20 – Evidence of What?

1Surprisingly and in repeatedly unexpected ways, historians have continued to debate in recent decades the central question of their craft: how can the past be described and explained? At each stage of the debate, the answers to this question have reflected and contributed to larger epistemological discussions across the disciplines. The following discussion examines selected aspects of the twists and turns of recent historical debate by using the example of research on census enumerations. From the time of the “new social history” of the 1960s and 1970s to the cultural history of the 1980s and 1990s, scholars have focused on census enumerations for quite different reasons and in quite different ways. At the core of this research have been changing answers to the question of evidence: census enumerations are evidence of what? In examining the different ways that scholars have addressed this question, particular attention will be paid to the interactions of historians with scholars in other disciplines across the social sciences and humanities. This analysis will focus on Canadian research although scholars in other countries especially the United States and Great Britain have similarly continued to debate how census enumerations can support historical interpretations.

2Before the 1960s, researchers did not view the census as a valuable historical source for analysing Canadian history. In contrast, the government officials who actually administered the census at the start of each decade since the mid-nineteenth century believed that they were creating a permanent record about Canadian society that would, in fact, be used by future researchers to analyse the patterns and trends of social, economic, cultural, and political change. As a result, government record-keepers and archivists were charged with preserving census enumerations and, despite pressures of space and resources, they did so to a considerable extent including the substantial microfilming in the early 1950s of original manuscript census schedules reaching back into the nineteenth century.1

3The conviction of census officials and archivists that the Canadian census should be preserved for historical study proved to be justified during the 1960s when scholars turned to the census for two key reasons: to study evidence of behaviour and to learn about the historically “anonymous.” Under the leadership of Michael Katz, David Gagan, and a few other researchers, the census came to be seen as a way to move beyond the literary sources that characteristically underpinned the established historical scholarship.2 These sources were now criticized as only providing “impressionistic” evidence of the thoughts, ambitions, and claims of a small number of official and unofficial leaders. The census promised to enable interpretations based not on the ideas of a minority but rather on the behaviour of the whole society. Partly as a rejection of the “history of ideas” and intellectual history that had gained considerable favour among historians by the 1950s, the new compelling questions in historical debate concerned what was being done across populations rather than what was being written by elites.

4The pioneering studies drew upon the social sciences both for concepts and methods. Sociology, for example, offered ways to study occupational structure while anthropology provided methods for analyzing family and household structures and networks. The basic approach was to count individual responses to the census enumeration questions and to inter-relate them with the responses of other individuals as listed in the same dwelling, neighbourhood, and community. The census thus became associated with quantitative history as the evidence of individual behaviour was counted up to identify patterns within class, ethnicity, and gender similarities and differences.

5At the time, the research possibilities seemed endless. Each census during the nineteenth century provided an increasing number of responses to an increasing number of questions ranging from those focused on personal attributes such as age, sex, and birthplace to those dealing with the means of living such as occupation and agricultural production. The ability to examine the manuscript census returns that were made available by the 92-year confidentiality policy allowed researchers during this period to move from the 1851-52 to the 1861, 1871, and 1881 enumerations. As a result, researchers could not only examine one historical moment but could also compare patterns as well as follow individuals from one census to the next. In this work, a North American version of micro-history became the preferred research strategy of scholars using the census as evidence of the historically anonymous. Focused on specific communities, townships, or cities micro-history sought to increase the level of interpretation and understanding by reducing the level of observation. Unlike local histories, micro-historical research treated the analysis of individuals and groups in a particular time and place not as an end in itself but rather as a means of understanding larger historical changes familiar elsewhere as well.3 By systematically examining the responses to census questions for certain enumeration districts, scholars sought to enhance their grasp of generalized social and economic transformations.

6In unprecedented ways, scholars claimed to be rewriting the history of modern society by describing significant features of the lives of both the “famous” and “anonymous.” Micro-histories emphasized the value of studying popular behaviour rather than elite perceptions, and they challenged established understandings of key features of the nineteenth century especially urbanization, commercial development, and early industrialization. In this way, the census became a familiar source in detailed studies of the specific and diverse ways in which social and economic change occurred during the nineteenth century.

