You are on page 1of 15

This article was downloaded by: [SENESCYT ] On: 20 February 2014, At: 06:40 Publisher: Routledge Informa Ltd

Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Discourse: Studies in the Cultural Politics of Education


Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/cdis20

Policy as Discourse: What does it mean? Where does it get us?


Carol Bacchi Published online: 01 Jul 2010.

To cite this article: Carol Bacchi (2000) Policy as Discourse: What does it mean? Where does it get us?, Discourse: Studies in the Cultural Politics of Education, 21:1, 45-57, DOI: 10.1080/01596300050005493 To link to this article: http://dx.doi.org/10.1080/01596300050005493

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution,

reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-andconditions

Downloaded by [SENESCYT ] at 06:40 20 February 2014

Discourse: studies in the cultural politics of education, Vol. 21, No. 1, 2000

Policy as Discourse: what does it mean? where does it get us?

CAROL BACCHI, University of Adelaide, Adelaide, Australia

Downloaded by [SENESCYT ] at 06:40 20 February 2014

The concept discourse has become ubiquitous in contemporary social and political theory. However, it is not always clear what different authors mean when they use the term. Moreover, it seems that at times the term discourse carries very different meanings. This paper examines the uses of discourse among a group of scholars who have taken to describing policy as discourse, either directly (see, for example, Ball, 1990, 1993; Watts, 1993/1994; Phillips, 1996; Torgerson, 1996; Goodwin, 1996; Bacchi, 1999 ) or by implication (Beilharz, 1987; Jenson, 1988; Yeatman, 1990; Shapiro, 1992). Michael McCann (1994, p. 6 ) refers to a related body of literature which describes law as discourse. I intend to investigate what these theorists hope to accomplish through the invocation of discourse and how their particular purposes affect the meaning of the term. I also intend to draw attention to a few lacunae in the uses of discourse by this group, which, to my view, need addressing if the term is to serve the purposes they desire. Meanings of Discourse The approach I take to discourse in this paper is part of a more general approach I develop to the political uses of concepts and categories which I call category politics (Bacchi, 1996). This approach builds on an understanding of language elaborated usefully by Tanesini (1994 ). According to Tanesini, we need to become aware of the fact that concepts are not descriptive of anything, but that they are proposals about how we ought to proceed from here. The purpose of concepts or categories is to in uence the evolution of ongoing practices. Hence, they can be de ned to certain purposes and rede ned to other purposes. As Tanesini says, paraphrasing Wittgenstein, to make a claim about the meaning of a certain word is to make a claim about how the word ought to be used, it is not to describe how the word is used. Along similar lines, Derek Edwards (1991, pp. 516 518 ) describes categorization as social practice, and language as primarily a medium for the accomplishment of social actions. Because of this he advises starting not with abstracted category content but with situated usage (1991, p. 520 ). While I wish to use Tanesini and Edwards to emphasize the active marshalling of discourses for political purposes, I would want to qualify this with a reminder that no one stands outside discourse. This raises an important question about the extent to which subjects are either discourse users or are constituted in discourse (Burr, 1995, chapters
ISSN 0159-630 6 (print)/ISSN 1469-373 9 (online)/00/010045-13 2000 Taylor & Francis Ltd

46

C. Bacchi

7, 8, 9 ), a question where I detect some ambivalence among some policy-as-discourse theorists (see section on Theoretical Lacune below ). With this proviso, I want to investigate brie y the meaning of discourse produced by an emphasis on the situated usage of the term. Paul Bove (1990, p. 53 ) helps here with his insistance that it is inconsistent to search for a correct de nition of discourse. In his view, to attempt to provide a de nition would contradict the logic of the structure of thought in which the term discourse now has a newly powerful critical function. I would support his argument with the claim that we cannot provide de nitions of discourse because the whole idea of discourse is that de nitions play an important part in delineating knowledge. Because de nitions have these effects, they require scrutiny, not replication. With Bove (1990, p. 51 ), I will be arguing that key terms are nally more important for their place within intellectual practices, than they are for what they may be said to mean in the abstract. Moving from this general perspective on language, I would suggest that policy-asdiscourse theorists de ne discourse in ways that accomplish goals they/we deem worthwhile. In the main, policy analysts who describe policy-as-discourse have at some level an agenda for change. They tend to be political progressives, loosely positioned on the left of the political spectrum. They de ne discourse then in ways that identify what they see to be the constraints on change, while attempting to maintain space for a kind of activism. Their primary purpose in invoking discourse is to draw attention to the meaning making which goes on in legal and policy debates. The goal is to illustrate that change is dif cult, not only because reform efforts are opposed, but because the ways in which issues get represented have a number of effects that limit the impact of reform gestures. The argument is that issues get represented in ways that mystify power relations and often create individuals responsible for their own failures, drawing attention away from the structures that create unequal outcomes. The focus on the ways issues get represented produces a focus on language and on discourse, meaning the conceptual frameworks available to describe social processes. I am suggesting then that policy-as-discourse theorists develop an understanding of discourse which suits their political purpose. This argument may also help explain the varieties of uses of discourse available in contemporary social and political theory. Ania Loomba (1998, p. 96 ) shows, for example, that another meaning of discourse, which emphasizes text rather than context, is common among those located in literary studies. Raymond Michalowskis (1993, pp. 378 389 ) identi cation of two branches of deconstruction assists in sorting out these different uses of discourse. Michalowski associates one tradition, literary deconstruction, with Barthes (1967 ) and Jameson (1972 ). This tradition places meaning-making in the hands of the readers rather than writers . The second, social deconstruction, associated with Bakhtin (1968), Foucault (1970, 1975, 1977, 1980 ) and Bataille (1985 ), approaches society less as an analog to the literary text and more as an integrated patterning of ritual performances. It is this latter tradition that dominates among policy-as-discourse theorists (Codd, 1988, pp. 240 242 ). Literary deconstruction tends to see everything as text, whereas social deconstructionistsamong whom I would include policy-as-discourse analystsemphasize the processes involved in the creation of text. Because literary deconstruction places meaning-making in the hands of the readers rather than writers , social analysis informed by literary deconstruction represents a radical return of the subject to social inquiry. By contrast, according to Michalowski (1993, p. 383, [T]hroughout analyses in uenced by social deconstruction there are rumours of power. Sociey is more than an

