and

May 1996

This is a revision and expansion of an earlier report entitled "Science Intensive Policy Disputes: A Review of the Literature," prepared for the People and Natural Resources Program, USFS, PNW Research Station, Seattle, WA, in February 1995, through a cooperative agreement on Science and Advocacy in Natural Resources Policy.

Alex Antypas is a Ph.D. student in the Social Sciences Program, College of Forest Resources, University of Washington. Seattle, WA 98195. (antypas@u.washington.edu)

Errol Meidinger is Professor of Law and Adjunct Professor of Sociology at the State University of New York, Buffalo, NY 14260. (eemeid@acsu.buffalo.edu)

This paper reviews the part of the extensive interdisciplinary literature on science, technology, and society (STS) dealing explicitly with science intensive public policy disputes. The aim of the paper is to synthesize scholarship on both the origins and resolution of policy disputes with scientific content. The review provides a broad overview of the literature on disputes, drawing from research in several field including political science, political sociology, and knowledge utilization. While subjects such as the institutional role of science in policy, social views of science, the history of policy institutions, the structural position of science in society, and risk and uncertainty are treated in the review, each of these areas has a large literature of its own that obviously cannot be adequately summarized in a reasonable space. Instead, salient features of these literatures that relate to the issue of science intensive policy disputes have been chosen to represent the intensive conversations going on in these interrelated fields. It is hoped that the discussion in this paper will serve as a valuable introduction to these conversations.

The frequency of scientific policy disputes reflects, in part, the degree to which policy of various sorts has become dependent on science. Science today does not conform to old images of the "pure" scientist doing research only for the sake of knowledge and the spiritual and intellectual uplift of humanity. Science is systematically applied to technology development, resource extraction and use, health policy, and many other public policy areas. Science has, in other words, become indispensable to social institutions. The more indispensable science has become, however, the more distant ordinary people have become from policy making that is highly technical. Outsiders to science who have an interest in the policy questions under discussion, may not be able to take part in policy processes because they lack the training to understand the concepts and terminology used.

Simultaneously, as Nelkin (1992) points out, the public has mixed emotions about scientists. Scientists may be able to discover truth, but can they be trusted with it? Perhaps more importantly, can other people be trusted with it?

Scientific policy disputes, do not occur in a neutral social field. Science is imbued with multiple meanings. The position of scientists in society is ambiguous, both central and contested. The following sections take a closer look at the literature on scientific policy disputes, how and why they occur, who is involved, and how some researchers think they should be handled.

II 1. Science-Policy Disputes

Ezrahi (1995) has identified two fundamentally different stances towards science for policy that help explain the sources of the cognitive dissonance that causes public attitudes towards failures of science intensive policies. Ezrahi calls these positions the pragmatic rationalist approach and the utopian rationalist approach. Pragmatic rationalists consider

Consequently, pragmatic rationalists accept "the inevitability of political ingredients in the making of policies" (112) and the equally inevitable fallibility of technically intensive policies. The fallibility of such policies is seen to be a result of both the limits and uncertainties of science and technology, and the inherently political mode of making decisions in which much more than technical rationality determines final outcomes.

In contrast, utopian rationalists regard "political aims and considerations in the making of policy to be obstacles to be overcome on the way to progressive rationalization of decisions and actions" (112). The failure of science intensive policies are blamed on "politics", which becomes the principle enemy that must be fought against. Scientists involved in policy relevant fields of work should therefore be "protected" from politics, leaving their knowledge and advice uncontaminated. The "best science" can then be expected to produce the best policies.

Both of these approaches to policy relevant science value systematic knowledge and aim to increase the scientific credibility of policy decisions. However, pragmatic rationalists recognize both the diverse interests that must be taken account of to make public policy in a democratic context and the inherent uncertainties and institutional and disciplinary biases of science itself. "Good" policy thus incorporated a variety of values, the values embodied by science being only some. On the other hand, utopian rationalists are idealistic about both science and politics in that they are viewed as perfectible through increased rationalization. To the extent that utopian rationalists prevail in the public discourse about appropriate expectations from science intensive policy, they predispose the public towards unreasonable expectations of these policies. The result of these disappointed expectations may be significantly deleterious consequences for scientists, politicians, and the hope for effective public policies.

Ravetz (1987) suggests that the modern environment itself predisposes the public to adopt utopian expectations in regards to science and technology, and to display a childish anger and betrayal when things go wrong. Ravetz thus describes scientific policy disputes in terms of stages. First, because science and technology have created a world of medical wonders, safety, and conveniences, people expect events to unfold in a predictable and beneficent fashion. This expectation of the benign usefulness of science and technology is thwarted when things go wrong, such as at Love Canal or Seveso, causing confusion, or "cognitive dissonance." Confusion quickly leads to conflict as an angry public, certain of malfeasance or at least incompetence, looks around for someone to blame.

False expectations, disappointment, and conflict becomes a recurring cycle in the late twentieth century. Moreover, conflict does not necessarily lead to resolution; often the result is paralysis and continuing political combat. In cases like nuclear waste disposal, local publics immediately oppose proposals to build repository sites near them, interest groups line up on either side, and scientists often take center stage as participants in the dispute. Various parties to the dispute parade "their" scientists in the press, and the battle for public opinion and legislative favor commences. Many disputes wind up in court, although even the courts often fail to bring adequate closure (Nelkin 1992).

II 2. Sources of Science-Policy Disputes: Knowledge Limitations

Disputes over science and technology policy often result from what Collingridge (1980) calls the "dilemma of control." First, the effects of new technologies often cannot be adequately determined. In Ravetz's words, "for policy purposes, they are not probability-based uncertainties but only thinly disguised ignorance" (Ravetz 1987:83). Then, as the technology develops and matures, some negative effects become evident (nuclear power is the quintessential example), but at that point intervention by critics is blocked by institutions that have developed a vested stake in the technology. Critics are held at bay by their lack of access to information. Eventually accidents and other disasters draw public attention and opponents of the technology can gain some leverage on policy, but not soon enough to prevent serious harm to people or the environment (Collingridge 1980). The general idea that Collingridge expresses is the widespread sense that technology inevitably leads to unforseeable consequences, what Winner has termed "technological drift" (Winner 1977). In addition, Collingridge seeks to emphasize the social effects of technology, and to show exactly why it is so difficult to control the effects of technology. In Robert Heilbronner's words, "forces of technical change have been unleashed, but when the agencies for the control or guidance of technology are still rudimentary" (Heilbronner 1970:163).

In Collingridge's usage technology is not limited to manufactured artifacts but also includes scientific tools and techniques such as genetic engineering and timber harvesting methods and regimes. In this sense knowledge itself is a technology which requires certain institutional arrangements in order to be put to practical use. The effects of these arrangements and the direct effects of the technology on society are unpredictable and difficult to control as they become institutionalized.

Clark and Majone (1985) cite the "criteria of adequacy" as an important issue in science-policy disputes. Often the science simply cannot provide the answers to urgent policy questions.

These weakness could be due to transcientific problems, but might also be due to unresolved but in principle resolvable scientific problems. In either case, Clark and Majone conclude that disputes "reflect a profound need for better education throughout society concerning the limited competence of scientific inquiry in policy contexts" (1985:12).