7One example of the robust and often surprising results of the census-based, micro-historical studies of the 1970s was the conclusion that fertility rates declined during the later nineteenth century in both rural and urban areas. Although the pattern of this decline varied considerably across the communities under study, the discovery of significantly changed fertility rates seemed to justify the enthusiasm for the “new social history.”4 Not only did the decreasing family size reflect a major behavioural change that was not documented in the maligned literary sources but it occurred despite the absence of any official or unofficial support among leadership groups. In this way, the declining fertility rates that accelerated the growth of mass schooling, new domestic relations, and numerous other social and economic reconfigurations encouraged scholars to continue studying the ways that historical change occurred from the “bottom-up.” Moreover, the study of fertility further motivated scholars to undertake micro-historical studies since the initial research projects revealed considerable variation in the ways and extent to which family size declined across different groups and settings. This diversity suggested that questions of family size were addressed within webs of relationships radiating out from the domestic to larger communities. By focusing on specific cities or townships, scholars believed that it would be possible to incrementally move toward comprehensive understandings by comparing patterns of behaviour across time and space. The manuscript census seemed to be an ideal source for such comparisons since each enumeration required responses to the same questions across different communities. Scholars were often frustrated by the fact that census officials did not simply add but also sometimes changed the wording or definition of specific questions from one enumeration to the next but, in comparison to other historical sources, the manuscript census seemed to offer rich evidence of mass behaviour to an extent that far surpassed other sources.

8But soon some nagging concerns moved from footnotes to the top of the research agenda of many scholars including some of those who had been most enthusiastic about census-based micro-histories. This transition gained speed during the late 1970s and through the 1980s and was propelled by research results that raised questions about each of the key distinguishing characteristics of the new social history. In the case of studies of declining fertility, for example, the complexity of behavioural patterns motivated scholars to focus increasingly on the possibility that attitudes, perceptions, and values played key roles in explaining the diverse trajectories across different groups. The continuing inability of researchers to fully explain fertility patterns by interrelating factors such as religious identity, occupation, and wealth inspired greater attention to changing ideas and priorities about which the census seemed to provide no evidence.

9Similarly, the repeated research finding that, with only limited exceptions, family size was declining at least to some extent across diverse communities during the later nineteenth century suggested that a macro-level process was unfolding in these years. Scholars increasingly suspected that this process involved conceptual changes that lead to changed behaviour.5 These historiographical developments called into question the conviction that the key to understanding historical change involved systematic study of mass behaviour through micro-historical research on sources such as the manuscript census. Rather, interest turned to the possibility that social and economic change had to be explained in terms of perceptions and attitudes that transcended the particularities of specific communities and that had to be apprehended through study of what was said and written.

10In this way, the research findings produced during the late 1960s and 1970s undermined the perceived value of the concepts and methods borrowed from the social sciences. In turn, these research findings encouraged the scholarly rehabilitation of the literary sources so thoroughly disparaged just a few years earlier. The result was that the new social history unexpectedly helped fuel the rise of the influence of the humanities on historians, and the popularity of cultural history during the 1980s and 1990s. This trend was reinforced by an increasing scepticism about the ability of the manuscript census to even provide reliable evidence about mass behaviour. The initial tendency to take at face value the responses to the various census questions written on the enumeration forms gave way to growing doubt that these forms deserved to be studied at all. Using a vocabulary of errors, inaccuracies, pitfalls, biases, and limitations, researchers increasingly documented the difficulties of carrying out a census in settings like nineteenth century Canada. Historical debate moved from early skirmishes among researchers about the usefulness of specific census questions to full-blown battles about the evidentiary value of any enumeration for understanding the “anonymous,” specific communities or social and economic transformations.6

11The increasing questions about how successfully census enumerations were conducted helped to discourage researchers from undertaking projects that would pursue the interpretations offered by scholars such as Katz and Gagan. Although some substantial efforts did continue during the 1980s, more and more researchers turned away from the study of popular behaviour in specific settings based on the census. Instead, increasing attention came to be focused on ideas, those in positions of influence, and macro-level developments. Unlike the earlier history of ideas and intellectual history, however, the new cultural history followed the “linguistic turn” toward the view that reality is not perceived but rather is constructed by mental processes. This view attributed an unprecedented scholarly importance to the articulation of thoughts not only as expressed in written form but also in all human creations. While some scholars argued that historical understanding could never move beyond the analysis of the various articulations of human thought, most historians engaged cultural history as a new strategy to address the same questions that attracted attention during the earlier “new social history” especially those posed in terms of the origins of modern societies. This new strategy characteristically made three assumptions: that culture significantly explains behaviour; that those in positions of official and unofficial power primarily cause historical change; and that larger forces significantly frame the histories of local areas. Clearly, these assumptions contrasted markedly with the previous emphases on what people did more than what they said, on the possibility of historical change originating “bottom-up,” and on the need to lower the level of observation in order to increase the level of explanation.

