Sociological Forum, Vol. 11, No. 2, 1996

The Beginning of the End or the End of the Beginning? The Theory Construction Movement Revisited Shanyang Zhao 1

Using book publication dates as markers, I divide the movement of (formal) theory construction into two periods. The first period started in 1954 and lasted for about a decade; the second period started in 1965 and ended in 1972. I argue in this paper that the publication of a dozen more books on how to construct scientific theories in the seven years after 1965 marked the beginning of the end of rather than the end of the beginning of, the theory construction movement in sociology. Although the formal mode of theory construction is dead now, efforts toward theory construction will never end. I conclude by delineating the major features of a new approach to theory construction that has been emerging in sociology. KEY WORDS: sociologicaltheory; theory construction; metatheory; methodology.

INTRODUCTION During the late 1960s and early 1970s, about a dozen books (Zetterberg, 1954/1963/1965; Wilier, 1967; Glaser and Strauss, 1967; Stinchcombe, 1968; Merton, 1949/1957/1968; Blalock, 1969; Dubin, 1969; Reynolds, 1971; Mullins, 1971; Wallace, 1971; Gibbs, 1972; Hage, 1972) on how to construct scientific sociological theories were published. But just as quickly as it came, the sudden surge of interest in modes of theory construction disappeared abruptly. With a few exceptions (Chafetz, 1978; Freese, 1980; Cohen, 1989), books on the same topic have not been seen since. In 1990, a national conference was held at the University of Maryland on the cause of the 1Institute for Social Research, Universityof Michigan, Ann Arbor, MI 48106 3O5 0884-8971/96/0600-0305509.50/09 1996 Plenum Publishing Corporation

306

Zhao

early "death" of what has come to be known as the theory construction movement. In a book that came out of that conference, Hage (1994:9) made the following comments: Seldom have there been so many contributions and so m u c h apparent consensus on the need for material and yet within only a few years such a total lack of interest at least as evidenced by the absence of any new articles or books on the topic . . . . Furthermore, many of the leading sociology departments added theory construction courses to their graduate programs and then abandoned them within a few years. It would appear that the topic of formal theory was in the same class as hula hoops: a fashion that almost explodes and just as quickly disappears.

Several factors have been identified as being at least partially responsible for the failure of the movement. Hage (1994) observed that the timing of the theory books could not have been worse because of the Vietnam War, which resulted in a high level of political conflict in sociology departments, and because of the emergence of antipositivism in the discipline. Turner (1994) held that a major reason for the movement's failure was the lack of strong organizational control in sociology over the process of theory production; besides, the goal of the movement was set too high to be reached, causing massive disappointment soon after. Gibbs (1994:65) argued that the failure of the movement was due to the fact that a formal mode of theory construction had never been fully adopted, as the field of sociology had always remained "little more than a debating club." The combination of all these factors presumably contributed to the quick end of the beginning of an otherwise promising movement. There is some element of truth in these explanations. In this paper, however, I want to advance a different argument. It is my view that the publication of the aforementioned books in the seven years after 1965 marked the beginning of the end of, rather than the end of the beginning of, the theory construction movement. Using book publication dates as markers, I divide the movement of theory construction into two periods. The first period started in 1954 and lasted for about a decade; the second period started in 1965 and ended in 1972. Although the first period was filled with excitement, enthusiasm, and high expectations, all the books mentioned above came out in the second period when the movement started to go down hill. These how-to-construct-theory books marked a major retreat from the grandiose goal set by the earlier advocates of the movement. In the following discussions, I will describe the initial objective of the movement and how it was modified and abandoned by the authors of the later books. I will close my argument by pointing out that efforts toward theory construction will never die, yet what seems to have died is, in fact, now the movement toward a formal mode of theory construction.

