Myths of the Chicago School

5. Chicagoans as atheoretical empirical researchers

5.5 Chicago theorising and sociology in the United States

5.5.1 The epistemological basis of 'standard American sociology'

5.5.2 The cumiulative-falsificationist model

5.5.3 The Conference on the Polish Peasant

The view that Chicago sociology was essentially concerned with empirical data collection, or the view that its theoretical work was confined to the pioneering stage of urban sociology, are both indicative of an idea that somehow Chicago stood outside the general development of American sociological theory. The emphasis on the empirical concerns of the Chicagoans in presentations of their work tends to act to separate them from the sociological mainstream. While the Chicagoans were in the vanguard of the shift from 'armchair theorising' to 'inductive theorising' they were not alone in this endeavour nor, indeed, were they the sole pioneers in empirical data collection.

The social survey movement, based on ameliorist concerns of British social surveyors in the nineteenth century, was well established in the United States by the time the Chicagoans adopted an empirical base for sociological theorising during the 'golden era' (Furner, 1975; Schwendinger and Schwendinger, 1974). Following in the footsteps of the Booth survey of London, a survey movement sprang up in the United States. The Pittsburgh Survey (1909-1914) was followed by the establishment of a survey department within the Sage Foundation and by 1928 its director reported 154 general surveys and 2621 specialist surveys. These were not sociological surveys: much of the survey tendency 'took its own course, diverging early from academic sociology and finding a close partnership with organised welfare activities' (Faris, 1967, p. 8).

Nonetheless, as Burgess wrote in 1916 while at Ohio State University, the example of the Belleville and Lawrence surveys directed by F. W. Blackmar at the University of Kansas was indicative of the increasing involvement of sociologists in social surveys. The social survey of a community is 'the scientific study of its conditions and needs for the purpose of presenting a constructive program for social advance' (Burgess, 1916).

Park's course at Chicago on the social survey (from 1915 to 1921) provided the critical framework in which empirical work was developed. His critique of the social survey was not only that it tended to overemphasise statistical data and reformist concerns but that, in so doing, it also distanced itself from sociological theorising. In this, he reaffirmed the stance of Small and Thomas and reflected the growing tendency towards 'inductive theorising' in American sociology.

Besides the empirical work done at Chicago, other sociologists were engaging with empirical data in their attempt to develop sociological theorising. Notable was the work at Columbia under the directorship of Giddings, which included Ogburn's early research; developments at the University of Southern California; as well as the research done at Michigan (Cooley, 1930; Rice, 1931). Indeed, Lapiere (1964) recalled that 'one man' departments all over the country were springing up with the aim of developing a credible empirical base for sociological theory. As Chicago sociologists became more and more concerned with empirically based study rather than 'armchair theorising', they reflected the emerging tendency in American sociology as a whole.

5.5.1 The epistemological basis of 'standard American sociology'

While the Chicagoans were among the pioneers in the development of a sociological theorising grounded in empirical data, they were far from unique in this endeavour. Their empirical data collection orientation should not be viewed, in itself, as indicative of a separation of 'Chicago sociology' from the direction of mainstream sociological theorising. Further, it is necessary to address the extent to which Chicago sociologists developed a distinctive 'interpretive' or 'phenomenological' approach [5]. The following sections attempt to set out the approach to sociology in the United States that emerged out of the 1920s and prevailed into the 1960s (following Mullins (1973), this will be called standard American sociology) and to assess the extent to which the epistemological perspectives at Chicago differed from it.

5.5.2 The cumiulative-falsificationist model

Over the course of the twentieth century, American sociology has emerged as an empirically grounded endeavour, adopting, in the main, what may be described as a falsificationist model of the production and validation of sociological knowledge [6]. In conjunction with such a model, there has been a tendency to accept the idea of the cumulative development of theory (Social Science Research Council, 1939; Merton, 1945; Mills, 1959; Willer, 1967).

