Knowledge and its applications

We have said in our introduction that in political speech knowledge is represented as the rational foundation of judgment and policy as the direct practical consequence of judgment. That is the typical "rational" model of decision-making generally taught in management schools, and indeed as a normative model it seems sensible, but as an analytical tool it is completely inadequate. Michel Crozier (1977 [1992]: 307-9), for instance, proposes a systemic approach to understanding organizational decision-making, underlining the fact that decision makers learn both how to decide and what their objectives are as they are deciding: the mental process is continuous. We tend to forget this because of our constant use of what he calls "a posteriori logic," the way decision-making elements seem to fall into place after the policy decision has been implemented. And in any case, "humans are incapable of optimizing", says Crozier (1977 [1992]: 54); freedom of action, as well as the quantity and quality of information to act upon, are always severely limited and coloured by their source. Individuals are always "misinformed" about their own goals, often conflictual organizational goals, available means, etc. At the same time, they are always engaged in various intertwined relations of power that help create these goals and means in the first place.

What does this do to our definition of knowledge? At this point, it is only useful to underline a few consequences: (1) information used for policy decision does not have to come chronologically before the decision has to be made; it is equally useful during and after implementation. (2) In that case, though, information has to be compatible with the logic of the policy being implemented or amenable to reinterpretation--the most obvious case of this is usually referred to as "spin." This does not necessarily mean that the new information must avoid raising questions or produce criticism, and that it will be rejected if it does; simply put, it will be made to fit in with the general way in which the "problem" and the "solution" have been defined through the use of an accepted language about the situation.

It is not our intention here to dwell in epistemological considerations on the nature of science. However, it is necessary to look at the appearance of science in the public domain, and this seems to lead inevitably to some form of philosophy of science, since most of us cannot help but be curious about exactly how many forms expertise can take in a continuously evolving society. Two points of view seem to take most of the room here: either it is a science and by clearing political/ideological brush one can always reach a core of true "objective" -- and hopefully usable -- knowledge; or on the contrary social science does not have any particular specificity and thus can simply be adapted to any political agenda. We will argue that both points of view are off the mark. We do not simply mean that we will situate ourselves somewhere in the middle of the continuum; rather, we think that it is not only possible but useful to altogether leave this level of analysis.

A good example of the first point of view is Daniel Moynihan's (1969) evaluation of the use of sociological theory in a governmental program in the U.S. (a community-based program part of the "war on poverty" policies of the fifties and sixties). It is a good example not because it celebrates science as the answer to society's ills, as the more naive tenants of this perspective would, but precisely because it is largely skeptical of the role of science in decision making, but without ever dismissing it. In his conclusions, Moynihan dissects the theories that were used to design the program, the implication being that the right theory would have worked, if only officials had done their homework properly. The cause of the failure is that the theory that was used, decided upon by experts who mistook their social consciences for objective data, was in retrospect unreliable, or not "good" objective science (Moynihan, 1969: 178, 199).

Ultimately, though, Moynihan (1969: 193), like James Wilson (1983: 53-6) for instance, recommends that since general theory is rarely sufficient for designing specific programs, experts should be limited to policy analysis⁽⁴⁾ , meaning that experts will be making recommendations after the fact. If a program is judged to be successful, this will mean "keep it", "renew it", "emulate it", etc. If not, experts will have to be clear as to why that is the case. But this only means that they are making recommendations for the next time: so the problem remains whole.

Wilson is particularly interesting on the matter of scientific advice. For him, finding "causes" (note that he does believe that to be possible) of crime never allows for direct policy implications⁽⁵⁾. The logical step towards policy is always a personal/subjective one on the part of the scientist and should be considered as such -- does this means that non-expert opinion is equally valid at all levels? Of course not: Wilson goes on to explain that since crime is a product of what he calls human volition, policy can only hope to tip the costs/benefits balance of crime on the side of the costs. His entire demonstration only serves to dismiss one type of knowledge to replace it with another, "truer" one.

On the opposite end of the continuum, if science is in fact always ideology, then our question on the relation of scientific knowledge to policy is settled; politics and expertise are in fact the same thing: they are normative statements about society, and policy is an attempt to make reality conform to ideals. Scientists simply use a different vocabulary: that of scientific method, which is supposed to test hypotheses. But they are in fact rearranging, or "resculpting" reality to make it fit cultural/ethical/political necessities. Now the problem with that second point of view is that it assumes that scientists never "sample" reality, not even as ordinary human beings. Because if they did, they would instantly realize that there is a huge difference between the normative and the existing, between wishes and realities, etc; in fact, there is circularity here, since normative statements by definition are not statements about reality, but what it ought to be, and making normative statements -- and policy -- is useless if reality is already perfect. In other words, even the possible observation that science always serves its political master does not allow the conclusion that it cannot do otherwise. And more concretely, it also ignores the many instances of theory not following, let alone directly criticizing policy, and creates a conspiratory/manipulative perspective that fails to pass the test of common sense.