Salter (1988) identifies four types of uncertainty that are encountered in what she calls "mandated" science, that is, science used in policy making. The first is an underdeveloped state of knowledge that will be resolved in time through the normal progress of disciplines. This type of uncertainty is usually found in the context of what Kuhn (1970) called the "normal" science consistent with a particular scientific paradigm. The second type of uncertainty results from inadequate resources, although there is an expectation of certainty on the part of non-scientists. The third type involves the limits inherent in certain kinds of research and research methodologies. The last type of uncertainty refers to the seemingly limitless complexity that can be discovered through scientific research, and the inability to produce final and comprehensive conclusions through open-ended inquiry. Although all four forms of uncertainty are encountered in the policy arena, the last form is likely to be prominent in highly contested policy contexts. This is due in part to the "real world" and interdisciplinary nature of technically intensive policy problems. In other words, the scientific issues involved in policy cannot be kept isolated in a laboratory where they remain uncontaminated by other variables.

Arlid Underdal (1989) argues that scientific information is often in greatest demand when cause and effect relationships are most obscure. This means that science often operates under a handicap in policy situations because it deals, using high standards of evidence, only with the most complex questions. In addition, science must operate under conditions of time pressure when decision makers are desperate for the expertise and legitimizing power of science. Under these circumstances science is likely to produce only probabilistic and tentative conclusions. The decision maker's need for certainty in the short term and the scientist's inability to deliver it largely explains the perceived "uneasy partnership" between science and policy. One consequence of this is that many resource management choices are crude and must be made in the face of fundamental uncertainty where the range of possible consequences is unknown. This supports the view that knowledge, including, scientific knowledge, is a necessary but not sufficient condition for making rational resource policy decisions.

II 3. Sources of Disputes: Hidden Values and Programmatic Conflicts

When scientists are on opposite sides of a dispute they tend to "argue past one another" (Mazur 1981) because their fundamental scientific assumptions are different. Scientists may appear to be arguing about the same thing, and it is not necessarily evident to observers that the seemingly technical dispute cannot be resolved through science. Usually "better" science is called for out of the belief that "objective" research will yield the truth, but in many cases science has already done the best it can, and scientists are arguing among themselves over values and political and moral choices. For example, disputes about siting nuclear plants often revolved around highly technical issues of risk analysis and nuclear engineering which submerged purely political concerns such as an inherent antipathy by members of the public to massive centralized institutions controlled by experts and distant owners. Yet this dynamic is not often discernible in the rhetoric scientists use, which in most cases is couched in the conventional language of science. One of the most frequently deployed rhetorical devices that scientists use is to say that there is "no evidence" to support their opponent's claims, or that it "cannot be proven that ...". Differences may come down to what Weinberg (1972) calls "transcientific" issues - issues that can be stated in scientific terms but cannot be scientifically resolved. Scientists may also disagree over what constitutes evidence for any number of reasons, some of which may have nothing to do with science. In most cases it is impossible for either side to conclusively prove its claims scientifically, therefore, it is not uncommon to witness scientists for both sides belittle each others' claims in the same rhetorical terms (Martin 1991; Mazur 1981).

--Kuhn, inter/intra disciplinary competition

II 4. Broader Social Forces

Science-policy disputes are never only about the technical issues involved. As noted above, the terms of the debate may be technical, but the underlying issues have wider ramifications. Dorothy Nelkin argues that scientific controversies "are a means of negotiating social relationships and of sustaining certain values, norms, and political boundaries at a time of important scientific and technological change" (Nelkin 1987: 284).This issue is taken up again below in the longer discussion about the sociology of science. When disputes erupt, they are rarely simply between two individuals or groups of scientists. They may begin when a scientist or group of scientists point out what they think are risks in certain technologies or practices. However, other social groups quickly become involved, and the public plays an important role in the conduct, resolution, or non-resolution of the debate. The issues involved in science-policy disputes can be diverse and complex.

Thus, in science-policy disputes often a "meta-discourse" can be found, the object of which is a struggle over very basic values, worldviews, and historical forces. It has been suggested, for instance (SECC, 1996) that nuclear power can only flourish over time in the context of an autocratic political system. The more democratic the system, the sooner the nuclear industry will decline. The reasoning is that the centralization and size necessary to have a profitable nuclear industry requires the long term absence or at least limitation of public review and decision making authority. In a system such as in the United States, nuclear power may have a fortuitous beginning because of trust in its extravagant initial claims, but the dangers and eventual public opposition to nuclear plants will serve to drive up costs and make construction politically very difficult. Only a streamlined process of permitting with little public review can keep costs under control and maintain the political will to expand the industry. According to this argument, local struggles over nuclear power plants should actually be interpreted as a continuation of a historic struggle over centralization of political and economic power versus decentralization.

II 5. Scientific Rhetoric

Disputes may be exacerbated by the clash of cultures when scientists try to communicate with policy makers and the public. Wynne (1987) notes that part of this clash involves the "implicit" and "explicit" languages that scientists use. In their daily practice, scientists recognize the necessity of using tacit knowledge and flexible methods to obtain results, yet in their writing and speaking they use an "idealized normative language" that gives the impression (to non scientists) that scientific work is tidy and objective, yielding truth. Wynne emphasizes that this is not a deceptive ploy by scientists but that the language is "used descriptively as a way of gaining prescriptive purchase on scientific practice, so that even though they are not fully achieved, the procedural ideals do have a real quality controlling effect" (1987:98). Traditionally, scientists were taught to write for and speak about science only to other scientists, and so misunderstandings were rare. Everyone "knew" that knowledge is not as fixed as the explicit language of science seemed to imply. This assumption cannot hold for non scientists.

Given the lack of any formal language clarifying the relationship between real and idealized science, it may not be surprising, as Mazur showed, that scientists will often point out the ambiguity of their opponents' arguments (for political purposes) and present their case in the language of objectivity, leaving themselves open to the same attack, and confusing policy makers and the public. If scientific disputes were confined to science, these contradictions might not be problematic, but rather constitute a normal process in science. (This issue is elaborated upon below in the discussion of Collingridge and Reeve's "over critical model" of science-policy disputes.) Because the public and policy makers are often not aware of the conventions of science, and because scientists are not aware of or cannot adjust to the conflicting demands made upon them in the policy arena, the fact that they "are trained to treat uncertainty as both a problem and an opportunity to advance knowledge" (Barke 1986: 143) becomes highly controversial. Uncertainty in science for policy is a problem because on matters of life and death (not uncommon in highly technical debates such as those over toxic chemicals, nuclear power, etc.) and large financial or social investment, policy makers demand certainty from scientists. Uncertainty in science for policy is also an opportunity - to be attacked by one’s enemies.

II 6. Institutional Cultures

Another point at which the culture of science clashes with another culture is in the courts, where many science intensive policy disputes are brought, ostensibly for settlement. As Wynne (1982:126) notes, judicial analysis tends to try to find single causes for all events, "so that clear associations can be made between lines of responsibility and legal rules of guilt, rights, and obligations." Related to this imperative is the attempt to attain certainty. Further, "The judicial ideology of certainty is bound up with an extreme empiricism, which assumes that reality can be exhaustively understood by reference to concrete facts" (Wynne 1982:129). Judges and lawyers expect science to yield objectively indisputable facts, and when scientific experts disagree, especially when each is speaking in the implicit language of science (perhaps in part because they do not want to reveal the vulnerability of their knowledge), then one side is presumed to be dishonest or simply wrong. The objective of the court is to determine which side is right. Naturally, because science is a "fragile, shifting network of interpretive and theoretical activity" (Wynne 1982:129) that does not lend itself to judicial settlement, disputes continue even after courts rule.

