Decisionmaking under uncertainty Is there any other kind

Decision-making under uncertainty: Is there any other kind? Naomi Oreskes History Department and Science Studies Program University of California, San Diego

“Decision-making under uncertainty” implies the existence of an alternative Presumably: decision-making under conditions of certainty

No such thing Statistician George Chacko (1991) defines decision-making as “the commitment of resources today for results tomorrow. ”

Because decisions involve expectations about the future, they always involve uncertainty

If people talk about “certainty” they can only be referring to certainty about what they want the outcome to be (desires)

Why would anyone even imagine certainty is possible? • Decision-making involves premises (assumptions, beliefs, conditions) • Logic tells us that if premises of a conditional statement are correct, then outcome is known (predictable). • Common assumption: the premises are correct. (Least examined aspect)

In environmental decision-making, the premises typically include “underlying scientific information” Examples? CO 2 is a greenhouse gas Lead is a neurotoxin Marine biodiversity is declining

We take these things to be true, and I’ve chosen examples that I think are true. But experience proves that widely accepted premises may turn out to be incorrect

Premises can always be disputed. • Marine biodiversity is declining Ransom Myers and Boris Worm, 2003. "Rapid worldwide depletion of predatory fish communities. " Nature 423: 280 -283. • Is it really? Or is it just a handful of heavily hunted fish and mammals species? • Response: Sloan Foundation Census of Marine Biodiversity. More information to test premise.

Certainty is a false idol Why would anyone imagine that scientific knowledge could be certain?

Erroneous and refuted conception of science: Positivism

Sea of “positivist expectation” Certain knowledge based on 1. Observable foundations 2. Verifiable implications

No need to be cruel… Positivist aspirations were laudable enough: Science as alternative to superstition, clericalism, confusion.

Positivists asked important questions • What aspects of scientific investigation account for the reliability of the knowledge produced? • Can those elements be adopted by others wishing to increase the robustness of their own investigations? • Can these elements be used as a criterion for judging information?

But the vision failed • Historically: it fails to account for major conceptual revisions in science. • Philosophically: it fails to account for the diversity of scientific methods and the flexible interplay between theory and observation. • Sociologically: it fails to account for the social dimensions of scientific proof and persuasion. Verification of knowledge is a social process.

Alternative? Science as an intellectual and social consensus of affiliated experts

Scientific consensus achieved by 1. Consilience of empirical evidence, achieved by tested methodologies. 2. Coherence between evidential frameworks and theoretical understandings 3. “Theoretical integrity” (relation to existing beliefs and commitments) 4. Social organization for establishing and declaring agreement on all of the above.

Does this process eliminate uncertainty? Of course not.

So how do we judge consensus? How should we response to the presence of vocal dissenters, and imperfect data?

Need to reject a second false idol: “The Kuhnian expectation” • Kuhn’s famous paradigm concept accounted for the social dimensions of scientific consensus and the historical reality of conceptual change. • Left us with an incorrect impression of “normal science”: dissent- & anomaly-free

Science as dissent-free In Structure (1962) Kuhn wrote: “What is surprising, and perhaps also unique in…the fields we call science, is that…initial divergences…disappear to a very considerable extent, and then apparently once and for all. ” What characterizes--even defines --science is unanimity

Science as anomaly-free • Kuhn characterized normal science as (essentially) anomaly-free, with emergence of an anomaly as the beginning of crisis. • Most “problems” are viewed as “puzzles. ” When puzzle changes to an “anomaly”--> crisis --> revolution.

“When…an anomaly comes to seem more than just another puzzle of normal science, the transition to crisis…has begun. ”

Left impression that normal science involves few if any meaningful uncertainties. Just filling in details.

A very mistaken view, as erroneous and damaging as the positivist view Generates impossible expectations. Gives fodder for exploitation of dissent.

Alternative? Living with uncertainty

If uncertainty, anomalies, and dissent are normal science, how can we learn to live with them? I. A “reasonable expectations” model II. Taxonomy of uncertainties to help to identify useful courses of action.

A “reasonable expectations” model • Consensus unanimity • There always dissenters. – Better: “outliers” • There always anomalies • Anomalies and outliers can (and probably will) be exploited.

A taxonomy of uncertainty (preliminary) I. Science not generally accepted--active scientific debate by scientists II. Science mostly accepted by scientists, perhaps some outliers. III. Science contested by parties outside the scientific community

Appropriate responses depend on the situation

I. Area of active scientific debate Response: More research. While no guarantee, increased knowledge base has potential to increase technical consensus.

II. Science mostly accepted by scientists, with some outliers More scientific research is unlikely to decrease uncertainty. In fact, it may increase it. More information less uncertainty Existing consensus can be destabilized.

III. Science contested by parties outside the scientific community • The issues at stake are almost certainly not technical (moral, political, religious, aesthetic). • More technical research will not resolve disputes. • Inclusionary processes essential.

More science is unlikely to help us make our most important decisions

References • George Chacko, 1991, Decision-making under uncertainty: An Applied Statistics Approach, New York: Praeger 1991, quote on p. 5 • Kuhn, T. S. , 1962. The Structure of Scientific Revolutions, University of Chicago Press, quote on p. 17, 82.