Downloaded by [SENESCYT ] at 06:40 20 February 2014

Policy as discourse

47

accumulation of private, subjective meanings. Meanings are bound to historical conditions. This distinction between literary and social deconstructionist provides a starting place to re ect upon the varieties of social theory often contained under the broad rubric postmodernism. Pauline Rosenau (1992, pp. 15 16 ) usefully distinguishes between af rmative and skeptical postmodernists. In her view skeptical postmodernists (Michalowskis literary deconstructionists ) represent the dark side of post-modernism, the post-modernism of despair, the post-modernism that speaks of the immediacy of death, the demise of the subject [compare with the radical return of the subject above], the end of the author, the impossibility of truth, and the abrogation of the Order of Representation. Rosenau characterizes their view of modernity as one of radical, unsurpassable uncertainty in which no social or political project is worthy of commitment. Her af rmatives, Michalowskis social deconstructionists, are also critical of modernity but remain open to positive political action (struggle and resistance) and do not shy away from af rming an ethic, making normative choices, and striving to build issue-speci c political coalitions. Theorists who describe policy as discourse line up closer to Rosenaus af rmatives than to her skeptics, and hence on the side of social rather than literary deconstruction, though some continue to re ect upon the possibilities of combining the two approaches (Ball, 1993 ). The attraction I see between policy-as-discourse analysts and af rmative postmodernism, using Rosenaus terminology, is a consequence of their political purpose, their commitment to an agenda for change. Skeptical postmodernists, by contrast, according to Rosenau (1992, p. 5), dismiss policy recommendations. Rosenau provides an insight into the connection between af rmative postmodernism and the emphasis on the social processes involved in the production of text/discourse. In her view, literary deconstructionconsidering the text in absolute isolationproduces a politics of despair, possibly because it makes it impossible to identify an enemy and hence a political project. On the other side, social deconstructionists become af rmatives because they are willing to identify sources of power and to propose projects to challenge them. In direct contrast, interestingly, for Michalowski, social deconstruction produces a politics of despair, probably because he sees in the identi cation of power blocs that need to be overthrown a kind of determinism, a rendering helpless of the individual against these power blocs. We have here very different understandings of agency, and different emphases that need to be teased out. Policy-as-discourse analysts, I would suggest, are primarily interested in identifying the reasons progressive change has proved so dif cult to accomplish. Hence, they tend to emphasize the contraints imposed by discourses, through meaning construction. In the process, the power to contest discursive constructions goes undertheorized. In some of the writing I detect an underlying tension in understandings of discourse, a tension I characterize as one between an emphasis on the uses of discourse, and an emphasis on the effects of discourse. Before examining this tension, I will illustrate the way in which discourse has found a place in policy analysis. Policy as Discourse: the construction of policy problems Policy-as-discourse analysts have found discourse useful, as I note above, in identifying the reasons progressive change has proved so dif cult to achieve. This is due, in their view, not simply because opponents of change quash attempts at reform but because issues get represented in ways that subvert progressive intent. This point is made through

Downloaded by [SENESCYT ] at 06:40 20 February 2014

48

C. Bacchi

drawing attention to the ways in which social problems or policy problems get created in discourse. The premise behind a policy-as-discourse approach is that it is inappropriate to see governments as responding to problems that exist out there in the community. Rather problems are created or given shape in the very policy proposals that are offered as responses. Rob Watts (1993/1994, pp. 116 118 ) elaborates most clearly the source of the turn to discourse in policy analysis in his rejection of the deep-seated assumption found in both social liberal and radical readings of the modern state that in state policy intervention there was/is a discovery process which uncovers/ed real social problems as a prelude to state policy interventions. Insightfully he points out that this exercise deploys categories in such a way as to ignore the possibility that the discovery of problems requires the discursive constitution and abstraction of categories of social practice (my emphasis). In Ness Goodwins (1996, p. 67 ) words, a policyas-discourse approach frames policy not as a response to existing conditions and problems, but more as a discourse in which both problems and solutions are created. Hence, the focus for policy-as-discourse theorists is not problems, which are often the presumed starting place for policy analysis, but problematizations (see Kritzman, 1988, p. 257 ). Murray Edelman (1988 ) was one of the rst to draw together a focus on discourse and policy problems. Edelman (1988, p. 12) acknowledges his debt to Foucault, describing Foucaults analysis of madness, crime and sexuality as tracing changes in discourse that constitute problems. In this view, problems are rarely solved, except in the sense that they are occasionally purged from common discourse or discussed in changed legal, social or political terms as though they were different problems. For Edelman (1988, p. 16 ), the recognition of the discursive construction of policy problems produces this de nition of policy: [A ] policy then is a set of shifting, diverse, and contradictory responses to a spectrum of political interests. The emphasis in policy-as-discourse analyses is upon the ways in which language, and more broadly discourse, sets limits upon what can be said. As one example, Stephen Ball (1990, p. 23) brings a study of discourse to his interpretation of education policy under Thatcher in Britain, noting the way in which these emergent discourses were constructed to de ne the eld, articulate the positions and thus subtly set limits to the possibilities of education policy. Ball also acknowledges Foucault, appealing to his de nition of discourse in Discipline and Punish: discourses are practices that systematically form the objects of which they speak; they do not identify objects, they constitute them and in the practice of doing so conceal their own invention (Foucault, 1977, in Ball, 1990, p. 17 ). Discourse in these accounts is meant to capture the ways in which bodies of knowledge, interpretive schema, conceptual schema and signs de ne the terrain in ways that complicate attempts at change. In traditional approaches, policy is what governments do. Some policy analysts (see Pal, 1992; Burt, 1995 ) have been willing to accede that what governments refuse to do can be as important as action. But in these accounts we are still encouraged to re ect upon only that which is addressed in political debate. In other words, it has to be deliberate refusal to act which we consider. Policy-as-discourse approaches, by contrast, encourage deeper re ection on the contours of a particular policy discussion, the shape assigned a particular problem. In many cases, it is argued, it is not a matter of governments deliberately refusing to act but of talking about a problem as if acting is simply inappropriate or not an issue. Watts (1993/1994, p. 119 ) elaborates the duality of the relationship between reality

