Toward Understanding Aspects of the Precautionary Principle

Transcription

1 Journal of Medicine and Philosophy 2004, Vol. 29, No. 3, pp Toward Understanding Aspects of the Precautionary Principle Carl F. Cranor University of California, Riverside, Riverside, CA, USA ABSTRACT The idea of a precautionary principle (or precautionary principles) is beginning to come to the wider attention of the environmental community, governmental agencies, regulatory agencies, and the regulated community. Different precautionary principles have not been specified in detail, and, of course, this is difficult to do. Yet some specification must be done in order to understand it better and, if it is to be used for specific action-guidance, to implement it. Moreover, it is important to understand more about the principle, its background assumptions and its comparison with other principles to which we might subscribe. This paper explores aspects of the PP and its background assumptions and presuppositions, comparing them with those for risk assessment and other statements of the PP. It also briefly indicates how it resembles legal principles in addressing problems of uncertainty. Finally, it recapitulates two possible versions of the PP and suggests an application of it for an emerging threat to the environment and public health. This review suggests the PP has plausible applications and is not the radical principle some have suggested. Keywords: precautionary principle, presuppositions, risk assessment, uncertainties I. INTRODUCTION The idea of a precautionary principle or precautionary principles is beginning to come to the wider attention of the environmental community, governmental agencies, regulatory agencies, and the regulated community (Raffensperger & Tickner, 1999, pp ). In many respects, different precautionary principles have not been specified in detail, and, of course, it is difficult to specify any principle in considerable detail. Yet some of this must be done in Address correspondence to: Carl F. Cranor, Ph.D., M.S.L., Department of Philosophy, University of California, Riverside, HMNSS Building 3220, 900 University Avenue, Riverside, CA 92521, USA. carl.cranor@ucr.edu / $16.00 # Taylor & Francis Ltd.

2 260 CARL F. CRANOR order to understand it better and, if it is to be used for specific action-guidance, in order to implement it. Moreover, it is important to understand more about the principle, its background assumptions and its comparison with other principles to which we might subscribe. There are a variety of reasons for this clarification. One is philosophical to understand it better. Moreover, in the current politically charged context in which the precautionary principle is often subject to sharp attack, it is perhaps important to say more by way of clarification to understand it and its presuppositions, to see whether it is an odd principle, perhaps the enemy of much that is good and serves social progress, or whether it is perhaps not so threatening after all. In addition, clarification suggests that, despite facial differences, in many respects it is not so different from other principles with which it is often contrasted, although it differs from them in general implications for action. In what follows, I take some steps toward understanding this principle, or perhaps aspects of a collection of principles that go by the name of the Precautionary Principle. In order to do this, I consider some of its background assumptions, its approaches to uncertainty, what might be attractive about it, and how it resembles in various ways other principles in our lives, especially principles in the law. When some of its features are clarified, it may appear to be less worrisome than some of its critics have claimed it to be. I am not enough of an optimist to think that its most committed opponents will be persuaded, but perhaps others will recognize its plausibility and significance for various problems. II. WHAT IS THE IDEA OF THE PRECAUTIONARY PRINCIPLE? There have been a variety of formulations of a principle that goes by this name, some by proponents, some by opponents. A variety of different but related principles have come from the European Community, especially from the Northern European countries (Raffensperger & Tickner, 1999, pp ). There have also been formulations of it proposed by opponents. For example, one author, clearly not a supporter, articulates it as the principle that government should take precautions to protect public health and the environment, even in the absence of clear evidence of harm and notwithstanding the costs of such action (Cross, 1996, p. 2, emphasis added). Such a principle appears to be an easy target as this author states it because it does not

3 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 261 require clear evidence of harm, and costs appear irrelevant. While someone may have urged adoption of such a principle, it seems too easy a target. Another author argues that, While there is no universally accepted definition for the doctrine, it s generally taken to mean that... we need new policies based on what might cause harm, even if there s no scientific evidence that a hazard exists (Whelan, 2000, p. 4). Again, this appears to be an easy target, but does not address one of the most widely endorsed statements of the Precautionary Principle. Thus, it is perhaps most fruitful to begin with a common and widely endorsed articulation of it by the United Nations in Agenda 21. Where there are threats of serious or irreversible damage, lack of scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation. (United Nations, 1992, p. 10) This statement does not obviously suffer from some of the above problems. Unlike the principle Cross criticizes, it does not preclude having clear evidence of harm, instead only urging that decision makers not require full scientific certainty before taking cost-effective action (thus obviously not ignoring costs). However, the U.N. principle does not obviously require clear evidence of harm, which may invite such attacks. Clear evidence of harm could be part of the threat to public health and the environment, but there is no requirement for it. While the U.N. principle does not obviously suffer from some of the problems attributed to the above caricatures, it has considerable vagueness and, like many principles, needs specification (Hickey & Walker, 1995). Before turning to this task, one might wonder why the U.N. principle might have generated such heat, since it is a largely a wimpy principle. Who could be against it? (Cranor, 2001) It is wimpy simply because it is a principle that is easy to implement (Cranor, 2001). One can comply with its requirements simply by not demanding full scientific certainty before taking cost-effective action against threats of serious or irreversible damage to human health or the environment. In many areas of our lives we take protective actions without having full certainty about the threats involved. Moreover, action need not consist in expensive measures to address the threats, only cost-effective ones. In the discussion that follows, I first consider some presuppositions and background commitments of the PP, using risk assessment, some of the time, as a foil to clarify some of the concerns (and also utilizing the principle suggested by Cross as a foil because it puts into relief some of the features of

