Tilly > Weblog

Introductory Essay to Charles Tilly’s Writings on Methodology

Under the direction of Johann Peter Murmann

Introduction
While Charles Tilly may be best known for his research of large scale social change and collective action in European history, he has also written extensively on research methodology. Over the years, roughly a quarter of his publications have concerned method, ranging from specific techniques to general considerations of logic, epistemology, and ontology. In his early methodological writings, Tilly spoke especially to social historians, urging them to embrace quantitative (formal) methods in historical analysis (Tilly 1972). Later, Tilly’s methodological writings focused more on the ontology of macro social change. In assembling a website of Tilly’s methodological contributions, our primary goal is to provide a convenient source where scholars can become acquainted with his methodological views on social research.

Core to Tilly’s writings is the assertion that history matters and is a critical component in analyzing social change. For Tilly, the fields of social history and sociology are parallel paths towards understanding social change, but both require adjustment in their approaches. Addressing sociologists, Tilly advises them to return to the sociology of Marx’s Capital and Weber’s Economy and Society, in which social processes and social structures are seen first and foremost as historically contingent (Tilly 1995b). Addressing historians, Tilly urges them to overcome what he regards as a false dichotomy between qualitative and quantitative research and instead apply formal methods in all areas of historical research. Tilly recognizes that many of the organizing questions of history as a discipline focus on trying to understand the experiences and conditions that underlie an important historical time period. As a result, historians typically assume that these questions are best answered through informal methods (Tilly 1985a). Tilly believes, however, that the use of formal methods, prevalent in the historical study of urban areas, labor, social mobility, and collective action, can aid all types of historical analysis by sharpening arguments and ruling out alternative explanations (Tilly 1984).

Addressing comparative historical sociologists, Tilly challenges analyses of macro social change where nations, states, or societies serve as the unit of analysis, a practice which he terms the Big Case Comparison method. Take as an example Spruyt’s (1994) analysis of the factors that led to the dominance of the sovereign state system in post-modern Europe through the comparison of three cases: the French territorial state, the Hanseatic city-league, and the Italian city-state. Tilly argues that the Big Case Comparison method is based on the misguided ontological assumption that individual states, societies, and cultures exist as autonomous entities (Tilly 1995a). Instead, Tilly proposes that researchers adopt a mechanism approach to the study of social change, where the goal is to identify the robust mechanisms which consistently make up large processes, albeit in various configuration and sequences.

After reviewing Tilly’s methodological writings, we have organized them for the purposes of this introduction (and the website) into three classes. In the first section, we provide an overview of Tilly’s writings concerned with the field of social history. Tilly has written a number of pieces that define unique characteristics of the field of social history in relation to sociology. The second section focuses on Tilly’s advocacy of formal methodologies in historical analysis, a consistent theme throughout his early writings. Tilly argues that formal methods offer scholars an analytic tool that can provide a deeper understanding of social phenomenon. The final section links Tilly’s view concerning formal methods to his ontological perspective of macro social change. On this topic, Tilly is critical of current comparative historical work in sociology that uses small n case comparisons to substantiate broad generalizations about large structures and processes. Tilly advocates that researchers should instead identify the smaller social mechanisms that in combination structure larger social phenomena. We conclude with a short guide on how to navigate this website.

The Field of Social History
In an article outlining the character of Historical Sociology, Tilly recounts Comte’s early nineteenth century conception of sociology as a field which “consisted largely of analyzing the development of humanity through historical stages” (Tilly 2001: 6753). It was after this point, Tilly argues, that the fields of sociology and history developed in separate directions. Historians segmented themselves by specializing in time and place (i.e. modern Latin America or Ancient China). What makes the field of history distinct, in Tilly’s view, are five characteristics: (1) its insistence on time and place as fundamental causes of variation, (2) its interpenetration of professional and amateur efforts, (3) its heavy reliance on documentary evidence, (4) its emphasis on identifying critical actors and their motivation, and (5) its presentation in narrative form (Tilly 1991).

By contrast, sociologists, Tilly writes, separated from historians by specializing in structures and processes (i.e. families, religions, industrialization), giving rise to an academic field that is characterized primarily by its explicit conceptualization and hypothesis testing, its use of systematic comparison, and its attempts to verify knowledge objectively (Tilly 1986). Tilly underscores that the interaction between the two fields is limited to relatively small sub-fields such as historical sociology, historical demography, and urban history, leaving sociologists and historians largely unaware of each other’s methods, models, ideas and discoveries.