Downloaded by [SENESCYT ] at 06:40 20 February 2014

Policy as discourse

49

and discourse. In one sense, he tells us, problems or issues only come to be that way when they have become part of a discourse. At the same time, this opens up the possibility of continuing debate and contest about what it is that is being de ned as a problem worthy of the interest of the state or of becoming the object of state policy. To this I would add the importance of turning an eye to those conditions deemed unworthy of this interest. This is where the idea of policy as discursive activity comes into its own, because it promotes consideration of the ways in which the terms of a discourse limit what can be talked about. Such an approach is markedly different from analyses that ask why and how some issues make it to the political agenda, while others do not (see, for example, Kingdom, 1995; Cobb & Elder, 1983; Bachrach & Baratz, 1963). Its starting point is a close analysis of items that do make the political agenda to see how the construction or representation of those issues limits what is talked about as possible or desirable, or as impossible or undesirable. In a forthcoming book (Bacchi, 1999 ) I elaborate the traditions of policy analysis comprehensive rationalism and political rationalism (see Dudley & Vidovich, 1995, pp. 16 18)against which policy-as-discourse analysts are reacting. Comprehensive rationalists, like Herbert Simon (1961 ) and Marshall Dimock (1958, in Braybrooke & Lindblom, 1963, p. 38) have a rm belief that administrators need to guide policy and to smooth out, if not over, the protests of citizens. In this position, problems are easily identi able in something called the decision space, and administrators have only to act to the best of their abilities to resolve these problems. By contrast, political rationalists, like Charles Lindblom (1980 ), David Dery (1984 ) and Aaron Wildavsky (1979 ), are highly critical of comprehensive rationalism. They see themselves as defenders of the citizenry against bureaucracy, and recommend political structures that increase input by citizens into the decision-making process. They highlight the way in which solutions necessarily ow from the kinds of problem de nitions that are produced, and hence they emphasize the politics behind problem de nition. Despite this sensitivity, they continue to assume that policy analysts stand outside this process and can identify and monitor the impact of their values. Moreover, their focus is pragmatic, unashamedly applied, to use David Derys (1984, p. 38) phrase. In their view, the task of policy analysts is to nd a problem about which something can and ought to be done. In a word, the solution is part of de ning the problem (Wildavsky, 1979, p. 3 ). The task, according to Wildavsky (1979, p. 3), is not to compile a list of all unful lled human needs (or even the shorter list of those which deserve ful llment), but to connect what might be wanted with what can be provided. For policy-as-discourse theorists, by contrast, no social actor stands outside the process as either technical adviser or policy planner. Moreover, there is an implied imperative to consider the impact of policy design on unful lled human needs (compare Wildavsky, 1979, p. 3). A model that suggests that (P )olicy analysis is an activity creating problems that can be solved (Wildavsky, 1979, p. 17 ) is considered narrow and implicitly complicit in maintaining the social status quo. For example, Michael Shapiro (1992, p. 99 ) re ects upon the framing of the problem of traf c congestion as a way of illustrating the effects of the typical passive grammar of decision makers faced with problems, rather than, for example, a more politically astute version that would inquire into the way public policy thinking tends to remain within certain narrow modes of problematization. He describes traf c congestion as a middle-class problem, which already accepts the segregation, housing, and shaping of the labor force that has arisen from the structures of real estate speculation, work-force creation, city planning, and so on.