Beginning of the End

307

EARLY YEARS OF THE MOVEMENT

The theory construction movement can be traced back to the publication of the first edition of Merton's Social Theory and Social Structure (1949), in which the idea of theory codification was conceived. The first call for a unified movement toward theory construction was made, however, by Hans Zetterberg, who published in 1954 the most well-known book on the topic of theory construction--On Theory and Verification in Sociology. This book in effect became the manifesto for the movement. Three years later Merton published the second edition of Social Theory and Social Structure (1957), in which the issue of codification of theory and theory building occupied a center stage. In 1959, Llewellyn Gross edited a symposium monograph, a significant part of which was devoted to the discussion of issues related to theory construction. By then the movement had already made headway and gained momentum. The theory construction movement originated from widespread dissatisfaction with the inability of the Parsonsian paradigm to generate scientific laws of society similar to the laws discovered in the physical sciences. The state of sociological theorizing prior to the movement was best described by Merton's famous paragraph: "We have many concepts but fewer confirmed theories; many points of view, but few theorems; many 'approaches' but few arrivals" (Merton, 1949/1957/1968:52). In Zetterberg's words, the field of sociology in that period was marked by "theoretical paucity" in the presence of "an embarrasement of riches of ordinary hypotheses" (Zetterberg, 1954/1963:36/1965). The theory construction movement, if simply put, was to turn "ordinary hypotheses" into empirically verified theories that reveal laws of the social universe. The basic tenets of the movement can be broken down into the following five components: 1. An ontological assumption: It was assumed that there were "laws which govern social events in the same sense in which the laws of physics and chemistry govern the events in nature." That is to say, "there are uniformities of development and of interactions of sufficient stability to warrant a logico-format statement of laws and the integration of such laws into explanatory theories" (Gross, 1959:490-491). This statement contains two major points: (a) there exist stablized uniformities in social life; and, more importantly, (b) these uniformities can be captured in the form of general theories that explain a wide variety of social phenomena. "The assumption here is that sociology will eventually discover a small number of propositions that are valid in several diverse contexts" (Zetterberg, 1954/1963:8-9/1965). This assumption, which was crucial to the movement, was based on the accomplishments in the physical sciences that "made it possible to describe all the phenomena of the physical world in terms of

308

Zhao

a limited number of laws, which we call the theories of physics" (Zetterberg, 1954/1963:1/1965). 2. The growing gap between sociology and the physical sciences: The physical sciences had made significant progress in search of laws. A wellknown example was Newton's discovery of the gravitation of the physical world. Sociology, in contrast, had lagged far behind the law-finding race, bogged down deeply in mere descriptive studies or endless philosophical debates. There was much for the sociologists to do to narrow the huge gap. As Merton (1949/1957/1968:47) frankly acknowledged, "between twentieth-century physics and twentieth-century sociology stand billions of manhours of sustained, disciplined, and cumulative research. Perhaps sociology is not yet ready for its Einstein because it has not yet found its Kepler--to say nothing of its Newton, Laplace, Gibbs, Maxwell or Planck." 3. The need to borrow the methods of the physical sciences: The failure of finding general sociological laws was attributed to the lack of a sound methodology for social research. The major factor that thwarts the development of a scientific body of knowledge of social a n d h u m a n p h e n o m e n a is t h e c h a r a c t e r of social s c i e n t i s t s themselves--problems within the social scientists, not within the phenomena. Two major deficiencies are lack of clarity in theoretical writings and ignorance about what scientific knowledge should look like and how it is created. (Reynolds, 1971:163)

The success of the physical sciences had pointed out the direction in which sociologists should go--"[t]he physical sciences have been widely accepted as the model after which to pattern all sciences, and because the nature of theory construction, best exemplified in our most advanced sciences, is the same everywhere. If the social scientist wants to emulate the physicist, then the pattern is now clear" (Gross, 1959:490). 4. Toward a verificational approach: The theory construction movement was to "suggest certain conventions for sociological research . . . . These conventions may be termed 'formalized derivation' and 'codification'" (Merton, 1949/1957:100/1968). The methods codified by the earlier advocates of the movement--later known as the verificational approach--focused on the testing of theoretical propositions against empirical facts. A set of procedures was formalized covering virtually every major aspect of sociological research, including hypothesis formulation, data collection, statistical analysis, and report writing. The idea was that if researchers used the right methods of testing, they would be able to tell which theoretical statement was correct and which was wrong; if they kept doing that, they would be able to accumulate a large number of tested theories and eventually obtain a general theory of society upon synthesis of all the tested materials.