II 7. Societal Views of Science

The origin of scientific policy disputes and the difficulty of resolving them may lie partly in a cultural understanding of science that is false, and that leads to inappropriate expectations that are sure to be disappointed. As Ravetz points out, the accomplishments of science are awesome and easily perceived in everyday life. Science seems to create objective facts that are hard to quarrel with: vaccinations against disease, flights to Mars, exotic forms of energy harnessed for everyday uses, computers and electronic highways - the list is almost endless. The language of science further contributes to the impression that scientists operate in the realm of certainty: "Facts", "evidence", "objectivity" are standards of assessment. The results: Deoxyribonucleic acid, radioimmuno assays, nuclear magnetic resonance spectrometer - the black boxed products of science presented to society as certain representations of nature or the infallible instruments used to discover facts about nature. Collingridge and Reeve (1986) argue that this sense of certainty engendered by the accomplishments and language of science be differentiated into a number of specifics: Science yields truth; experts can be expected to agree; science is one; scientific ideas are not influenced by the use to which they will eventually be put. These myths provide the foundation for two additional myths that justify relying on science in policy making:

This view of science, as Nelkin (above) pointed out, is rooted in a long history. Ravetz argues that there is more than one source for this view.

Folk memory, or folklore, has also led to expectations that do not fit reality.

Inevitably, science cannot deliver on its promise, often leaving the public and policy makers disappointed and angry, and scientists equally angry about the harsh treatment they received in the vicious policy process. Basing their arguments largely on recent advances in the sociology of science, Collingridge and Reeve hold that the realities of science are exactly opposite of what the myths state. Science does not yield truth, but rather provisional, temporary, and socially negotiated theories that are always open to challenge. Experts can be expected to disagree, usually in the language of facts but actually driven by moral and political values. Science is not unitary, but forms a multiplicity of disciplines and practices that impart very different worldviews, cognitive competencies, and professional interests on its practitioners. Finally, the scientific ideas are always shaped by the use to which they are intended to be put in policy relevant sciences. Consequently, Collingridge and Reeve argue that scientific findings do not constitute valid criteria by which to make policy - science based policy is doomed to fail. Policy makers will always demand "hard" answers from scientists, and will blame the scientists when their facts turn soft in the heat of a controversy. The fundamental flaw in trying to formulate policy solely on the basis of scientific evidence is that this evidence will invariably fail to provide final "proof" about the correctness of an action and will therefore invite further controversy over technical issues that cannot be solved on technical grounds.

The dissonance between the "myths" and "realities" of science are most consequential when policy is based on science in an open political system. When science is either implicitly trusted by the public or when the public cannot participate in political debate, scientific elites can work rather well with government officials to make policy. As Nowotny writes,

II 8. Science for Ideology

The political imperative to hold facts and values separate, at least in public rhetoric, and to promote science as a neutral arbiter of the truth, has, as Brickman showed, not raised the political clout of scientists in the U.S. Only when a scientific consensus can be achieved, or dissenting scientists silenced or discredited, can science hope to be a dominant force in policymaking. Because American policy processes are open to any actors who can organize themselves they often result in bitter ideological contestations in which scientists, as a practical matter, do not necessarily have more authority than other actors. Contrary to the expectations of policy analysts who hoped to rationalize policy making (relying on "myths" of science), "scientific expertise was treated like any other input into the political process: as a political resource to be used by both sides, negotiable, and not necessarily "true"; in any case not endowed with higher political credibility than other inputs" (Nowotny 1987:66). In Nowotny's view this is not an unreasonable response by the policy process because the boundary between science and policy cannot be clearly defined.

The first point seems almost common sense now, although this was not always the case. To the extent that scientists help in revealing and making social choices, such as creating or evaluating a new technology, they are engaged in a political process. The second point is illustrated by Weinberg's concept of the transcientific. For instance, Western toxicologists argue that biological harm does not occur to an organism if its exposure to a toxic substance does not exceed a threshold level, while Soviet scientists argue that exposure to any level of toxic substance causes potentially damaging perturbations in an organism. Neither side can prove its position right or the other wrong (Majone 1984). Both positions are argued in scientific terms, but science cannot resolve the difference of opinion because it occurs on the level of fundamental assumptions. Data can always be found to fit either model, making one as plausible as the other. Choices have to be made. Transcientific issues are inherent in all science, and when a science has policy implications they are inherently political.

However, there is another explanation for why science is often treated as just another input into the policy process, and why many political actors, publicly in favor of "good science", privately are interested only in science that supports their positions. Politics is inherently ideological, and although science can help form broad ideologies it cannot hope to be more than one among several factors. With many competing ideologies, it becomes increasingly difficult to gain general public support for any set of ideas that would bring legitimacy to government decisions. The American system is filled with ideologically based interest group competition and often has no broad consensus about what values and ideas should form the basis of decisions. "Ideology" and "politics" are disparaged. Science can then be seen as merely a strategic weapon in ideological battles. Interest groups benefit from the centuries old image of science as the quest for truth and certainty and hope to translate that perception of certainty and definitive truth to their cause. This does not mean they are genuinely interested in the pure pursuit of truth, as science claims to be, but only in truth that benefits them. This is another aspect of the irony that the forces that elevate science in politics also diminish it.

Sometimes the ideological forces that constrain and direct science in politics encompass national cultures, myths, and aspirations. Jasper (1992) found that at the national level science-policy disputes tend to develop and be argued according to pre-existing political alignments and ideological platforms of the political parties. In other words, the controversies are "driven" by the institutional political arrangements of a given country. The explanation for this is that political parties are opportunistic and take positions on issues according to the advantages they might gain over competitors. "Every policy debate becomes an opportunity for parties to gain ground over their competitors, or to reaffirm their own philosophical positions for their supporters" (Jasper 1992: 98). Thus, in the case of nuclear power, the debate in the United States was "framed" in terms of the value of the free market over strong government intervention. In France the debate was framed as a referendum on the ability of France and the French state to maintain prestige. In Sweden the debate became a referendum on the rule of the Social Democrats, who had been in power since the 1930s, and the large bureaucratic state they had erected. Jasper claims that at this level of politics the scientific and technical issues simply become tools politicians conveniently use in their fights with each other.

II 9. The Social and Institutional Context of Science

In some European countries science policy disputes have been muted or more easily resolved due to the privileged status of science and a more closed, authoritarian style of decision making. Gillespie, Eva, and Johnston (1979) found that in the U.K. scientific elites could set pesticide policy without much effective public review because they still occupied privileged positions of authority in society that limited external review of their decisions Wynne summarizes these findings:

Brickman (1984) concurs with these conclusions, arguing that both political culture and institutional structure tend to shield scientists and administrators from intense scrutiny over decisions that would be highly controversial in the United States. In Europe administrations have the support of parliamentary majorities, bureaucracies are insulated and enjoy a tradition of deference from the public, and courts rarely review administrative decisions. In this context, where decisions are not highly contested or even visible to the public, science actually plays a lesser role as a legitimating tool and is less of a focus of political debate than in the U.S. Rather, it is merely one consideration among many. Administrators and scientists work out decisions such as regulatory standards side by side, and political, economic, and scientific considerations are woven into the decisionmaking process from the start, with no attempt to clearly delineate their boundaries.

