1The concept of the “argumentative turn” in policy analysis was first introduced by Frank Fischer and John Forester in 1993. They set out a new orientation in policy analysis that represented a shift away from the traditional empirical-analytic approach to problem-solving to include the study of language and argumentation as essential dimensions of theory and analysis in policy-making and planning. The book was instrumental in stimulating a large body of work in policy research both in the United States and Europe over the subsequent decades. Since the publication of the book, the emphasis on argumentation has converged with other developments in the social sciences focused on discourse, deliberation, social constructivism, narration, and interpretative methods. [1]

2In pursuit of an alternative approach to policy inquiry the “argumentative turn” links post-positivist epistemology with social and political theory and the search for a relevant methodology. [2] At the outset, the approach emphasized practical argumentation, policy judgment, frame analysis, narrative storytelling, and rhetorical analysis, among others. [3] From the early 1990s onward, argumentative policy analysis matured into a major strand in the contemporary study of policy-making and policy theory development. As one leading policy theorist put it, this perspective became one of the competing theoretical perspectives. [4]

3Over these years, as Fischer and Gottweis write, the argumentative turn expanded to include work on discourse analysis, deliberation, deliberative democracy, citizen juries, governance, expertise, participatory inquiry, local and tacit knowledge, collaborative planning, the uses and role of media, and interpretive methods, among others. Even though these research orientations are not synonymous, they share the special attention they give to communication and argumentation, in particular the processes of utilizing, mobilizing, and assessing communicative practices in the interpretation and praxis of policy-making and “analysis”. [5] Whereas neo-positivist approaches generally embrace a technically oriented rational approach to policy-making – an attempt to provide unequivocal, value-free answers to the major questions of policy-making – the argumentative approach rejects the idea that policy analysis can be a straightforward application of scientific techniques. [6] Instead of a narrow focus on empirical measurement of inputs and outputs, it takes the policy argument as the starting point of analysis. Without denying the importance of empirical analysis, the argumentative turn seeks to understand the relationship between the empirical and the normative as they are configured in the processes of policy argumentation. It thus concerns itself with the validity of empirical and normative statements, but moves beyond the traditional empirical emphasis to examine the ways in which they are combined and employed in the political process.

4This orientation is especially important for an applied discipline such as policy analysis. In so far as the field exists to serve real-world decision-makers, policy analysis needs to be relevant to those whom it attempts to assist. The argumentative turn, in this regard, seeks to analyze policy to inform the ordinary-language processes of policy argumentation, in particular as reflected in the thought and deliberation of politicians, administrators, and citizens. [7] Rather than imposing scientific frameworks on the processes of argumentation and decision-making, theoretical perspectives generally designed to inform specific academic disciplines, policy analysis thus takes the practical argument as the unit of analysis. It rejects the “rational” assumptions underlying many approaches in policy inquiry and embraces an understanding of human action as intermediated and embedded in symbolically rich social and cultural contexts.

5Recognizing that the policy process is constituted by and mediated through communicative practices, the argumentative turn therefore attempts to understand both the process of policymaking and the analytical activities of policy inquiry on their own terms. Instead of prescribing procedures based on abstract models, the approach labors to understand and reconstruct what policy analysts do when they do it, to understand how their findings and advice are communicated, and how such advice is understood and employed by those who receive it. This requires then close attention to the social construction of the normative – often conflicting – policy frames of those who struggle over power and policy.

6These concerns take on special significance in today’s increasingly turbulent world. Contemporary policy problems facing governments are more uncertain, complex, and often riskier than they were when many of the theories and methods of policy analysis were first advanced. Often poorly defined, such problems have been described as far “messier” than their earlier counterparts – for example, climate change, health, and transportation. [8] These are problems for which clear-cut solutions are missing – especially technical solutions, despite concerted attempts to identify them. In all of these areas, traditional approaches – often technocratic – have proven inadequate or have failed. Indeed, for such messy policy problems, science and scientific knowledge have often compounded problem-solving, becoming themselves sources of uncertainty and ambiguity. They thus generate political conflict rather than help to resolve it. In a disorderly world that is in “generative flux”, research methods that assume a stable reality “out there” waiting to be discovered are of little help and prone to error and misinterpretation. [9]

7The essay that follows explores a number of aspects of the approach, particularly as they pertain to giving policy advice. While discussion of policy expertise and advice-giving in politics is in no way new, the practices of expertise and advice changed over the latter part of the twentieth century. The essay seeks to show that the argumentative turn can be usefully employed to interpret these changing patterns, as well to offer a number of potentially useful prescriptions. Toward this end, the discussion begins with a brief discussion of the long history of thought about political expertise and advice, moving from political philosophy to social science. It then examines the limits of the social scientific approach and presents the new argumentative approach that has emerged as its challenger especially as reflected in the field of policy analysis. In the third part of the essay, the advantages of a post-empiricist argumentative approach to policy inquiry are presented and several practical contributions to policy deliberation based on its tenets are outlined. The discussion closes with an examination of the further post-positivist challenges posed by policy deliberation and the argumentative approach.

From political wisdom to policy analysis

8The role of political wisdom and policy advice – and argumentation about both – are scarcely new topics in government and political science. [10] The importance of politically relevant counsel, and who should offer it, in the realm of politics can be traced back to the earliest discussions on statecraft. In Western political thought, it is a prominent theme in the writing of Plato about the need for knowledgeable advisors, in Machiavelli’s discussion in The Prince, Francis Bacon’s New Atlantis, St. Simon and Auguste Comte’s theory of technocracy, Thorstein Veblen’s emphasis on the engineers in the price system, the “Brain Trust” of Franklin Roosevelt’s New Deal, the writings of the policy intellectuals of the Great Society of the 1960s, and modern-day think tanks since Reagan and Thatcher. [11] During the past three decades, moreover, this need for expertise has been a central theme in the literature of policy analysis.

