one were interested in making a study of one thousand legislators from past congresses one might well find that data was readily available for only half of them, and the task of attempting to dig out the relevant information for the remaining five hundred, even assuming it exists, might well be sufficiently for. bidding to lead to the termination of the study. However, if a random sample of one hundred of those legislators were drawn, and if as one might reasonably expect, data was readily available for approximately half of them, the task of digging out the relevant information for the remaining fifty looks much more manageable and if successful would yield estimates respecting the total population of interest which would probably be quite precise enough to solve the research problem. More generally, we often find ourselves facing a situation in which information relevant to our problem exists for only a few thousand individuals from a past population of several millions. Even where the magnitude or direction of the relevant biases cannot be estimated, it is clearly more significant t(. know the characteristics of the few thousand "survivors" than o? -one or two "typical examples," where "typical" means special cases chosen for unspecified reasons by the investigator. Such a description of the set of the few thousand "survivors" is quite within the range of standard sampling theory. There are a number of cascs in which we are so totalty ignorant of the reasons which led tn the preservation of the data respecting these individuals that nothing can be said about the bias of that group. This number is I suspect, much smaller than one might at first think. A furthcl example of the application of statistical procedures to historical data, and one of great significance, concerns the case in which we find ourselves in possession of statistical data collected in the past but reported for some aggregated unit such as the county or parish. We have all learned from Robinson et al that "ecological correlations" that is, correlations on such aggregated units are not acceptable estimates of the correlations which obtain for the individual members of those aggregated units. Nevertheless, more sophisticated estimating procedures have become available in recent years which partly offset this difficulty, and we have also learned to live with the ecological unit as the unit of interest as well as observations and to draw meaningful conclusions from ecological correlations, conceived as correlations upon areal units rather than .on residents of those units. The extraordinarily interesting and important work of Thomas Alexander on Southern voting behavior before the Civil War (in The labama Review, 1966) is a l)rime example of what intelligence, hard work, and determination can accomplish using such techniques. In the area of methodology, therefore, one can look forward with confidence to an increasing , >. of statistical and mathematical methods in the near future. Finally, although few such studies have yet been done, it is clear that American Civilization offers a prime opportunity for the use of stochastic models. As our endeavor is to describe and analyze processes of change over time, formal techniques which permit the study of sequential processes, where any given outcome depends upon some specified number of prior outcomes, are clearly made to order for our purpose. Studies of intergenerational, occupational and class change and of voting behavior have already provided indications of what may be hoped for in this area, and of course there has recently been a vast development in the application of such models in psychology. Hand in hand with such increments in mathematical methods has come the increasing use of the computer as a means of carrying through the requisite procedures. So far as the processing of large quantities of data is concerned, the computer is to the student of culture what the telescope was to the astronomer; when used with skill it will make possible the delineation of the general outlines of aggregate behavior in ways which have heretofore been impossible It would be impossible here to list the applications of the computer already made or in progress. The vast archival project undertaken by the Inter- University Consortium for Political Research at Ann Arbor will soon make available in machine-readable form all American county election returns for major offices from 1820 to the present, together with certain related demographic variables, and the studies which will issue from that project in the near future will be legion. Nor are the applications limited to voting behavior: Joel Silby's recent application of scalogram techniques to the analysis of the behavior of Congress between 1840 and 1850 ( The Shrine of Party, 1967) shows what can be done by such means with legislative behavior, and Robert Zemsky's forthcoming study of the Massachusetts Assembly from 1730 to 1755 (unpublished dissertation, Yale University) shows that such techniques are equally applicable to the colonial period. Nor are they limited to political or economic subjects Anthony Garvan is now engage(l in a computerized study of the relation between church silver anši church practice during the colonial period (Some Statistical Aspects of American Church Plate, forthcoming), a project which illustrates nicely the application of mathematical techniques to the study of material culture. Indeed, the ability to make effective use of the computer will in the future be a necessity for all researchers in this field. , In the domain of qualitative methods, American Civilization includes all those methodological