This lack of interaction, in Tilly’s view, impairs scholarly progress in both areas. Tilly proposes research that combines the two approaches, arguing that this will help address two important features of social phenomenon. First, given the protracted period of time it takes for important social and political processes to unfold, historical knowledge is required to understand these phenomena. Tilly argues that historical knowledge of sequences and events is essential to any study of large scale social change such as war-making, capital accumulation, population growth, or international migration (Tilly 1991). Second, past social relations and their residues constrain present social relations and their residues. For this reason historical sociologists contend, for example, that understanding the construction of modern institutions of credit and property rights requires first understanding how historical network ties became transformed into tangible structures and resources (Adams, Clemens and Orloff 2005; Verdery 2003). Tilly argues that social history (or historical sociology) is one of the few areas that acknowledges and explicitly studies the constraints that history imposes on how institutions can develop. According to Tilly, compelling historical sociology, then, does more than simply incorporate historical knowledge into abstract sociological models. It strikes a delicate balance between abstracting enough so that the analysis is generalizable to more than one case and maintaining sufficient specificity so that the analysis identifies recognizable features of the social world (Tilly 1986).

Tilly has also written about key ongoing philosophical debates in the field of social history; in the article "How (and What) Are Historians Doing?" (1991) Tilly highlights four debates in particular. The first concerns the question of whether the central phenomena in macro-social change are large social processes or individuals. The second debate focuses on whether the study of macro change requires the observation of human action or the interpretation of motives behind that action. The third debate tries to sort out how closely sociology and history are related to one another. The fourth debate centers on how to best present data for macro change, as an explanation or as a narrative. Recognizing that each set of options represents a continuum rather than dichotomous choice, Tilly believes that most historians lean towards the latter alternatives in each debate while social scientists prefer the former. Historians in general, Tilly writes, favor research that represents a narrative-based interpretation of individual experience, and they perceive their endeavor to be quite distinct from sociology.

Formal Methods
Tilly has continuously advocated formal methods in historical research. Tilly defines formal methods as the explicit representation of a set of elements and the relations between them. In the 1960’s formal methods were hailed as the big breakthrough in modern historical analysis. According to Tilly, historians collectively billed the new effort as the “New Social History” and classified research using formal methods under three major headings. Statistical studies that accumulated standardized biographies of individuals, households, other small-scale social units, and events into collective portraits within particular historical contexts were labeled as Prosopography. Many studies under this category are found in labor history; they were used to analyze topics such as the determinants of fluctuations in national strike activity and the demographic correlates of different forms of industrial organizations (for more examples see Tilly 1985a). Collective Biography, defined as “the assembly of comparable files concerning the lives of many individuals, followed by the regrouping of those files into a collective portrait of the population involved” (Tilly 1985b: 22), has been used extensively in studying European history to “trace the impact of capitalism […] and changes in the character of national states on day-to-day behavior” (Tilly 1991: 94). Event Catalogs has been frequently used in the study of political contention. Event Catalogs, defined as a set of descriptions of multiple social interactions collected from a delimited number of sources using a uniform procedure (Tilly 2002), has provided the key to uncovering many of the underlying mechanisms related to political protests, strikes, and revolutions.

According to Tilly’s own recollection however, the “New Social History” never quite lived up to expectations. Tilly notes that the proponents of these methods in 1960’s later faulted studies that employed such methods as having lost “their wit, grace, and sense of proportion in the pursuit of statistical results” (Tilly 1984: 369). Tilly details how in the 1970’s, the field moved away from the complicated numerical modeling required in formal analysis and to focus again on the traditional method of historical analysis, namely, the creation of comprehensive narratives. Despite this shift back to the narrative form of analysis in the field of Social History, Tilly remained an advocate of formal methods, with the exception of Invariant Modeling, which Tilly strongly rejects (we will say more about this in the next section).

Time and again, Tilly’s writings argue that formal methods are useful for the analysis of both qualitative and quantitative evidence. Tilly urges historical analysts to use formalisms in their research, particularly in the stage between the initial archival data collection and the final production of a narrative. This, Tilly writes, provides two key advantages for historical research. First, formalisms discipline the encounter of argument and evidence. “Good formalisms”, Tilly writes in a recent article, “make explicit the analyst’s claims about relations among the elements under observation” (Tilly 2004: 598). Tilly implies that forcing researchers to make these explicit claims improves the logical structure of their argument. Second, having an explicitly stated argument improves the ability of other researchers to formulate testable competing explanations.