Joseph Gusfield (1989), argued that the applied social sciences (e.g. social work, criminology, psychology) are essentially defining "problems" in terms of what the current mode of government is assumed to be responsible for and capable of addressing. In the case of the "welfare state" for instance, situations are systematically constructed as "social problems", because welfarism is based on the idea that social progress and individual well being can be facilitated by official social intervention. By finding/studying a social problem experts automatically set the framework in which the solution will be found because they define it with the language of the solution. Furthermore, Gusfield concludes that experts have clear incentives in "finding" (creating) social problems and long-term treatments: it earns them a living. The public in general also finds what it wants: "social problems" and "sick" people become new myths about the human condition and provide vivid images that help constitute what we call a simplified "story of the social" (or "worldview"). All of this activity (whether we agree with it or not) helps us make sense of our world and we like it when the world makes sense.

Unlike Gusfield, while we recognize the obvious benefits for experts that defining and appropriating a problem may represent, we believe that (1) all social situations are always socially defined (and thus constituted), whether as family problems, community problems, social/public or personal ones: before situations are given a meaning, they do not (socially) exist (more on this, next section ). This means that in itself no one mode of definition is inherently better or worse than the others; what is important is the underlying process or rationale at the root of the definition. (2) The image of experts scheming to "create" problems is again misleadingly conspiratorial or machiavellian. As we will see, there is absolutely no need for experts -- or bureaucrats for that matter -- to engage in any form of "mutual understanding" or plot or what have you for reality to conform to Gusfield's observations. In fact, they do not even need any kind of "common interest". It is only important that they have a common language.

Gusfield is certainly right in his analysis of the welfarist culture and the unquestionable certitudes it produces: it rests on (assumed) consensus, (perceived) progress and (apparent) scientific knowledge, and yet has managed to disqualify other forms of political and scientific discourse. None of these basic concepts is seriously questioned at the political or scientific level. However, Gusfield does end in a typical recommendation for skepticism, not an outright rejection of expertise (1989: 439; see also Cohen, 1990). Accordingly, we should "move the study of social problems closer to the study of how social movements and institutions affect and are affected by the interpretations, the language, and the symbols that constitute seeing a situation as a social problem" (Gusfield, 1989: 439). That seems dangerously close to saying that the real problem is politics and language, or bureaucracy, in other words that there really are no "problems" other than the ones created by government and experts, a typically neo-liberal/conservative stance. Social problems do exist -- as social constructions: but that is how anything "exists" in the social world. It is misleading to differentiate between a language of problems and a reality of problems.

So how is it that science and politics can interact, be both subject to the same language, symbols, etc., and still remain science and politics? This is where we turn to Foucault.

Let us begin this section with a typical foucaldian misunderstanding: during a conference at the Université de Montréal at the beginning of the eighties, Foucault was discussing "alternatives to incarceration" and started his lecture by saying, essentially, "what alternatives?" In his mind, community work, probation, halfway houses, educational programs, etc. may very well be alternatives to the physical building called a prison, but certainly not to the dominant rationale of carceral discipline. What is done, as well as how and why it is done, in the framework of these alternatives, was for him exactly the same as incarceration at the level of the required mentality. Elsewhere (1975), Foucault for instance explained how the movement from isolation and prayer in prisons to more scientific techniques of rehabilitation was in fact based on exactly the same disciplinary/control/reform discourse. Actually, part of the reason why prison has never changed, despite criticism that has been constant since its very invention a the beginning of the last century, is that both critics and supporters have always used the same language⁽⁶⁾.

For Foucault, both knowledge and the language used to express it are part of larger unit of culture he calls "discourse". Discourse is always true, even when it says the opposite of another discourse, because it comes complete with the intellectual tools necessary to recognize "truth." It "creates" social reality by giving situations, objects and individuals meaning. It also creates our individual subjectivity as discourse "bearers." as individual thinkers capable of discerning right from wrong. This is how knowledge is power: it creates social reality and its observers.