12Surprisingly, perhaps, given the rapidly changing scholarly orientation, the census soon became a focus of historical research once again. Now, however, the census was defined as providing evidence of elite perspectives, values, and ambitions. In this view, it was the census questions and not the answers that were of foremost historical interest. Census enumerations were examined not as offering ways to reveal the hidden histories of communities but rather as evidence of state projects in which those in leadership positions attempted to enhance their power over large jurisdictions.7 Rather than depicting enumerators writing down the information provided by householders to questions of general importance, scholars now described how census officials imposed elite conceptions on the rest of society by designing both the questions and acceptable answers. While researchers in the 1960s and 1970s had focused on the “taking” of the census, other scholars in the 1980s and 1990s wrote about the “making” of the census.8 Each enumeration was studied not as evidence of behaviour but rather as evidence of elite political, economic, cultural, and social ideas. And rather than counting responses to census questions, researchers studied the questions themselves as illustrations of the ways in which governments were attempting to increase and solidify their power. In this way, the census came to be associated with “qualitative” research as well as with the “quantitative” research of the earlier new social history.9

13During the 1990s, some scholars remained fully attached to epistemologies associated with quantitative or qualitative, micro or macro, social or cultural approaches. Toward the end of the decade, however, many researchers were attempting to go beyond such dichotomies. Two questions were at the heart of their efforts: how to build on the strengths of both the new social history and of the subsequent cultural history to develop an integrated socio-cultural epistemology, and how to use micro-historical convictions to underpin macro-historical interpretations. The pertinence of these questions arose from a growing conviction that ideas and behaviour had to be studied in holistic ways since they were inherently interrelated in the histories of individuals and groups. Similarly, historians increasingly perceived deep connections between the histories of the “anonymous” and the “famous” with differential flows of influence both up and down, depending upon the historical context including the relevant distributions of power and influence.10

14In the case of the census, by the late 1990s the result was new efforts to examine various enumerations as providing evidence of individual lives that were both linguistically-constructed and materially-based. The census was now seen as enabling insights about diverse perceptions and realities relating to specific communities as well as larger jurisdictions.11 Scholars agreed with the importance of analyzing the enumeration process and census questions as related to elite priorities and perceptions but they also increasingly emphasized that census officials could not simply ask whatever they liked; enumerations depended upon, at least to some extent, a shared grid of understanding about both the questions and the expected answers. Such understanding was not always obtained, of course, but rather than viewing the contradictory and confused results in these cases as indications of the inadequacy of the census as a historical source, scholars began seizing upon such results as opportunities to probe the mentalities of those being enumerated by reading the schedules “against the grain.”

15In the same way, researchers tended to move away from the idea that the census enumerations only provided evidence of the ability of those in power to impose their concepts and to define individuals and groups according to their own preferences. Rather, the census suggested a sense of inadequacy and ignorance among officials as well as a sense of comfort and control.12 In other words, the convictions of both the new social history and the subsequent cultural history seemed warranted. But the first question remained: how could the census be read as providing evidence in keeping with integrated sociocultural approaches to historical change?

16The second question probed connections between the deep complexity of specific times and places and the overall similarity of key historical trajectories related to the expansion of mass schooling, declining fertility, and other features of social, economic, cultural, and political transformation in numerous (though certainly not all) societies. Scholars became increasingly convinced that particular times and places were characterized by both a specificity and a generality within which such specificity had to be contextualized.13 As researchers acquired increased appreciation of the diversity and complexity of everyday life, they also repeatedly concluded that larger forces determined the limits within which the ideas and behaviours of individuals and groups had to be situated. In other words, micro-history and macro-history represented different but interrelated observational levels.14

17One way to pursue the ambition of an integrated socio-cultural and micro-macro approach to historical evidence is to apply the concept of multi-authored sources. In this concept, each source is created by numerous authors in direct and indirect ways. The appropriate analysis of the census, for example, can be seen to depend upon understandings of the multi-layered political, social, economic, and cultural contexts within which enumerations took place. In recent years, scholars have paid greater and greater attention to the challenge of developing such understandings as they have become increasingly convinced of the complexity and diversity of the ways in which censuses were conceptualized, and how specific questions were formulated and responses given. Specifically, the nineteenth and early-twentieth century census in Canada can be analyzed in terms of eight distinct authors:

1. the international community of census officials and advisors who developed approaches within their own jurisdictions by sharing, comparing, and debating their plans and experiences;
2. the politicians who consistently approved the census but who also argued about and influenced its design and operation;
3. the religious leaders who supported, contested, and affected the census in numerous ways;
4. the business and community leaders who sought to profit from the census for their own purposes;
5. the journalists who created and reflected elite and popular interest in the census through considerable coverage especially in enumeration years;
6. the census bureaucracy in Ottawa who prepared for and followed up each enumeration including the editing, revising, and compiling of results;
7. the enumerators who conducted the census;
8. the respondents whose statements affected to varying degrees what was written on enumeration forms.