Downloaded by [SENESCYT ] at 06:40 20 February 2014

50

C. Bacchi

The point here is to recognize the non-innocence of how problems get framed within policy proposals, how the frames will affect what can be thought about and how this affects possibilities for action. Marie Danziger (1995, p. 438) draws attention to this very point with a lovely example from Neil Postmans (1992 ) Technopoly. Postman relates the story of two priests who were having dif culty deciding about the appropriateness of a certain behaviour and wrote to the Pope for a de nitive answer: One priest asks, Is it permissable to smoke while praying, and was told that it is not, since prayer should be the focus of ones whole attention. The other priest asks, Is it permissable to pray while smoking, and was told that it is, since it is always appropriate to pray. Postmans point, as Danziger says, is that the form of a question may block us from seeing solutions to problems that become visible through a different question (Postman, 1992, p. 126 ). To take a more serious example, consider Michael Shapiros (1992 ) analysis of an Australian government investigation to discover why Aborigines seemed to have a high infant mortality rate. The reports conclusion blamed the semi-nomadic life of some of the aborigines. Here the problem is represented to be the Aborigines way of life and the solution, by implication, was for them to change their lifestyle. Consider by contrast a suggestion that the medical system adjust its delivery facilities to keep up with aboriginal migration. Here the problem becomes the mode of delivery of the medical system and this is what must change. It is apparent that one of these options is more expensive than the other and this might explain the framing of the problem in the way it appeared. The point is to recognize that this might indeed be the case whereas the language of the report made the conclusion reached seem self-evident. The politics of funding was hidden in this language. The purpose of a policy-as-discourse approach is to bring such silences in problematizations out into the open for discussion. Such revelations are deemed to be an important part of a political process of challenge. That is, while policy-as-discourse analysts are intent upon revealing the ways in which discursive constructions of problems make change dif cult, they tend to believe that exposing these constructions is a useful political exercise. This is because it marks a rst step in demysti cation. While discourse limits what can be said, therefore, there remains a place in these accounts for discursive reconstruction. There is an insistence that social actors can make a difference to the ways in which problems are constituted. This move is accomplished by insisting that discourses are plural and contradictory. Typically, Watts (1993/1994, p. 123, footnote 63) criticizes work that displays little sense of change or embeddedness in history and relatively little evidence of the real work of actors in revising and amending and using discourses, and little sense of the contest between discourses. Some recent policy studies recognize the role of values and competing interests in the shape of policy and policy evaluation. Many have moved on from the model of policy making as rational decision making and planning (see above). The study by Dalton et al. (1996, p. 16 ), for example, explicitly elaborates a counter-model, which sees policy as strategic and political process. A policy-as-discourse approach agrees that policy is strategic and political process. However, it sees the battles not simply at the level of wanting or resisting a particular policy initiative, but at the level of constituting the shape of the issues to be considered.

Downloaded by [SENESCYT ] at 06:40 20 February 2014

Policy as discourse Theoretical Lacunae: power, ideology, subjectivity

51

Within the social deconstructionist, and hence policy-as-discourse position, there are some ambivalences or theoretical lacunae that need addressing. As mentioned above, in some of the writing I detect an underlying tension in understandings of discourse, a tension I characterize as one between an emphasis on the uses of discourse, and an emphasis on the effects of discourse. A focus on the uses of discourse produces an emphasis on the agentic marshalling of discourses, including concepts and categories, for political purposes (see Tanesini and Edwards above). Theorists who talk about rhetoric tend to emphasize the uses of discourse. A focus on effects produces an emphasis on discursive location and the constraints this imposes on our political analyses. Stephen Balls (1993 ) article What is Policy? makes it clear, for example, that to him policy as discourse emphasizes the constraints imposed by discourse, while an analysis of policy as text places more control in the hands of the readers. His ambivalence about how to combine the two approaches points to a need to clarify the relationship between discourse and subjectivity. In this section, I want to illustrate where this tension between the uses and effects of discourse causes problems for some policy-as-discourse theorists. According to Purvis and Hunt (1993, p. 486 ), the lacunae I identify here form part of a wider gap in theorizing on discourse: [O ]ne aspect of discourse that has received insuf cient attention is the relation between the conditions of their production and the manner of their deployment. To put my argument brie y, I would suggest that in policy-as-discourse theory both the effects and uses of discourse are described, but that these tend to be applied selectively. That is, there is a tendency to emphasize the effects upon those who are considered to be lacking power, and an equal tendency to insist that discourses can be used but by those holding power. I think this tendency, which appears in the continued use of terms like ideology, suggests that policy-as-discourse theorists are still wedded to aspects of critical theory which need to be discussed. Pauline Rosenau (1992, p. 36, footnote 12) offers these insights into what was at stake in Foucaults decision to retheorize discourse. She points out that Foucault wished to criticize Derrida for attributing too much autonomy to language and for missing the historical and political implications of the text. Hence he produced an understanding of discourse which emphasized the relationship of the text to power and to the many forces that in uence its production and its nal form. Terry Threadgold (1988, p. 50) usefully describes the Foucauldian problematic as twofoldwhat the subject is able to say, and what the subject is permitted to say. Theorists who describe policy as discourse generally employ this dual problematic. Stephen Ball (1990, pp. 17 18) describes discourses as about what can be said, and thought, but also about who can speak, when, where and with what authority. This necessarily draws attention to both the power of discourse to delimit topics of analysis (e.g. effects ) and the power to make discourse (e.g. uses ). On the power of discourse, Ball contends, discourses construct certain possibilities for thought. They order and combine words in particular ways and exclude or displace other combinations. On the power to make discourse, he (1990, p. 18) explains, [M]eanings thus arise not from language, but from institutional practices, from power relations, from social position. Words and concepts change their meaning and their effects as they are deployed within different discourse. James Codd (1988, pp. 240 242 ) develops a similar position. He contrasts the idealist and materialist views of language, in which the latter recognizes that words, whether in speech-acts or texts, do more than simply name things or ideas that already

