Two Ways of Looking at Theory, Exemplified by The Dynamics of Bureaucracy by Peter M. Blau

1 Theory as a Text

The conventional approach to theory, to repeat, is to view it as embodied in a book or an article. Here you want to extract the theory through a close reading of the text. In doing so you need to get a sense for the general content of the study, but what is of special interest are the places where the theory is explicitly discussed. This is especially true for its full formulation at the end.

If we approach The Dynamics of Bureaucracy in this manner, we end up with an account of the following type, based on a reading of Blau’s book as it appeared in 1955. The volume, as the reader would notice, was published by the University of Chicago Press, and its author assistant professor at the Department of Sociology at the University of Chicago. Blau begins his work, which is 269 pages long, by stating that he will be studying bureaucracy; and that he will do this with the help of the theories of Max Weber and Robert K. Merton. Weber, he says, has produced the classical theory of bureaucracy: it is an organization characterized by such features as abstract rules, strict discipline and hierarchy; and these make it work with the speed and precision of a machine (Weber 1946:214–16, 196–98). Weber did not, however, base his analysis on systematic data, Blau says, something that is necessary for a reliable theory. Furthermore, Weber also made the mistake of describing the features of a typical bureaucracy at the same time as stating what the effects of these features are. This, however, is not possible, Blau says; only through empirical research can you state with assurance what the result of a social structure is.

Merton, Blau writes, has deeply influenced his sociological thinking, especially his ideas on middle-range theory and functionalism. Merton was very critical of the use of general theories; and advocated in their stead theories with a considerably more narrow scope, based on systematic data. Merton had also developed his own version of functionalism, according to which dysfunctions, latent functions and manifest functions, all play a role and have to be taken into account.

After these introductory remarks on the ideas he starts out from, Blau gives a detailed account on the next 150 pages of his research. In order to find out how a bureaucracy works in actual reality, Blau repeats, you need systematic data; and for this reason he has conducted a case study. Surveys can not provide the kind of detailed and precise information on interaction in work groups that he wanted to study. Blau then describes how he studied two small public bureaucracies, paying particular attention to the way that people went about their tasks in these. About 20 people worked in each of the two organizations.

One of these these two bureaucracies was a state agency, with the task of helping unemployed individuals find a job; the other a federal agency, which checked that employers followed a recent labor law. In 1948 and in 1949 Blau spent three months in each of these two agencies, during which he engaged in systematic observation of how people carried out their tasks. As often as possible he tried to quantify the interactions in the work groups, especially how people reacted to instructions from the administration.

Of particular interest in the employment agency, Blau says, was the attempt of the administration to increase efficiency. This was done by publishing statistics on how many referrals each employee had made; and the idea was that this would make the employees work harder. The measure was considered successful since it resulted in higher productivity. But it also intensified the competition among the employees, and this meant that the cases that were extra hard were not attended to. This negative effect, however, was partially offset by the administration’s decision to let a special unit handle them.

In the second agency Blau’s attention was drawn to the fact that the employees often consulted each other when they worked on a difficult case. To do so, however, was explicitly forbidden; the employees were only allowed to consult a supervisor if they needed help. In the social pattern of consultation that now emerged among the employees, some employees were consulted very often, with the result that their status increased and the cohesion of the work group was threatened. This was however offset by the tendency of those who sought advice to become extra social, and in this way restore equality in the interactions.

This represents the end of Blau’s account of his research in his book. In the final chapter Blau pulls together his analysis of the two agencies and presents the theory that is the result of his study. What he says here is the following. While Merton had argued that dysfunctions disturb and upset the workings of an organization, this is not always the case. In some cases the dysfunctions may instead result in counter measures which can be successful. Blau refers to the latter in functionalist terminology as activities in response to “emergent organizational needs” (Blau 1955:251–55).

As examples of this, Blau cites the two cases just cited: when statistics was introduced in one agency, to increase production; and when the consulting pattern in the other agency first disturbed the cohesion of the group of employees but was then restored. Blau concludes from these and some similar cases that Weber’s theory of bureaucracy has a flaw. Bureaucracies are not static but in constant change and therefore dynamic. Hence the title of his book: The Dynamics of Bureaucracy.^[2]

2 Theory as Activities, Part I (Only Using the Text for Material)

What has been presented so far can be described as a fairly conventional view of how a sociological study is read and how its theory extracted.^[3] You note all the places where theory is referred to but focus especially on the theory as it is presented at the end of the study. Blau himself presented the theoretical core of his study along these lines since he thought that a theory could best be described in a very concise way, as a set of propositions.

This, however, does not represent the only way to look at the role of theory in a study; and it is neither a very helpful one if you want learn how to work with a theory and further develop it. There are two major reasons for this. First, the type of reading of a study that has just been described is reductive in its view of what theory is. The focus is squarely on the final version of the theory as the crowning achievement of the whole work – but isolated from method and facts. The theory is also reduced to a few sentences or to an even smaller number of propositions. “Sociological theory is a system of interrelated explanatory propositions” (Blau 1956:27).

The second reason is that the theory, which is presented in the study, has come into being in a way that is only partially described in the study. Quite a bit of material that is needed to get a full picture of the theory is therefore not available. Both of these reasons mean that the theory, as understood according to the conventional view is underspecified and hard to get a good grip on.

Instead of looking at a theory exclusively in this way, that is, to view it as embodied in a piece of writing, and to see this as all there is to a theory, there also exists another way. This is to view theory not as a text, analogous to a literary creation in several ways, but as a set of activities (e.g. Cartwright 2020, Swedberg 2021, 2022). This makes it easier to use the theory, to understand how it came into being, and to see what its full implications are. In brief, this is a fuller as well as more practical approach to theory, than to treat it as a text.