Arguably, American sociology has also been dominated (if not exhausted) by two general theoretical perspectives, 'interactionism' and 'functionalism'. While not entirely compatible these two perspectives overlap and do not constitute the theoretical base of two distinct traditions. Rather they concur on many elements of the sociological process. Both are essentially nomothetic, both are anti-behaviourist, both focus on group processes, both are empirical and non-critical. The one approach assess the function served within a social process by a particular phenomenon, the other assesses the interactive process in order to see the way action is mediated by social processes. Clearly they are interrelated, to some extent the opposite sides of the same coin.

The prevalent model of sociological knowledge in American sociology emphasised cumulative theory, falsificationism, and nomological concerns incorporating meaning adequacy (Harvey, P., 1982). The tension that existed concerned the extent to which the problems of establishing adequacy at the level of meaning should inhibit nomological concerns. The Conference on the Polish Peasant, 1938, reflected these concerns and acted as an indication of the pervasiveness of the prevalent model. No alternative, 'anti-positivistic' model emerged from it; despite the existence of an embryonic critique in the work of Blumer and of Wirth.

The conference on the Polish Peasant was an important, well-documented and widely known debate; even if the participants were not 'representative' of the entire gamut of American sociology. A 'centralist' approach to sociology emerged from it despite some scepticism over the nomothetic possibilities voiced by some of the Chicagoans. The discussion on the Polish Peasant clearly indicates how the main institutionalised perspective in sociology saw the nature of the discipline. While the analysis of the conference that follows shows the points of debate, it also reveals a strong commitment to a more-or-less agreed view of the nature and aims of sociology. Chicago does not stand outside the general approach.

5.5.3 The Conference on the Polish Peasant

This was called by the Committee on Appraisal of Social Research of the Social Science Research Council, in New York, on December 10th 1938. The main contributors of those present were three Chicago professors, Herbert Blumer, Louis Wirth, Samuel Stouffer in addition to G.W. Allport, Read Bain, Max Lerner, W.I. Thomas and W.W. Waller.

The debate followed Blumer's written critique of the Polish Peasant. Initially the debate was directed to the efficacy of personal documents as a device for testing theoretical assertion. In developing his written submission Blumer pointed to the premises underlying the Polish Peasant. These were, first, the need of a plan of research suited to a complicated changing society that may be applied to any society undergoing transition, and, second, the declaration that the understanding of human life necessitates the grasping of the subjective factor. Thus any sociological study should involve both the external factors (social values) and the internal factors (attitudes).

Given the above, Blumer argued that two things were necessary. First, to develop a guiding theoretical scheme that will set hypotheses, that is, provide a framework for interpretation and analysis. Blumer reckoned that in the Polish Peasant all the theories are developed in 'intrinsic relation to these basic concepts of attitude and value'. Second, the scheme needs source data that will reveal the 'subjective factor in human experience and which, at the same time, will meet the usual requirements for scientific data, viz., that one can always go back to these data and that other workers may have access to them' (Social Science Research Council, 1939, p. 108). The resulting debate thus tended to turn attention from the particular problems of personal documents as evidence to address the problem of proof, causal attribution and objectivity in sociology given the 'need of recognizing and considering the subjective factor in human experience'.

The conference spent considerable effort looking at the implications of Blumer's critique for the human document as data and, in so doing, broached the nature of scientific enquiry of society and of alternative approaches to sociology. Throughout, however, certain elements were taken for granted. Most notable was the primacy of inductive theorising. Interestingly, however, there was also a constant concern to ensure that sociology adopted approaches that do not regress to atheoretical fact collection any more than it should to accept non-empirical abstract theorising. In short, the idea that sociology should progress through cumulative development of theory grounded in empirical data, in a manner that Popper was to codify in the natural science debate as 'falsificationism', underwrote the discussion.

The concern was with how different types of approach could provide 'objective', that is 'reliable', data. The accepted approach closely resembled what Merton (1948) came to call middle-range theorising [7]. Indeed, prefacing this, Lerner noted

Here we have the empirical data, here we have the abstraction into which we are attempting to fit them. Modify the abstraction to fit the data and go and collect more data to fit the new abstraction; there is a constant interaction between them. (Social Science Research Council, 1939, p. 177).