Beginning of the End

309

5. Starting from the middle-range theory: The ultimate goal of theory construction in sociology was to unveil the entire hierarchy of social laws. But, obviously it was impossible to achieve that all at once. According to differences in the level of abstraction, theories were divided into two categories: (a) miniature or middle-range theories and (b) grand or inclusive theories. The proposed strategy for theory construction was to start with miniature or middle-range theories and gradually move up the hierarchy of scientific laws until a grand or inclusive theory was obtained. "In physics, the theory of relativity and the quantum theory are inclusive theories in terms of which most laws of physics can be explained. The final goal of the scientific enterprise is to know such a theory (Zetterberg, 1954/1963:2/1965). In a nutshell, the theory construction movement as launched by Zetterberg and Merton was a redirection of the practice of sociological research by way of introducing a verificational approach in an attempt to discover gravitational laws of the social universe. In subsequent revisions of his most famous book, Zetterberg spelled out the details of the proposed theory-testing methods such that the theory construction movement had not only an influential manifesto and a well designed strategy, but also a set of codified procedures. It was then that Zetterberg (1954/1963:4/1965) announced with a note of envy that "Some fortunate members of the generation now being trained in sociology will be the first ever to orient themselves from the very start of their careers toward actual theory construction." In the early years of the theory construction movement, many sociologists regained their hope for the establishment of a truly scientific sociology. As Lipset (1994:201) recently recalled in an article, the "sense of revolutionary excitement" people experienced during that period is hard to reproduce: " T h e early period was m a r k e d by an e n o r m o u s surge of intellectual creativity and energy. We literally believed that we, or rather our elders, were creating, for the first time, a scientific sociology . . . . "

LATER YEARS OF THE MOVEMENT The "sense of revolutionary excitement" did not last long, however. The reason appears to be simple: the proposed approach failed to bring about the promised outcome. Before the onset of the movement, Merton (1949/1957:96/1968) described the paucity of scientific laws in the sociological field as follows: Despite the manyvolumesdealingwith the historyof sociologicaltheory and despite the plethora of empirical investigation, sociologists (including the writer) may discuss the logicalcriteria of sociologicallawswithoutciting a single instancewhich fully satisfies these criteria.

310

Zhao

Ten years later when Wilier (1967:vii) began writing his book on theory construction, he discovered to his dismay that still "there are no scientific laws of society." Blalock (1969:28) reached the same conclusion after he had tried to look for materials for his mathematical formulations: "Thus far, in sociology at least, we do not have enough good examples of reasonably systematic verbal theories to give useful insights." In other words, the state of theory formation in sociology was nearly as bad as it had been before the start of the movement. Failure to produce what had been promised caused widespread doubts about the theory construction movement. As Stinchcombe (1968:5) commented in the beginning of his book, '~kn argument about practical methods for constructing social theories is a pretty poor show if it cannot construct any." Wilier (1967:xix) was even more explicit about his impatience with any methodological discussions about theory construction: Real advancement in sociology, and not a mere discussion of legitimation, is now needed. Indeed, the only a n s w e r . . . i s in significant scientific development--not in another critique of existing methods, nor even in their refinement . . . . Nor would yet another call for a new direction be of value now. Instead, immediate progress must be made.

By the time the surge of book releases began, many of the authors already had lost their confidence in the original objective of the theory construction drive and started to redefine the direction of the movement in their own ways. Earlier advocates of the movement had hoped that, by focusing on middle-range theories at the outset, sociologists would eventually end up obtaining more general theories that have global implications. The ultimate success of the theory construction movement was to be seen in the establishment of "a small number of general propositions" (Zetterberg, 1954/1963/1965). This cumulative progress toward high level generalizations of sociological theory was essential, for "underlying this modest search for social uniformities, there is an enduring and pervasive concern with consolidating the special theories into a more general set of concepts and mutually consistent p r o p o s i t i o n s " (Merton, 1949/1957:10/1968). The importance of the generality of sociological theory was, however, significantly downplayed by many of the later authors. Blalock was probably the most notable explorer of general laws of society among the later authors. Blalock's goal was to search for the gold of universal laws in the garbage can of sloppy verbal theories. But it did not take very long for Blalock (1969:141) to realize that (1) it was indeed very hard to discover general laws about society and, perhaps more importantly, (2) general social laws were usually not very useful as they were most likely common-sensical:

Beginning of the End

311

Social scientists want their theories to have as wide a range of applicability as possible, while still being sufficiently specific to provide useful predictions and explanations of "middle-range" phenomena. This objective is very worthwhile and laudable, though it should be recognized that it may be difficult to formulate highly general theories that imply predictions taking us very far beyond the common-sense level of analysis.

