330 335 Kinzig

Throughout their history, humans have devised a number of approaches and strategies to reducing (perceived) uncertainty or the psychological impacts of uncertainty, and to function and thrive in a partially unknown world. A certain world begets certain ‘best strategies’ for survival; an uncertain world necessitates a larger portfolio of options and approaches, and means for choosing among them. We cannot hope to cover the whole of human history here. Instead, we focus on the ways in which uncertainty and approaches to reducing uncertainty have played out in recent history in two arenas; science and policy. We focus on these because it is the interplay between these that has relevance for global environmental futures, and because the characterizations of uncertainty, and strategies for coping with it, differ so dramatically between the two. The scientific process is built on the goal of advancing knowledge; penetrating and reducing the reaches of what is not known. Each advance is built on knowledge acquired earlier. The cost of ‘incorrect knowledge’ is therefore quite high, affecting not only that building block in the foundation but those that follow. Science has therefore evolved procedures whose primary goals are to protect against being wrong. The policy process is built on the goal of addressing societal ills or challenges. Timeliness is frequently of the essence; action may have to precede knowledge. The errors that need be avoided are those associated with undue political or social costs, including harm to constituents, the economy, national security, or the environment. Because of these fundamentally different goals, science and policy have evolved significantly different ‘evidentiary standards’. Scientists apply relatively high evidentiary standards—insisting, for instance, that there be only a small probability (sometimes 5% but frequently lower than 1%) that an incorrect conclusion has been drawn. This means that many conclusions concerning the workings of the world that have a reasonable probability of being correct—say 75%—may not pass muster. The policy process employs both stricter and looser evidentiary standards, largely based on the (perceived) costs of being wrong. If we are certain a patient is going to die shortly, for instance, there is little hazard in prescribing a drug whose efficacy is largely unknown, but offers some hope of life extension. Here the unknown interacts with our perceptions of risk in such a way as to allow us to employ relatively loose evidentiary standards. A drastically different evidentiary standard should apply for instance, in the case of cosmetics, where society insists on a vanishingly small chance that one might die from the application of mascara. The manufacturer must demonstrate the highest confidence that the product works as intended and is very unlikely to have undesirable side effects. Difficulties arise due to these different standards in 4 ways: i) Failing to recognize that different standards in policy and science exist; this can result in a failure to communicate properly. Policymakers may believe that scientists are using a stricter or, more likely, looser evidentiary standard than is actually being employed. Thus, an outcome that a scientist believes is highly certain may be (mis)interpreted as only probable or likely in policy circles. ii) Difficulty arises because scientists may fail to provide information that could be considered useful or even crucial in the policy arena, in large part because they may have failed to study the appropriate phenomena. The success of scientists depends on their ability to advance their field under the relatively strict evidentiary standards that apply. This selects for fairly reductionist approaches to studying phenomena whose drivers can be tightly controlled and manipulated. Until recently, this meant largely avoiding study of precisely the kinds of environmental systems that society is influencing most profoundly—global climate systems, entire ecosystems or landscapes. Fortunately, our capacity to study these complex, and often highly uncertain, systems has advanced tremendously, but the strong scientific tradition of applying strict evidentiary standards can still delay the introduction of information that could be considered crucial by policymakers. iii) Scientists may find it difficult to quantify the uncertainty associated with highly uncertain, and to some extent unknowable, futures. This has led, in some circles, to an avoidance of characterizing uncertainty at all. Witness, for example, the recent emissions scenarios of the IPCC, each of which was presented without an assessment of likelihood, and each of which was thus treated as equally likely by the public and policymakers. No rational assessment of probabilities of the various scenarios would conclude equal likelihood; rather such a process would conclude (potentially drastically) different likelihoods for each scenario, albeit with very low confidence (1). We will discuss the need for likelihood assessment and relevant methodology in the next section. iv) The neutral language of evidentiary standards can be used to mask what is ultimately a debate over values. This tactic is not limited to the (supposedly subjective) policymaker, but is employed by the (supposedly objective) scientist as well. Most if not all of the environmental problems faced by society are complex all the way down. They are rife with uncertainty and there can be numerous plausible solutions, leading to numerous possible futures. These futures will differ in terms of environmental protection, social justice, economic growth, and political freedoms, among other things. Scientists can help illuminate those futures and trade-offs, but they are no more expert than Coping With Uncertainty: A Call for a New Science-Policy Forum Ann Kinzig and David Starrett et al.