The debate was concerned with the realist position that is central to the Polish Peasant's methodological note, that the subjective factor be incorporated into social analysis. The concern of the conference when assessing human documents was that while they provide some 'attitudinal' insights, they may not be 'objective' in as much as they may prove to be inadequate for consistent inductive inferential or testing purposes.

There were differing views on the degree to which human documents could be treated as reliable, 'objective', or self-evident facts in the sense of data in the physical and biological sciences. This assumed that the model of science rested on a factual base. Human documents, Bain contended, were non-specific instances, reflecting a whole culture and could not be seen as data in the sense of data in the natural sciences. Blumer and Wirth defended the principle that documents, although specific instances, may provide abstracted data. They argued that the document can be used as a specific instance through an abstractive process that indicates what aspect of the document is being considered through an application of a prior conceptual scheme. Wirth suggested that, for example, 'Thomas does this by saying that this is a case of restlessness or new experience. He says, 'Now I am focusing on those elements in the human document which to me incorporate this particular motive or this particular attitude' (Social Science Research Council, 1939, p. 119).

Wirth and Blumer argued that no data is as simply abstract as the ideal model of the physical sciences pretends, and that in the social sciences, one must incorporate meaning. However, these points are subverted by the underlying falsificationist objectivist position.

At no time during the discussion was there any suggestion that data has meaning only through its theoretical base, (thus the approach resembled what Lakatos (1970) refers to as naive falsificationism [i.e. it ignores the concept of the theory-laden nature of observation]), nor that the culturological frame raises problems of a hermeneutic nature. On the other hand, the realist position was rarely threatened and only Stouffer and Bain had any reservations about the necessity to include the subjective factor; and that only because of  the possibility of some valid sociology having put aside the subjective aspect. Stouffer, in asking whether the taken-for-granted subjective factor really is so vital, pointed to the prediction studies of Glueck, Burgess and Vold that are 'very important for social action' and that provide 'fairly valid conclusions and generalizations' but are independent of the 'meaning of any of this activity to the individual'.

Nonetheless, there was agreement on the view that social science does involve, for whatever reasons, problems different to the natural sciences, and that these problems must somehow be overcome if sociology is to enhance its objectivist, falsificationist, scientific credibility. There was a general concern that sociology should avoid a 'nihilistic' attitude, given that social scientific research can only be seen as plausible not definitively validated, and that no laws (or approximations to laws) as directive of action can be drawn up at anything but a trivial level.

Bain summed up the cumulative-falsificationist orientation to science and its application to social science.

Those grand generalizations always get tested by being broken up into a great number of simple problems. What we call 'progress' in all the natural sciences, among which I would include the social sciences, has come about through the development of the art of stating simple or unequivocal propositions, or hypotheses, which are capable of empirical test. When enough such propositions have been tested and retested and all of them are logically consistent with the grand generalization, it may be said to be verified. The empirical verification of no one single hypothesis relevant to the general theory of organic evolution can be said to be adequate proof of it, but when thousands of such simple single hypotheses have been verified, and they all hang together—none of them are clearly negative cases—we eventually come to accept the general theory of organic evolution as an actual valid scientific fact. (Social Science Research Council, 1939, p. 161).

One aspect of the debate is an overall desire for a synthesis of grand abstraction and scientifically precise empiricism, a view that methodological monism is not particularly desirable, and that eclecticism should be further developed rather than discouraged. This is reflected in Thomas' own reappraisal of the Polish Peasant in which he commented that he would put methodological considerations out of view if doing a similar study and restated his view that a mixture of approaches is most suitable, i.e. life histories and statistics, including factor analysis.

In an appended summary to the transcript, Bain (the transcript editor) indicated what he saw were the main divisions among the conferences, and which reflected differences in American sociology. There was disagreement over the nature of social phenomena, methods by which they can be studied, the possibility of laws and of the testing of generalizations. Bain suggested two opposing views of validation, which he saw revolving around the issue of the possibility of social laws. On the one hand the view that social laws were impossible given that they could not grasp 'values'. The alternative view, accepting the possibility of social laws emphasizes the idea that 'laws are possible because there is considerable uniformity and permanence in the occurrence of observed and observable social phenomena, whether they be called 'objective' or 'subjective''.