4 262 CARL F. CRANOR the PP). Following that, I consider further presuppositions of it by using legal models to suggest that the PP is not unusual compared with legal approaches to uncertainty. When one considers some of the uncertainty and scientific ignorance about the universe of chemical substances and the legal context in which they are assessed, it becomes clear why some believe that a different approach to addressing such substances needs to be considered. III. PRESUPPOSITIONS AND BACKGROUND ASSUMPTIONS OF PP AND RISK ASSESSMENT Elsewhere I have considered plausible presuppositions of the U.N. principle and compared those with risk assessment (Cranor, 2001). Here I will briefly review some of those findings, develop some of them further, and go beyond them to articulate other background assumptions. Neither the PP nor risk assessment is a scientific principle; both are recommendations for utilizing scientific considerations in decision-making contexts. In some respects risk assessment may be seen as more similar to a scientific principle, since, on at least some understandings, it might be seen as recommending a variety of scientific procedures and principles to identify and estimate risks to the environment, public health or other valued things (Cranor, 2001, p. 315). Both PP and risk assessment appear to be concerned with risks or threats that have not materialized into harm. On this dimension, they are similar, although there may be subtle differences between the two. Threats, while they must be serious or irreversible might be seen as somewhat more vague than the risks with which risk assessment is concerned. Risk assessment, at least ideally, aims to utilize various scientific techniques to identify and quantify actual risks, something with which the PP is perhaps less concerned. There are contexts, however, in which threats might be more specific than risks: the threat of rain on a mountain climb seems like a more specific chance or prediction of an adverse effect than a (mere) risk of rain (Cranor, 2001, p. 315). Even on the issues of risks and threats, there are differences. Presenters at recent conferences have pointed out that risk assessment has some quite particular presuppositions under which it operates: the range of outcomes must be well defined and one must be able to reasonably assign a quantifiable probability to each of the outcomes. If the range of outcomes is not well

5 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 263 defined or quantifiable probabilities cannot reasonably be assigned to each of the outcomes, it is more difficult to speak of risk assessment; rather one should better speak of making decisions under uncertainty or in ignorance (Savitz & Funtowicz, 2001). On its face, the PP has a greater normative aspect to it than risk assessment, although this seeming difference may be overemphasized. What appears to underlie the PP is a judgment about the substantial value of a state of affairs or object that is threatened (Cranor, 2001). That is, the idea of precaution presupposes some assessment of the status quo and a judgment that it is quite valuable, but threatened in some manner, such as a valued painting, e.g., Leonardo Da Vinci s The Last Supper, so that one in turn can take protective actions toward it to try to avoid the threat or prevent it from materializing into harm (Cranor, 2003a, p. 306). The idea of the PP suggests, at least on a facial statement, a reasonably specific view about what should be specified by means of norms or a matter of policy and what should be left to factual confirmation by the requisite empirical or scientific fields. In particular, the PP appears to take the view that the public health and the environment in which we live are valued states of affairs. And when these are threatened, ceteris paribus, they merit protective attention (Cranor, 2001). Although public policy decisions will be the product of normative assessments of things valued, disvalued or sought to be protected, the PP appears to give greater prominence to threats toward valued things, while risk assessment often appears to emphasize more the empirical data needed for a decision (Cranor, 2001, p. 316). However, as I consider below, these facial differences conceal some underlying similarities. Consider an example from the criminal law that could have two different approaches to it. In designing a law about possession of illegal drugs, Congress created legislation requiring that people possessing illegal drugs, such as heroin, be deemed to have illegally imported the drugs. Moreover, this presumption held, unless the defendant explained the possession to the satisfaction of a jury to the contrary (Turner v. United States, 1970, p. 402). The reason for this appears to be that Congress had an interest in deterring trafficking and, thus, assigned trafficking much higher penalties than mere possession. On one approach to the trafficking problem (analogous to a finding of risk or harm), Congress could have required prosecutors to establish by empirical evidence presented in court that the illegal drugs possessed were imported and, thus, the accused should be held accountable for trafficking. On the approach