techniques developed by social scientists for dealing with the contemporary situation and in addition the full range of historiographic techniques. For example studies recently completed or now underway in this department range from Bruce Biever's Crisis in Catholicism (forthcoming) which was based upon interviews with several thousand Catholics both in the United States and in Ireland, to Stuart Blumin's study of occupational mobility in Philadelphia in the Jacksonian era, and from James Flink's study of the social processes involved in the introduction of the automobile to Norman Yetman's attempt to test Stanley Elkin's thesis respecting slave personality through the use of personal documents collected from slaves. Thus no division in principle need be drawn between contemporary and historical research: the general canons underlying all empirical inquiry are essentially the same and the differences which arise are essentially those which result from the differences in the nature of the data. Interviews, questionnaires, and direct observation are not possible for students of the past, who in this respect rather resemble the astronomer they see but a limited portion of the universe they wish to describe, they cannot experiment directly upon the objects in which they are interested, and they must in most cases wait upon observation to bring before them the information that they seek. Yet significant advances are being made in this area. If we cannot interview the dead, we can nevertheless develop more and more sophisticated means of extracting information from the traces they have left us. Much of my own work in recent years has been involved in just this question of eliciting information concerning attitudes from documents written in the past.4 Furthermore, as the social scientists provide us with increasingly sophisticated theories against which we can interpret historical data, more and more reasonable estimates can be made concerning characteristics of those past populations. Nor should the importance of the methodological contributions of classical historiography be obscured in the rush to embrace more formal, but not necessarily more accurate, tools. Classical historiography has made three great contributions which I believe even historians have rarely taken as seriously as they should. First, it laid great stress upon the criticism of the particular item used as evidence usually the document. That type of criticism, both internal and external, is of the highest importance, not only with respect to documents, but with respect to every type of data. The discovery that the Piltdown skull was a forgery, and the recent redating of Zinjanthropus are as much triumphs of external criticism as the proof that the Horn papers were fallacious all resulted in significant alterations of our concept of the past. Second, methodology in general owes an immense debt to those generations of Biblical scholars whose patient devotion to the problem of the gospels established the classic canons of historiography and still afford the paradigm case of the introduction of explanatory constructs into classical historical study. Thirdly, historians have traditionally insisted that qualitative and subjective factors are indispensable in historical interpretation. Indeed, if one examines those monumental attempts of German scholarship to codify historical method, what one finds is essentially a set of rules by which it was hoped that the historian could minimize subjective caprice an attempt, if you will, to discipline and subject to rational constraints the qualitative and subjective convictions which scholars derive from their materials. Positivistically oriented philosophers and so-called scientists have railed against this sort of methodology nevertheless, I believe that the historians have been on the right track, and the rapid development within the last two decades of Baysean statistics and subjective probability as fields of serious study point to a potential meeting of minds on these ques- tions. Although such statements are necessarily conjectural, l strongly suspect that the methodology which will emerge in this field will constitute a new blend of traditional Baysean techniques which will form a meeting ground for the rigorous and often too positivistic methods of that social sciences and the qualitative an(l often too undisciplined judgments of traditional history. A third and major methodological domain is of course that of archaeological method. As noted above, archaeological data is just beginning to receive the attention which it deserves as a source for the study of American culture; nevertheless, archaeology has long been a major branch of research in all classical studies, in ancient history, and in physical anthropology. We have, therefore, a great amount of basic methodological expertise which requires only modification and minor revision to make it applicable to the problems of American Civilization. The work of John Cotter at Jamestown (Archaeological Excavations at Jamestown, Va., 1958) and the further archaeological work now underway under the direction of Cotter and Garvan are already rewriting certain chapters of early American history, and will serve to establish this as a major area of research within the field. It remains to speak of the methodology appropriate to the study of literature from the standpoint of American Civilization. For reasons already indicated, this is not an issue over which dispassionate discourse is likely, but certain remarks are, I believe, in order. In the first place, the existence within a given culture of certain aesthetic norms is clearly