9Today, the topic of policy expertise and advice-giving has also attracted new interest, in a number of European countries in particular. There has been considerable political and academic discussion about the role of policy advice from unelected experts to whom top politicians have turned for ideas, even in some cases for defining components of their political agenda. In Germany, for example, former Chancellor Gerhard Schröder engaged a group of external experts, many of whom were from the business community, to come up with a new design for overhauling the country’s social welfare system. This expert-oriented “Hartz Commission”, named after the former personnel manager of Volkswagen, developed a controversial program that was more or less adopted outright by the Social Democratic government, despite much public opposition and dissatisfaction with the onerous burdens it was to subsequently impose on the unemployed and the poor. Described by some as a new pattern of governance, this approach has been dubbed as the politics of the “Raterrepublic”, referring ironically to the earlier Soviet model of government by councils (which had a brief life of its own in Munich after the Second World War). The concern has emerged as well in discussions of constitutional development in the European Union, seen to be carried out by experts far too removed from the citizens of the countries that the proposed constitution would bind together. It is a discussion all the more intensified by the rejection of the proposed document by the French and the Dutch citizenries, more or less on the grounds that it was an idea put forward by distant technocrats with little knowledge of, or interest in, everyday European life.

10Historically, a central theme running through the literature on the subject has focused on minimizing or replacing political argumentation with more rigorous forms of thought. Within this tradition, we also find a shift from wise men, such as Plato’s philosopher-king to scientifically grounded advice, emerging first in Bacon’s New Atlantis, a model utopian society planned and administered by scientists. [12] Somewhat later, St. Simon and Auguste Comte spelled out a theory of technocratic governance that more specifically featured social scientists. [13] Underlying these later writings was a “positivist” epistemological challenge to earlier forms of philosophical knowledge, characterized as a form of “negativism”. Rather than hard core empirical analysis upon which testable theory for building a better theory of society could be founded, such earlier modes of inquiry were seen to rest on speculation and critique. They offered no basis for “social progress”.

11The emphasis on positivism and technocratic policy-making in government and politics took on practical dimensions in the early twentieth century, especially in the United States. Much of continental Europe, by contrast, remained wedded to the more philosophic Germanic understanding of knowledge (extending back to Kant and Hegel) and thus adverse to the narrower techno-empirical orientation of the technocratic movement. What is more, the focus on policy analysis emerged more as an American than a European project, in part because of a less philosophical, more pragmatic approach to knowledge and its uses. Indeed, these differences continued to separate the two communities well into the later part of the century.

12The story of technocracy in the United States has its roots in the “Progressive Era” around the turn of the twentieth century. Leading Progressive writer, Herbert Croly, influenced by the writings of the French sociologist Auguste Comte, advocated the role of expertise in government as a reform strategy for addressing the economic and social problems confronting a rapidly industrializing nation. Indeed, he advocated the use of the principles of Taylorism and “scientific management” as the basis for governmental reform. [14] In line with the enthusiasm of the era, leading University of Chicago sociologist Lester Ward even argued that legislators should have training in the social sciences to be qualified for office, if not be social scientists themselves. [15] It was, in short, science rather than political deliberation that counted in the Progressive vision of the good society.

13Drawing support from the theories of positivism emerging at the time, the movement constituted a “mentality” as much as it did a set of scientific principles. [16] The resulting empirical or “behaviorial” approach to the newly developing social sciences was at first closely related to the problem-oriented social reform movements of the 1920s and 30s. Indeed, throughout this period, the objective of writers such as Walter Lippman was to replace the “irrational public” and corrupt political parties with technocratic administration based on the finding of the new sciences of governance. [17]

14The influence of this “Progressive” creed is attested to by the election of two Progressive presidents in the US, one a Republican – Theodore Roosevelt – and the other a Democrat – Woodrow Wilson. These elections and the politics surrounding them reflected the would-be “value-free” nature of scientific management and the “non-ideological” approach to good government through expertise that it prescribed. Indeed, Wilson, himself a political scientist (and often considered one of the founders of the American discipline of public administration), called for scientifically explicating the efficient practices of Prussian bureaucracy and applying them – independently of culture or context – to American government. In the process, this new field of inquiry and its practices would replace the traditional emphasis on legal-rational bureaucratic authority with social scientific principles of organization and management. [18]

15Technocracy received renewed support in the 1920s from Thorstein Veblen and his call to replace the capitalist price system with the decisions of efficient engineers. [19] The goal for Veblen was to eliminate waste and encourage efficiency in the American economy through methods and practices of experts. Pinning his hopes on the political possibilities posed by a newly emerging professional class, he called on the “absentee owners” of corporate America to transfer their power to reform-minded technocrats and workers. Throughout this early work, the term “technocracy” was seen as a positive new direction for government. The message was straightforward : replace the talk of politicians with the analysis of the experts. Some technocratic thought was influenced by emerging socialist governments emphasizing planning, although it was often difficult for this enthusiasm to be explicitly expressed, especially when it related to the newly formed Soviet Union.