6 264 CARL F. CRANOR Congress chose (more analogous to a precautionary approach), it presumed that the drugs were imported because heroin is typically not produced domestically, and, thus, that the possessor was a trafficker, unless that legal presumption could be rebutted. 1 Congress created a presumption of trafficking perhaps because of a normative assessment of the frequency of importation and smuggling, the fact that heroin is not produced domestically and the dangers associated with them, and even perhaps because of the difficulties of proof of importation. It made the presumption rebuttable, however, because persons could be wrongly accused or perhaps could explain or justify the possession. The background normative assessment, thus, helped create the legal presumption without explicit case-by-case evidence being offered in support of the claim of importation (although there could have been Congressional hearings and findings on the general problem). Evidence could be offered in rebuttal, but this might require the accused to take the witness stand, be subject to cross-examination and jeopardize his or her right against self-incrimination (Cranor, 2001, p. 316). However, differences in facial demands for evidence in turn suggest further differences in presuppositions that are philosophically revealing. Demanding empirical data to change the status quo, which would have been required under the hypothetical alternative drug laws, would convey a view that something about the status quo has sufficient value as it stands or changing the status quo should be so difficult (as the explicit case-by-case prosecution for trafficking would require), that a case for change must be supported by strong evidence. By contrast, a criminal law that presumes that a possessor of heroin is a trafficker suggests a more explicit normative judgment that something (else) about the status quo is of sufficient concern, e.g., a possessor of heroin (a significant disvalue) in a world in which there is too much importation of illegal drugs, that this obviates the need for substantial evidence to support the prosecution for trafficking. Moreover, this presumption can be overturned only if the defendant offers empirical evidence to the contrary. I have developed this example from the criminal law in some detail because it suggests some analogues for the PP compared with other principles. Advocates of the PP have substantial concerns about specific aspects of the status quo concerning the state of the environment, the public health and perhaps trends that are already in evidence. For example, concerning the environment they can point to a body of evidence that not all is well. Concerning the public health, evidence that there are problems tends to be less clear, in the sense that there are not currently high rates of disease and

7 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 265 epidemics. Rather there are substantial grounds for unease. Readers may have shorter or longer lists. With regard to the environment, even a brief review of aspects of the state of the world is potentially quite disturbing. Sources of biodiversity forests, grasslands, croplands, and Brazilian rain forests are shrinking, stressed, and often replaced by degraded land or wastelands (Rodgers, 1994, pp. 8 11, 23-24). Water tables are falling around the world and humans are over-pumping aquifers in China, India, North Africa, Saudi Arabia, and the U.S. by about 160 billion tons of water per year. Humans are essentially mining water from non-renewable resources sufficient to produce the food for 480 million people per year, an unsustainable rate (Brown, Flavin, & French, 2000). The worldwide cropland per person has decreased dramatically from.24 hectares to.12 hectares in the last 50 years; it could shrink to.08 by 2050 (Brown et al., 2000). About two-thirds of major marine fisheries are fully exploited, overexploited or depleted (Lubchenco, 1998). Oceans have surprising amounts of pollution and plastics in them, with much more serious pollution near coastal regions (Rodgers, 1994, pp ). About 11 percent of bird species, 25 percent of mammal species, and an estimated 34 percent of all species are vulnerable or in immediate danger of extinction (Brown et al., 2000). Humans are making a substantial contribution to global warming (McKibbon, 2001). World population, which has more than doubled in the lifetime of readers older than 40 and is projected to increase about 50 percent in the next 50 years, is a significant force behind many of these problems (Rodgers, 1994, pp. 2 4). Moreover, those who endorse PP may have a picture of the planet, not with nearly unlimited frontiers and vast natural resources in which human actions are in general a benign influence on their own life-sustaining base, but perhaps more as a partially polluted fish tank with definite limits in which the fish foul their own bowl. Such a partially polluted, partially damaged world may be more susceptible to harmful insults than a less damaged world. Concerns such as these are likely motivating, at least in part, those who advocate the PP with respect to environmental protection. If their assessments of the state of the world are plausible, it appears that there are already