By the term activity, it can be added, is roughly meant an action that is mentally directed at some specific task and includes the element of meaning, as in Weber’s interpretive sociology. Activities can be primarily mental or “internal” (say the reading a text) or primarily practical or “external” (say the carrying out of an interview; Weber 2012:273–74). The term “activity” is similar to that of “practice”, but differs in its use here in that it is more individual and less collective in nature. Since practice is currently used in sociology for other purposes than those in this paper, it has also been seen as important to use a different term (for use of the term practice, see e.g. Pickering 1992; Schatzki, Knorr Cetina and Savigny 2014; for a discussion of these issues, see Swedberg 2022:14–15).

The first consequence of making the shift from text to activities is that the basic unit of focus now becomes the research process and not the theory, extracted and presented separate from the method and the facts of the study. Theory, method and facts are the three central parts of a sociological analysis. They are also organically united in the research process, in the sense that they are all part of what the researcher does. Facts are partly theory and method; methods are partly theory and facts; and theory is partly facts and methods. Exactly how the three are linked and interpenetrate differs from case to case and therefore also needs to be investigated in each case.

A second consequence of shifting to the perspective of theory as activities is that we now become interested not only in the content of a theory but also in how the theory is used. The main reason for this is that we can learn from this, and then use it ourselves in a more efficient manner. We look, in brief, at theory as a practical activity.

Third, we also want to determine what is old and what is new in the theory. The key question here is the following: Which parts of a theory are just taken over from the existing store of theory, and which have required the researcher to engage in a new way of acting and thinking? The latter constitutes the main contribution and therefore needs to be extra carefully examined. Again, we want to be able to see exactly how the contribution was made and learn from this. By referring to what is new and what is old in a theory, it also becomes possible to see how well the researcher handles existing theory. We especially want to know if existing theory was just taken over and repeated without much thought, or if it was used in a creative way first after having been carefully inspected. The chance to make a real contribution increases the more the pre-existing theory has been studied, understood and absorbed.

Let us now return to Blau’s book and try to determine what the perspective of theory as a set of interconnected activities can make us see, that the conventional perspective of a theory as a text cannot. This will be done in two steps. First we will carry out this exercise by only using Blau’s book for our material, but this time look at it from the perspective of theory as a set of activities. As will be shown in the next few pages, this will give us a different picture from that of theory as a text. The result of this exercise will not, however, give us the full picture that we are after. For this more material is needed than just Blau’s book. What we can say about Blau’s theory when we have access to such additional material will be presented in the following section.

There exist several reasons to look at The Dynamics of Bureaucracy from the activities perspective even if we only have access to the book itself. One is that we very often only have access to the work and not to any additional material. Another is that many times we simply do not have the time to work with anything more than the work itself. The same goes for articles.

In approaching Blau’s book as a set of activities, the first thing we want to do is to establish what Blau’s research project is, and this also means to see how theory, method and facts are related to one another; how they overlap and how they link up. Blau’s research project, to recall, was to see what would happen to Weber’s view of bureaucracy if it was confronted with systematic data. Note that Blau was well aware of the fact that he might not come up with anything interesting in his research; he knew that all serious research is a gamble. To cite the quote which he placed at the beginning of his book: “None of the precautions of scientific method can prevent human life from being an adventure, and no scientific investigation knows whether he will reach his goal” (Morris Cohen as cited in Blau 1955:v).

Blau seems also to have been well aware of the links that exist between theory, method and facts in his study, even if he did not always spell these out in detail. When he, for example, mentions how he used Merton’s functionalism, he adds that it had helped him to sort out and organize the data. But that is also all; no details are given of how this was done. To cite: “Conceptions of functional analysis are used to organize the data of this study, taking as a starting point the conceptual scheme developed by Robert K. Merton” (Blau 1955:6).

Blau handled Weber in a different way. Here the link between theory and facts was clear to him from the beginning. Weber’s theory, as Blau repeatedly points out, was not properly grounded in systematic data. In confronting Weber’s theory of bureaucracy with this type of data, Blau decided to focus on work groups in a bureaucracy with the help of a case study. In doing so he used some quantitative methods of observation that had been developed in industrial sociology, in studies of workers in their work places. Blau also decided to follow Merton’s advice to focus on patterns of social behaviour rather than on isolated acts by the administration and what effect these might have. This would also allow him to look at the ways in which these patterns of interaction affected the structure of the bureaucracy, according to functionalist theory (manifest function, latent function, dysfunction). Blau’s full research project was in other words to confront Weber’s theory of bureaucracy with systematic data, and to use Merton’s theory of functionalism to sort out and ultimately make sense of the data.

To look at Blau’s research project also shows how closely his theoretical ambition (to study a bureaucracy empirically, and to adjust Weberian theory accordingly) was linked to the kind of method he had selected (a case study of interactions in work groups) as well as the social facts that he wanted to trace (patterns of behavior). We are also reminded of the importance of keeping in mind the practical side of Blau’s project, since to carry out his project he would have to learn how to apply a method developed in one setting (a factory) to that of another setting (a bureaucracy), something that was likely to lead to a number of practical problems for him.

For his general theory Blau proceeded, as we now know, by adopting Weber’s approach to bureaucracy in combination with Merton’s version of functionalism. What he found out, however, was that Merton’s theory was unable to handle some phenomena that he observed in the two agencies he studied. More precisely, it could not explain why the dysfunctions of some measures by the administration sometimes did not lead to a disturbance of the organization but instead to its change. According to Merton, dysfunctions upset by definition the organization; they do not lead to changes and they leave the organization intact. What Blau termed emergent organizational needs represented his solution to this problem; and it implied a revision of Merton’s theory of functionalism. It also meant that Weber’s theory of bureaucracy was partly wrong. Bureaucracies are not static, as Weber had argued; they change and are dynamic in nature. One mechanism through which this change can take place is the following: when actions of the administration misfire, they are sometimes followed by new actions by the administration which introduce changes in the organization that allow it to continue to operate well.