16A major expansion of the Progressive ideas came with the New Deal “Brain Trust” of the 1930s, also adding a new dimension of expert advice. Particularly influential were the writings of John Maynard Keynes, holding out the possibility of technically managing the economy. [20] This period saw a large-scale influx of economists and social scientists in Washington, employed in policy planning and related activities throughout the growing administrative state. It also featured a change in the politics of the policy process, or what came to be called the New Deal “Liberal Reform Strategy”. [21] The political strategy brought experts and politicians together, through so-called “Blue Ribbon” advisory commissions, to identify and analyze contemporary social and economic problems. The reports of these commissions were then turned into central components of the liberal reform agenda by party politicians, subsequently offered to voters in electoral campaigns. In the European context, it is interesting to note that in the years preceding the development of this strategy, Max Weber advanced the idea of the expert-oriented Enquete Commission in Germany for what would be the Weimar Republic, which was later to take on some of the same features.

17In the postwar period the emphasis on the role of empirical social science in political advisement took full form in the development of the field of policy analysis, from Harold Lasswell to Aaron Wildavsky. Although policy analysis emerged in less grandiose terms than Lasswell had envisioned it, including his recognition of the need to include more than empirical investigation, the field and its practices that emerge in the 1960s was dominated by empiricist positivist-based methods. [22]

Technocratic expertise and the rise of counter-expertise : policy argumentation

18In the 1960s, during the Democratic Party’s “Great Society” reform program, the same expert-oriented liberal reform strategy of earlier decades was found to be operating inside the White House. Policy experts gave birth to an outpouring of poverty programs, under both Presidents Kennedy and Johnson, as well as the technocratic planning and execution of the Vietnam War, directed by the technocratic Secretary of Defense, Robert McNamara. [23] Enjoying easy access to President Johnson, these policy professionals produced a nearly unprecedented array of programs and bills that gave shape to the liberal administration’s War on Poverty. Daniel Moynihan, a social scientist in the administration, called it the “professionalization of reform”. [24] The leading presidential historian, Theodore White referred to the period as “the golden age of the ‘policy intellectuals’ “. [25] For White, it was nothing less than a new system of power in American politics. Acting in concert with political leaders in both the White House and Congress, these new policy advisors were the “driving-wheels” of the Great Society. As a new generation of policy experts with special problem-solving skills, they sought to shape the country’s defenses, redesign the cities, put an end to poverty, reform the schools, and more. They represented a bridge across the gap between government and the producers of ideas. The White House served as “a transmission belt”, packaging and processing scholars’ ideas to be sold to Congress as programs.

19Of necessity, the scope of these efforts gave rise to new tools for program management and analysis, including the systematic development of “policy analysis”. Universities, more than enthusiastic in response to the turn to policy expertise, rushed in to take advantage of an outpouring of Congressional monies for both the development and evaluation of such public programs. This led to countless doctoral dissertations, and in turn, extensive theoretical and methodological discussions that became the foundation of the emerging field. Designed to bring social science out of the “ivory tower”, policy analysis was to confer “social relevance” through problem-solving and informed advice. Public policy rapidly become the most fashionable disciplinary specialization in political science. Indeed, it became the “in” topic in the social sciences.

20This liberal policy strategy interested many foreign observers, as both a positive and necessary development. In Germany, for example, leading social scientists such as Fritz Scharpf and Renate Mayntz adopted and promoted these ideas and policy scientific techniques as part of a new steering apparatus for the state. [26] In Denmark and Sweden scholars similarly turned to policy analysis as a tool for managing and controlling the social welfare state, emphasizing in particular the problems of policy implementation and evaluation. [27] But the positive benefits didn’t emerge so rapidly. By the late 1970s, the practice of policy analysis appeared to many as a disappointment : if the goal was to solve social problems, there was little to show in the way of a clear-cut payoff. This perception emerged in Europe as well as the United States.

21The result was more than a measure of academic soul-searching and gave rise to two responses. One was methodological ; the problem was first seen to be a lack of empirical rigor and the need for better methodological tools and theories. Systems analysis and computers were held out by many as the way to rectify the limited success of the endeavor. But others started to rethink policy analysis more in terms of “enlightenment” than problem-solving per se. This view recognized that policy was as much about the discussion of values as it was about facts, if not at times more so.

22And the late 1970s and 80s were just such a period, as values became a central political theme of the neo-conservative governments that emerged in both the US and Europe. The neo-conservative theorists of the time took the development of policy advice in a different direction. For them, the liberal policy experts constituted a “new class” that imposed its own liberal agenda on citizens who hadn’t asked for their advice. [28] Instead of rejecting the strategy, though, it was redirected in the US and Britain to what came to be called the “War of Ideas”, carried out by a network of high-visibility, well-funded conservative think tank experts. [29] Emphasizing deregulation, privatization, and supply-side economics, the “counter intelligentsias” in these think tanks began to perform two primary functions. One was to organize regularized discussions between conservative economic and political leaders and leading conservative academics. And the other involved helping to shape the conservative policy agenda through these interactions, an outcome of which in the US was to bring Ronald Reagan to the White House. An unmistakable success, the effort resulted in the conservative political climate that persists to this day. It was a process also to be found in Britain in the 1970s, where it played an important role in helping to bring Margaret Thatcher to power. It was also copied in Germany. [30]

23To the degree that conservatives took an early interest in formal policy analysis, they turned the focus from policy development to implementation and evaluation, as post-hoc evaluation research could be conveniently employed to document the failures of liberal policy programs, even when that was not its intended purpose. But the main thrust of think tanks was the politicization of policy advice, which tended to undercut the technocratic legitimation of policy science. As ideological “counter-argumentation” returned in full public view, techno-scientific expertise was increasingly demystified. Softening their claims to non-partisianship – or dropping it altogether – many policy experts began to operate in an argumentatively contentious adversarial style that openly featured their political biases. [31] Some even described themselves as “hired guns”, likening the role more to that of the lawyer with his or her client, rather than the scientific expert giving objective advice.