Bain then identified four theoretical positions expressed at the conference, first, illuminative insight, second, organising concepts, third, logico-systematic analysis and fourth, delimited empirical research. The last is opposite to illuminative insight, it is reductionist and demands framing propositions for testing, therefore requires relatively simple problems, few controlled variables, accessible and permanent data which are uniform and repeatable.

It is the most highly abstract way of dealing with concrete, i.e., experienceable, reality. It stresses verification by repetition, prediction, application, external and internal logical consistency. It is based on the probability calculus, it is actuarial or statistical. It advocates the development of precision instruments for use in observation, recording or manipulation, as the indispensable prerequisites for sound scientific work in any field. It holds that the history of science is the history of scientific technology.... The general methods and point of view is the same for all science, though necessarily the particular methods, techniques and technological devices used will vary greatly with the data being studied... All scientific data are abstractions... It is out of the cumulative findings of such simple, particular, highly abstracted empirical researches that the material for valid general scientific theories must come. It is by such research only that 'causal validity' can be ascertained and upon it, at long last, that all 'meaningful validity' must depend. (Social Science Research Council, 1939, p. 201).

Bain concluded that none of the Conference participants saw any of these positions as adequate alone, all agreed that some kind of synthetic position was required.

However the trend is towards the type of research called 'delimited empirical'; logico-systematic analysis is increasingly dependent upon such research; organizing concepts tend to grow out of such research and to be tested by it in the general manner described. This is a continuous process. The specific researches make imperative the revision of organizing concepts and general theories, and such revision by logic-systematic analysis sets new problems for further empirical research which requires the development of new or improved precision procedures which depend upon the invention of new or improvement of old technological devices of observation, recording and manipulation along with new or improved methodological skills and procedures. (Social Science Research Council, 1939, p. 202).

Thus, Bain clearly laid out the cumulative-falsificationist model that had emerged as prevalent in American sociology. Ironically, this 'delimited empirical' approach with its reductionist emphasis actually provided the potential for a division in American sociology. Its failure to take into account the theory-laden nature of observation, meaning adequacy, the nature of historical evidence or the cultural frame which Bain referred to provided the basis for a fundamental critique of the falsificationist approach to sociological research. This critique existed in embryonic form at Chicago in the position advocated by Wirth, Blumer and even in the 'culturological' approach of Chicago's major quantitative practitioner, Ogburn. Nonetheless, as will be examined below, the Chicagoans tended to remain within the prevailing tradition rather than engage it in a radical way.

While the Polish Peasant conference was a major event in American Sociology and illustrative of the development of a consensus orientation towards sociology, it is, of course, not a definitive statement. However, similar issues were raised and discussed within very similar constraints when, for example, the American Sociological Review was inaugurated and the issue of operationalisation engaged (Lundberg, 1936; Waller, 1936). Similarly, the earlier discussions about the relative efficacy of case study and statistics were contained within a framework of debate that took for granted the essentially nomothetic concerns of the cumulative theoretical approach [8].

Notes

[5] No attempt is made here to define the terms 'interpretive', 'phenomenological' or 'positivistic' as they are not used as a basis for comparison but merely indicative of the type of general contrast implied by some commentators when comparing, for example, 'Chicago sociology' with the structural-functionalism of the Columbia sociologists (Bogdan and Taylor, 1975). Return