Policy institutions in the United States are much less insulated. Decisions must be publicly justified, and able to withstand judicial review. Interest groups will press their case in any number of available political forums, and will attempt to undermine the arguments and credibility of their opponents any way they can. Technical weaknesses are exploited and the legitimating authority of opponents' science is challenged. The American political system invites, even insures, frequent and prolonged controversy and factionalization, which in the present context insures a central political role for science as factions try to gain an advantage over each other with more authoritative, "unpolitical" truth claims. For these reasons issues regarding the role of science in policy and the courts have been more pressing here than elsewhere. For instance, the relative success of nuclear power in France has often been attributed to that country's almost autocratic governance structures which separate a governing elite from the rest of society and do not permit the public access to the decision making apparatus (see Jasper 1992, Jasper 1990, Nelkin 1981, Nelkin and Pollack 1980, Richardson 1982).

Ironically, the greater prominence of science in the U.S. policy making endangers the public authority and status of scientists. In Europe, where scientists enjoy greater deference, they are scrutinized only by peers and are thereby able to maintain their status. In the U.S., the high level of scrutiny that policy relevant science receives easily leads to bitter arguments over methodology, validity of findings, and the competence or integrity of the scientists themselves. The status of science for policy in the U.S. is inherently contradictory: "[I]nstitutions both place science on a pedestal and work to knock it off" (Brickman 1984). Such high expectations are placed on science to provide unpoliticized and definitive knowledge to legitimate decisions in an overpoliticized environment, that when it fails to deliver, or appears to fail, a reverse kind of process begins in which the reputations and careers of the involved scientists may be severely damaged.

However, there are important advantages to the American system. Here science for policy is driven to higher levels of development and certainty. In other words, the intense political competition sees to it that better, surer knowledge is produced. Greater emphasis is put on separating facts from values, and there are structural protections against the monopolization of knowledge for power. While it can be argued, especially in policy relevant science, that facts and values cannot ultimately be clearly separated, the attempt to do so does provide a significant measure of protection from the direct politicization of the knowledge producing process, or the judgments that scientists are asked to make. The American system idealizes science because political actors hope to exploit it as the only available unpolitical source of authority, and simultaneously reveals the limitations of science, thus undermining its authority. It should thus be expected that controversy and contention will be a ubiquitous part of this system, and can be counted as evidence of an open democracy rather then political and social dysfunction.

II 10. Legitimacy of Policy Institutions

Participation of other critics in the disputes is often better explained by what Nelkin calls "a decline in public trust" (Nelkin 1992: xvi).

Some observers of such conflicts see opponents to certain scientific and technological projects as contemporary Luddites, overwhelmed by anxiety in the face of changes brought about by modernity. Others see them as welcome watchdogs over government and industry. Theodore Roszak claims that protest groups play an essential part in a society that "has surrendered responsibility for making morally demanding decisions, for generating ideals, for controlling public authority, for safeguarding the society against its despoilers" (In Nelkin 1979: 11).

II 11. Regaining Legitimacy Through Science

The prestige of science that politicians like to draw upon is in part the prestige of science idealized. Idealized science is simply the image of science as certain knowledge, "hard" facts, and predictable futures, with vague or no reference to uncertainty, ignorance, trans-science, or the values inherent in the sciences. When uncertainty is considered in the discourse it is only as a temporary obstacle in the implacable march of scientific progress. It may be a feature of human nature to seek certainty, and to a greater extent than anything else, science has replaced religion as the source of certainty in modern society. When the word science is attached to a knowledge claim an implicit statement is made about the legitimacy of the claim. To counter the claim only further scientific argument can be used. To reject the claim on other bases is to be irrational, ideological, or political. In politics, the use of science always entails a legitimation strategy, whether it succeeds or fails. In other words, politics cannot be avoided and a pragmatic rationalist approach is the only effective orientation.

In some cases, administrators and politicians deploy science as a way to regulate the behavior of interest groups (Salter, 1988). This is especially true when interest groups are polarized and political institutions are unable to make decisions as a consequence. Politicians and agencies may come under intense pressure to address a conflict but find themselves caught in the untenable position of either having to choose between mutually exclusive alternatives or to develop a middle position which enjoys no support at all. In such cases, scientists may be called upon to provide an "objective" analysis of the facts, which can then help decision makers decide. In order to succeed in legitimating a decision under these highly contested circumstances, science must be constructed as "the" science, the unqualified consensus of the "best" available scientists. This can be achieved by, for instance, appointing a blue ribbon committee, or by institutionally excluding potential dissenters or undermining their credibility. However, as Collingridge and Reeve have shown, only a temporary state of stability is likely to be achieved. Interest groups and other dissenters, motivated by complex ideas that include moral and other considerations, can be counted on to eventually produce technically competent challenges that undermine the credibility of "the" science used to legitimate decisions they don't like, continuing the cycle of science-intensive policy disputes leading to the loss of legitimacy of both policy and science institutions.

II 12. Social Movements

Often critics of science and technology are engaged in what they themselves see as primarily a moral dispute. Yet critics are usually part of a pre-existing social movement or form a social movement to articulate their concerns over science and technology. Examples include the environmental, animal rights, labor, peace, and feminist movements. At various times these movements take on new issues which might spawn an independent movement. For example, the anti-nuclear movement developed out of the environmental and peace movements. Today, many anti-nuclear groups concentrate on nuclear issues only, but nuclear divisions can also be found in peace organizations and environmental organizations such as the Sierra Club and Friends of the Earth. Nelkin argues that the moral dimension of anti science and technology movements is prominent, and is deeply rooted in our cultural history.

Rights arguments tend to take on a shrill and inflexible character as participants view theirs as universally true claims, not subject to compromise. In many cases, of course, the purported rights of one party conflict with those another party is advancing.

Moral, political, and psycho-social issues involved in a dispute are often mixed and difficult to distinguish individually. For instance, Nelkin (1987) cites fear of risk and fear of misuse as the cause of some opposition to scientific and technological projects. Nuclear power and biotechnology are obvious examples. Yet, both cases also involve a complexity of other issues. These include the rights of local communities, worker's rights, concern about the centralization of power and the rights of citizens to participate in technological policy making, the rights of future generations, and the morality of "playing God". Fear of misuse may also be tightly connected to rights issues, such as in the case of research into possible biological foundations of criminal behavior, the results of which could be used by the government to infringe upon the civil liberties of law abiding citizens who are identified as having a "predisposition" to criminality (Nelkin 1987: 285-286).

The role of protest movements in contemporary science-policy disputes is, as noted above, the most common form of public involvement in such issues. Mazur (1981: 86) identifies three steps in the growth of protest over a particular technology: 1) A public warning about the technology, which might lead to site level and Washington, DC. protests; 2) an adoption of the protest by a core group of activists and organizations, which leads to recruitment efforts; and 3) a mass protest. Scientists often participate as advisors to protest movements.

Recently it has become increasingly common for scientists to be employed by protest movements. The reliance on scientists to wage battles on behalf of protest movements has led to a "fundamental irony" in science intensive policy disputes.

Far from simply being tools of special interests or protest organizations, scientists often choose (or are at least drawn into) the role they will play in science-policy disputes based as much on "ideological" considerations as science. Mazur (1987 and 1981) found that scientists tend to fall into certain categories of political preferences based on discipline and institutional affiliation. For instance, social scientists were found to be almost three times as likely to be "liberal" as engineers. For all disciplines for which a large enough sample was taken (physics, medicine, mathematics, biology, chemistry, geology, and engineering) liberal views were positively associated with employment at elite academic institutions. Thus, 66% of physicists employed at elite institutions were found to be liberal while only 33% of physicists employed at the lowest prestige category institution were liberal. Thirty nine percent of engineers at elite institutions were liberal compared with 16% of engineers at the fourth ranked institutions (see Mazur 1987: 271).