While the idea that bureaucracies are dynamic constitutes Blau’s main theoretical contribution, it is also clear that he was only able to make this contribution because he had accepted Weber’s general theory of bureaucracy. The same can be said for Blau’s relationship to Merton’s theory of functionalism. To rely in this manner on theories by others is something that every sociologist does, so there is nothing remarkable about this. What deserves to be emphasized, however, is that Blau had made very good choices when he selected which theories to use in his research; Weber is a very strong theoretician and so is Merton. Blau’s use of their ideas also shows that he had not just automatically taken them over; he had thought deeply about them and absorbed them into his thinking. This was also a precondition for figuring out which parts of the theories of Merton and Weber needed to be changed and also could be changed. Still, it should be pointed out that for something like 95 % of his theory, Blau relied on the collective heritage of sociology.

3 Theory as Activities, Part II (The Full Version)

What has just been said is roughly as far as it is possible to get in understanding Blau’s theory of bureaucracy by reading his book from the perspective of theory as activities. The result is a fuller and more practical version of Blau’s theory than was the case when his study was read from the perspective of theory as a text.

It is, however, possible to get an even fuller and also considerably more practical picture of Blau’s theory. This can be done by looking at it from the perspective of theory as activities in combination with more material than just Blau’s study. This extra material can be of different types. We know, for example, that a published text in sociology, as any scientific text, contains a highly stylized and incomplete version of what actually took place during the research process. According to Merton, the following five items are typically omitted from scientific articles but are part of the actual research: “intuitive leaps, false starts, mistakes, lose ends, and happy accidents” (Merton 1968:4; emphasis added). It can also be added that there is something deceptive about the published version of an article, since instead of describing what actually happened during the research the writer has to follow certain norms for what to include and what to leave out.

The five factors that according to Merton are typically left out of scientific articles apply to theory as well as to method and facts. You can e.g. make mistakes when you handle a theory; and there are often a number of false beginnings when you work with the theory part. You can also stumble over some theoretical insight; and you may not push all of your ideas on theoretical issues to their logical conclusion. And sometimes you succeed in making a good theoretical move without knowing exactly how you did it. But, to repeat, nothing of all this is mentioned in published articles and monographs.

Merton’s list may seem complete but there is in fact quite a bit more that happens in the research process which does not make it into print. For one thing, most of the theorizing process is left out, such as how you settle on, define and handle items such as the research object, the categories and the concepts; why some alternative theories were not used; the type of explanation that was chosen, and why other types were rejected, and more.

Left out are also theoretical assumptions, tacit theoretical knowledge, much of the context, and how the personality and personal history of the researcher (including her training in theory) affects the way she deals with theory. Every sociologist is also part of the collective unfolding of the discipline and the tradition of sociology. One can go on, but the point should be clear: if you really want to know how and why the theoretical part of a study was handled the way it actually was, you will need to look at a much larger number of factors than those that are discussed in published articles and books.

The type of material is many times not available, but in the case of Blau we happen to be lucky; and this means that we are in a position to go well beyond the view of Blau’s theory based exclusively on a study of his book itself. The material that allows us to do this comes from several different sources. For one thing, Blau’s theory grew out of his dissertation at Columbia University; and Merton, who was his thesis adviser, has written an article about his work with Blau on the dissertation (Merton 1990). The two also corresponded during the time that Blau worked on his thesis; and many of the letters from those years (and later) have survived.^[4] A couple of years after Blau’s book was published, he was asked to write an article on how his study had come into being; and this resulted in a very helpful document. The person who asked Blau to do this was another Columbia sociologist, Phillip Hammond, who in the early 1960s put together a book on the ways in which well-known sociological studies had come into being. This was a pioneering enterprise; and the title of Hammond’s book nicely captures its content: Sociologists at Work (Hammond 1964).

The authors were asked to portray their own research activities as it was experienced during some specific investigation. Their instructions were to particularize, not generalize … Finally, and in defense against any who may hold that even a variant of scientific reporting must necessarily be impersonal and impassionate, the authors were encouraged to write in the first person. - Phillip Hammond, Sociologists at Work (1964), p. 4.

There also exist some personal factors and data that need to be taken into account, in order to get a better handle on how Blau worked with theory in his study. One would especially want to know something about his life before he started to work on his dissertation. This, unfortunately, is a topic that we do not know very much about. The reason for this is that Blau has left behind very little information about himself.^[5]

All that is known about Blau before he started to work on his dissertation is roughly the following (see especially Blau 1995; R. Blau 2002, 2016; Scott and Calhoun 2004). Peter Michael Blau was born in on February 7, 1918 into a middle-class Jewish family in Vienna which made its living through the bookbinding business. He studied to become a medical doctor at the University of Vienna; he was also involved in socialist politics and showed a great interest for the ideas of Marx and Freud. After Hitler’s annexation of Austria in 1938, Blau tried to flee to Czechoslovakia, was caught and tortured by the Nazis. After many difficulties he arrived in the United States in 1939. For a long time he did not know what had happened to his parents who had remained in Vienna. Eventually he found out that they had been murdered in 1942 at Belzec, an extermination camp in Poland.