Downloaded by [SENESCYT ] at 06:40 20 February 2014

52

C. Bacchi

exist. Rather, he calls for a conception of how the use of language can produce real social effects. In this materialist theory of language, discourse embodies both the formal system of signs and the social practices which govern their use. In this interpretation, discourse refers not only to the meaning of language but also to the real effects of language-use, to the materiality of language. In policy-as-discourse analysis, there is a tendency to concentrate on the ability of some groups rather than others to make discourse, and on some groups rather than others as effected or constituted in discourse. To put the point brie y, those who are deemed to hold power are portrayed as the ones making discourse, whereas those who are seen as lacking power are described as constituted in discourse. For those with power, policy-as-discourse theorists tend to suggest an agentic marshalling of discourse, along the lines of Tanesini and Edwards (see above). For example, Edelman (1988, p. 36 ) describes the construction of problems as as much a way of knowing and a way of acting strategically as a form of description. The use of the term strategically is at the very least suggestive of intentionality. Watts suggestion (1993/1994, p. 108) that the discovery/response model of traditional policy analysis neglects the multifarious ways in which governments and other groups, such as the medical profession, shape our understandings of social life again produce certain groups, those with power, as shapers of discourse. Susan Phillips (1996, p. 256 ) states explicitly that (F )or policy studies, however, it is important that discourse as meaning be linked with a subject as the supplier of that meaning and that language and text be understood in relation to the actions of these subjects. This supports my earlier contention that policy-as-discourse analysts by default favour an understanding of discourse which allows for agentic production of discourse, and which implies the use of discourse by dominant groups in their efforts to remain dominant. There is an attempt, however, to refuse conspiracy theories. Stephen Ball (1990, p. 155 ), for example, refers to the purposive actions of individuals while insisting (1990, p. 155 ) that the discursive process cannot simply be reduced to the intentions and ambitions of a few key actors. Attempts to smooth over any tension here tend to use the language of discursive and extra-discursive factors. This is meant to direct attention to the social-institutional context in which discourse is located (see Fraser, 1995, p. 287; Bove , 1990, p. 57 ). A balance then is struck among policy-as-discourse analysts. Discourses are described as conceptual schema attached to speci c historical, institutional and cultural contexts, making it clear that no agent is completely free to construct or reconstruct them (see Bosso, 1994, p. 189 ). At the same time, recognizing the institutional location of discourses draws attention to the differential power of some actors in their production. These theorists are crucially concerned with discussing who produces discourse, who controls the enunciative position (Maroney, 1992, p. 239 ). Discourses then are not the direct product of intentional manipulation by a few key political actors, but neither are they transhistorical structures operating outside of human intervention. While this balance seems to work at one level, it leaves undiscussed the meaning of power implied in this kind of analysis. Michalowski identi es the rumours of power operating in social deconstruction (see above). There is a tendency among some policy-as-discourse analysts to assume some of the old connotations of power over which reside in critical theory. They tend to use language that talks about domination and implies a group or groups that are dominant (see for example Fegan, 1996, p. 78). Watts (1993/1994, p. 118) directs attention to the special role of the intellectually and professionally trained, whether in state employment or in civil society, who are implicated in processes of what can be called constitutive abstraction . Here he is

Downloaded by [SENESCYT ] at 06:40 20 February 2014

Policy as discourse

53

attempting to identify the actors engaged in the role of meaning making, speci cally in the realm of policy and policy problems. For Jane Jenson (1997, p. 294 ), [T]his understanding of representation implies that a power relationship is at the root of all political discourse. Jenson is explicit that for her the important point is that [A]ll actors do not exercise the same power over meaning systems. This particular concern is re ected in the continuing use of the term ideology by many policy-as-discourse theorists, despite the fact that Foucault explicitly rejected the usefulness of the notion (Loomba, 1998, p. 34; McLeod, 1993, p. 113 ). Vivian Burr (1995, chapter 5) explains that there are at least four meanings of ideology operating in contemporary social analysis, and that the old Marxian notion of ideology as false consciousness is generally rejected by contemporary theorists. Still, at some level, there continues to be slippage around some of these issues. Those who continue to use the term ideology seem to want to be able to identify an enemy, a focus of attack, the ideologues. Trevor Purvis and Alan Hunt (1993, p. 476 ) make this point more subtly: (I )deology thus implies the existence of some link between interests and forms of consciousness. Interests, or power blocs, operate as sometimes unnamed actors in policy-as-discourse analyses. On the other side, in some authors there is a tendency to see discourses as expressions of values, values picked up by the by through what can only be described as processes of socialization. In Jane Jenson (1988, p. 156 ), for example, discourses are assumptions, or at least this is all we can conclude given Jensons insistance on the role of discourse in delimiting policy and the comment: (P)olicy-makers assumptionsalong with those of other signi cant political actorsset limits on the alternatives considered feasible for policy implementation. Similarly, Mary Hawkesworth (1988, p. 82) focuses on the in uence of presuppostions in the shaping of policy debates and policy. These, we are told, are acquired through a process of indirect learning inseparable from immersion and socialization to a particular culture. Vivian Burr (1995, chapter 1 ) points to the inconsistency in positing a socially constructed world while attributing decision making to the attitudes people hold. Her point here is that this kind of explanation accepts an essentialist view of human beings as value holders. To contrast with this view, she (1995, p. 116 ) offers the idea of interpretive repertoires, developed by Potter and Wetherell (1987 ). These accounts return the emphasis to the volition of discourse users, where values are not necessarily held in some internal sense but may be invoked or appealed to in order to produce particular effects. I offer an example from my own work to illustrate these two approaches. Whereas Hawkesworth sees the debate about af rmative action as due to a con ict between deeply embedded, competing conceptions of the individual, atomistic individuality and socialized individuality, I (Bacchi, 1996 ) highlight the way in which political actors, including theorists, offer competing conceptions of the individual to defend their interpretation of af rmative action. What is missing from accounts like Hawkesworths is an awareness that indeed some people pro t from the visions of reality they offer, that they have good reasons to defend their views of reality, and good reasons to oppose contending positions. I try to capture this aspect of the uses of discourse in my notion of category politics, which refers to the deployment of categories for political purposes. The suggestion that policy actors simply implement policies that align with values they have absorbed through socialization has another problem. As a number of poststructuralists (McLeod, 1993, p. 112; Thorne, 1993, p. 107; Davies, 1994, p. 76 ) have pointed out, it creates individuals as the helpless objects of processes of socialization. For these