8 266 CARL F. CRANOR substantial harms to the environment, with obvious further threats suggested by present evidence and trends therein. Thus, like trafficking in drugs, if there is reasonable evidence that ecosystems have been substantially damaged, with further nonquantified threats seemingly suggested by trends, it is not unreasonable to seek to address these threats with greater speed and before having extensive certainty that they will materialize, in order to protect an environment that resembles a fouled fishbowl. With regard to threats to the public health posed, for example, by exposures to potentially toxic substances, such documented adverse trends are not as readily apparent (indeed, life expectancies in the developed world have been increasing). Of course, there have been substantial harms as a result of poorly controlled exposures to toxicants in the past. A variety of substances have caused public and workforce health harms, as well as environmental damage: DDT, chlorofluorocarbons, PCBs, lead, mercury, cadmium, nickel, vinyl chloride, benzene, asbestos, and other toxicants, as well as poor disposal practices. (Cranor, 2003b) However, beyond documented high-profile harms, there is a disturbing pattern that has developed concerning chemical substances. Huff and Melnick (2002) have noted that there are about 20,500,000 unique organic and inorganic chemicals. Moreover, there are about 100,000 substances and their derivatives registered for commerce in the U.S. with another added to the list each year with minimal or no testing (U.S. Congress, 1987). Scientists, the public and regulatory bodies appear to know little about this universe of 100,000 substances. In 1984, the U.S. National Academy of Sciences found that there were 12,860 substances produced in volumes exceeding one million pounds per year for 78% of which there is no toxicity information available, 13,911 chemicals produced in volumes of less than 1 million pounds (76% with no toxicity data), 8,627 food additives (46% with no toxicity data), 1,815 drugs (25% with no toxicity data), 3,410 cosmetics (56% with no toxicity data), and 3,350 pesticides (36% with no toxicity data) (National Research Council, 1984). In the early 1990 s, there was insufficient change in the data to justify updating the National Academy Report (Bailar, 2002). For the 3000 substances produced in the highest volume, there remained substantial knowledge-gaps for about 75% of them as recently as 1998, when the U.S. EPA entered into an agreement with the producers to close the knowledge gaps (EDF, 1997; Business Publishers, Inc., 1998). As recently as

9 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE , there were another ,000 substances for which extensive toxicological information would be quite important but which is not available (U.S. Congress, 1995). There may, of course, be some harms at the moment or that are likely to be identified soon as a result of the extensive ignorance of the universe of substances (Hooper & McDonald, 2000). A much more serious generic concern, however, is the specter, as it were, of commercial and legal institutions that appear to be out of control with little knowledge and seemingly little concern about the effects of thousands of unstudied substances on the environment and public health. Under current institutional arrangements, the environment, as well as the public and workforce health, is left vulnerable to any exposures that result with action taken only if there are demonstrated risks or harms. Just like Congressional concerns that possessors of heroin are traffickers, it does not seem unreasonable for the public to have concerns about a universe of poorly understood substances being released into commerce resulting in public and workforce exposure. Someone concerned about such ignorance might not unreasonably surmise that the public and workforce are something like guinea pigs in a great unknown commercial and legal experiment. Moreover, some of the substances that have been studied in fact may pose documented health issues, but have not been subjected to regulation (Hooper & McDonald, 2000). Against this background, it does not seem unreasonable for knowledgeable individuals to endorse something like the precautionary principle; the trends revealed by documented research some actual harms and much ignorance about substances and their effects on ecosystems and health do not portend well for the future. Such concerns are only heightened when one compares some further presuppositions of risk assessment with the precautionary principle. Beyond that, it becomes clearer why more proactive and preventive approaches might seem reasonable when risk assessment is placed in the context of existing legal structures. Within this context, risk assessment can be seen as having substantial value commitments; different background assumptions and different presuppositions change the picture suggested by facial differences between PP and risk assessment. Let me explain. There are different laws that govern the entry of substances into commerce. Some, such as pesticides, drugs, food additives, and cosmetics, are governed by pre-market approval statutes. That is, the substances are subject to some

10 268 CARL F. CRANOR degree of pre-market testing (depending upon the statute) as well as subject to review by a regulatory agency for safety (or for not posing an unreasonable risk ) before the substances can enter commerce (U.S. Congress, 1987). Drugs, for instance, are subject to extensive pre-market testing. Drugs proposed for commercialization are first tested on animals and, then, if they are shown safe and effective in animal studies, are subject to trials on small groups of people under very carefully controlled conditions. Following that, they are then subject to clinical trial testing on somewhat larger, but still comparatively small groups of humans to ensure that they are safe and effective for human use (U.S. Department of Health and Human Services, 1986, pp ). 2 Even substances entering commerce under stringent premarket scrutiny are not without problems. Pre-market testing and review does not adequately identify all the harms that can arise when large heterogeneous populations are exposed (Green, 1998). The safety of most substances, however, is governed by means of postmarket statutes, e.g., substances that might contaminate surface waters, drinking water, the air, the workplace, consumer products, to name a few (U.S. Congress, 1987, p. 199). Post-market laws permit substances to enter commerce with little or no legally required testing and then are subject to legal review, if a governmental agency has evidence that they pose harms or risks of harm. For example, for substances to which the workforce are exposed, the Occupational Safety and Health Administration (OSHA) must establish that there is more likely than not a significant risk to the health of the exposed workforce before exposures can be reduced (and, in the Supreme Court s original Benzene decision, there was also the suggestion that there must be a showing that reducing the exposure was more likely than not to reduce adverse effects on health) (Industrial Union Dept, AFL-CIO v. American Petroleum Institute, 1981). One major point to take from this decision is that OSHA has both the burden of proof to establish that there is a risk and it must be carried by the standard of proof known as the preponderance of the evidence. Thus, for substances that are utilized in manufacturing processes or in many other products that enter commerce with little or no required testing, the burden falls on someone else, e.g., OSHA or the U.S. Environmental Protection Agency, to establish that there are harms or risks of harm. There is one general qualifier to the previous point. Since the early 1980 s, all substances proposed for manufacture in the U.S. are subject to a premarket, pre-manufacturing review: Firms must notify the U.S. EPA that they intend to manufacture a substance and submit any data they have about it in