How deep traumas affect the work of social scientists is rarely discussed, neither by those who have experienced them nor by those who comment on their works. It is also hard to understand the relationship between these trauma and, as in this case, the specific work of a social scientist. Still, it is clear that deep traumas of the type that Blau experienced do play an important role in a person’s life and therefore need to be considered.^[6]

Soon after arriving in the United States, Blau got a scholarship to attend Elmhurst College in Illinois where he graduated with a major in sociology in 1942. He would later say that he had chosen to study sociology because of its progressive character: “I was originally attracted to the field [of sociology] by a concern with inequality of power and opportunity” (Blau 1952:ii). He also said that “my main sociological interest was in theory” and that his undergraduate thesis was on the relationship between the theories of Mead and Freud (Blau 1995:2).

After graduating, Blau joined the U.S. army for three years, during which he did combat in Europe and also worked as an interrogation officer, thanks to his fluency in German. In February 1946 he began graduate studies in sociology at Columbia University, where he had applied primarily because of the “progressive ideology” of Robert Lynd (Blau 1995:3). That there also was a professor on the sociology faculty called Robert K. Merton he did not know.^[7]

At this point we need to leave Blau the individual for a moment and instead turn to sociology as a collective and historical enterprise, in order to get a better idea of the ways in which Blau was later to make his theoretical moves in The Dynamics of Bureaucracy. One key fact of this type is the following. In the mid-1940s two young assistant professors in sociology had teamed up at Columbia University in an effort to develop sociology further. This was Paul Lazarsfeld and Robert K. Merton. “They joined forces and established a coalition that made the first, and quite possibly most successful, research-cum-theory team in sociology”, Blau later wrote (1995:4).

Merton was a first-class theoretician and Lazarsfeld a first-class methodologist, and their joint project was to transform sociology by setting it on a firm theoretical and empirical foundation. Data, they argued, must not be anecdotal, as was common at the time; it should be systematic and so should theory. The key term here is systematic, as in systematic data and systematic theory. With theory Merton and Lazarsfeld mainly had the classical European sociologists in mind, such as Weber, Durkheim and Simmel.

Merton and Lazarsfeld were also helped in their efforts to realize their project by the presence at Columbia of several cohorts of brilliant students who were attracted to their ideas and wanted to work with them. Blau would later write, “The two years I was in residence, Selznick, Lipset, and Gouldner were completing their dissertations; Rossi, Inkeles, Rose Coser, and Wrong were in my cohort; and the next included Lewis Coser, Coleman, Trow, and Katz” (Blau 1995:4). All of these were to become prominent sociologists.

The sociology department of Columbia University at midcentury was an exciting place to be. The intellectual academic atmosphere that permeated it engendered the feeling in many of us that we were on the verge of new developments and advances in sociology. This atmosphere had its source in a challenging faculty and an exceptionally large student body with disproportionate numbers of very bright students. Many of them had been held back by the war and were now eager to make up for lost time by learning quickly and starting to make contributions of their own. The outstanding established sociology faculty was epitomized by Lynd, whose two books, written in collaboration with his wife, on Middletown (Lynd and Lynd 1929, 1937) are still classics. The faculty had recently been augmented by two young assistant professors, Paul Lazarsfeld and Robert Merton, and soon thereafter by a third, C. Wright Mills.

– Peter M. Blau, “A Circuitous Path to Macrosocial Theory”, 1995, p. 4.

Blau took several methods courses for Lazarsfeld but felt closer to Merton and defined himself, now as later in life, primarily as a theoretician.^[8] As already mentioned, he had developed an interest in social theory already as a youngster in Vienna and it continued at Elmhurst College. When he came to Columbia Blau made sure to take all the theory courses he could, especially in classical theory:

I took virtually all theory courses from Abel, MacIver, and Merton, and my reading concentrated on the classical theories. I read nearly all the major works (translated or in German) by Weber, Durkheim, Pareto, and Simmel as well as American theorists. (Blau 1995:3)

Under the influence of Merton, Blau became convinced that general and speculative theories were not useful in sociology and should be replaced by middle-range theories, based on systematic empirical research. Blau was also lucky enough to take Merton’s course in theorizing during his first term at Columbia. In this course, which was unique at the time, he was taught how to handle theory in actual research, including several helpful techniques. He learned, for example, how to add to the way in which a phenomenon is conceptualized in sociology and thereby improve it (respecification), and to do the same with concepts (reconceptualization).^[9]

Blau also worked closely with Merton during 1948 when the latter was writing his programmatic essay on functionalism, adding such terms as dysfunction, latent function and manifest function (Merton 1949:21–81; Merton 1968:136; Merton 1990:52; see Sohlberg 2021a; Sohlberg 2021b). The main focus of this essay, Merton emphasized, was to show how functionalism could be used in research; “how one goes about the business of functional analysis”, as he put it (Merton 1949:21; Merton 1968:73; emphasis added). Blau writes, “I was Merton’s research assistant [in 1948] while he was framing his essay on functionalism, and the opportunity I had of working with him while he developed this new theoretical framework undoubtedly contributed to the great impression it made on me” (Blau 1964:19; Merton 1949:21–81).

It was probably also by being around Merton that Blau came up with the idea to make a sociological study of bureaucracy for his dissertation, and to use functionalism as part of the analysis. At the time when Blau was at Columbia, the growth of bureaucracy was regarded as one of the major social trends of the time. A few years earlier Merton had also published an important essay on bureaucracy, in which he launched the idea of “the displacement of goals” (Merton 1940). He also used his ideas on functionalism to analyse bureaucracies and spoke, among other things, of the role of “dysfunctions” in these (e.g. Merton 1940; Merton et al. 1952:11).