Dubin was obviously less enthusiastic about endorsing the search for general laws. After noting that general laws had been defined either as "the law of interaction pertaining to a given universe composed of fLxed units" or as "a law that has application in several universes," Dubin (1969:117) denied the importance of either definition, arguing that it was not the business of sociologists to worry about the generality of their theories: If the theorist-scientist is interested in universal laws in either of these senses, there is certainly no prohibition against his pursuing them. What is important to recognize is that the category universal law is not used here as being a law different in kind from laws of interaction. For most theorist-scientists, the primary focus of attention will be on laws of interaction, leaving the issue of universality to the great generalizers and the philosophers.

This retreat from the pursuit of general laws was legitimated by Wilier in his introduction of the concept of universal conditionals. Citing Hempel and Oppenheim (1960), Wilier (1967:102-104) differentiated the scope of theory from the range of application of theory, arguing that, although scientific theories vary in their range of application, they do not differ in scope: "Scientific laws have to be constructed to satisfy the 'condition of nonlimited scope.' Nonlimited scope does not mean that the law must be stated without conditions, but that the conditions for the law may not be in the form of particulars." In other words, a scientific theory is conditional and universal at the same time: "conditional" in the sense that a theory is true only under the specified conditions, and "universal" in the sense that a theory must be true for an infinite number of cases meeting the specified conditions. Thus, by redefining the concept of universality, Wilier shifted the focus of interest in theory construction from generality to conditionality. The later period of the movement also witnessed the emergence of a negative attitude toward the employment of causal terms in theory construction (Klein, 1987). The concept of causation played a pivotal role in providing the philosophical underpinnings for the search of general laws of society. To the earlier advocates of the theory construction movement, it was never an issue that the world is causally structured. The ultimate goal of sociological theorizing was to unlock the secrets of the causal nexus of the social world. This attitude, however, began to change in the 1960s. Many authors of the later books appeared to be very cautious about the issue of causation. Some obviously refrained from employing any causal

312

Zhao

terms; others treated causation as one of the many relationships, rather than the relationship, to be studied; and still others simply d u m p e d the concept of causation altogether. Dubin (1960:92), for instance, m a d e it clear that "scientific law" denotes a specified interaction and only that: It does not improve our understanding or make more accurate our prediction to say that our law of interaction, Y=a+bX+cW, is a causal law. The operations by which we test the relationship between theoretically predicted values and empirical values differ in absolutely no respect whether we label the relationships among units of a model as laws of interaction or as causal laws. We conclude that no purpose is served by gratuitously employing the label causal law. A rather extreme position was taken by Gibbs, who flatly rejected the use of causal terms in theory construction. Gibbs held that since sociological research was primarily nonexperimental in nature, sociological data could tell us nothing other than space-time relations. This was precisely the reason that "a causal language is excluded from the proposed m o d e of theory construction": Should the word be used, tests of the theory are precluded unless the theorist stipulates rules by which causation is to be inferred from space-time relations, and that stipulation entails all of the issues and problems previously considered. (Gibbs 1972:24-25). It must be mentioned that these later authors were by no means denying the existence of causation in the social world; rather, they became m o r e aware of the difficulties associated with nonexperimental research in studying causation. Unlocking the secrets of the causal nexus of the social world was definitely a daunting task to many of the later authors. The change of the overall climate in theory construction in the later years of the m o v e m e n t was well indicated by the appearance of discussions about meta-theoretical issues. Meta-analysis is a friend of doubt. It has been the case that when theorists are sure about what they are doing, they do theories; but when they are in doubt, they do meta-theories (Zhao, 1991). Meta-theoretical discussions occurred with an increasing frequency toward the end of the m o v e m e n t and became a focus of attention in Hage's book on theory building. Hage divided his book into two parts: the first part was devoted to techniques of theory construction, and the second part to problems of meta-theory. In his incisive meta-theoretical discussions, H a g e reflected upon many important issues of theory construction, including the future of the movement. Hage (1972:186) noted quite philosophically that Theories are never true or false. They are partly true, part of the time. Just as a theory is an approximation, it remains something that is never completely right or wrong. Thus, we must strive continuously to improve our theories. The task of theory construction never ends.

