What should the interplay between scientists and policy makers look like? Should it be permitted to base a policy on uncertain knowledge? One of the speakers at this year’s Conference on Foreign Affairs was Sven-Ove Hansson, professor in philosophy at the Royal Institute of Technology in Stockholm, who helped the audience make sense of these questions. This article sums up some of the most important points from his talk.

Scientists and policy makers often look at the world in quite different ways. Policy makers put up guidelines to help achieve certain outcomes while scientists try to find out how things really are.When a policy decision is made on an issue that bears on scientific knowledge, there is a coordination problem between policy makers and scientists: Scientists first gather data, and when there’s enough evidence to draw a certain conclusion, that conclusion becomes part of the corpus – the scientific knowledge which at a given time could be written in a textbook without reservation.Then there’s the task of policy makers: to decide on a policy given the scientific evidence. The reality, however, isn’t always as clear-cut.

Whenever scientists do science, there’s always some uncertainty involved; in some issues more than in others. Especially if they are dealing with scientific knowledge that is not part of settled science, the corpus, there sometimes exists a rather large interpretative scope within which scientists believe that the truth lies. This interpretative scope should be treated as a range of possibilities that different scientists believe to be true, and not–as policy makers sometimes do–as a spectrum within which one can freely cherry-pick whatever data that happens to suit one’s preferences.

Partly as a consequence of policy makers cherry-picking data for their arguments, some have argued that policy decisions should only be based on knowledge that is part of the corpus, i.e. complete knowledge. But basing decisions on incomplete data is a common thread in policy making, and not allowing more uncertain knowledge as a basis for policy decisions doesn’t seem reasonable. Consider a case where you are deciding whether a potentially harmful chemical that leaks from feeding bottles should be legal or not. Let’s also assume that half of the scientists researching this chemical think that it’s harmful while the other half think it’s not. This scientific knowledge clearly doesn’t belong to the corpus, but does that mean you shouldn’t make the decision? Given the potential harm of the substance, it seems reasonable to use knowledge that hasn’t passed the entry requirements for the corpus in this case.

Another problem is that policy makers sometimes disregard the expected value of a decision. The expected value is a weighted average of the possible outcomes of a decision, and is often used by economists to decide which decision out of many possible ones is favorable. The greater the expected value is, the more favorable is the decision. But policy makers often neglect the expected value if the probability of an outcome is small. If given the choice between an option that with a small probability has very bad consequences and an option that with certainty has slightly bad consequences, policy makers often treat the small probability as zero, and conclude that the first option is preferable. Even when the expected value is greater for the second option, policy makers sometimes prefer the first option.

Acting rationally, and in accordance with one’s goals, can sometimes be a hard task. In policy making, part of the problem is that the interaction and division of labor between scientists and policy makers isn’t well-defined. By making these distinctions clear we can hopefully witness a more fruitful interplay between these professions in the future.