One of the main roadblocks to the increased use of formal methods in historical analysis, in Tilly’s opinion, is the widespread belief that quantitative and qualitative research is fundamentally different. Tilly contends that this distinction misidentifies historical work as either (1) the collection of evidence or (2) the writing of narratives. For Tilly historical analysis is always a combination of both. Tilly adds that this misidentification fails to recognize that the choice of using formalisms in the transformation and analysis of evidence is independent of the choice of using formalisms in the presentation of the evidence (Tilly 2004). Tilly believes using formal methods during the analysis phase is important for any research studying social processes. Tilly points out that the eventual presentation of the data can be either more quantitative by including tables, mathematical formulas and graphic representations, or can be more qualitative by presenting detailed narratives.

Despite the value that formalisms provide in understanding social processes, Tilly believes that formal methods will remain at the periphery of many domains of social historical analysis until members of the discipline begin to address research questions that involve explicit models, systematic variation, and the comparison of many cases. For Tilly, the usefulness of quantification in historical research increases as a function of the complexity of the explanatory model, the importance of variation to the argument, and the number of units observed (Tilly 1987).


Ontology of Macro Social Change
There is a deep conceptual connection between Tilly’s advocacy of formal methods and his view of the nature of social processes and social change. Tilly does not believe, however, that all formal methods are created equal. One formal methodology that Tilly has critiqued sharply is Invariant Modeling, typically used in Big Case Comparison (BCC) studies. His primary objection is ontological in nature: Tilly rejects the assumption in invariant modeling that society consists of self-contained and self-directed elementary units which exist in recurrent structures and processes. He labels this criticism “misplaced concreteness” (Tilly 1995a: 3). Informed by his own empirical work on political processes, Tilly argues that current BCC efforts that construct invariant models that are continuously altered when new evidence does not fit have failed to provide cumulative knowledge of how social processes work (Tilly 1995b).

Tilly also criticizes BCC for drawing polemically misleading conclusions about how social processes take place. BCC, Tilly writes, is based on John Stuart Mill’s methods of agreement and disagreement which requires that the researcher has the ability to observe all possible cases. Because, Tilly continues, resource limitations deny any researcher this luxury, social scientists that use this method spend most of their time demonstrating how other models do, or do not fit a newly discovered case of a specific phenomenon. The responses to Theda Skocpol’s classic States and Social Revolutions: A Comparative Study of France, Russia, and China are a prime example. Since its publication in 1979, various scholars have attempted to improve Skocpol’s general invariant model of social revolutions to account for increasing number of more recent revolutions in countries such as Iran, Vietnam, and Nicaragua (Goodwin 2001; Parsa 2000).

Tilly believes that this exercise of “improving the model” is a waste of time because macro-social processes do not operate in the form of recurrent structures and processes. Tilly sees macro-social processes as a complex combination of smaller social episodes, processes, and mechanisms. Tilly outlines this ontological perspective and provides examples in his article entitled “Mechanisms in Political Processes” (2001). Within the “mechanism view”, social mechanisms, a delimited class of events that change the relations among elements, are the basic elements of social phenomena. One example is brokerage, a mechanism which joins two social parties more directly by a third party and a cornerstone of sociological theorizing since the writings of Simmel’s social forms. Frequently recurring combinations of these mechanisms make up social processes, the second level of Tilly’s hierarchy. Tilly cites the example of the process of scale shift, defined as a change in the number of sites that engage in a coordinated action. This process, Tilly writes, is a common result of “a concatenation of brokerage with the mechanisms of diffusion, emulation, and attribution of similarity” (Tilly 2001: 26).

The largest and most malleable of the building blocks are social episodes, defined as bounded streams of social life. There is no consensus about what constitutes the bounds of a social episode, Tilly writes, because social episodes are socially constructed by its participants and observers. Episodes can be grouped for purposes of coherent comparison, as with the study of social revolutions, or they can be grouped based on conventions of the analysts who study them. In the end, for proponents of the mechanism-view, the coherence and significance of a social episode is “to be proven rather than assumed” (Tilly 2001: 26). For an example of a social episode, Tilly draws on Mexican politics and captures three distinct events. His episode begins with the period of presidential opposition mobilization in Mexico from 1988-2000, includes the presidential campaign from 1999-2000, and concludes with the Mexican presidential election of 2000.