Merton felt strongly around this time that sociological studies of bureaucracy were needed; and he noted with satisfaction toward the end of the 1940s that “a series of empirical studies of sociological problems in bureaucracy” were under way at Columbia or had already been published. As examples of this he mentioned the dissertations of Alvin Gouldner and Philip Selznick (Gouldner 1954; Merton 1949:118–19; Merton 1982; Merton 1990:n.33; Selznick 1949). When Merton noticed that Blau wanted to make a study of bureaucracy he decided that also Blau’s dissertation should be “one of the empirical building blocks for the sociological analysis of bureaucracy” (Merton 1982:919). Merton, in brief, was deeply invested in what was eventually to become Blau’s book; and as we shall see, he also kept a very close watch over everything Blau did. That this also meant that Merton pressured Blau and had some power over him is likely.

During the spring of 1948, which was his last semester in residence at Columbia, Blau started to write down various problems that was related to his prospective thesis. His general idea in generating these problems was to start out from the theories of Weber on bureaucracy and Merton on functionalism. “Guided by Weber’s theory and Merton’s functional conception, I conceived of my investigation from its inception as a case study of bureaucratic work groups” (Blau 1964:56).

At this early stage Blau was especially interested in the distinction between formal and informal aspects of a bureaucracy, and also how work groups in a bureaucracy are affected by orders from the administration. For a while he thought of using a survey to get the data he needed, since this was the favoured method at Columbia at the time. He feared, however, that he would not be able to get the kind of detailed information he needed through a survey. And this meant that he had to find some other method.

He eventually found one in the following way. As part of his course work, he had taken a class at Columbia in industrial sociology for anthropologist Conrad Arensberg. What the students were taught in this course was how to make quantitative studies of the interactions among workers in factories. Blau found this to be a congenial approach for his own idea to study work groups in a bureaucracy. Nonetheless, it was also clear to him that to do so, he would have to improvise since white collar work and factory work are quite different.

At this stage of Blau’s dissertation Merton suggested that he should study two bureaucracies, one private and another public, and then compare the two. Blau agreed; but as things turned out he could not get access to a private organization and therefore decided to study two public organizations. The reason for this, he says, was that public organizations were challenging enough to understand. Merton handled the negotiations with the heads of the two agencies and made it possible for Blau to get access to them and do his field work there.

At the beginning of his research Blau had great difficulty in figuring out how to do the kind of systematic observation that he felt was necessary to get the data he needed. Field work was not taught at this time at the sociology department at Columbia, nor at Lazarsfeld’s Bureau of Applied Social Research (Merton 1990:47–50). Apart from what he had learned in the class on industrial sociology for Arensberg, Blau had in other words to improvise and figure things out for himself. And as one would guess, this also meant that he committed several errors (e.g. Blau 1964:28–9, 31–3). Merton would later call him “the autodidact field worker” (Merton 1990:47).

The close link between Blau’s method and his theory comes out very clearly in his account of how his study came into being in Sociologists at Work. One example of this is the following. In looking at the way in which the orders of the administration affected the groups of employees, Blau decided not to trace the impact that these had on the organization. More precisely, he did not try to figure out if the orders per se helped to maintain or to disrupt the organization, that is, if their function was positive or if some dysfunction was involved.

Instead Blau decided after some hesitation (and probably inspired by Merton’s teaching), to do something else. This was to first establish the patterns of social behaviour that the orders resulted in (Merton 1949:60–1). Only when this had been done, did Blau look at the effect that these had on the organization; that is, he looked at the ways in which the patterns (and not just the initial orders) affected it.

As mentioned in the book, Blau ended up by arguing that some dysfunctions of the orders of the administration were followed by new orders or other events that offset these. And as we know, Blau presented this as the main evidence for the argument that bureaucracies are dynamic in nature and not static, as Weber had argued. In the account that Blau gives of his research in Sociologists at Work, however, we get a fuller picture of how Blau came to this conclusion. For a long time, he here says, he did not know what to make of the fact that dysfunctions could sometimes be offset by new orders or other events. This clearly contradicted Merton’s theory of functionalism, according to which dysfunctions disrupt and harm the organization. After quite a bit of time however, Blau says, he finally came up with the idea of “emergent organizational needs”; and when he did this things fell into place (Blau 1955:251–55). “The explicit application [of this idea] enabled me to see connections between various segments of the analysis that I had not previously noted” (Blau 1964:54).

What we have here, in other words, is an example of one of those intuitive leaps that according to Merton do not make it into scientific articles but nonetheless constitute an integral part of the research. Since creative insights are central to successful theorizing, Blau’s reflections on his various leaps in Sociologists at Work are of special interest. We know, for example, that during the research for his dissertation Blau created an enormous amount of material. He often worked 16 h a day and nearly filled three filing cabinets with memos and field notes, many of which never made it into his book. Why did they not? Blau says that he does not know why he selected to work on certain topics and not on others. In a way, this statement is not so surprising since few people are fully aware of why they make certain choices during research.

Even fewer, however, make an attempt to think back on what actually happened during the research process, in an effort to see what made them act in a certain way. Blau, however, did precisely this, and his most important reflection on how he came up with his ideas is the following:

For most of these [major conceptual] problems … we can find an early trace, be it ever so faint [in my initial thoughts from 1948 on bureaucracy], but the prefield-work conception typically underwent fundamental changes as the result of the research experience [later in 1948 and in 1949]. The field situation is ripe with serendipity: new insights are gained, and they can often be corroborated with empirical evidence; yet these insights can typically be traced back to earlier theoretical conceptions and thus serve to refine them. (Blau 1964:37)

By 1951 Blau had finished a first draft of his dissertation and immediately sent it off to Merton for comments. He got his manuscript back with an enormous amount of corrections and suggestions for changes. According to a friend of Merton who has seen the manuscript, “It has more red ink than original type, with side notes and comments [by Merton] everywhere” (Cole 2021). After making what Blau describes as “intensive revisions”, based on Merton’s “detailed and helpful criticisms”, he presented his dissertation at Columbia in 1952 under the title The Dynamics of Bureaucratic Structure (Merton 1990:59, Blau 1952). He then reworked his manuscript for publication; and the book version appeared in 1955, under a slightly different title: The Dynamics of Bureaucracy.