11 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 269 order to receive permission to proceed to manufacturing. However, any testing is not extensive in order to keep pre-market development costs low (U.S. Congress, 1987). Under this provision of the Toxic Substances Control Act (TSCA), about new substances enter commerce each year with such minimal testing all that is legally required (U.S. Congress, 1987). Moreover, the EPA has limited time to conduct a paper review of each substance before it goes into commerce (90 days), unless it identifies a feature of the substance that raises a toxicity red flag, which then justifies the agency in requesting more data and conducting a more extensive review. Nonetheless, TSCA premarket review of substances is not extensive and is unlikely to detect and identify even many risks and adverse effects of substances if they have them. With the above digression as background, we can understand better some of the differences between PP, risk assessment, and even demands for documented harm before regulatory intervention is permitted. On its face, the PP has a commitment to a greater value component than the risk assessment paradigm. Thus, one might expect that PP and risk assessment differ on how much particular empirical data must be generated by scientific means in order to take some legal action on a specific substance. Indeed, this appears to be the case. Some risk assessors appear to demand greater chemical-specific information about substances that might pose risks to humans or that actually cause adverse effects (Cranor, 2001). The point can be made concerning benzidine, a known human carcinogen and three benzidinebased dyes, Direct Black 38, Direct Blue 6 and Direct Brown 95. There is no direct human data on any of these dyes alone showing that they are carcinogenic to humans (International Agency for Research on Cancer, 1998). These dyes are judged to be carcinogens on the basis of the known carcinogenicity of benzidine together with the fact that some portion of benzidine-based dyes appear to reduce to benzidine in mammalian bodies and the dyes themselves appear to cause mutations in mammalian cells (International Agency for Research on Cancer, 1998). Consensus scientific bodies and regulatory agencies have identified all of these dyes as probable human carcinogens (International Agency for Research on Cancer, 1998). In effect, this means that IARC is virtually certain that the substances are human carcinogens (Rice, 2000). Despite this evidence and these relationships, some scientists and some affected industries have urged the regulatory bodies to perform individual risk assessments on each dye before legal action is taken (Zeise, 2002).

12 270 CARL F. CRANOR Putting the best philosophic interpretation on such demands, one might suppose that individuals who urge such action have the view that there is something so important about not declaring these dyes as carcinogens that individual risks from each one must be established before legal action is taken. Indeed, sometimes small differences in chemical structure can result in sufficiently significant differences in biological activity that, while one substance might be a carcinogen, a similar, but not quite identical, dye is not a human carcinogen or possibly a much weaker carcinogen. Whether this would have been the actual motivation, or something less vaunted like commercial self-interest, I do not judge. The consensus scientific bodies, by contrast, might have taken the position that the reason for treating the chemically similar dyes as sufficiently similar to benzidine to take action was based upon fundamental biological similarities between substances and the carcinogenic potential of the dyes. However, the committees could also have judged that the substantial similarity between the dyes and benzidine, though not mathematically certain, was similar enough, given the public health risks at stake, to judge the dyes to be probable human carcinogens, even if they could not be categorized with total certainty as known human carcinogens. This latter position would much more closely resemble an application of the PP. The dyes appear to pose serious, even irreversible, threats to human health, i.e., cancer; thus, they should be considered to be probable human carcinogens. Even though the PP may appear to have a greater value component than a risk assessment paradigm, this now appears to be more an apparent than a real difference. Lying behind both risk assessment and the PP are somewhat different normative pictures of the world. The risk assessment paradigm suggests that the status quo ante is acceptable, valuable, perhaps even prized, until risks are discovered. There is also a tendency for those advocating risk assessment to insist on clearly identifiable risks and to require their documentation with considerable empirical data. This requirement to document clear risks tends to elevate the importance of not disturbing the status quo until the documentation is complete. In addition, there appears to be a strong temptation for some risk assessment advocates to demand quite certain and very strong scientific evidence to support the assessment of a particular risk. This would more strongly preserve the status quo. Demanding strong empirical justification can be and has been utilized as a weapon against legal action, resulting in regulatory paralysis by analysis (Markowitz & Rosner, 2002, p. 203).