Downloaded by [SENESCYT ] at 06:40 20 February 2014

54

C. Bacchi

reasons, Bronwyn Davies (1994, p. 76) insists on the need to distinguish between the humanist concept of socialisation and the poststructuralist concept of subjecti cation. The challenge is to the humanist vision of one who essentially is, rather than being positioned as one who can or cannot speak in this way or that. So, subjects are positioned in relation to multiple and contradictory discourses, opening up a space for change. Jenson and Hawkesworth would doubtless agree with this analysis, but their use of the language of socialization and assumptions tends to produce subjects who can be expected to speak in only one way. Moreover, Jenson and Hawkesworth direct their attention only to the meaning of discourse among those who are seen as decision makers. I detect a distinct undertherozing among these and some other policy-as-discourse analysts of the meaning and uses of discourse among those at the receiving end of decisions. For those at this end the emphasis rather is upon how they are constituted in discourse. Above, it was mentioned that Foucault describes discourses as practices. Most theorists who describe policy as discourse accept this focus and direct attention to the material effects of discourses (see Frazer & Lacey, 1993, p. 179). We have already discussed the emphasis on the ways in which discourses place constraints on what can be said. Relatedly, policy-as-discourse analysts are interested in the way in which groups are assigned positions and value within policy discourses, as needy (Fraser, 1989 ), for example, or as disadvantaged (Eveline, 1994; Bacchi, 1996 ). Gillian Fulcher (1989 ) examines the way in which the disabled are constituted as the problem in policy on disability. The argument here is that these positionings leave the power to de ne need and disadvantage in those designing the policy. They can also disempower groups who are thus created as supplicants. There is a need to be wary here about how this emphasis on the discursive positioning of outgroups produces them as disempowered in ways reminiscent of strong socialization theory (Bacchi, 1990, p. 236). The only ones getting to use discourse in this approach, it appears, are those holding power. In the desire to insist upon recognizing the lived effects of discourse, possibilities for challenge and change go undertheorized. The parameters of this problem are illustrated in a debate between Foucault and authors more concerned about the material effects of discourse. In Foucault and some adaptations of him (see Marcus, 1992), once we acknowledge that the way we think about things is delimited by socially constructed meanings, the way forward appears to be simply to challenge those meanings. Regarding rape, Foucault insists that its power resides in the discursive construction of sexuality as integral to personhood, and that therefore we need to encourage women to challenge this meaning, to think about rape as like shoving a st in someones mouth (quoted in Change, 1977, in Plaza, 1980, p. 31 ). Foucaults attempt to challenge the constructed nature of sexuality produces the decision, on his part, to deny the sexual character of rape. This particular interpretation of the way forward seems to allow subjects to step outside their discursive positioning to challenge discourses. Policy-as-discourse theorists tend, by contrast, to insist upon the lived effects of discourse (Bordo, 1993; Frazer & Lacey, 1993 ). Monique Plaza (1980, p. 36 ) develops this position, insisting that we must con rm that rape is sexual, to the extent that it refers to social sexing, to the social differentiation of the sexes. Whereas Foucault offers a purely discursive response, Plaza emphasizes the relevance of nondiscursive factors, such as the social location of women and men. In her view, for the woman raped, the experience is not equivalent to having a st shoved in her mouth. Her feelings will re ect the lived effects of discourse and it is sadly inadequate to suggest that she simply start to think differently about sexuality.

Downloaded by [SENESCYT ] at 06:40 20 February 2014

Policy as discourse

55

There are real bodies and real people living the effects of discursive conventions, and it is essential to attend to the harms they experience. Again, this kind of explanation emphasizes the constraints upon possible responses. The point of this discussion is to suggest that policy-as-discourse analysts need to consider explicitly their claims about discourse and its effects. The tendency to identify only some groups as able users of discourse needs to be reconsidered. So too the emphasis on the discursive positioning of outgroups produces these groups as disempowered in ways that work against the political agendas of policy-as-discourse theorists. We need then to strike a balance between the constitutive effects, including the lived effects of discourse (see Davies, 1994; Bordo, 1993 ), and a recognition that discourses can be used to effect (see Tanesini, 1994).

Conclusion

Downloaded by [SENESCYT ] at 06:40 20 February 2014

I began this paper with the proposal that discourse, like any term, can have many meanings. Any imposition of de nition is related to a range of factors. I emphasize the connection between de nition and political project. The point is that there is no single or correct de nition of discourse; we de ne it to suit our purposes, though this usually happens without conscious intent. What we have uncovered in contending positions on discourse, its origins and its impact, is not just a matter of confusion or even intellectual disagreement. There are political stakes in the contending positions. In order to understand disagreements over the word, I nd it useful to uncover, to the best of our ability, the political projects they represent. I have suggested that policy-as-discourse scholars have an often unstated commitment to political projects that challenge current relations of domination (see Thompson, 1984 ). Consequently, for them, discourses are powerful; they provide meanings that assist particular groups to maintain positions of in uence; but they are not an overarching structure operating outside of history. People use discourses in these accounts; some shape discourses that help maintain their positions of authority and in uence; others intervene and contest representations that uphold the power relations they want to challenge. The invocation of the term discourse is, rst and foremost, an expression of the desire to say something about how dif cult it is to accomplish progressive social change. At the same time, a reluctant optimism about the possibility of change produces a de nition of discourse which emphasizes contradiction and multiplicity, in order to create the space for challenge. In the last section of the paper, I suggested that policy-as-discourse analysts need to spend more time theorizing the space for challenge. I nd an overemphasis on the constraints imposed by discourse/s and a tendency to concentrate upon some groups, those described as having power, as the makers and users of discourse. My argument is that discourse will not adequately serve the strategic purposes of policy-as-discourse theorists as an analytical tool until this overemphasis is corrected. To this end I would suggest pursuing further the lines of investigation opened up by Balls (1993 ) description of policy as discourse and policy as text, keeping open the fertile tensions between theoretical perspectives (Mallon, 1994, p. 1515, in Loomba, 1998, p. 253 ).