24Strictly speaking, though, the American-style of policy analysis, focused on problem-solving and normative argumentation, has remained suspect in many European countries. Indeed, in Britain, and even more so in Germany, the discipline has only been slowly embraced. In France, where the focus has tended to be on the sociology of the policy-making process, policy analysis as problem-solving has also been slow to develop. In more recent years, especially among younger European scholars, there has been increasing interest in constructivist policy perspectives. [32] Among other indications, evidence for this can be found in the growing attendance of the Annual Conference on Interpretive Policy Analysis, in particular the more recent conferences in Grenoble (2010) and Cardiff (2011).

The argumentative turn as epistemology

25Whereas policy analysis had emerged enshrouded in a neo-positivist conception of knowledge – and the “overdetermined” technocratic understanding of policymaking long associated with it – the new emphasis on adversarial argumentation unwittingly helped to open the door to a very different kind of post-positivist orientation based on a social constructivist conception of knowledge, a dialectical model of argumentation, and the interpretive methods common to both. [33] If the new politics suggested the need for a different understanding of knowledge, the influential works were those of Habermas (both his critique of scientism and the theory of communicative competence), Foucault (his theory of disciplinary control and discursive power), social constructivism (e.g., Kuhn in the US, Woolgar in the UK, Latour in France, and Knorr-Cetina in Germany), the practical discourse and rhetoric of Toulmin and McClosky, and the pragmatism of Dewey and Pierce. Theorists working within these theoretical perspectives have interacted in the literature in various ways over the past three decades to offer more critical ways of looking at policy expertise, the processes of analytic inquiry, policy argumentation, and public deliberation. [34]

26Some of this work has involved a micro focus on particular communicative practices ; other efforts have taken a larger theoretical perspective examining discourse and public deliberation. None of it has evolved in a linear fashion, but taken together it has offered a wealth of new ideas and ways of thinking that were brought to bear on policy analysis. Most fundamentally, these perspectives recognize that social reality is constituted through language rather than merely mirrored back by it. All of them appreciate how discourse and practical rhetoric depict and characterize reality, including some parts of it and excluding others.

27As an interpretive orientation, the argumentative turn in policy analysis understands social reality and the empirical observations of it to only exist in relation to intellectual constructs used for thinking about them. In so far as these constructs are grounded in values that shape our perceptions of reality, the researcher’s findings are not a report of that which is “out there”, as neo-positivism would have it, but rather part of a process that creates that version of reality. Knowledge in the social world, then, “is a human construction never certifiable as ultimately true but problematic and ever changing”. [35] In this view, theory and knowledge can never be fully probed. Not only can knowledge never be value-free, there can be no definitive criteria for choosing one theory over another. Furthermore, knowledge in and of the social world is the product of a dialogic logic that understands knowledge to be the result of a confrontation between differing interpretations, the outcome of which can be a constructive synthesis leading to a new intersubjective understanding. [36] This understanding, then, is the basis for the consensus that comes to be called knowledge, at least until some point in time when the consensus is successfully challenged.

28Unlike the neo-positivist conception of knowledge, such consensus does not rest on a reality independent of those who shape and share it. In so far as the possibility of further confrontations with differing points of view always remains open, the construction of a consensus is never fully finished or complete. There can be progress in the production of consensus, but such knowledge can never by proved in the standard sense of the term. The process by which the production of such a consensus is shaped becomes the special interest of social constructivist investigation.

29The social constructivist perspective thus describes a world that is richer and more complex than the empiricist theories offered to explain it. Rather than break reality down into variables to be measured, it strives to coherently incorporate the multiplicity of theoretical perspectives and explanations that bear on a particular event or phenomenon. In Toulmin’s words, post-empiricist coherence theory brings to bear “the range and scope of interpretive standpoints that have won a place”. [37] Alongside quantitative analysis, the orientation includes the historical, comparative, philosophical, and phenomenological perspectives. Quantitative empirical research, in the process, loses its privileged claim among modes of inquiry. Although it remains an important component of theory construction, it no longer offers the crucial test.

30Whereas scientific policy-making has always been hostile to the “irrationalities” of political argumentation, those developing the “argumentative turn” have posited the argument itself as the basic unit of real-world policy-making and analysis. Policy becomes understood here as “crafted argument”. [38] The goal of such post-positivist policy analysis shifts then to improving arguments through the illumination of the “contentious dimensions of policy questions, explaining the intractability of policy debates, identifying the defects of supporting arguments, and elucidating the political implications of contending prescriptions”. [39]

Policy analysis as deliberative craft : integrating empirical and normative inquiry

31Policy analysis, from this argumentative perspective, is better defined as a “craft” than as a science in the conventional understanding of the term. In this view, the professional policy analysts employ knowledge and skills acquired as much or more through imitation, experience, and practice than formal methodological training. The task of the policy analysis depends, as Majone puts it, “more on “knowing how” than “knowing that”. [40] Rather than a purely logical activity, following the social constructivist perspective, such policy-analytic work is a social process. The repertoire of craft knowledge and skills exercised by the analyst constitutes procedures, conventions, and judgments that combine social, institutional, and personal factors. In deciding whether specific data is of acceptable quality the policy analyst “applies standards that derive from his own experience but also reflect the professional norms of teachers and colleagues, as well as culturally and institutionally determined criteria of adequacy”. [41]