Discourse is not ideology, and as this is an important nuance, we will give the following example: let us imagine the exiled Vladimir Ulianov turning marxism into leninism. To adapt Marx's ideals to Mother Russia, invent soviets, etc., Lenin created a new ideology, even if the general principles remained the same. But more interestingly, some basic truths were not even questioned: concepts of historical objectivity, causality, human nature, country, sovereignty, power, production, industry, etc. were never part of what Lenin though of changing while adapting Marx. These were held as objective, universal and permanent truths, and giving the impression of doubting them would have seemed downright silly -- and not only to other marxists. Lenin would have had a Revolution, but no one would have come. The problem is that even these truths have a history: a beginning, an end. Basic truths are discourse. They are in a way pre-conceptual, because they are the very tools of thought. Ideology, on the other hand, consists of ideas that are on the intellectual table, so to speak; offered as true, of course, but because they are said by their proponents to be the obvious or only way to extract political deductions from unquestionable truths. Unlike discourse, ideology is always expected to be disputed, and needs to be defended.

What we mean by "truth having a history", hopefully, is clear. It has been said before -- relativism is not exactly new -- that values change with time; this however is naive by comparison: what a foucaldian perspective implies is that the very fabric of human thought is constituted by culture through exposure to competing/cooperating discourses. As a historian like Paul Veyne (1978) can make clear, Foucault's point of view turns conventional historical explanations on their heads. Typically, historians express causality in a chronological chain of events, for example explaining changes in government practices in terms of changing circumstances (e.g. religious reforms, invasions, modifications in the modes of production) and thus tend to present an evolutionary view of governance. Foucault's message here, essentially, is "beware of false continuities": changes in modes of government in fact do not come from administrative experience, "learning from past mistakes", but from changes in the dominant discourses and the redefinition of the past they produce. For instance, the transformation of the modes of governance in the West towards the "governmentalization of the state" (Foucault, 1978) was based on a fundamental change in the metaphorical representation of governance in culture. The Prince was no longer a "ruler", using his authority as a tool for keeping his throne, but during the Enlightenment became a "head of family" "governing" "things" towards their convenient ends. The subsequent advent of statistics changed the family into a "population" (and governance became population control), and made scores of entirely new "social facts" available to analysis -- and natural focal points of state intervention. It is crucial to understand that this is not a change that occurs at the government's level: it permeates all of culture.

It thus becomes very difficult to draw a clear line between science and culture, between scientific facts and discursive "truth". Ultimately, such a conceptual boundary is not necessary, nor is it especially useful. The role of experts, already, is taking an entirely new shape: they are bearers of a specific discourse. At this point it should be underlined that this perspective in no way rejects the existence of a "reality" or adopt other rather problematic by-products of post-modernism. Foucault himself was well aware that the material environment influences the way we live. But the subject here is our construction of a language to speak of ourselves and act in the social: not the material world itself but the conceptual tools we use to give it human-scale meaning.

We have seen briefly what explanation of the welfare state is proposed by Gusfield. While his perspective is not entirely compatible with governmentality studies, it nonetheless offers a very interesting analysis. Especially when we compare it with O'Malley's (1996a) exposé on the current neo-liberal practices in the field of illegal drug consumption in Australia. O'Malley essentially describes a move away from the rationale of welfarism towards a redefinition of individuals as autonomous and responsible for their own life circumstances. In short, there are no "social" problems anymore: only individual ones. In the case of drugs, consumption by individuals is seen as a simple, "natural" fact that government cannot do very much about; the new strategies therefore revolve around "informing" users and potential users (note the word "user" instead of "abuser", or "addict": use is a "free choice") about the risks they take and hoping to modify their cost/benefit analysis towards "responsible" use (remember Wilson's use of punishment as a deterrent: here, the harmful effects of the drugs are the deterrent). In this case experts are actuarial managers of "risk"; study of drug consumption and harm minimization is done in strictly probabilistic terms. Note that this movement has very little to do with the political preferences of specific governments; comparable insurance-company style strategies come from both left and right traditions.

O'Malley (1996: 35) points out that criminology is not simply becoming, as we have seen, more and more administrative; it is also being pushed in the direction of actuarial analysis at the detriment of any sociologically oriented tradition. This is presented as the discipline becoming less "theoretical", and more "applied", but in fact what is taking place is not so much a refocusing towards concrete applications than a general dismissal of all social theory as being non-productive in the current neo-liberal definition of expertise (remember Wilson, again). In the discourse of individual risk, "they are almost bound to find that, in their terms and within their time-frames, programs based on social theories produce few measurable benefits" (O'Malley, 1996: 35). Garland (1996: p. 461) goes further: " a criminology which trades in images, archetypes and anxieties, rather than in careful analysis and research findings -- more a politicized discourse of the unconscious that a detailed form of knowledge-for-power". The fit with current political discourse is almost perfect; could that be a coincidence?