The text in the dissertation is very similar to that of the book. It does, however, also contain some new material that is of interest for the account in this paper. One is that a new person who deeply influenced Blau in his work, now enters the stage for the first time. This is his wife, Zena Smith Blau. Wives and husbands are typically excluded in accounts of why sociologists do the kind of work they do; and one reason for this is that items like emotional support and family are presumably not very important. In this particular case, however, Zeno Smith, who married Blau in 1948, was also a graduate student in sociology at Columbia with Merton as her thesis adviser.^[10] How she influenced Blau’s study we do not know, beyond the brief statement (in the preface to the dissertation) that he owes her “much” and that she “criticized my ideas and manuscripts”.^[11] A preface, it can be added, is one of the few places in a scientific work where personal information of this type is allowed to be mentioned.

It is also in the preface to the dissertation that we can find the fullest account from Blau’s side on what Merton had meant for his work. Blau writes at one point that “Professor Merton has advised me at every stage of its’ development [that is, the thesis]” (Blau 1952:iii–iv). At another point he thanks “Professor Merton, whose lectures, writings, and discussions have influenced my sociological understanding more than any other social scientist.”

In these two short statements, which all in all consists of some 25 words, Blau has compressed what must have been hours and hours of discussion between himself and Merton plus all the work that he carried out inspired by Merton’s comments on his outlines and manuscript.

So why did Merton watch so carefully over Blau’s thesis? We know already that Merton around this time wanted to promote sociological studies of bureaucracy. His favorite way of teaching was also one-to-one with specially selected students; he also liked very much to edit the work of others (Caplowitz 1977; Merton 1998:196–97). Blau was furthermore one of Merton’s favorite students. Merton wrote many years later to Blau, “I knew from the start yours would be a scholarly life of the very first class” (Merton 1988).

There also exists another difference between the text in Blau’s thesis and his book; and it has to do with its political stance. Blau emphasized the political dimension of bureaucracy much more sharply in his thesis than in his book. In the former he starts out by mentioning some of the dangers that the growth of bureaucracy poses for democracy; he also ends the dissertation in a similar fashion. Democracy, he says, emerged historically before bureaucracy; and the tendency of the latter to concentrate power in a few hands represents a serious danger to democracy.

There exists, on the other hand, very little mention of the political dimension of bureaucracy in the main text of the dissertation or in the analysis itself, which makes you think that Blau had either failed to include his own ideas about the political role of bureaucracy or he had decided not to do so. Both may be true. The idea of power and related concepts are not part of Blau’s theoretical argument; and he tried very hard to present a logical argument in which all parts of the theory were closely related to one another. More generally, and perhaps also of importance, was the fact that Blau did not like to deal with inflammatory topics since they tended to draw attention away from his main concern as a sociologist, which was always theoretical. He writes as follows in Sociologists at Work:

I am quite wary of – biased against, if you will – the study of intrinsically interesting subject matters where a journalistic fascination with the engrossing substantive issues easily diverts the analyst from focusing on problems of theoretical significance. I chose an unglamorous bureaucracy for study, not the United Nations or the Pentagon, and my analysis ignores the sensational facets of problems in favor of those aspects I consider of theoretical significance.

(Blau 1964:53)

When Blau was turning his thesis into a book, he removed the statement about the danger of bureaucracy at the beginning of his thesis. He, however, left what he had said on this topic at the end. By proceeding in this way, however, the latter statement lost much of its force and seemed more of an afterthought than something that was central to the thought of the author.^[12]

Once a book or an article has been published, its formulations will remain the same and cannot be changed; and that also includes its theory part. Or to be more exact, this is nearly true, for while articles cannot be revised, it is in principle possible to publish a new and revised edition of a book. And in a second edition, the author often wants to add some new material; and this means, from the perspective of this paper, that she first has to engage in some new activities.

There is also the fact that just as a study is always the result of the activities of other people than the author herself (just think of the list of references!), its content can also change after the publication because of the activities of others. The text itself will remain the same, but the theory and its results are likely to be evaluated differently after a while since new research is usually produced. What was once seen as a theoretical idea to emulate, may some time later be regarded as superseded by newer and brighter ideas. At a higher general level, there is also the fact that the whole field of sociology may slowly change direction and draw attention to certain subjects, concepts, theories and so on, while others are seen as less relevant. This may seem as an abstract argument but as we soon shall see, this is exactly what happened to The Dynamics of Bureaucracy; and the reason for this is that the field of sociological studies of bureaucracy was soon to be replaced by the emerging field of organizational studies.

In the case of Blau’s book, a second edition appeared in 1963, eight years after the initial publication. Blau notes in the new preface:

The analysis of research findings never reaches a final state of completion. It is always possible to derive new inferences from the data, to carry the interpretation a step further, to make additional comparisons, to use information from other sources to refine the analysis, and to view the research from a broader perspective.

(Blau 1963:vii)

The second edition contains the original text of The Dynamics of Bureaucracy plus a number of new additions, all together about 75 pages of new material. Some of these represent additions that have their origin in observations, thoughts and analyses that Blau had made after the publication of his book in 1955. During the next few years a replication had, for example, been made of his study; also other relevant literature had been published. Blau commented carefully on all of this new material in his additions to the second edition. He also included much (but not all) of what he had written for Sociologists at Work in the form of two new chapters, one on methods and one on theory (Blau 1963:269–86, 287–305).