Beginning of the End

313

Doubtlessly, the "sense of revolutionary excitement" experienced by many sociologists in the early years of the movement had been long gone by then. In general, the later authors became more sensitive to the methodological issues associated with theory construction, more conservative in judging the merits of their own approach, less confident of the direction of the movement, and even less optimistic about the cumulative process of discovering the gravitational laws of society. In a sense, the earlier advocates saw theory construction as an urgent calling, a lofty mission, and a huge responsibility while the later authors saw it as an interesting topic for a book, an intellectual itch to be scratched, or, probably, a long-term debt to be got rid of. This difference was clearly a sign of the coming of the end of the theory construction movement.

CONTINUATION OF THEORY CONSTRUCTION However, Hage was right in pointing out that theory construction would never end, for the movement toward the search for general laws by means of empirical verification was just one type of exercise in theory construction, other means of theory building are constantly being invented, practiced, and improved. Retrospectively, it is more accurate to say that the books on theory construction published in the seven years after the mid-1960s played a transitional role in putting an end to the ambitious search for general laws of society, and in introducing a more modest attempt to look for social regularities that occur under specified conditions. The latter approach to theory construction is well alive today and, in my opinion, is becoming a received view in the discipline. A cornerstone for the new approach is the concept of "universal conditionals" (Wilier, 1967; Cohen, 1980; Walker and Cohen, 1985). This concept is powerful for the following reasons: (1) Defining the universality of theory in terms of "nonlimited scope," which is analytical, rather than nonlimited range, which is spatial, effectively places all scientific theories on an equal footing, making generality less essential to theory building, as all scientific theories are by definition universal in the sense that they must be true for an infinite number of cases under the conditions specified by the theorists. Though theories may differ in the number of conditionals (generality) needed to ensure their universality (validity), it would make no sense to reduce conditionals at the price of universality. (2) Carrying this line of argument a bit further, it can be seen that the difference between the study of a unique historical instance and the survey of a large number of recurrent events does not lie in the principle of investigation but in the availability of cases to be examined. As Wilier (1967:104) ex-

314

Zhao

plained, "'Nonlimited' refers only to the form of the conditionals and to the applicability of the law to all cases thus contained. In the extreme case it would be possible to so limit the phenomena by appropriate conditionals that only one example ever existed." When this situation does occur, it is not the fault of the researcher, as there is not much that can be done to make it otherwise. The generality of a theory is in this sense determined by the nature of the relationship under study rather than by the method of investigation. The idea of nonlimited scope was later elaborated by Cohen (1980) and Walker and Cohen (1985) in their formulation of the concept of scope conditions. Walker and Cohen noted that the field of sociology has long been plagued by the paradox of a theory being both true and false under empirical examination, for a theory is often supported by one set of evidence and falsified by another, making theory testing fruitless. To resolve this problem, they suggested the use of scope statements to make explicit the conditions under which a theory is believed to be valid: A theory's scope, or domain of applicability, is a set of conditions such that, if the conditions are satisfied, the theory will not be found false. The scope of a theory can be described by a set of statements which we refer to as scope statements. Scope statements share the property of universality with statements of general principles. That is, they are comprised of constructs that do not refer to particular times and places. (Walker and Cohen, 1985:291)

A theory is, therefore, universal only under the specified conditions. If a theory is falsified by facts, or its universality is challenged by the presence of exceptions, the conditionals of the theory need be respecified to further delimit the theory's scope such that all the cases within the scope conform to the regularities theorized by the theorist. If, on the other hand, a theory has been repeatedly supported by evidence, the theorist may want to relax the conditional restrictions to make the theory more general in its range of application. A theory is, in this sense, assumed to be restrictive unless proved otherwise. The extent to which a theory can be applied across diverse contexts is in the final analysis determined by the ontology of the world. While the universality of a theory is under the control of theorists, its generality is at the mercy of God. Whether it is possible to construct a gravitational theory that explains the entire range of social phenomena is an ontological issue that can not be answered methodologically. The concept of scope conditions, however, needs further modifications to accommodate the fact that most sociological research is nonexperimental. Due to the inability of creating experimental settings that resemble the specified conditions, sociologists are not able to collect uncontaminated cases that fit perfectly the theoretical profiles. As a result, obtaining pure universality in sociological theorizing is highly unlikely--assuming that we

Beginning of the End

315

do live in a deterministic world--even if the scope condition of a theory is properly delineated. One solution to this problem is to take the advice of Lieberson about the use of probabilistic terms in theory testing. Rather than requiring every single case to be consistent with a theory under the specified conditions, we look for the "relative likelihoods" of the occurrence of an event under different conditions. As Lieberson (1992:7) put it, If theories are posed in probabilistic terms, i.e., specifying that a given set of conditions will alter the likelihood of a given outcome, not only will the reality of social life be correctly described, but we will also be freed from assuming that negative evidence automatically means that a theory is wrong (A deterministic theory posits that a given set of conditions will lead to a speeified outcome, pure and simple).