13 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 271 Beyond these points, return to one of the criticisms of PP offered at the outset, namely, that it does not require actual evidence of harm. Here there is a substantial difference between PP and a view that requires documentation of clear harm as proposed by Cross (1996). Demands for clear harm suggest that the status quo ante is so highly elevated in importance that nothing short of clear harm provides a reason for disturbing the status quo. The status quo is so important that it must be preserved even if, for example, humans that work in close proximity to benzidine-based dyes that are probable human carcinogens need not be protected, until there is clear evidence of harm to them from the dyes. Something else about the status quo is so prized in such a case that the workforce need not be protected until harm to actual persons results; this is far from a precautionary approach to protecting workforce health. It now appears that the PP is no more normative than a risk assessment paradigm; it is just that the norms of PP are carried more overtly on the face of the principle. Both have presuppositions of valued states of affairs; both address the status quo, but perhaps differ in their judgments of the extent of the threat posed by substances, in the value placed on the status quo and in how much scientific evidence it takes to justify disturbing the status quo. Moreover, the two approaches differ in what they value and, very likely, the value placed on the particular things valued. Each sees something threatened in the status quo. Those advocating risk assessment see threats to commercial activity, to particular products that might be threatened, to the value of a predictable status quo, to a market in which firms are free to innovate and distribute products in commerce until risks or harm result, or to something vaguer like overall social benefits or social progress. Risk assessment advocates often indicate that they want social policy appropriately based upon science. Do they so highly value a particular approach to science-based regulation that this is favored over health protections? 3 However, even if this were their real reason (and sometimes it appears not to be), we should examine the presuppositions of their views and these presuppositions indicate that such views have deeper value commitments. 4 Advocates of the PP, by contrast, appear to be concerned, on the basis of somewhat different evidence, about threats to public or workforce health or adverse impacts on the environment, and sufficiently concerned that they are willing to take cost-effective action based on less than full certainty concerning particular substances. The argument just presented may be much too charitable to risk assessment advocates, however. Consider how these differences come to light concerning post-market regulatory strategies. As noted above, the vast majority of

14 272 CARL F. CRANOR substances enters commerce with little or no testing and remains in commerce until harm or risks are established by appropriate evidence. The PP implicitly calls attention to the costs of removing uncertainty; agencies need not wait for full scientific certainty before taking action (hardly a strong requirement). There has been a temptation under post-market statutes to require greater certainty of a risk or harm before agency action is taken. This encourages greater analysis and study of a problem, as well as more data in support of an agency action. The PP, by contrast, while not being precise about when action is justified, nonetheless issues a general caution against requiring too much certainty before providing protections against a serious threat. In the environmental and public health arenas, as well as in science more generally, there will always be uncertainty; it is pervasive. PP and risk assessment paradigms have different attitudes toward these uncertainties as well as different views of the values at stake, as noted above. The PP urges addressing threats to valued states even though not all scientific uncertainty is removed. Risk assessment has a different approach: other aspects of the status quo are so valuable, and not so obviously threatened by a substance, that more certainty must be demanded before governmental action is taken to address any risks posed by it. There can be mistakes from either taking or not taking action under uncertainty. Taking regulatory action may result in regulatory false positives treating non-toxicants as toxic or in over regulation (regulating a known toxicant overly stringently for the actual threat it poses), while not taking regulatory actions may result in regulatory false negatives (failing to regulate a toxicant) or in under regulation (having too little regulation for the risks and harms actually posed by a toxicant) (Cranor, 1993). The PP can remind us that there are costs to removing scientific uncertainty. Taking extensive time to remove uncertainty for substances regulated under post-market statutes (a) permits them to continue in commerce, and (b) takes time, personnel, and resources from other harms or risks that merit attention. (Cranor, 1993, p. 146) Most commentators (and risk assessors) recognize the first problem and some exploit this point, but the second is more hidden. IV. LEARNING FROM THE LAW As I have argued elsewhere, we can learn from several legal devices about how to address uncertainty, as well as the division of labor between facts and values