18Certainly, these authors as well as others contributed to the content of each census in vastly different ways and to significantly different extents depending upon their positions of influence and involvement. And, of course, the visible results of such authoring in the extant census documents are a profound simplification of the input of the diverse influences. By viewing the census as a multi-authored source, an emphasis is placed on the importance of contextualizing the enumeration process and the specific questions and other writing on actual census schedules. Each enumeration resulted from complex interactions of “authors” both directly and indirectly involved in census work. In this view, the administrative history of each census enumeration becomes crucial to the appropriate analysis of the evidence produced in various communities as well as in centralized offices. The important questions range from decisions about when and whom to enumerate to form design, census questions, respondent reaction, and subsequent administrative processes. Research on such questions sheds light on the diverse and competing concepts, definitions, and objectives that are associated with each census enumeration at the various levels from the interactions at specific dwellings to international discussions.15

19Beyond analyzing the roles of the multiple authors of census enumerations is the challenge of interpreting the traces of this authorship on the actual documents. One way to address this challenge is to systematically consider the criteria that appear to underpin each evidentiary trace under examination. In this approach, the ambition is not to evaluate census questions, responses, or the enumeration process in an abstract way but rather to relate this evidence to the criteria used in their creation. Five questions can be posed about each of the groups of authors in this regard:

1. What are the (often) competing and distinct criteria being used by different individuals, groups and institutions in determining the census objectives, enumeration process, questions, responses, and subsequent handling and analysis:
2. Why are these criteria being used and not others?
3. How do the various criteria change over time?
4. Why do these changes in criteria occur and not others?
5. What are the consequences of the use of various criteria for different individuals, groups and institutions?

20By posing questions about the diverse and competing criteria that underpin the multiple authoring of the census, scholars are moving beyond the notion of “right” and “wrong” census questions or responses just as they are abandoning distinctions between “qualitative” and “quantitative” evidence. Viewing sources such as the census as multi-authored makes clear that the key epistemological question is when to take a realist stance rather than how to choose between constructivist and realist perspectives. Similarly, this approach enables analysis of the embeddedness of individual and collective histories within larger contexts that go beyond specific communities.

21Questions about authorship and criteria can be posed in the case of the census to a vast array of sources from the actual enumeration schedules to the archives of administrative history to newspapers and records of political debate. The challenge is to situate and relate each of these sources within their historical setting; in other words, to contextualize them by identifying their interconnections.16 At the same time, it should be emphasized that the ability to interrelate the linguistically-constructed and materially-based character of historical documents will never do justice to the full complexity of the past. Nonetheless, the move toward socio-cultural epistemologies that attempt to contextualize different observational levels promises to build effectively upon the significant research findings that have so considerably enhanced our historical understandings since the 1960s.

22The example of the changing ways in which research using census enumeration schedules has unfolded in recent decades emphasizes that metaphysical and epistemological questions are indeed at the heart of historical debate. Although most scholarly writing reports, compares, and contests specific research findings, it is the use of new concepts, new sources, and new research strategies that changes the larger context within which specific historical studies are undertaken. It is often argued that such innovations come from outside the discipline of History where researchers are characteristically depicted as theoretically weak and methodologically unimaginative. In recent decades, for example, sociologists, anthropologists, philosophers, and literary scholars have all claimed to have had significant impacts on the changing ways in which historians practice their craft. The dominant image of historical debate since the 1960s is of a discipline being driven by external forces beginning with the social sciences and moving to the humanities by the 1980s.

23In contrast, the preceding discussion suggests how surprising research findings in recent decades inspired historians to reconceptualize historical change and to develop new ways of addressing historical questions. Such innovations have certainly taken place in light of developments in other disciplines but the result has been more of an exchange among social scientists and humanists than a borrowing by historians; indeed, the greater influence may be from History to the other human sciences which have become increasingly preoccupied with the importance of change over time.17 One conclusion is that, in order to practice interdisciplinarity, we cannot only exchange concepts and methods but must also study the same sources whenever appropriate; no particular type of evidence is the property of a particular discipline. In this sense, one way to move historical debate forward toward socio-cultural and micro-macro approaches is to encourage interdisciplinary encounters through a focus on the same sources and on the key question: evidence of what?