The most important novelty in the second edition, when it comes to theory, was however Blau’s statement that when he worked on the first edition he had not fully realized the full potential of one particular concept, namely that of social exchange. He now realized more fully that in many situations a person does something for someone else and typically wants something back, in exchange. The most important example of this in Blau’s book from 1955 was the consultation pattern in the federal agency; if you ask for help from someone with a case, you will have to pay for this in some way.

Blau also stated in the new edition of his book that recent developments in how to carry out surveys had now made it possible to produce analyses that went well beyond in-depth case studies of bureaucracies. With their help you could now study organizations of all types, even social organization itself. To cite: “The use of such methods by the theoretically sophisticated investigator promises to spur the development of a general theory of formal organizations and, ultimately, an even more inclusive theory of social organization” (Blau 1963:305).

As we know, there is always the collective level that needs to be taken into account when you study theory as a set of activities; and in this particular case that means sociology as an evolving discipline. By the time that Blau’s second edition appeared in 1963, something called organization studies had begun to replace studies of bureaucracy in sociology. Blau himself would from now on also try to contribute to the development of organization studies as a separate academic field, rather than to studies of bureaucracy as part of general sociology which is what Merton had had in mind (e.g. Blau and Scott 1962).

The notion of social exchange that can be found in the two editions of The Dynamics of Bureaucracy would be developed even further, and also take a different direction during Blau’s long and distinguished academic career (University of Chicago, 1953–1970; Columbia University, 1970–1988; UNC at Chapel Hill, 1988–2001). Inspired by some very positive comments on The Dynamics of Bureaucracy by George Homans during a talk at the University of Chicago in the late 1950s, Blau would a few years later formulate his own version of so-called exchange theory (Blau 1964; Blau 1986:viii, 1989). According to a statement by Blau, “The basic ideas underlying the theory of social exchange originally occurred to me … while I was engaged in a case study of government officials for my Ph. D. dissertation” (Blau 1986:vii). Exchange theory, it can be added, was also part of the plethora of new theories in U.S. sociology that appeared after the downfall of functionalism in the 1960s.

4 Discussion

In science, each of us knows that what he has accomplished will be antiquated in ten, twenty, fifty years. This is the fate to which science is subjected … Scientific works certainly can last as ‘gratifications’ because of their artistic quality, or they may remain important as a means of training (Mittel der Schulung). Yet they will be surpassed scientifically – let that be repeated – for it is our common fate and, more, our common goal. We cannot work without hoping that others will advance further than we have.

– Weber, “Science as a Vocation”^[13]

It is true that the main contribution to sociological theory in Blau’s book – that bureaucracies are dynamic – has by now been absorbed into the sociological tradition. But as Weber reminds us of in the quote above, some outdated works can be useful for didactic purposes or, to use his own phrase, as “a means of training” (Mittel der Schulung).

This idea also goes well with the main exercise in this paper, which will now be summarized. Two different approaches to theory have been presented. According to the first, which can be described as conventional, a theory is seen as embodied in a text: theory as a text. You work your way through the item in question, say Suicide by Durkheim or an article in Administrative Science Quarterly. You study it carefully and try to extract the theory and state it in a concise manner. The result is a short theoretical statement, say that the nature of bureaucracy is dynamic. This formulation is what you try to remember for possible use in your own research.

The perspective of theory as activities is more in the spirit of theorizing. Here you both study the work in a different way and try to develop a fuller version of the theory by drawing on additional material. When you study a work from this perspective, the focus is not on the theory as a separate and distinct i.e. but on the research process as a set of activities, with theory being part of this. In the research process, the three parts of a sociological analysis – theory, method and facts – are linked together and overlap; they form an organic whole. You cannot study or speak of the theory without also looking at the methods and the facts. While the book or article is central also for this type of approach, there is more material that needs to be considered since a written theory is always embedded in a large number of activities.

In the approach of theory as a set of activities, in brief, you do not dismiss the work but study it in a different way than in the conventional approach, more precisely as a set of interconnected activities. A written theory is here seen as the result of research activities in a broad sense, including the act of writing. It is also important to try to locate the theoretical novelty that is involved in some research and how it came about. To do so you need to distinguish between the theory that the author started out with, and what she has added on her own. It is the latter that constitutes the theoretical contribution, which needs to be evaluated and closely inspected so one can learn from its content as well as the way it came about.

To get a handle on all of the activities that together constitute the full theory, it is imperative to also go beyond the published work since it excludes much relevant material. There are, as we know, the process of theorizing including the five items that Merton says are excluded from scientific articles (“intuitive leaps, false starts, mistakes, lose ends, and happy accidents” – Merton 1968:4). There is also the immediate microenvironment of which the author was part when she did the research (colleagues, students, teachers and so on). To this should be added what went on in society as well as sociology at the time and in the past. And finally, there is the personality of the researcher. All of this is part of the activities that went into a theory.

But there is even more; also what happens after a work is published will affect its content or rather how it is understood. Some parts of a theory may turn out to have been wrong, others will perhaps be improved through research conducted after the text was published. Research in other social sciences than sociology can also affect the standing of a theory. Science never stands still.

How did the example of Blau’s book fare when it was analyzed according to the two different perspectives on theory that have been presented here? In the first approach, the main effect was to focus on Blau’s theory as this was portrayed especially towards the end of his book, that is, on his main thesis that bureaucracies are dynamic. This theory reads: bureaucracies are not static, as Weber had argued; they are dynamic in nature. They change all the time.

This, however, represents a minimalistic and underspecified version of Blau’s theory. The perspective of theory as activities presents a considerably richer and more precise picture of Blau’s theory. It is also more practical in nature; and the reasons for this are the following. First, this approach focuses on the research process as the basic unit, and not just on theory as something that is self-contained and exists separate from method and facts. This means that the close links between Blau’s theory, on the one hand, and the method and facts he worked with, on the other, now are specified as well as inspected.