32Much of the experimental knowledge on which such judgments rest is often as tacit as it is explicit. Thus successful practice, as in the case of traditional crafts, depends on a highly personal knowledge on the part of the craftsperson of tools, materials, and tasks. Rather than dealing with materials such as stone or wood, the policy analyst as craftsperson works with technical tools, data, concepts, and theories to structure evidence and arguments that support particular conclusions. Although the basic principles and precepts of the craft can never be fully spelled out, especially those that constitute a form of tacit knowledge based on experience, the good craftsperson can generally determine good from bad work. [42] In order to properly understand and appreciate these craft-oriented aspects of policy analysis and to be able to competently judge the quality of the end product, the analyst needs to learn how to explicate and explore the microstructures of a policy argument. Such exploration of the activities of the policy analyst would be of little more than academic interest if it were enough simply to examine the policy-analytic conclusions against real-world outcomes. But complexity and uncertainty make it necessary to supplement outcome criteria with process criteria oriented considerations. In the final analysis, a policy analysis has to be judged in terms of the nature of the problem, the context in which it is situated, the use of appropriate methods, the degree of uncertainty surrounding the related issues, and the like.

33In this perspective the focus shifts to the policy argument and its features. In this regard, Majone suggests that the structure of a policy argument can be understood as a complex blend of factual statements, interpretations, opinion, and evaluation. [43] The function of the argument itself is to provide the links that connect the relevant data and information to the conclusions of an analysis. Beyond connecting data and information, from the deliberative perspective, the policy argument also has to include and clarify the normative dimensions that intervene between findings and conclusions. The extension of this approach needs to establish the nature of the interconnections among the empirical data, normative assumptions that structure our understandings of the social world, the interpretive judgments involved in the data-collection process, the particular circumstances of a situational context (in which the findings are generated or the prescriptions applied), and the specific conclusions. [44] The scientific acceptability of the conclusions depends ultimately on the full range of interconnections, not just the empirical findings.

34Whereas empiricists see their approach as more rigorous and therefore superior to less empirical, less deductive methods, this model of policy argumentation actually makes the task more demanding and complex. [45] Not only does it encompass the logic of empirical falsification, it includes the equally sophisticated normative questions within which it operates. The researcher still collects the data, but now has to situate or include them in the interpretive framework that gives them meaning. No longer is it possible to contend that such normative investigations can be ignored, as if they somehow relate to another field of inquiry. This involves a multi-methodological framework for integrating these component parts. In Evaluating Public Policy, I have outlined such a framework based on insights drawn from the informal logic of practical reason. In this model, empirical findings are normatively explored in an interrelated set of normative contexts, from local, to societal and ideological, each with their own logical requirements.

35Policy analysis, as deliberative craft, thus seeks to bring a wider range of contextually sensitive empirical and normative criteria to bear on the argument under investigation. As Hawkesworth explains, [46] the reasons provided in support of alternative policy arguments organize evidence, marshal data, apply explanatory criteria, address multiple levels of deliberation, and employ various strategies of presentation. But the reasons given to support one theory over another seldom, if ever, offer definitive proof of the validity of a competing alternative. Through the processes of deliberation and debate, a consensus emerges among particular researchers concerning what will be taken as valid explanation. Although the choice is sustained by factual and normative reasons that can be articulated and advanced as support for the inadequacy of alternative interpretations, it is the practical judgment of the community of researchers and not the data themselves that establishes the accepted explanation. To be sure, the informal logic of practical reason cannot guarantee the eternal verity of particular conclusions, but the social rationality of the process is far from haphazard or illogical.

36From this perspective, the role of the deliberative expert is that of an interpretive mediator operating between the available analytical frameworks of social science and competing local perspectives, including the local situational knowledge of the relevant citizens. The dialectical exchange can be likened to a conversation in which the understandings of both the policy analyst and the citizens are extended through discursive interactions. Thus, exchanges among policy experts, citizens, and policy-makers are restructured as a discussion with multiple voices. Recognizing the epistemic importance of these deliberative relationships, as well as the importance of the citizen’s perspective, Bogason advocates a bottom-up approach to policy-making and offers an approach for bringing together the relevant parties. [47]

37The task is to enlarge the range of political possibilities through a greater recognition of the dialectics of contestation and its implications for expanding the spectrum choice, although the judgment of the analyst can never substitute for the choices of the political community. The approach thus requires a participatory practice of democratic deliberation. “By encouraging policy-makers and citizens to engage in informed deliberation upon the options confronting the political community”, as Hawkesworth has put it, such inquiry “can contribute to an understanding of politics which entails collective decision-making about a determinate way of life”. [48]

38The argumentative turn also emphasizes the productive capacities of deliberation – namely, its ability to generate ways of thinking and seeing that open new possibilities for problem-solving and action. Or in Habermas’s language, it focuses on “communicative power”. [49] Well-designed argumentative processes are seen to facilitate communicative competencies and thus citizen learning, which draws attention to the democratic potential of policy analysis. It thus extends Lasswell’s earlier call for a “policy sciences of democracy”.

Deliberative policy inquiry as dialectical argumentation : theory and methods

39From the kind of work outlined above, we can begin more practically to construct new ways of thinking about and designing alternative argumentation-oriented inquiry systems for policy deliberation and decision-making. The starting point for such a system is the adoption of a dialectical understanding of inquiry and the recognition that the most fundamental aspect of inquiry is the identification of ways of seeing. [50] While the technical aspects of problem-solving still have an important role to play in policy formulation and evaluation, even more significant from the perspective of dialectical argumentation is the way in which the actors see a particular problem and thus shape the issue agenda. In contemporary methodological vernacular, this is discussed in terms of problem-framing. The process of framing is the basis for not only identifying the problem, but also for defining it. In this sense, framing predetermines the direction and nature of the technical analysis that might follow. From the argumentative perspective, then, inquiry has the broader function of enlightenment than the narrower problem-solving orientation traditionally associated with policy science.