15 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 273 in decision-making contexts (Cranor, 2001, p. 319). Once these legal devices are reviewed, it becomes clear that the PP is not unusual compared with other principles we utilize in our legal system. The law is simply one normative system that provides a variety of models for how much and what kind of evidence needs to be provided before taking some legal action (Cranor, 2001, p. 319). Legal burdens of proof, legal standards of proof, and legal presumptions are three ideas that provide approaches to coping with uncertainty. Since I have argued for this in greater detail elsewhere, they are merely summarized here (Cranor, 2001, p. 319). Legal burdens of proof are devices to predispose legal outcomes against one party to a dispute until that party offers sufficient contrary evidence (Cranor, 2001, p. 319). In civil litigation the plaintiff has the burden of proof to show that the defendant s action harmed her, while in the criminal law, the state has the burden to prove that a defendant is guilty of a crime. A burden of proof is a normative statement that in effect preserves the legal status quo ante until sufficient evidence has been offered to persuade the decision-maker or until sufficient uncertainty has been removed (Cranor, 2001, p. 319). Legal standards of proof specify the degrees of certainty that a decision maker must have before judging that a party with a burden of proof has carried the burden (Cranor, 2001, p. 319). In the criminal law, the state must establish claims for which it has the burden of proof beyond a reasonable doubt, while typically in civil litigation, plaintiffs must show their claims by a preponderance of the evidence (Strong, 1992, p. 952, 964). A scientific analogue that is similar to the criminal law s beyond a reasonable doubt standard of proof is an implicit standard of proof that statistical significance for scientific studies must be established by 95% confidence (95% statistical significance seeks to rule out by 95% confidence that statistical chance is an alternative explanation to a true positive result in a statistical study). Legal presumptions impose required inferences on one party or the other or shift the burden of producing evidence once certain required facts have been established (Strong, 1992, p. 966). Rebuttable presumptions specify that the establishment of fact B is sufficient to satisfy a party s burden of producing evidence with respect to fact A, unless the opposing party establishes evidence to the contrary (e.g., possession of heroin is sufficient to establish trafficking in heroin) (Strong, 1992, p. 965). Non-rebuttable presumptions do not permit rebutting evidence, in effect establishing a rule of law (Strong, 1992, p. 966).

16 274 CARL F. CRANOR Burdens of proof, standards of proof, and legal presumptions have all had normative justifications developed for them over the years (James & Hazard, 1977). Sometimes these might have been explicitly articulated in legislation or legal decisions; sometimes the justification might only have been seen in retrospect. Burdens of proof specify who must bear the risk of losing if sufficient uncertainty is not removed about a legal decision. Standards of proof indicate how much uncertainty must be removed before a decision may be legally taken. Presumptions predispose legal decisions for specified parties to disputes for some particular normative or social purpose, e.g., the presumption that a letter that has been mailed is presumed to have been delivered aims to limit debates about the delivery of legal documents. The PP might be seen as incorporating aspects of all of these legal devices. Although it leaves the burden of proof on an advocate of acting against a serious threat to the public or workforce health or the environment, it does vaguely specify a slightly reduced standard of proof before cost effective action can be taken. Moreover, it appears to presume that threats of serious or irreversible damage to valued states of affairs create a reason for considering action to address the threat. Nonetheless, it remains quite vague on these several dimensions. In this, it is hardly as radical as the Food, Drug and Cosmetic Act that implicitly presumes that drugs pose risks to human health until established otherwise in pre-clinical and clinical trials and before agency approval is given (U.S. Department of Health and Human Services, 1986). There is an additional reason for considering the similarities between PP and these legal devices. The legal devices explicitly or implicitly have normative justifications for their use and application to particular issues; they explicitly call for justification. So does the PP. By contrast, risk assessment is often articulated as if it does not need normative justification. However, as the argument to this point has suggested, it should now be clear that we should have an explicit normative debate about the implicit normative presumptions, burdens of proof and standards of proof associated with each of the approaches. Were this done, it would become clearer that each has normative commitments, and we could consider which approach with its normative commitments we want to guide which areas of our lives. Finally, in light of the above we should note that some seem to suggest there is something generally illicit or suspect about the PP. However, as I have just argued, in effect just as we make normative decisions in the law about who has the duty to remove uncertainty, how much uncertainty must be removed, and what presumptions might be socially justified, in a democracy the citizenry are

17 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 275 entitled to make these judgments about protecting the public and workforce health, as well as the environment, if it judges these are sufficiently threatened. Moreover, different countries may make these judgments differently, e.g., the Northern European countries appear to be much more concerned at present about such threats than the U.S., judging from the fact that the PP has its origins and considerable use there (Cranor, 2001). V. POSSIBLE STATEMENTS OF THE PP AND AN APPLICATION With the legal devices presented above we can reformulate the U.N. version of the PP by replacing lack of scientific certainty with two different burdens of proof. PP1: Where there is scientifically credible evidence of threats of serious or irreversible damage, failure to establish the threat of damage beyond a reasonable doubt shall not be used as a sufficient reason for postponing cost-effective measures to prevent environmental degradation. PP2: Where there is scientifically credible evidence of threats of serious or irreversible damage, failure to establish the threat of damage by a preponderance of the evidence shall not be used as a sufficient reason for postponing cost-effective measures to prevent environmental degradation. (Cranor, 2001, p. 321) The first principle seems easy to comply with and hardly controversial; after all we take many actions in our lives involving risks for which we do not insist on having evidence of proof beyond reasonable doubt. PP2 may be somewhat more subject to debate. That is, just because the balance of evidence cannot establish the threat of harm, this would not be a sufficient reason for inaction concerning the status quo according to PP2 (Cranor, 2001, p. 321). Thus, simply credible evidence of threats would be sufficient for taking action. Such a principle would no doubt be reserved for highly valued things around us or for very serious threats. Both principles merely remove reasons for action and do not state very precise or affirmative reasons for taking action. I have suggested more positive versions of the two principles might be the following:

18 276 CARL F. CRANOR PP1 0 : Where threats of serious or irreversible damage are supported by a preponderance of the available evidence, this provides a presumptive reason to take cost-effective measures to prevent environmental degradation (to change the status quo ante to prevent degradation). PP2 0 : Where threats of serious or irreversible damage are supported by credible scientific evidence and this is appropriately verified, this provides a presumptive reason to take cost-effective measures to prevent environmental degradation (to change the status quo ante to prevent degradation). (Cranor, 2001, p. 322) Each states a rebuttable presumption to try to prevent the threat in question from materializing into harm, with the second principle more protective than the first. One would need to consider different circumstances and objects to be protected in order to implement them. In the spirit of these principles, there might be patterns of evidence that would provide scientifically credible threats of human health or environmental harm to guide scientists and regulatory bodies. Consider just one example, a class of little studied substances that appear to have toxicity mechanisms similar to known toxicants: polybrominated diphenyl ethers (PBDEs). PBDEs have a variety of uses in electronics, building materials and plastics, largely as flame-retardants. Worldwide production exceeds 80,000,000 pounds with more than 1 million pounds produced in the U.S. in some recent years (Hooper & McDonald, 2000). Their volume of production alone might provide a reason for greater study (Cranor, 2003b), but more worrisome is the fact that they appear to be chemically very stable and to bioaccumulate, much like DDT, polychlorinated biphenyls (PCBs) and dioxins and are highly likely to migrate like several of these (Hooper & McDonald, 2000). Their structural and biological activity resembles these better known classes of substances; they appear to act by similar mechanisms, and appear likely to have similar toxicity properties, e.g., likely to have carcinogenic, developmental and thyroid toxicity potential (Hooper & McDonald, 2000). Moreover, according to recent reports, PBDEs are rising rapidly in the breast milk of mothers with the highest concentrations seen in the U.S. (Cone, 2003). There is, thus, considerable data that these substances likely pose problems similar to PCBs, dioxins and DDT. What should be our approach to them? Should we insist on documenting risks or even harm for each different congener of PBDEs, e.g. PBDE-47 compared with PBDE-99, as risk assessors might be tempted to do (some congeners will likely be more toxic than others).

19 TOWARD UNDERSTANDING ASPECTS OF THE PRECAUTIONARY PRINCIPLE 277 Or does the body of evidence accumulated over thirty or more years about quite similar compounds provide sufficient evidence of a threat to justify taking social action on these substances without congener-specific data about whether and to what extent they pose specific risksto human health and the environment? Another possibility might be to address all PBDEs as a class, unless there are very good and compelling reasons for exempting a particular congener (it is obviously non-toxic and for some reason does not bioaccumulate, there are no benign alternatives to it and it is critical to the world economy for some reason). Does the scientific community know enough about them for different regulatory purposes? Moreover, as recently as three years ago, the U.S. did not have a breast milk-monitoring program, which might document the extent to which these substances have bioaccumulated in women s bodies (Europe has such data). As already indicated, spotty monitoring in the U.S. now suggests that U.S. women have among the highest body burdens of PBDEs (Cone, 2003). Anticipatory foresight and research (being alert for substances that are stable, bioaccumulate and toxic) might have led to earlier identification and, one hopes, to avoidance of these potential problems. VI. CONCLUSION Formulating defensible and more specific versions of the PP will not be problem free and much remains to be done. Moreover, they may not always be applied without mistakes or unintended consequences. However, we can learn much from reflecting on the generic idea of precaution as well as various versions of the PP and their presuppositions and assumptions, comparing them with risk assessment with its various presuppositions and assumptions, and with approaches taken in the law to address problems of uncertainty. When we make these comparisons against the background of what is currently known about ecosystem status quos and the extensive ignorance about the universe of chemical substances, perhaps the PP has plausible statements and applications without seeming as radical as some have suggested. ACKNOWLEDGEMENTS This paper was written while the author was supported by research grants from the National Science Foundation (Grant # ) and the University of California Toxic Substances Research and Teaching Program. I am grateful to David Strauss for research assistance on this paper.