Moral and other value considerations may be at the center of certain science-policy disputes, but this does not necessarily help in predicting which policies will erupt in disputes, or when it will happen. Not all dangerous technologies are highly controversial. Examples include the automobile, airplane, and x-rays. Mazur (1981) and Collingridge and Reeve (1986) argue that certain characteristics of technologies and policies are helpful in predicting whether or not they will become controversial. Mazur found that a technology's newness had something to do with the likelihood of its being protested. He also found that if a technology is forced upon a local populace it is more likely to meet with resistance. Most important, however, is whether there is high public concern over the broader issues of which the technology is a part. For instance, if public concern over environmental issues is already high, resistance to nuclear power will be great. If public concern over the environment is low (during times of economic trouble, for instance) resistance to nuclear power is likely to be low. (An analogous case could be the consternation over high executive salaries, which may occur only when the economy is in recession. Without recession, people are unlikely to care much about relatively minor inefficiencies and injustices.) Mazur summarizes his model in three consecutively related hypotheses.

II 13. The Policy Salience of Science

Of course, this model begs the question, what determines the level of national concern over the wider issues? Collingridge and Reeve (1986) develop what they call the "over-critical model" which purports to explain why science intensive policy disputes develop, and to give researchers and policy makers a predictive tool as well. The over-critical model begins with three necessary conditions for "healthy, flourishing science: autonomy, disciplinarity, and a low level of criticism associated with the low error cost of scientific conjectures" (1986: 29). Autonomy is necessary in that science works best when left to scientists to progress as their professional judgment sees fit. Disciplinarity is necessary because quality scientific research necessitates the use and transmission of a great deal of tacit knowledge and this is most easily done within disciplinary settings that have standardized routes of communication. A low level of criticism is necessary because all scientific theories can be refuted. No scientific finding is immune to refutation, and therefore progress on a research program can only proceed when scientists are able to ignore some of the contradictions and incoherencies of their basic assumptions. This is one of the basic tenets of Karl Popper's philosophy of science (1959, 1969). When science is used to make public policy, all three requirements are violated.

In the first case, autonomy is diminished as scientists work either directly for, at the behest of, or to gain the attention of public authorities. Science-policy issues usually require interdisciplinary work, or at least pose issues in multiple disciplines in contradiction to the second principle. Finally, the stakes in policy issues are often very high, violating the third. This mismatch between the requirements of science and the requirements of policy making lead to what Collingridge and Reeve believe is an iron dynamic subject to prediction.

To varying degrees the authors reviewed here all present us with a dilemma: Many public policies are of an increasingly technically complex nature, pressuring policy makers to rely ever more heavily on scientific advice to guide them, yet science is uniquely unsuited to successfully manage this role. In recent decades the side effects of scientific enterprises such as nuclear power and the use of pesticides and other chemicals have aroused public concern and a demand that scientists remedy the situation. In the ensuing confusion of science-policy controversies in which scientists were seen to disagree on fundamental matters of knowledge, trust in science has diminished. However, no one is under any illusion that future policy choices will be any less technically complex than present ones. Therefore the issue of how to deal with such conflicts has received prominent attention in the literature. The next section reviews some of the key proposals that have been discussed.

II 14. Science and Advocacy

Weiss (1991) argues that there are good reasons for researchers to use their knowledge as advocacy in order to have a greater impact on policy. Several views are plausible on the issue of deliberately using research as advocacy. One could say that research is already inherently position laden and it is simply more honest to be explicit. The partisan give and take involved in using research as argumentation may ultimately also contribute to consensual decision making as research findings and ideas become common property and the "competition of ideas" allows good choices to become evident. One can also argue that ideas need to be tested against a set of logical criteria, thereby not eliminating their indeterminacy entirely but escaping the trap of utter relativism. Logically structured argumentation may also serve as a model for public discourse over policy options. Problems with this model include the likelihood that while some policy researchers may impose that kind of rigor on their presentation, other policy actors will use research in more selective ways. Also, biases will creep into the research regardless of what steps are taken to insure rigor (including biases inherent in the discipline and in the training of researchers).

Weiss suggests that policy researchers are obliged at a minimum to make sure that their data are not deliberately distorted and that they are at the disposal to all parties involved in the policy debate. Under conditions of fair presentation and open access to the research, researchers can take positions without taking unfair advantage of their role as experts.

Primack and von Hippel (1974) argue that scientists should act as advocates "in the public interest" because scientists are often the first to perceive problems and, because of their technical competence, the best qualified to identify the range of feasible options. Given its privileged status as a potential early warning system, the scientific community must act to protect public interests when these are threatened by existing practices or trends. By forging links to politicians, scientists can allow top decision makers to bypass bureaucracies with their established interests, operating procedures, and institutional relationships. Bureaucracies generally filter information flowing to the top, thereby helping to determine the range of possible policy options decision makers consider. Scientists who bypass bureaucracies can create large, informal communications networks through which ideas and information can flow from the scientific community to decision makers with minimal bureaucratic filtering. Moreover, politicians tend to choose science advisors who agree with them ideologically, thus potentially decreasing the value of their technical advice. It is therefore responsible behavior for outside scientists to advocate their views and initiate contact with decision makers in order to penetrate the ideological network that normally surrounds them.

III 1. The Science-Policy Connection

In the last several decades the sociology of science has made significant contributions to the understanding of the role of science in society. Sociologists have found that by studying disputes within science itself and in science intensive policy they can document the "construction" of elements of the social order itself. That is, they have found that such disputes reveal many of the fundamental values, beliefs, ambitions, and structures within society, and are therefore windows into basic social processes.

One can conceptualize science and politics as distinctly separate institutions of modern society that at times cross paths and may have effects on each other. This has been the traditional approach in the analysis of science and society (Brooks, 1964). Because the federal government provides the lion's share of the funding for basic science research, politicians on select Congressional committees help make science policy. Scientists also influence policy making by providing expert advice to policy makers on technically intensive issues. The sociology of science takes different view of science, politics and policy-making, emphasizing the connections and showing how the institutions of science, politics, and policy making interpenetrate and co-determine each other. Moreover, the boundary between science and politics is seen to be a consequence of the intricate interactions among actors rather than a cause of action. In other words, any boundary that actors agree exists between the science and politics must be reconstructed after the fact.

Bruno Latour's concept of technoscience can be helpful here in elaborating the dynamic between science, politics, and policy making. Latour uses the term technoscience "to describe all the elements tied to the scientific contents no matter how dirty, unexpected, or foreign they may seem..." (1987:174). Latour is making the case that while there is an "inside" to science--the laboratory--this inside is only possible because of its connections to the outside, and cannot be understood except in the context of these relationships. Thus, a scientific fact is not an isolated knowledge event but is linked through social structure and history to many elements that seem "dirty, unexpected or foreign", such as politics and other social activities. The connections between the actors involved in what are differentiated but interdependent activities can be conceived of as networks. Actors are interdependent because they need each others' resources in order to accomplish their goals.