40The insight is not altogether new to the field of policy studies. Although working in a different context, Carol Weiss (1990) advanced the “enlightenment function” as the goal for policy analysis in an earlier critique of traditional policy analysis methods. [51] In an effort to explain the failure of policy decision-makers to readily adopt the analysis of policy analysis, she argued that the contribution of much of the work was misunderstood. While it did not offer unquestioned policy solutions, namely the conventional goal of the discipline, it did supply enlightening perspectives to the relevant decision-makers. And in many ways, this contribution to the broader processes of deliberation and argument counts as much or more than the narrower focus on achieving an administrative goal. This important insight makes clear that the limitations and failures of rigorous scientific policy research in no way renders the enterprise useless. Although Weiss did not draw it out per se, her emphasis on enlightenment also implied alternative ways of seeing a problem, potentially introducing a phenomenological understanding of social and political reality. Based on this and other similar insights, deliberative policy analysis shifts the focus to the processes of policy argumentation. How does argumentation work ? And how can it be better informed ?

41Once we stress different ways of seeing as the essential contribution of the enlightenment function, it becomes easier to recognize the dialectic role of conflict as opposed to empirical consensus in post-positivist inquiry. It is not that consensus is unimportant, but rather that it tends to operate more within normative frameworks than it works to open them up. This in turn facilitates the technical orientation that continues to dominate policy analysis. From the dialectical perspective, it is the clash of ideas that leads people to a deeper and potentially more enduring consensus, although the path to such consensus is generally longer and more time consuming. But such argumentative conflict takes the inquirers beyond an easy consensus based on conventional ways of viewing things, typically embedded in a given set of power relations and their discursive constructs.

42A dialectical/argumentative approach, then, begins with the normative rather than the empirical task. Instead of fitting the norms and values into the empirical framework, the task is to test empirical findings within normative frameworks. In this view, normative-based analysis can be facilitated by an organized deliberation among competing normative positions. Designed to both identify potential conflicts and create consensus, the model emphasizes the interactive and productive role of communication in cognitive processes. Unlike the more abstract or academic modes of reason removed from the real world, the power of critical judgment depends on potential agreement with others. In fact, such judgment relies on, even anticipates, communication with others.

43In such a scheme, policy analysts and decision-makers each take on the assignment of preparing arguments for and against particular policy positions. In the process, they spell out their positions in the deliberation and leave the task of taking apart the arguments of those on the opposing side, including the alternative empirical evidence used to support them. Such policy argumentation begins with the realization that the policy analysts and decision-makers do not have solid answers to the questions to be deliberated or even unambiguous methods for obtaining the answers. They seek to organize the established data and fit it into the normative frameworks that underline and support their own arguments. Each then confronts the other with counter-proposals, comparing the underlying assumptions and findings being used, both normative and empirical. The grounds or criteria for accepting or rejecting a proposal is the same grounds for accepting or rejecting a counter-proposal and must be based on precisely the same data. The goal is to synthesize the competing positions into a meaningful perspective capable of generating a workable consensus. The problem posed by the absence of appropriate evaluative criteria can be mitigated by designing rational procedures to govern the formal communication between the various points of view that bear on the decision-making process. Toward this end, some have suggested that legal courtroom procedures can offer insights into the development of rules for governing such discursive policy deliberations. [52]

44In so far as the argumentative model reverses the analytical process – from the standard task of fitting qualitative data about norms and values into an empirical model through quantification to the process of judging the empirical data against a normative framework – the central locus of the judgmental process is expanded from the expert community to include the inquirers of the practical world of action. In this model, each participant would cite not only what he or she take to be the relevant causal relationships, but also the norms, values, and circumstances to support or justify a particular decision. The final outcome of evaluative inquiry is determined by the giving of reasons and the assessment of practical arguments rather than technical demonstration and verification. As in interpretive explanation generally, the valid interpretation is the one that survives the widest range of criticisms and objections. Such interpretive evaluations, as practical arguments, connect policy options and situations by illuminating the features of those situations that provide ground for policy decisions.

45In this scheme, the formalized deliberation is itself seen as the most instructive part of the analytical process. The technique is designed to clarify the underlying goals and norms that give shape to competing positions, and enables qualitative judgment to be exercised in as unhampered a way as possible. The free exercise of normative judgment, released from the restrictions of the formal policy model, increases the chance of developing a synthesis of normative perspectives that can provide a legitimate and acceptable basis for decisions and actions based on the strongest possible argument. Even if analysts cannot agree, the argumentative approach provides a procedure for probing the normative implications of recommendations and for indicating potentially consensual conclusions that can offer productive ways to move forward. In the process, it also makes clear the basic points of dissensus that stand in the path of agreement.

46A principal advantage of this approach is that it better informs the policy decision-making process than does traditional empirical policy analysis. It mirrors, in particular, the way real-world policy deliberation actually works. In politics, politicians and policy decision-makers advance proposals about what to do based on normative arguments. Empirical questions seldom drive the process in politics and policy. They come into play, to be sure. But they mainly do so when there are reasons to doubt factual aspects of the argument.

47And here again there have been policy-relevant moves in this direction, mainly emerging in organization theory. The path-breaking but now mainly neglected work of C. West Churchman and his colleagues on dialectical inquiry systems for managerial decision-making is based on a dialectic understanding of deliberation. [53] Unfortunately, this important line of theoretical work has gotten lost in the return to traditional decision-practices in more conservative times, such as rational choice theory and “evidence-based policy-making”. The basic methodological observations are also found in work by George and Porter dealing with the task of organizing expertise advice in policy areas such as national security and economic policy-making at the presidential level. [54] The basic assumption underlying what George calls “multiple advocacy” is that a competition of ideas and viewpoints, rather than reliance on analysis and recommendations from advisors who share the perspectives of the policymakers, is the best method for developing security policy. As such, multiple advocacy is described as a process of debate and persuasion designed to expose the policy-maker systematically to competing arguments made by the advocates themselves. Through the efforts of an “honest broker”, his approach attempts to ensure that all interested parties are genuinely represented in the adversarial process, and that the debate is structured and balanced.