While there may be discussion, and even controversy over specific programs, there is to be less and less about the basic notions underlying them such as individual choice, risk management, harm minimization, deterrence, etc. Again, it is important to note that this is not itself a policy, an obligation that governments impose on experts; it is a cultural change, a shift in the discourse of governance that permeates society: politicians, bureaucrats, experts and citizens alike. One only needs to look at the type of advice currently given to government: for instance, Normandeau and Leighton (1990) present in a document submitted to the Ministry of the solicitor General the point of view of hundreds of "experts" in police work: all agree on a reorientation towards statistical/strategic approaches to policing (see next section).

How experts of "crime" can free themselves of the dominant discourse is perhaps a hollow question. Anyone has to adhere to some discourse; there is simply no "truer" or more "real" alternative. However, something is to be said for creative/productive discussion, i.e. "conflict" instead of blind following. But this conclusion, obviously, flies in the face of the accepted logic that "science" is objective and thus universal. There seems to be something of a catch-22 here: either all follow dominant discourse to maintain the illusion of coherence no matter what consequences follow, or we recognize that in the end universality is out of reach and still lose credibility. Experts, if they are to infuse a dose of reality into the political and administrative discourse, should then concentrate primarily not on the actual practices of government but on deconstructing and subverting the popular language.

d) Crime and order in neo-liberal discourse

Traditionally, "crime" and "law and order" rhetoric are said to have officially entered politics at the end of the sixties in the U.S. As is obvious in Moynihan's (1969) analysis, the "war on poverty" had as a principal objective a reduction in that of criminality which was believed to have its source in a lack of socio-economic opportunities. Whether or not the programs were successes, it is clear that "root causes" were targeted, that the "social" was the field on which the battles would be fought. In a neo-liberal rationality, crime is reconstructed as a problem for its victims, and the state is to concentrate on teaching them new ways to protect themselves, to organize into neighborhood communities, to "retake their social space" (no shortage of buzzwords here), to insure their own security (Garland [1996] calls this "supply side criminology"; p. 451). That because crime is now a statistical inevitability, a routine phenomenon that can be prevented not through looking at its sources, but by establishing levels of risk for individuals and ask them to modify their behavior accordingly. The state can control this behavior not through punishment -- which is too expensive -- but with the threat of victimization -- which is free. It also modifies the penal system to make offenders more responsible towards their victims: simple punishment does not repair the harm done. And when there is indeed punishment, parole release is granted on the basis of statistical analyses of "risk for for the community".

Experts are reconstructed as statisticians studying crime through actuarial management of risk, and evaluating crime prevention programs in terms of their costs and benefits. Experts who refuse face a shortage of financing. Police officers become security experts, and guardians of personal "empowerment" through "partnerships" with the community.

Offenders are reconstructed as decision-makers whose decisions happen to harm others. With rational decision-makers, deterrence becomes the main objective of punishment, as it is assumed to add weight to the "cons" column of rational choice.

Finally, society is reconstructed as a patchwork of communities joined by free agents recognizing their own interests in group life. That may sound like old Hobbesian contractual society to the reader, with reason. Individuals again are seen as contracting parties of a social pact. It is easy to see how this type of "prevention" is a redirection of social control towards the entirety of the population, i.e. that "crime" is becoming a new (unspoken, obviously) excuse for general population control. "Where the state once targeted the deviant for intensive transformative action, it now aims to bring about marginal but effective changes in the norms, routines, and the consciousness of everyone (Garland, 1996: p. 454).

NOTES

4. In fact, it is interesting to see how much Monyhan and Wilson, although they seem to agree on the role of experts, are diametrically opposed. While Moynihan still believes that poverty, for instance, is a "social problem" that the state can adress, and, given the right recipe, solve, Wilson sees poverty as a given fact no one can do very much about, except deter its criminal elements through punishment.

5. His main argument is that root causes, being "independant variables", are independant and therefore unchangeable. Obviously, this is one of Wilson's less inspired moments; it is like saying "the cake will not bake itself, therefore we cannot bake it either".

6. Other example: labor movements in Canada, and especially in Ontario, basically agree with the governments' neo-liberal stance on the importance of deficit reduction, productive excellence, individual responsibility, etc. and thus are left to protest on the details of implementation, on the same conceptual grounds. The very language they use legitimizes policy decisions.