Applying his mechanism-based approach to social revolutions, Tilly drives his point home with an oceanic analogy (Tilly 1995b: 1601). He describes proponents of the BCC method as conceptualizing revolutions as a phenomenon such as the ocean tide where regularities can be deduced from a singular or short list of causes (i.e. celestial motion). Instead, Tilly argues, revolutions should be conceptualized as a great flood where the causes of the occurrences are fairly standardized but the variance observed in each occurrence is dependent on recurrent causes in differing circumstances. Just as the actual unfolding of a flood is a function of the combination of various elements such as existing terrain, previous precipitation, and human response, the unfolding of a social revolution is a function of the combination of various elements such as the strength of state power, the alienation of the elite class, and the popularity of common grievances.

Tilly notes that aside from the mechanism-based view, social scientists have articulated four other ontological perspectives: (1) the “covering law” perspective in which explanations of social phenomena consist of empirical generalization which at the highest levels of abstraction become standing laws (invariant modeling falls in this category); (2) the “skepticism” perspective which holds that social life is “so complex, contingent, impenetrable, or particular as to defy explanation” (Tilly 2001: 22); (3) the “propensity” perspective which focuses on the motivations of human actors embedded within the phenomenon (Tilly 1999); and (4) the “system” perspective, a functionalist view in which social phenomena are explained by their consequences for the system in which they exist (Tilly 2000). Tilly favors the mechanism and process view, which explains salient features of large social processes by identifying the smaller components within. Throughout much of his career he has been a staunch proponent of this approach particularly in his own areas of expertise such as political processes (Tilly 2001) and collective action (Tilly 1989).



References

Adams, Julia, Elisabeth S. Clemens, and Ann Shola Orloff (Eds.). 2005. Remaking Modernity: Politics, History, and Sociology. Durham and London: Duke University Press.

Goodwin, Jeff. 2001. No other way out : states and revolutionary movements, 1945-1991. Cambridge ; New York: Cambridge University Press.

Parsa, Misagh. 2000. States, ideologies, and social revolutions : a comparative analysis of Iran, Nicaragua, and the Philippines. Cambridge, UK ; New York: Cambridge University Press.

Spruyt, Hendrik. 1994. The sovereign state and its competitors : an analysis of systems change. Princeton, N.J.: Princeton University Press.

Tilly, Charles. 1972. "Quantification in History, as Seen from France." in The Dimensions of the Past, edited by Val Lorwin & Jacob Price. New Haven: Yale University Press.

—. 1984. "The Old New Social History and the New Old Social History,." Review 7:363-406.

—. 1985a. "Neat Analyses of Untidy Processes." International Labor and Working Class History 27:4-19.

—. 1985b. "Retrieving European Lives." in Reliving the Past. The Worlds of Social History, edited by Olivier Zunz. Chapel Hill: University of North Carolina Press.

—. 1986. "The Tyranny of Here and Now." Sociological Forum 1:179-188.

—. 1987. "The analysis of popular collective action." European Journal of Operational Research 30:223-229.

—. 1989. "History, Sociology and Dutch Collective Action,." Tijdschrift voor Sociale Geschiedenis 15:142-157.

—. 1991. "How (and What) Are Historians Doing?" Pp. 86-117 in Divided Knowledge Across Disciplines, Across Cultures., edited by David Easton and Corinne S. Schelling. Newbury Park: Sage.

—. 1995a. "Macrosociology Past and Future,." Newsletter of the Comparative & Historical Section, American Sociological Association, 8:1, 3, 4.

—. 1995b. "To Explain Political Processes." American Journal of Sociology 100:1594-1610.

—. 1999. "Wise Quacks." Sociological Forum 14:55-62.

—. 2000. "Errors, Durable and Otherwise,." Comparative Studies in Society and History 42:487-493.

—. 2001. "Mechanisms in Political Processes." Annual Review of Political Science 4:21-41.

—. 2002. "Event Catalogs as Theories." Sociological Theory 20:248-254.

—. 2004. "Observations of Social Processes and their Formal Representations." Sociological Theory 22:595-602.

Verdery, Katherine. 2003. The vanishing hectare : property and value in postsocialist Transylvania. Ithaca: Cornell University Press.