48These argumentative approaches represent important steps toward the development of a dynamic methodology designed to facilitate complex dialectical exploration of facts and values, empirical and normative inquiry, throughout the policy decision process. [55] At minimum, the work of these theorists goes a considerable distance toward removing the ideological mask that has often shrouded policy analysis and the resultant expertise drawn from it. Like any step forward, however, it only brings us to the next set of hurdles. One of them has to do with the scope of participation. Although Churchman and his followers focus rather narrowly on managerial decision-making, such a communications approach would not need to be confined to the interactions between organizational policy-makers and policy analysts. Ideally, it could be extended to the range of differing interests and political viewpoints drawn from the policy environment. Importantly, it is a consideration taken up by a major scientific organization in the United States, the National Research Council, an arm of the US Academy of Science.

Extending deliberation : citizens and stakeholders

49Over the past decade or more there has been an increasing recognition of the need for deliberative interaction with the broader public in policy-analytic inquiry. An important sign of the extent of this development has been the National Research Council’s call for an “analytic-deliberative” method for bringing together a wider range of stakeholders. Although the Council focuses on policy issues related to science, technology, and the environment, its proposals are broadly applicable to policy analysis more generally. Specifically, the Council has taken the position that coping with technological and environmental risks requires “a broad understanding of the relevant losses, harms, or consequences to the interested and affected parties, including what the affected parties believe the risks to be in particular situations”. [56] It thus becomes necessary to find ways to incorporate the perspectives of these groups into the processes of policy analysis and decision-making.

50Toward this end, the organizations are counseled to make special efforts to ensure that the interested and affected parties find reasonable the basic analytic assumptions about risk-generating processes and risk estimate methods. Although recognized to often be time-consuming and cumbersome, at least in the near term the Council argues that it is wiser “to err on the side of too-broad rather than too-narrow participation”. [57] Organizations are advised “to seriously assess the need for involvement of the spectrum of interested and affected parties at each step, with a presumption in favor of involvement”. [58]

51The methodological challenge is described as the need to develop “an analytic-deliberative method” [59] capable of bringing together citizens and experts. Such a deliberative method is required to guide a participatory process capable of “broadly formulating the decision problem, guiding analysis to improve decision participants” understanding, seeking the meaning of analytic findings and uncertainties, and improving the ability of interested and affected parties to participate effectively in the risk decision process’. [60] The process must have an appropriately diverse range of participation representing the spectrum of interested and affected parties, of decision-makers, and of specialists in risk analysis at each stage of the process. Most important is the need for participation in the early stages of problem formulation.

52In this view, analysis and deliberation are presented as complementary approaches to gaining knowledge about the world, forming understanding on the basis of knowledge, and reaching agreement among participants. Whereas analysis “uses rigorous, replicable methods, evaluated under the agreed protocols of an expert community”, [61] deliberation is a process “in which participants discuss, ponder, exchange observations and views, reflect and attempt to persuade each other”. [62] Deliberation, moreover, doesn’t just come at the end. It is important at each step of the process that informs decisions, from deciding which problems to analyze to how to describe scientific uncertainty and negotiate disagreement. Such structured deliberation contributes to sound analysis by adding knowledge and perspectives that improve understanding, and contributes to the acceptability of problem characterization by addressing potentially sensitive procedural concerns. As the Council puts it, “deliberation frames analysis, analysis informs deliberation, and the process benefits from the feedback between the two”. [63]

53For organizing such a deliberative process, the Council takes the additional step of advocating that organizations reach out with technical assistance to unorganized groups, or those inexperienced in matters of risk analysis and regulatory policy. As the Council puts it, “if some parties that are unorganized, inexperienced in regulatory policy, or unfamiliar with [the] related science are particularly at risk […] it is worthwhile for responsible organizations to arrange for technical assistance to be provided to them from sources that they trust”. [64] In this regard, the Council suggests that experts must at times assume the role of facilitators.

54Such deliberation cannot, to be sure, be expected to end all controversy. It does not guarantee that policy decision-makers will attend to the results of such deliberation, or prevent disappointed participants from attempting to delay or walk away from the outcomes. Controversies, in this view, are seen as constructive in helping to identify weak points from which scientific expertise can benefit rather than merely as impediments in the path of expert decision-making. Not only do controversies encourage in-depth analysis to identify and explicate the social implications of a policy solution, they can also force partly conflicting assessments of programs and policies to the surface that can then be further articulated and consolidated in the course of a controversy. In this view, the proper function of a controversy is the identification and evaluation of potential problems ; that is, as a complement to conventional methods of policy analysis.

55Examined against the Council’s earlier technocratic beginnings in risk analysis and management, this newer emphasis on public-oriented deliberation can only be judged as an impressive advance. Although technocratic practices still remain dominant, the fact that many of the contemporary methods in technology policy analysis and risk assessment were initially influenced by the Council itself holds out hope for change, even if it only comes slowly and reluctantly. The acknowledgement of the centrality of participation and deliberation by this prestigious body should not be underestimated. In this regard, the shift of the Council from its earlier technocratic perspective to a deliberative-analytical model is important documentation of a move toward an argumentative approach to public policy.