Correspondence : Carol Bacchi, Politics Department, University of Adelaide, Adelaide, South Australia 5005, Australia. Email: cbacchi@arts.adelaide.edu.au

56

C. Bacchi

REFERENCES
BACCHI, C. (1990) Same Difference: feminism and sexual difference (Sydney, Allen & Unwin). BACCHI, C. (1996) The Politics of Af rmative Action: women, equality and category politics (London, Sage). BACCHI, C. (1999) Women, Policy and Politics: the construction of policy problems (London, Sage). BACHRACH, P. & BARATZ, M.S. (1963) Decisions and nondecisions: an analytical framework, American Political Science Review, 57(3), pp. 632 642. BAKHTIN, M. (1968) Rabelais and His World (Cambridge, MA, MIT Press). BALL, S.J. (1990) Politics and Policy Making in Education: explorations in policy sociology (New York, Routledge). BALL, S.J. (1993) What is policy? Texts, trajectories and toolboxes, Discourse, 13(2), pp. 9 17. BARTHES, R. (1967) Elements of Semiology (London, Jonathan Cape). BATAILLE, G. (1985) Visions of Excess: selected writings, 1927 1939 , A. Stoekl (Trans.) (Minneapolis, University of Minnesota Press). BEILHARZ, P. (1987) Reading politics: social theory and social policy, Australian New Zealand Journal of Sociology, 23(3), pp. 388 406. BORDO, S. (1993) Unbearable Weight: feminism, western culture, and the body (Berkeley, University of California Press). BOSSO, C.J. (1994) The contextual bases of problem de nition, in: D.A. ROCHEFORT & R.W. COBB (Eds) The Politics of Problem De nition: shaping the policy agenda (Lawrence, KS, University Press of Kansas). , P.A. (1990) Discourse, in: F. LENTRICCHIA & T. MC LAUGHLIN (Eds) Critical Terms for Literary Study BOVE (Chicago, University of Chicago Press). BRAYBROOKE, D. & LINDBLOM, C.E. (1963) A Strategy of Decision: policy evaluation as a social process (Free Press of Glencoe. London, Collier-Macmillan). BURR, V. (1995) An Introduction to Social Constructionism (New York, Routledge). BURT, S. (1995) The several worlds of policy analysis: traditional approaches and feminist critiques, in: S. BURT & L. CODE (Eds) Changing Methods: feminists transforming practice (Peterborough, Ontario, Broadview Press). COBB , R. & ELDER, C. (1983) Participation in American Politics, 2nd edn (Baltimore, Johns Hopkins University Press). CODD, J.A. (1988) The construction and deconstruction of educational policy documents, Journal of Education Policy, 3(3), pp. 235 247. DALTON, T., DRAPER, M., WEEKS, W. & WISEMAN, J. (1996) Making Social Policy in Australia: an introduction (Sydney, Allen & Unwin). DANZIGER, M. (1995) Policy analysis postmodernized: some political and pedagogical rami cations, Policy Studies Journal, 23(3), pp. 435 450. DAVIES, B. (1994) Poststructuralist theory and classroom practice (Geelong, Deakin University). DERY, D. (1984) Problem De nition in Policy Analysis (Lawrence, KS, University Press of Kansas). DIMOCK, M. (1958) A Philosophy of Administration (New York, Harper & Bros). DUDLEY, J. & VIDOVICH, L. (1995) The Politics of Education Commonwealth Schools Policy 1973 95 , Australian Education Review No. 36 (Melbourne, Australian Council for Educational Research). EDELMAN, M. (1988) Constructing the Political Spectacle (Chicago, University of Chicago Press). EDWARDS, D. (1991) Categories are for talking: on the cognitive and discursive bases of categorization, Theory and Psychology, 1(4), pp. 515 542. EVELINE, J. (1994) The politics of advantage, Australian Feminist Studies, special issue on Women and Citizenship, 19(Autumn), pp. 129 154. FEGAN, E.V. (1996) Fathers, foetuses and abortion decision-making: the reproduction of maternal ideology in Canadian judicial discourse, Social and Legal Studies, 5, pp. 75 93. FOUCAULT, M. (1970) The Order of Things: an archeology of the kuman sciences, A.M. Sheridan (Trans.) (New York, Vintage). FOUCAULT, M. (1975) The Birth of the Clinic: an archaeology of medical perception, A.M. Sheridan (Trans.) (New York, Vintage). FOUCAULT, M. (1977) Discipline and Punish (New York, Pantheon). FOUCAULT, M. (1980) Power/Knowledge: selected interviews and other writings 1972 1979 (New York, Pantheon). FRASER, N. (1989) Unruly Practices: power, discourse and gender in contemporary social theory (Minneapolis, University of Minnesota Press). FRASER, N. (1995) Politics, culture, and the public sphere; toward a postmodern conception, in: L. NICHOLSON & S. SEIDMAN (Eds) Social Postmodernism: beyond identity politics (Cambridge, Cambridge University Press). FRAZER, E. & LACEY, N. (1993) The Politics of Community: a feminist critique of the liberal communitarian debate (Toronto, University of Toronto Press). FULCHER, G. (1989) Disabling Policies? A comparative approach to education policy and disability (London, Falmer Press).