At a very general level then, networks are constructed out of a mutuality of interests. But the interests themselves need to be constructed, actors need to be "convinced" that they can achieve their goals by sharing resources with those actors doing the convincing. The result of network building activities are more or less durable sets of relationships among human and non-human actors. "The relative stability of social boundaries and networks over a long period of time gives rise to systems for which we can determine degrees of autonomy" (Restivo, 1995:53). Thus, for example, over the course of decades forestry became a system that incorporated networks of knowledge, organizations, interest groups, politicians, communities, and forests. The system was the result of these networks, slowly built and stabilized through increasing mutual dependencies. As a system it would be stable so long as it was autonomous, and autonomous so long as it was stable. Stability and autonomy would be threatened when actors outside of the system successfully broke into the networks composing it and began to redefine the interests of the actors that made up the networks.

III 2. Constructing Technoscientific Networks

The categories "science" and "politics" are social constructions both of which appeal to culturally based understandings of legitimate action. What kinds of actions fall within the domain of science versus politics is a matter of constant negotiation between actors using enrollment strategies to achieve their goals. In other words, the boundary between science and non-science in general, and science and politics in particular, is negotiated and shifts both locally and as cultural categories over longer periods of time (Gieryn, 1995; Jasanoff, 1990). Similarly, what constitutes good science and bad science, relevant and irrelevant science, and also legitimate and illegitimate policy questions is unstable and socially negotiated. The effort to stabilize (define and delimit) relationships and the domain of what is meaningful and thinkable is the fundamental step in constructing the social order (Latour, 1987). A study of the role of science in policy, and science-policy disputes in particular, is a study of how scientists participate in a vital negotiation in the ongoing construction of the social order. For scientists, as for other actors, this negotiation involves much more than the advocacy of dry facts that may be applied to social problems but embodies moral values and a vision of a preferred order of social relationships. As is true of engineers, scientists, especially those working in an obviously applied field such as natural resources management, are "social activists who design societies or social institutions to fit" their scientific prescriptions (Law and Callon, 1995). Specifically, it takes one kind of society to do traditional scientific management and another to do ecosystem management. The social is embodied in the technical and scientific.

In the actor-network school of the sociology of science the concept of translation plays a central part in understanding how networks are formed and how the technical can embody the social and political. In this conception, a statement, such as a scientific truth claim has no initial force or persuasive power but is spread only because it is carried along by people. This is especially true in a policy context. Each of these people may act upon the statement in different ways: dropping, modifying, deflecting, betraying, appropriating, or adding to it. That is, shaping and presenting it in a way that fits their own tasks and ambitions. The statement acquires whatever energy it has through those who handle it. The statement does not accumulate energy by being passed from one person to the next but has only that energy imparted to it at any given moment. Most decisively, in the translation model, the statement does not remain intact. Rather, "everyone shapes it according to their different project" (1991a:268), thus changing it along the way.

III 3. The "Sociology of Translation"

Translation is the central concept in the theory of how networks are formed, what shape they take, how long they maintain that shape. Translation is at the core of an actor's strategy, the process by which an actor tries to enroll other actors in a project, and then control those who are enrolled so the project is carried out faithfully (Latour, 1987). Callon describes translation as a process "during which the identity of actors, the possibility of interaction and the margins of maneuver are negotiated and delimited" (Callon, 1986:203). Latour adds that translation is "the interpretation given by the fact-builders of their interests and that of the people they enroll" (Latour, 1987:108). Translation, then, encompasses all the diverse strategies and methods that actors use to enroll and join forces with other actors by stabilizing and enforcing a configuration of interests and identities that suits them. Enrollment and counter-enrollment may occur simultaneously, and thus all actors are engaged in processes of translation, the outcomes of which are at any time uncertain. Studying translation is a way of ordering and analyzing the actual interaction of actors, the aggregates of which then make up networks of association. In other words, translation is the "mechanism by which the social and natural worlds progressively take form" (my emphasis) (Callon, 1986:224).

Callon (1986) identifies four moments, or stages, of translation. Since this is a study of science and scientists, the examples given will be from the point of view of scientists. The first stage of translation is problematization, or becoming indispensable. This stage has several analytically distinct features. Through the various ways that scientists express themselves, they connect some natural phenomenon they are studying to other actors, thus positing a network with themselves at the center. This involves a double movement: defining actors and setting themselves up as an "obligatory point of passage" for all of them. Defining actors means naming those actors who are implicated in a scientific endeavor, and imputing interests to them. A preliminary paper on old growth forests, for instance, might define the interests of loggers, state governments, the Forest Service, spotted owls, and scientific colleagues. The second movement involves making oneself indispensable by setting up an obligatory point of passage. Researchers might pose a question such as, how much old growth do spotted owls need to survive as a species, which now becomes an obligatory point of passage for all the actors they have named and given interests to. The loggers want to earn a livelihood but to do so this question must be settled or the forests may be off limits to them. The owl wants to survive. The Forest Service wants to log and stay in the good graces of Congress and the public, and so forth. In the network that the scientists have posited, all the actors must go through them in order to get what they want. They must become allies of the scientists.

The second stage of translation involves locking allies into place. Imputing the interests of actors is not enough to enroll them as allies, as the potential allies have as yet no compelling reason to comply with the program of the scientists. While they are trying to define the interests of the actors in their network one way, those interests are simultaneously being defined otherwise by other actors. For instance, silviculturalists who have been working in the region for many years may define the interests of the Forest Service quite differently. The scientists must interpose themselves between the silviculturalists and the Forest Service. In French inter-esse means to stand in between, and in order to make an ally of the Forest Service the scientists must simultaneously weaken the links between the Forest Service and the silviculturalists and place themselves in between the Forest Service and its goals. The strategies that can be used to accomplish this Callon calls interessement.

To succeed in making alliances requires trials of strength (Latour 1987) such as scientific studies and peer review, or lawsuits, which test claims and create potential entities that link actors to one another (e.g. injunctions to cut timber, a result from a study showing a decline in owls). In these trials of strength the identities of actors are negotiated and tested, and if the interessement is successful, it results in enrollment, the third stage of translation. Enrollment is achieved when actors accept the identities and interests attributed to them, involving "the group of multilateral negotiations, trials of strength and tricks that accompany the interessements and enable them to succeed" (Callon, 1986:211). The enrollment of non-human actors such as spotted owls or old growth forests may be necessary as a fundamental move in establishing a durable network that then grows to include various human elements as well.

The fourth moment of translation identified by Callon is mobilization. Here entities that were once fixed in place are literally made mobile. This stage hinges on the representation of populations by spokespersons, and the distillation of representation into fewer and fewer hands. All the actors in the nascent network have representatives: the owls have those few owls who have actually been observed and studied, the scientific community has those scientists who read the crucial papers and attend conferences, the Forest Service has a few administrators, the political community has a few interested politicians and their slightly less interested committees, and so forth. The aim of the intervening scientists must be to become the spokespersons for all or most of these actors at key stages of the policy development process. If they withstand the requisite trials of strength the colleagues who read their papers and attend the relevant conferences agree that they are credible spokespersons for the scientific community. To mobilize actors means they have to be "first displaced and then reassembled at a certain place at a particular time" (Callon, 1986:217). Science is one of the most powerful sources of translation because of its unique abilities to displace and reassemble by representing entities and phenomena on paper (Latour, 1985). Thus, the few representatives of the owls can be transformed into numbers and graphs and tables that are easily moved and reproduced. The same is true of communities of loggers and the Forest Service with its interest in having a supply of timber to cut. If the scientists can become spokespersons for these actors at certain times and in particular places--key moments of translation and policy formation--they can severely delimit the "margins of maneuver of each entity", including the decision space for policy makers. In the context of this dissertation such key moments would include the Congressional Gang of Four process and the FEMAT process.