56But again the step forward brings us to the next issue. From a post-positivist constructivist perspective, the question turns more specifically to the divide between empirical and normative deliberation. Must analytic work and normative deliberation remain separate, complementary processes ? The challenge of the deliberative approach is to bring them into closer interaction with one another.

Policy epistemics : concluding perspective

57Whereas the “analytic-deliberative” approach advanced by the National Research Council remains attached to conventional notions about science, in particular the epistemic division between the empirical and the normative inquiry, the post-positivist perspective seeks to move beyond this divide by opening up to critical scrutiny the practices of science itself. Of essential importance here is the fact that the normative elements lodged in the construction of empirical policy research rest on interpretive judgments and need to be made accessible for examination and discussion. Recognizing that the social meanings underlying policy research are always interpreted in a particular sociopolitical context – whether the context of an expert community, a particular social group, or society more generally – a fully developed post-positivist deliberative approach focuses on the ways such research and its findings are themselves built upon normative social assumptions that, in turn, have implications for political decision-making. That is, they are embedded in the very understandings of the objects and relationships that policy science investigates. Indeed, the very construction of the empirical object to be measured is at times at stake.

58For this reason, empirical policy science cannot exist independently of normative constructions. While introducing deliberation as a complement to the analytic process is an important advance over a narrow technical orientation, from the post-positivist perspective it can only be understood as a platform from which the next step can be taken. Beyond a complementary approach, deliberation has to be moved into the analytical processes as well. Beyond understanding empirical and normative inquiry as separate activities that can potentially inform one another, they need to be seen as a continuous process of inquiry along a deliberative spectrum ranging from the technical to the normative. As deliberation and interpretative judgment occur in both empirical and normative research, they need to be approached as two dimensions along a continuum.

59It is important to concede that post-positivist deliberative policy inquiry enters here relatively uncharted territories. From a range of practical experiments, such as consensus conferences and citizen juries, it is evident that citizens are more capable of participating in deliberative processes than generally recognized. But questions about the extent of participation, as well as when and where it might be appropriate, pose complicated questions and thus need to assume a central place on the research agenda. Particularly challenging is the question of the degree to which citizens can actually engage in the analytic aspects of inquiry. Various research projects demonstrate that they can participate in at least parts of it, but where should we draw the line ? [65] Reconceptualizing empirical and normative inquiry along an interpretive continuum offers an approach for pursuing the answers to these questions. Elsewhere I have proposed that such investigation be taken up as a component of a new research specialization called “policy epistemics”. [66]

60The development of policy epistemics would draw extensively on work in social constructivism, the newer interpretive sociology of science and the post-Kuhnian work in the philosophy of science. However, where the constructivist approaches in sociology and philosophy focus on the conduct of science, policy epistemics and the turn to arguments extends the perspective to the goals of an applied social science – policy analysis – and the task of giving advice to decision-makers. The orientation, as such, stands between professional policy research and action-oriented policy-making. It thus explores the ways in which research speaks to different normative perspectives in the world of action – what kinds of knowledge(s) are relevant to particular situations, what are the normative implications of particular empirical findings, how can empirical policy findings and normative perspectives be brought together in a deliberative process, and the like ? These are questions which draw on a constructivist perspective but, at the same time, require different practical modes of thought and deliberation.

61Policy epistemics, then, brings together the relevant work in interpretive policy analysis, deliberative experimentation, discursive practices, and policy narration to explore ways in which the empirical/normative divide can be opened up to facilitate a closer, more meaningful deliberation among citizens and experts. As such, policy epistemics focuses on the ways people communicate across differences, the flow and transformation of ideas across borders of different fields, how different professionals groups and local communities see and inquire differently, and the ways in which differences become disputes. Following the lead of Willard, it takes the “field of argument” as a unit of analysis. [67] As polemical discussion, argumentation is the medium through which people – citizens, scientists, and decision-makers – maintain, relate, adapt, transform, and disregard contentions and background consensus.

62Whereas traditional policy analysis has focused on advancing and assessing technical solutions, policy epistemics investigates the way interpretive judgments work in the production and distribution of knowledge. In particular, it would examine the social assumptions embedded in research designs, the specific relationships of different types of information to decision-making, the uses of information, the different ways arguments move across different disciplines and discourses, the translation of knowledge from one community to another, and the interrelationships between discourses and institutions.

63Elaborating our understanding of the epistemic dynamics of public controversies would allow for a more enlightened understanding of what is at stake in a particular dispute, including the evaluation of both competing viewpoints and the effectiveness of policy alternatives. In doing so, the differing, often tacitly held, contextual perspectives and values could be juxtaposed, the viewpoints and demands of experts, special interests groups, and the wider public directly compared, and the discursive dynamics among the participants could be scrutinized. This would by no means sideline or exclude scientific assessment ; it would only situate it within the framework of a more comprehensive evaluation.

64These are questions that underlie the kinds of concerns and problems that confront policy decision-makers and citizens. There is no shortage of literature illustrating the ways in which failures in policy-making are often attributable to simplistic technocratic understandings of the relationship between knowledge and politics. It is to just these sorts of differing rationalities underlying citizens’ and experts’ responses to each other that policy epistemics would turn our attention. While such work would not need to absorb the energies of the field as a whole, it could become one of the important specializations in policy analysis. Those who engage in it could work to facilitate democracy while at the same time studying more particularly the processes that are involved in attempting to extend it. Short of providing policy solutions per se, it would go some considerable distance toward making an applied policy discipline more relevant to the needs and interests of both political decision-makers and citizens. And not least important, as Rorty has put it, it would help us “keep the conversation going”. [68]