Downloaded by [SENESCYT ] at 06:40 20 February 2014

Policy as discourse

57

GOODWIN, N. (1996) Governmentality in the Queensland Department of Education: policies and the management of schools, Discourse, 17(1), pp. 65 74. HAWKESWORTH, M. (1988) Theoretical Issues in Policy Analysis (New York, State University of New York Press). JAMESON, F. (1972) The Prison-House of Language (Princeton, NJ, Princeton University Press). JENSON, J. (1988) The Limits of and the discourse: French women as marginal workers, in: J. JENSON, E. HAGEN & C. REDDY (Eds) Feminization of the Labor Force: paradoxes and promises (New York, Oxford University Press). JENSON, J. (1997) Competing representations: the politics of abortion in Canada, in: C. ANDREW & S. RODGERS (Eds) Women and the Canadian State (Montreal, McGill Queens University Press). KINGDOM, J.W. (1995) Agendas, Alternatives and Public Policies, 2nd edn (New York, HarperCollins). KRITZMAN, L.D. (Ed.) (1988) M. Foucault: politics, philosophy, culture: interviews and other writings 1977 1984 (New York, Routledge). LINDBLOM, C.E. (1980) The Policy-making Process, 2nd edn, Prentice-Hall Foundations of Modern Political Science Series (New Jersey, Prentice-Hall). LOOMBA , A. (1998) Colonialism/Postcolonialism (New York, Routledge). MCCANN, M.W. (1994) Rights at Work: pay equity reform and the politics of legal mobilization (Chicago, University of Chicago Press). MCL EOD, J. (1993) Review essay: impossible ctions? utopian visions and feminist educational research, in: L. YATES (Ed.) Feminism and Education, A Special Edition of Melbourne Studies in Education (Melbourne: Australian Education Council). MALLON, F.E. (1994) The promise and dilemma of subaltern studies: perspectives from Latin American history, American Historical Review, 99(5), pp. 1491 1515. MARONEY, H.J. (1992) Who has the baby? Nationalism, pronatalism and the construction of a demographic crisis in Quebec, 1960 1988, in: M.P. CONNELLY & P. ARMSTRONG (Eds) Feminism in Action (Toronto, Canadian Scholars Press). MARCUS, S. (1992) Fighting bodies, ghting words: a theory and politics of rape prevention, in: J. BUTLER & J. SCOTT (Eds) Feminists Theorize the Political, pp. 385 403 (New York, Routledge). MICHALOWSKI, R.J. (1993) (De)Construction, postmodernism, and social problems: facts, ction, and fantasies at the end of history, in: J.A. HOLSTEIN & G. MILLER (Eds) Reconsidering Social Constructionism: debates in social problems theory (New York, Aldine de Gruyter). PAL, L.A. (1992) Public Policy Analysis: an introduction, 2nd edn (Scarborough, Ontario, Nelson). PHILLIPS, S.D. (1996) Discourse, identity, and voice: feminist contributions to policy studies, in: L. DOBUZINSKIS, M. HOWLETT & D. LAYCOCK (Eds) Policy Studies in Canada: the state of the art (Toronto, University of Toronto Press). PLAZA, M. (1980) Our costs and their bene ts, m/f: a feminist journal, 4, pp. 28 39. POSTMAN, N. (1992) Technopoly: the surrender of culture to technology (New York, Vintage). POTTER , J. & WETHERELL, M. (1987) Discourse and Social Psychology: beyond attitudes and behaviour (London, Sage). PURVIS, T. & HUNT, A. (1993) Discourse, ideology, discourse, ideology, discourse, ideology , British Journal of Sociology, 44, pp. 473 499. ROSENAU, P.M. (1992) Post-Modernism and the Social Sciences: insights, inroads, and intrusions (Princeton, NJ, Princeton University Press). SHAPIRO, M.J. (1992) Reading the Postmodern Polity: political theory as textual practice (Minneapolis, University of Minnesota Press). SIMON, H.A. (1961) Administrative Behavior: a study of decision-making processes in administrative organization, 2nd edn (New York, Macmillan). TANESINI, A. (1994) Whose language? in: K. LENNON & M. WHITFORD (Eds) Knowing the Difference: feminist perspectives in epistemology (New York, Routledge). THOMPSON, J.B. (1984) Studies in the Theory of Ideology (Cambridge, Polity Press). THORNE, B. (1993) Gender Play: girls and boys in school (Buckingham, Open University Press). THREADGOLD, T. (1988) Language and gender, Australian Feminist Studies, 6, pp. 41 70. TORGERSON, D. (1996) Power and insight in policy discourse: post-positivism and problem de nition, in: L. DOBUZINSKIS, M. HOWLETT & D. LAYCOCK (Eds) Policy Studies in Canada: the state of the art, pp. 266 298 (Toronto, University of Toronto Press). WATTS, R. (1993/1994 ) Government and modernity: an essay in thinking governmentality, Arena Journal, 2, pp. 103 157. WILDAVSKY, A. (1979) The Art and Craft of Policy Analysis (London, Macmillan). YEATMAN, A. (1990) Bureaucrats, Technocrats, Femocrats (Sydney, Allen & Unwin).

Downloaded by [SENESCYT ] at 06:40 20 February 2014