Another important feature of the emerging approach to theory construction is methodological pluralism. Early advocates of the theory construction movement believed that the verificational approach was the only correct strategy for doing theory, and that approach must be imposed on the discipline from top down. The process of imposition of the "right" approach was called "codification of theory and research" (Merton, 1949/1957/1968). Failure to generate general scientific theories of society, however, effectively destroyed the credibility of this approach. As a result, authors of the later books developed a more open attitude toward different methods of theory construction. Stinchcombe (1968:4), for example, stated that, "I have a firm conviction that some things are to be explained one way, some another. Trying to explain a phenomenon by a strategy inappropriate to the empirical terrain, because one thinks that a strategy is a 'theory' which must be either true or false, leads into ambushes." The later years of the movement were characterized by the advocacy of different methodological approaches to theory building: Glaser and Strauss' (1967) grounded theory methods, Willer's theory-model strategy, Stinchcombe's interest in the logical forms of theories, Blalock's effort toward mathematical formulations, Gibbs' insistence on the formal mode, and Hage's elaboration of theory building techniques. Although the methods proposed by these authors were different both in philosophical orientation and technical sophistication, they seemed to have converged on several basic dimensions that constitute what now Lenski (1988:1966) called the "minimal standards" for good theory construction: (1) unambiguous concepts, (2) specified relationships between or among the concepts, and (3) the formulation of the theory in such a way that the theory is empirically falsifiable. In other words, theorists are free to construct theories in any way they want as long as the concepts are unambiguously defined, the relationships are clearly specified, and the theoretical propositions are empirically testable (Cohen,

316

Zhao

1994). This liberal attitude toward theory construction was advocated long ago by Mills (1959:72): We should be as accurate as we are able to be in our work upon the problems that concern us. But no method, as such, should be used to delimit the problems we take up, if for no other reason that the most interesting and difficult issues of method usually begin where established techniques do not apply.

There is no correct method as such for theory building. A method is only as good as it can resolve problems in theory construction. Given the variety of problems that take place in theorizing, it is impossible to codify a procedure that works in all situations. The best that a methodologist can do is to state a set of minimal criteria any good method must satisfy, and this is indeed the direction in which the new approach has been going. The basic tenets of this new approach to theory construction can be outlined as follows: 1. No assumption of causal invariance in society: It is assumed that regularities exist in social life in different forms and degrees. These regularities do not need to be causal, nor are they necessarily invariant. The aim of sociological research is to find out the extent to which regularities exist in a given aspect of social life and the conditions under which the regularities are maintained and/or altered. 2. No imposition of codified methods: As there are different forms and degrees of regularities in social life, there must be more than one way to study them. Established methodological criteria are useful only in the sense that they provide guidelines for practice; sociologists should be encouraged to tackle research problems in whatever methods they think are the best for the tasks at hand. It does not matter whether these methods are borrowed or invented, inductive or deductive, formal or informal: they are good methods as long as they can be used to accomplish what other methods fail to do. 3. Empirical tests as the ultimate criterion: In an empirical science such as sociology, the difference between a good hunch and a good theory lies in the fact that while the former is based on intuition, the latter on empirical evidence. Theory construction methods, therefore, should facilitate the validation of theoretical propositions on the basis of empirical facts. No method can guarantee the discovery of gravitational laws of society, but a good method of sociological theorizing must ensure empirical testability of what is claimed to be discovered. In summary, I have argued in this paper that the theory construction movement in sociology can be divided into two phases, with the second phase marking a major retreat from the first one. Contrary to the common belief, the publication of books on theory construction that took place in the late 1960s and early 1970s signaled the end of the movement rather

Beginning of the End

317

than its beginning. Failure of the movement is attributable to two main factors: (1) ontologically, it was unwise to assume the existence of gravitational laws of society and their obtainability; and (2) methodologically, it was a mistake to advocate only the verificational approach to theory construction. The belief that the formalized verificational approach would lead to the discovery of general scientific laws of society effectively killed the movement after repeated failure to realize it. These two mistakes were corrected to a large extent by the later authors, who paved the way to the rise of a new approach to theory construction characterized by ontological neutrality and methodological pluralism. The existence of gravitational laws of society is no longer assumed, let alone its obtainability. The formal mode of theory construction is treated as one of the many research strategies that sociologists may use to investigate the lawfulness of the changing social world. Although the theory construction movement is now dead, efforts toward theory construction will continue.