Callon's four moments of translation are not necessarily definitive or complete. One would expect to find variation in each individual case. However, the essential feature of translation is that "whatever you do, and wherever you go, you have to pass through the contenders' position and to help them further their interests" (Latour, 1987:120). When this happens the order of society and nature is the result. However, even successful translations can be contested and turned into controversies. "Controversy is all the manifestations by which the representativity of the spokesman is questioned, discussed, negotiated, rejected, etc." (Callon, 1986:219). A controversy is closed only when the spokespersons are not challenged. In other words, allies have to be kept in line (Latour, 1987).

The fundamental feature of this sociological framework is the notion that actors, especially when they are engaged in a controversy, are in the process of "propos[ing] different versions of the social and natural worlds" (Callon, 1986:200). Controversy and disputes are thus fundamental to social processes, without them society would remain static, possible only in idealized theories. Just as there is no "climax" state of an ecosystem that lasts forever, so no social condition will remain the same. The sociology of science has found that science, far from merely describing the natural world, is one of the most potent agents of social change and conflict. However, it does not act in isolation, but rather acts in and through and is changed by heterogeneous networks.

In many ways disputes over science and technology policy are similar to other kinds of policy disputes. They involve questions of value and morality, political power, ideological preference, and economic interest. They are fought using political tactics in rhetorical language. What makes these disputes an area of special concern is the belief that the traditional separation of science and politics has been violated to the detriment of both science and the policy making process. Policy analysts have generally maintained that science can help rationalize policy making by basing it on objective truth. This belief in the truth of science also helped sustain the elevated status of science in society. Recent experience has demonstrated that the production of more scientific knowledge for policy often leads only to more questions and more controversy in areas that are already controversial. Rarely has science settled science-policy disputes. The effort to settle such disputes by turning to science and scientists is seen by some authors as having made policy making more difficult than need be. Involvement in policy making and subsequent disputes has also made the lives of some scientists more difficult, and may have affected the reputation of science generally. Authors have struggled with these contradictions and tried to conceive of ways to make science and policy fit that are more satisfying to members of each community and more beneficial to the public. In some cases authors have thought about how science might better work for policy, in other cases authors have tended to be more concerned with modifying how the policy process works so that it can better accommodate science. In either case, both science and policy are subject to change as the boundary between the two, in this increasingly more technically complex society, becomes blurred or even non-existent.

IV 1. Changed Public Understanding of Science

Many authors agree that science needs to be differently understood in order to avoid or more successfully negotiate science-policy disputes. For instance, Ravetz (1987) suggests that a "new understanding" of science would allow us to more effectively cope with the uncertainty and ignorance of science that complicates its application to policy. Ravetz would like to replace what he calls the "nuggets of truth" metaphor of science with the metaphor of science as a tool. He wants to emphasize that this reconceptualization does not admit that scientific knowledge is subjective, but rather the recognition of its origin, intention, and limitations.

Ravetz conceives of the policy process as being composed of several component systems, including the "sociotechnical", the "information system", the "decision process", the regulatory process, and the "system of devices used for amelioration and monitoring." (87) When science is thought of as a tool, keeping in mind its design, then the need to carefully, self consciously, translate information materials from one system to another will become evident. Ravetz argues that the goal of careful translation should be to avoid misunderstanding, thus increasing flexibility as expectations of the performance of the scientific tools remain reasonable. Decision makers should realize that where to place the burden of proof will always be a question that cannot be definitively answered and will therefore recur. This should be expected and not lamented. The development of policy will necessarily be incremental, recognizing that scientific inferences are strongly dependent on the context of their production and are subject to "continuous modification through ongoing experience." (88) Mistakes are inevitable but correctable and not socially expensive when policies evolve slowly, and politicians and the public recognize that no matter how much is known, significant areas of uncertainty and ignorance remain.

It has also been suggested that improving the science literacy of the American public would help create a more realistic view of science and therefore reduce the politicization and controversial role of science in policy making (Howell, 1992). With a better understanding of science the public could not so easily be swayed by emotion into either idolizing of demonizing science and technology. The alienation of laypersons from science and technology, and thus from many of the core institutions and discourses of modern society, would be reduced.

Howell suggests that it is of first importance to eliminate the mysticism associated with science and technology. Citizens must be educated in the social foundations of science and technology, the human and fallible nature of scientists and technologists, and the inherent uncertainties of scientific knowledge. Citizens must then be re-enfranchised through science literacy training. This can be accomplished through existing institutions. For instance, interest groups such as environmental organizations can publicize the importance of scientific issues as they relate to the environment; federal support for science education can be funneled through the National Science Foundation; private industry can support early education through scholarships and awards; and initiatives can be taken by professional organizations to forge links with outside institutions.

IV 2. Improved Peer Review Procedures

Liora Salter (1985) argues that openness is a central feature of regulatory and other policy making processes. Peer review is one method through which openness and accountability can be ensured. Science has traditionally relied upon peer review to maintain standards, and peer review is common though far from universal in science-policy settings. Salter believes that peer review procedures should be further institutionalized in policy making though, as Hadden and Katz (1986) note, she does not deal with issues that cannot be publicized due to national security or legal barriers.

Jasanoff (1985) argues that using peer review to correct legitimation deficits incurred by agencies and other policy making institutions is more complex than it seems. Policy contexts are very different from the normal circumstances under which peer review takes place. The public nature of policy decisions demands, first, that peer reviewers be accountable to the public and the courts as well as to their agency clients. Second, Jasanoff argues that peer review cannot succeed in legitimating policy decisions unless debates about institutional design, timing and intensity, and accountability are resolved. The institutional design issue involves questions around how peer review panels should be structured, especially in terms of their composition, disciplinary representation, and institutional affiliation. The timing issue revolves around when peer review should take place and how often it should occur (e.g. at the hazard identification, analysis of data, and/or risk assessment stages). Further, should the intensity of review differ according to the subject under review? The accountability issue involves questions about the forum of the peer review itself? Given the public nature of policy making, should the review be conducted in public, contrary to traditional norms? Should members of the public participate in the review? This might be appropriate as the science used for policy always must intertwine technical and political issues, data and objectives. This could be used as a strategy to prevent the discussion of objectives from being hidden in a technicized debate.

Jasanoff (1992) has found that agencies can reclaim lost legitimacy and settle disputes by relying on outside scientific bodies to either produce or review the science they use to make policy decisions. The credibility of such bodies rests on legal, institutional, and procedural controls such as "novel funding arrangements, interest group representation, and open meetings." The latter point up Jasanoff's belief that science can only help settle highly contested disputes and restore legitimacy to policy institutions if the accountability structure (such as peer review) is made "transparent" and is "publicly witnessed." However, to the extent that this also makes the science transparent the problem arises that "science too closely witnessed becomes indistinguishable from the political goals to which it is harnessed" and thus does not serve to settle the dispute or lend legitimacy but rather may lead to further political stalemate.