Blau needed a special method to study what he wanted; and the facts he discovered by making systematic observations in small groups was central to the theory he ended up with. The emphasis on drawing theory and facts closely together, which was very strong at Columbia during Blau’s time, clearly played a role here. Blau was also much helped in how he dealt with theory by being exceptionally well grounded in classical sociological theory.

Second, being a student at a university which had a first-rate theoretician like Merton had a profound impact on someone like Blau with his deep interest in theory and desire to grow as a theorist. Blau, to recall, was singled out by Merton for his talent and worked for a while as his research assistant. Merton also supervised Blau’s dissertation very, very closely, partly because he was himself interested in developing a sociological approach to bureaucracy but also because he saw Blau’s great potential.

Blau was finally affected in his sociological thinking by what was happening at a macro level, in society (Austria, WWII, the United States) as well as in social science, including sociology (the push towards quantification, logical empiricism as the main philosophy of science). Sometimes the effect of these macro level forces on Blau’s life and thinking was very direct, sometimes less so or mediated in very complex and hard to understand ways, such as his deeply traumatic war experiences.

Third, and as we know, what happens after a theory has been published also affects its content. “You cannot step into the same river twice”, as Heraclitus says. In Blau’s case, this can be exemplified by the fact that his study was soon viewed less as a contribution to the sociological study of bureaucracy, and more to the emerging field of organizational sociology and organizational studies. As we know, Blau also produced a second edition of The Dynamics of Bureaucracy, which included quite a bit of information on how his study had come into being, from his article in Sociologists at Work. Many years later Blau would also develop some of his insights from this study into a full-scale theory of social exchange.

It would seem, to sum up, that the example of Blau’s study suggests that the perspective of theory as activities is able to provide a considerably richer and more helpful view of theory than the conventional approach, according to which theory is treated a bit like a literary text. But besides doing this, does the specific example of Blau’s study perhaps also add some additional insights with practical implications for how to use theory? This is the question we now shall try to address; it is also the one with which the paper will end.

One practical insight that can be derived from Blau’s study is the following; and it is of special interest for organizational sociology and organization theory. Blau was at first interested in seeing what effects the orders of the administration would have on the work that was carried out in the two organizations he was studying. After a while, however, he realized that he should not study the immediate impact of these orders, but instead try to see how they resulted in new social patterns of behavior among the employees. In brief, instead of trying to establish the effect of official orders per se on work in an organization, sociologists may want to study the social patterns of behavior that may come into being as an effect of these.

Second, Blau comes close to suggesting an interesting addition to Merton’s theory of serendipity. In discussing his research experience in Sociologists at Work, he writes the following:

For most of these [major conceptual] problems … we can find an early trace, be it ever so faint [in my initial thoughts from 1948 on bureaucracy], but the prefield-work conception typically underwent fundamental changes as the result of the research experience [later in 1948 and in 1949]. The field situation is ripe with serendipity: new insights are gained, and they can often be corroborated with empirical evidence; yet these insights can typically be traced back to earlier theoretical conceptions and thus serve to refine them. (Blau 1964:37; emphasis added)

Blau is here commenting on the role of serendipity in empirical research. While Merton, who had come up with the idea of serendipity, suggests that you sometimes stumble over discoveries, Blau is saying something slightly different. And it is this slight difference that adds to Merton’s theory. It does this by transforming serendipity from being something that simply happens to you because of luck, into something that you can to some extent invite, even if not steer.

What Blau says is that when you stumble on some interesting phenomenon in your research, you may eventually be able to transform it into a theoretical insight; and this is done by integrating it into the main body of the theory you work with. What Blau does not say, and which needs to be clearly spelled out, is that Blau worked with very powerful theories (Weber, Merton). One can therefore restate and add to what Blau says on this point as follows: you can invite serendipity by working with strong theories; and these can also in this way be refined and added to.

Third and last, what may well represent Blau’s most interesting contribution to practical theory is something he was himself only vaguely aware of. The reader may recall that when Blau wrote the article on how he had done his research for The Dynamics of Bureaucracy, he had come to realize that when he did his research he had used the theories of Weber and Merton as guides. This is what he wrote in Sociologists at Work: “Guided by Weber’s theory and Merton’s functional conceptions, I conceived of my investigation from the beginning as a case study of bureaucratic work groups” (Blau 1964:56).

In fact, in his article for Sociologists at Work Blau repeatedly uses the term “guide” for the way that he used the theories of Weber and Merton during his research (e.g. Blau 1964:21, 52, 56). So the question that needs to be asked is this: Why did Blau use this particular term, and what exactly did he mean by it? One answer would be the following. A guide is something that helps you to find your way through a terrain that is unknown to you. It tells you where to go, what to avoid, what obstacles to expect, and more. Without a guide you are on your own and may soon be lost. In practical terms, and applied to theory, Blau implies that a good theory is one that helps you to move forward in your research. It indicates to you where to start and roughly how to make it to the end. It does this by showing how to locate the object of study, how to approach it, characterize it, and even explain it. It shows you how to make it through the thicket of facts and ideas.

The idea that you may want to look at existing theory not only as the result of practical activities, but also as a guide for what activities to engage in, represents an interesting and helpful view of the role that theory plays in research. It also integrates the earlier two contributions by Blau that were just mentioned. The first is that you should not use just any theory as your guide; you should select one that is licensed and experienced (type Weber and Merton). This kind of guide helps you to get underway and in the right direction. The second has links to Blau’s theory of serendipity: a good guide helps you not only to find your own way, but in doing so sometimes also to make discoveries that add to the general knowledge and insights of the discipline.