ACKNOWLEDGMENTS I would like to thank David Wilier and other three anonymous reviewers for their helpful comments on the earlier version of this paper. I also want to thank editor Richard Hall for his understanding and support.

REFERENCES Blalock, Hubert M., Jr. 1969 Theory Construction: From Verbal to Mathematical Formulation. Englewood Cliffs, NJ: Prentice Hall. Chafetz, Janet S. 1978 A Primer on the Construction and Testing of Theories in Sociology. Itasca, IL: Peacock. Cohen, Bernard P. 1980 "The conditional nature of scientific knowledge." In Freese Lee (ed.), Theoretical Methods in Sociology: 71110. Pittsburgh, PA: University of Pittsburgh Press. 1989 Developing Sociological Knowledge: Theory and Method, 2nd ed. Chicago: Nelson-Hall. 1994 "Sociological theory: The half-full cup." In Jerald Hage (ed.), Formal Theory in Sociology: 66-83. Albany: State University of New York Press.

Dubin, Robert 1969 Theory Building. New York: The Free Press. Freese, Lee, ed. 1980 Theoretical Methods in Sociology. Pittsburgh, PA: University of Pittsburgh Press. Gibbs, Jack P. 1972 Sociological Theory Construction. Hinsdale, IL: The Dryden Press. 1994 "Resistance in sociology to formal theory construction." In Jerald Hage (ed.), Formal Theory in Sociology: 90104. Albany: State University of New York Press. Glaser, Barney G., and Anselm L. Strauss 1967 The Discovery of Grounded Theory. New York: Aldine. Gross, Llewellyn, ed. 1959 Symposium on Sociological Theory. New York: Harper & Row.

318 Hage, Jerald 1972 Techniques and Problems of Theory Construction in Sociology. New York: Wiley. 1994 "Sociological theory: Complex, fragmented, and politicized." In Jerald Hage (ed.), Formal Theory in Sociology: 52-65. Albany: State University of New York Press. Hempel, Carl G., and P. Oppenheim 1960 "Problems of the concept of general law." In Arthur Danto and Sidney Morgenbesser (eds.), Philosophy of Science: 188-204. Cleveland, OH: Meridian Books. Klein, David M. 1987 "Causation in sociology today: A revised view." Sociological Theory 5:1926. Lenski, Gerhard 1988 "Rethinking macrosociological theory." American Sociological Review 53: 163-171. Lieberson, Stanley 1992 "Einstein, Renoir, and Greeley: Some thoughts about evidence in sociology." American Sociological Review 57: 1-15. Lipset, Seymour Martin 1994 "The state of American sociology." Sociological Forum 9:199-220 Merton, Robert K. 1949/1957/1968 Social Theory and Social Structure. New York: The Free Press. Mills, C. Wright 1959 The Sociological Imagination. New York: Oxford University Press.

Zhao Mullins, Nicholas C. 1971 The Art of Theory Construction and Use. N e w York: Harper & Row. Reynolds, Paul D. 1971 A Primer in Theory Construction. Indianapolis, IN: Bobbs-Merrill. Stinchcomhe, Arthur L. 1968 ConstructingSocial Theory. N e w York: Harcourt Brace Jovanovich. Turner, Jonathan 1994 "The failure of sociology to institutionalize cumulative theorizing." In Jerald Hage (ed.), Formal Theory in Sociology: 41-51. Albany: State University of New York Press. Walker, Henry A., and Bernard P. Cohen 1985 "Scope statements: Imperatives for evaluating theory." American Sociological Review 50:288-301. Wallace, Walter 1971 The Logic of Science in Sociology. Chicago, IL: Aldine. Wilier, David 1967 Scientific Sociology: Theory and Method. Englewood Cliffs, NJ: Prentice Hall. Zetterberg, Hans L. 1954/1963/1965 On Theory and Verification in Sociology. Totowa, N J: The Bedminster Press. Zhao, Shanyang 1991 "Metatheory, metamethod, meta-dataanalysis: What, why, and how?" Sociological Perspective 34:377-390.