IV 3. New Policy Science Institutions

Nowotny (1987) goes further than a new understanding of science and calls for a new institution of science in public policy. She wants to avoid what she calls the "orthodox response" to the problem of science-policy disputes, which is to reassert the separation of science from policy in which case science works for policy. This "utilitarian-instrumental" model in which science is institutionally autonomous but in which its products are assumed to be directly applicable and beneficial to society is undermined by the very real problems science and technology have themselves helped create in the areas of environmental degradation and health effects on humans. Science is in no meaningful sense separate from policy or society. Nowotny also wants to distance herself from what she calls the reformist response, which seeks to emphasize the relativistic and subjective aspects of science, educate the public to them, and open decision making processes to greater public participation. Nowotny agrees that the public image of science is not accurate, but argues that political conflict will not diminish simply because the public recognizes the inherent uncertainties of scientific knowledge. Nowotny's solution is to propose an "institutional split" between a new public policy branch of science and an academic branch. She lays out some basic requirements for the new branch:

Nowotny calls this new conception of policy science the "managerial conception" of science. Rather than solving problems it would manage them. Institutionally there would be a "multi-leveled hierarchy of responsibility" that would work to impose as much control as possible in an inherently uncertain environment. Importantly, the new conception calls for a "high degree of autonomy ... to the manager--in this case, to scientists" (72). Nowotny wants to bring science and policy together, modifying scientific institutions and methods so that policy can more effectively be served by the scientific knowledge producing apparatus.

IV 4. Assigning Responsibility for Handling Uncertainty to Scientists

Thompson (1986) proposes a far less ambitious plan, but the sentiment of handing more authority to scientists and technologists is the same. While not envisioning new institutions, Thompson recognizes reactions to scientific uncertainty as the main obstacle to a smooth relationship between science and policy. His idea is to rationalize uncertainty by quantifying it using Bayes' theorem; that is, on the basis of professional judgment. Although uncertainty is by definition unknowable, it can be estimated by experts in the field, yielding a better than arbitrary number which can help decision making. Thompson further proposes to put the burden of proof on any party who attempts to use uncertainty arguments to derail a proposed action or new technology. In his system scientific debates would not be open to the public until they are resolved, thus minimizing controversy. On all three counts then - the quantification of uncertainty, putting the burden of proof on critics, and keeping the public out of scientific disputes - Thompson favors enhancing the power of scientists and technologists at the expense of lay persons and interest groups who are not properly scientifically credentialed. To facilitate this distribution of power, the public should be made more scientifically literate (he does not explain what he means), and scientists should make an effort to present their work in such a way that the public can understand it.

IV 5. Distinguishing Technical from Social Uncertainty

Wynne (1987) also argues that uncertainty and ignorance (a conceptually distinct category) are important sources of conflict. He argues that many uncertainties that appear to be technical are actually social in origin. When the derivation of perceptions of risks and uncertainties is made more explicit we are able to decide with a greater degree of rationality what uncertainties should help shape decision making, and which we can simply forget about. That is, some uncertainties are genuinely technical and constitute real material restraints on policy, others are a consequence of social anxieties that should be addressed as such. The implication of this point is that social anxieties masked as concern over technical uncertainties are a kind of false consciousness that needs to be revealed for what it is so that more rational decisions can be made. He agrees that policy decisions, especially about technology policy, should be formulated in such a way and at such a pace that revision is possible as problems and new knowledge emerge. However, he recognizes that the implications of this kind of a vision for science and policy would have large scale effects on economic and political interests. It would constitute a rationalizing step that gives society greater intentional control over its development, but it would also upset the current distribution of power. He is not altogether sanguine about the prospects for such a development.

IV 6. Adopt Flexible Policies and Monitor Effects

Collingridge (1980) also wants to tackle to question of uncertainty and ignorance in science for policy. Recall his two horned "dilemma of control" of technology described previously. The first horn was the fact that when a technology is in early development stages its harmful consequences are not known. Therefore insufficient grounds exist to control the technology. The second horn is that by the time the consequences are known the technology is institutionalized and therefore difficult and expensive to control. Collingridge attempts to deal with the second but not the first horn, arguing that ignorance at early developmental stages is more prevalent than uncertainty, making it impossible to apply the kind of Baysian model that Thompson advocates. For decisions under ignorance Collingridge advocates using a Popperian decision process that makes correction more possible and likely to succeed. Decisions which can more easily be recognized to be wrong, if this is the case, and which lend themselves to correction are favored under this scheme. Decisions are thus seen as a process. Technologies should be designed to be as flexible as possible, and monitoring should assume an important role. Review and revision should be institutionalized. Decision options with low error costs should be favored over ones with high error costs. Similarly, the time needed to respond to unfavorable developments (gross response time) should be factored in when designing technologies. So, for instance, in the case of nuclear power, Collingridge would have recommended that generating stations be designed as small modular units rather than the massive central units that were in fact built. Small modular units would have made the technology much more flexible, and the error cost would have been greatly reduced (both in terms of the consequences of accidents, and also the financial error cost of having to make retrofits and other modifications). The response time would also have been cut dramatically. Collingridge's model for technology decision making leaves current institutions unchanged but modifies the assumptions of planners and administrators. Instead of trying to forecast all possible harmful effects of a technology (an impossible task), flexibility should be built into the decision making process, thus reducing the costs of errors.

Collingridge and Reeve (1986) developed the over-critical model of science applied to policy as described above. When science is looked to as a guide for policy making the conditions for good science are invariably violated, resulting in potentially endless debate and policy gridlock. To remedy the situation the authors do not suggest that science not be used in policy, but that all policy actors be scientifically conscious and employ their own scientists. In such a situation, no one interpretation of data dominates, thus pushing the underlying political issues to the fore and encouraging an incremental type of policy formation. The authors explain

Collingridge and Reeve, then, are not at odds with the way the policy process more or less works now. Their reforms are small, starting with a recognition of the limitations of science, extending to a revised policy making process which institutionalizes monitoring and mechanisms that allow for overhaul of policy as knowledge and the relative distribution of political power change. In their words, "The answer is not to undertake some root-and-branch reform of the policy process, so that it can better utilize the discoveries of science (the aim of policy analysis, and the knowledge utilization field of academics), nor to seek fundamental changes in the conduct of scientific research which would make its products more acceptable to policy makers (also the aim of knowledge utilization researchers, and taken to a limit by Nowotny as discussed above)" (Collingridge and Reeve 1986: 157). If our expectations of science are in agreement with what science can actually deliver, the prospects for avoiding policy gridlock will improve.

The authors point out that the problem of scientific policy disputes is exacerbated in an open political system like that in the United States. It is more difficult for policy makers to legitimate their actions (as opposed to in Europe where bureaucratic agencies enjoy a level of legitimacy and discretion not found in the United States), leading them to look to science for objective support of their positions. However, the openness of the system leads to endless disputes as opponents hire their own scientists and mount technical attacks. The inherent instability of scientific knowledge allows disputes to continue indefinitely as policy actors focus on technical details, all the while ignoring the more socially meaningful underlying issues. Science cannot meet the demands placed on it, and policy making proceeds erratically and often gets stuck in interminable disputes.

The prevalence of science-policy disputes has called attention to the complex interactions between science and policy, and the need to make sense of this multi-faceted phenomenon. It is to be expected that as society continues to become more and more dependent on both complex technologies and information of all sorts that these types of conflict will continue to become even more common. We are familiar with the prominent science intensive policy conflicts of the past twenty years -- over nuclear power, lead in schools and homes, asbestos, Agent Orange, fluoridation -- and have already begun to experience conflicts in the context of emerging issues such as genetic screening, biodiversity, and ecosystem management.