[6] The use of the term falsificationism here is indicative of the kind of approach that predominated in American sociology throughout the period of this study. This is not intended to represent an assertion about the nature of sociological enquiry through a discussion as to whether American sociology is characterised by inductivism or hypothetico-deductivism, both of these are subsumed within the term falsificationism, especially as characterised by Lakatos (1970). Lakatos defined, in effect, three levels of falsificationism, naive, sophisticated and the refined version of sophisticated falsificationism to be found in his own methodology of scientific research programmes. The use of the term here is not meant to refer to any one of these in particular, but to embody the central tenets of falsificationism, viz. conjecture and refutation, cumulative progress through empirical validation of theoretical, falsifiable statements, and the acceptance of the impossibility of deductive or inductive proof. The niceties of the debate as to how science progresses, which underpins Lakatos' distinctions of types of falsificationism is not germane to the use of the label here. As has been shown elsewhere (Chalmers, 1978) all falsificationist models ignore, in the last resort, the value laden nature of observation; see science as ultimately self-legitimating through its own protocols; and divorce scientific knowledge production from the wider scientific milieu. It is this critique combined with the general characteristics of falsificationism that makes the term appropriate as a descriptor of American sociological endeavours in the period under consideration. Return

[7] Merton codified the cumulative theory approach in various articles in the 1940s, which became the basis for the middle-range theorising perspective so important to structural functionalism. The reference to middle-range theory throughout the book is directly to Merton's formulation, although it is argued that such a formulation reflected sociological practice to which the Chicagoans subscribed.
 Merton (1948) noted the continuity of theory and cited various instances including some integral to Chicago, notably the 'conflicting self' or 'marginal man'. He pointed to developments in this sphere of theorising but suggested that the central problem of conflicting roles 'has yet to be materially clarified and advanced beyond the point reached decades ago. Thomas and Znaniecki (1918) long since indicated that conflicts between social roles can be reduced by conventionalization and by role-segmentation' (Merton, 1948, p. 515).
 Merton (1945, 1948, 1949) in laying out the basis of middle-range theorising attempted to forge a clear link between empirical research and social theory. Rather than 'the social theorist high in the empyrean of pure ideas' being replaced by the researcher 'equipped with questionnaire and pencil and hot on the chase of the isolated and meaningless statistic', he saw the interaction of theory and empirical research with empirical data informing theory and vice versa. In practice, however, he maintained that there were still those sociologists who did not link theory with research. Merton (1949) reflected Bain's contribution to the Polish Peasant debate of a decade earlier (Social Science Research Council, 1939) when he identified six approaches to theorising: methodology, general sociological orientation, analysis of concepts, 'post factum' interpretation, empirical generalisation and sociological laws. Methodology, he argued has nothing to do with substantive theorising. Conceptual analysis, Merton argued, is indispensable if confined to clarification of key concepts. However, to think of conceptual manipulation and definition in itself as theorising is spurious. Applying conceptual schemes in a 'post factum' and (heinously) ad hoc manner to data, similarly does sociological theorising a disservice. Indeed, Merton scathingly attacked approaches that collected data and then subjected them to interpretive comment. He regarded such approaches as having the logical structure of clinical enquiry because they do not test pre-designed hypotheses. This applied to both statistical and case-study data. The result is a merely plausible explanation. General orientations merely indicate the approach, such that Durkheim's orientation was that social facts should be sought in the facts that preceded it, and Znaniecki and Sorokin (amongst others) invoked a 'humanistic coefficient' as orienting principle. They are non-empirical generic orientations and must be specified as empirical generalisations. In isolation such generalisations are nothing more than summaries of observed uniformities in observational data. It is the combination of concept clarification, orientation and empirical generalisation within a theoretical frame that provides sociological theory. At the extreme this manifests itself as sociological laws. While this status is rarely achieved, Merton argued that it is possible to work towards it through the cumulative development of theory. Middle-range theorising provides that possibility. Merton saw 'middle range theory' as the pragmatic answer to the continuing development of sociology, which, he admitted, must 'ultimately meet the canons of scientific method'. In this respect, Chicago sociologists would not have disagreed. Return

[8] Cooley is a possible exception. 'It was more and more borne in upon me that I could never really see the social life of man unless I understood the processes of mind with which it was indissolubly bound up. I saw that there was a gap between the ideas of structure and function I had so far been working on and the actual motives and behavior of men, which left the former somewhat hanging in the air…' (Cooley, 1930, p. 30). Referring to his work in the first decade of the century. Return

Next 5.6 The Chicagoans general theoretical and epistemological orientation