one of the major facts of interest about such a culture, and this is a question which can be approached through the study of those objects which people in the society have regarded as having aesthetic value. Yet if one conceives of literature more broadly, so as to include dime novels, magazine fiction, and even comic books or as including movies and TV shows it cannot be assumed that the role of such materials in the culture is solely or even primarily aesthetic, or at least aesthetic in the usual sense of the term. The point may perhaps be made most sharply l~y contrasting symphonic music with rock and roll: leaving aside for the moment the question of how the various members of the society value these forms of music "aesthetically," one may well ask ~hat the functions of these particular types of music are for different groups in the culture. As Neil Leonard has pointed out (Jazz and the white Americans, 1962. Ch. 2), it is of great interest that jazz music was attacked by certain groups in America, not on musical grounds, but on moral grounds. This at once suggests that, quite apart from any questions of "aesthetic" worth, there are here involved deep-seated moral values and attitudes toward emotion and its expression which are perhaps most sharply revealed in these cases of innovation in the arts. To put the matter more generally, literary and artistic work is done by certain groups of people who stand in certain relations to other groups. They pro. duce certain types of products which do or do not affect definable audience groups in various ways. What are these relations among the relevant groups ? What are the characteristics of a given production which determine how it affects different groups of audience? What can the production tell us about its creator, its audience, or other groups? These are questions concerning literature and the arts which are significant and which are highly relevant to the student of American Civilization. There are, of course other questions about literature questions about the aesthetic merits of the work as a work of art but these are questions for the student of aesthetics or literature per se. They are questions of great importance, but they lie beyond the purview of the discipline of American Civilization. Perhaps an inevitable question is how American Civilization differs from other "disciplines." As I have tried to indicate above and indeed as social scientists have repeated to the point of weariness, there is in a broad sense but one field of social science the endeavor to provide a rational account of human behavior. Nevertheless, "fields" do differ in the nature of their central emphasis their focal point of interest. One might therefore say that what most concerns the workers in the field of American Civilization is the dynamic interaction of the components of the socio-cultural system over extended periods of time. Many recent studies, notably those of John Whiting and Irvin Child (Child Training and Per sonality, 1958) and David McClelland (Tlte Achieving Society 1961), have made us very aware of the relation between the kind of personalities that a socio-cultural system produces through its child-rearing techniques and the subsequent effects of those personalities upon other institutions in the culture, including those very institutions which condition the socialization of the next generation. One is dealing here with long-term processes of change processes obtaining over many generations and therefore with problems which are at least in part intrinsically historical. Nor will it do to say that we can simply begin now to collect the data so that in a hundred years our successors will have the materials needed to answer our questions: if any one fact is certain it is that the questions our successors will want to ask one hundred years from now will not be the questions for which we are now prepared to collect the data. What we in this field seek to do is to develop way and means by which such processes of long-term change can be studied now, not only in this particular society, but in general, and to find ways to describe and understand the character of those changes. That in this search we find many others following parallel courses is for us a heartening fact historians and economists, anthropologists and political scientists, psychologists, sociologists, and civilizationists can all learn from each other, and the more friends we find walking with us the likelier we are to reach our common goal . NOTES Anthony Garvan, "The New England Porringer, An Index of Custom," Smithsonian Institution Annual Report, 1958, "The New England Plain Style," Comparative Studies in Society and History, October, 1960; "Proprietary Philadelphia as Artifact," in " Historian and the City, eds., Handlin and Burchard, 1963; Neil Leonard, Jazz and The White Americans, 1962. 5N'. S. Robinson, "Ecological Correlations and the Behavior of Individuals," American Sociological Review, 1950. There have been innumerable discussions of this problem in the literature since Robinson wrote. See for example Leo Goodman, "Ecological Regressions and Behavior of Individuals," ibid., 1953; O. D. Duncan and Beverly l)avis, "An Alternative to Ecological Correlation," ibid., 1953. '>. e John Kennedy and J. L. Snell, Finite Markov Chains, 1960, pp. 190ff; S. S. Ulme, Harold Guetzkow, William Riker, Donald Stokes, Mathematical Applications in Political Science, Arnold Foundation Monographs XII, (Southern Methodist University) Murray G. Murphey, "An Approach to the Historical Study of National Character," Context and Meaning in Cultural Anthropology, ed., Melford Spiro, 1965.