The 5th Food Packaging Forum (FPF) workshop “Scientific challenges in the risk assessment of food contact materials” took place on October 5th, 2017, in Zurich, Switzerland. The talk by Karim Bschir from the Department of Humanities, Social and Political Sciences, ETH Zurich, Switzerland, offered a philosophical insight into why good science is not, and does not have to be, ‘value-free.’
The ‘value-free’ ideal of science postulates that “the goal of science is to produce robust, objective knowledge about empirical reality,” therefore the produced results “should not be influenced by social or moral values,” Bschir explained. However, in his talk he showed that scientists do make value judgments on a regular basis. Therefore, Bschir argued, “the value-free ideal of science has to be rejected,” because “scientific objectivity need not preclude value judgments.” However, to ensure that their independence and credibility are not compromised, scientists should be requested to “make their values explicit,” Bschir proposed. Furthermore, scientists should be held at least “morally responsible for foreseeable harmful consequences of potential errors.”
Scientists are often forced to make normative (as opposed to purely descriptive) statements about their work, and the justification of normative statements does involve the consideration of values. Furthermore, as a consequence of the nature of research and evidence collection methods applied in empirical sciences, most of the inferences made by scientists based on collected data are per se inductive, not deductive. This tremendously increases the chance of coming up with a wrong judgment, because there is an inherent risk related to inductive inference. The consequence of this, formulated in 1953 by Richard Rudner in his “The scientist qua scientist makes value judgements” article, states that, when deciding whether to accept or reject a hypothesis based on the (necessarily limited) evidence available, the scientist inevitably has to rely on subjective values. For “how sure we need to be before we accept a hypothesis will depend on how serious a mistake would be,” Rudner wrote. Rudner’s work has been taken up and further developed by Heather Douglas in her book “Science, policy, and the value-free ideal.”
Bschir summarized that scientists’ daily work concerns accepting or rejecting hypotheses based on evidence. This act involves a “decision as to when the evidence is strong enough.” Because the inference that needs to be made is usually inductive, this decision suffers from an inductive risk of making a mistake. Therefore, the process of meeting one or another decision always involves the “consideration of [the] consequences of potential errors.” In the case of toxicology, there is a chance that errors made may lead to “serious foreseeable consequences.” Thus, here scientists need to make value judgments in coming up with normative considerations regarding the acceptance or rejection of a postulated hypothesis. Particularly for data of a continuous type, where borderline cases are quite frequent, the data necessarily need to be interpreted, and “interpretation of data involves judgment.” Further, “the interpretation of empirical results can change depending on background assumptions,” and choosing these background assumptions is again “not value-free.” Thus, the fact that “good science is not value-free” has to be acknowledged and properly dealt with.
Karim Bschir (October 5, 2017). “Why good science is not value-free.” (Youtube)
Karim Bschir (October 5, 2017). “Why good science is not value-free.” (pdf)
Richard Rudner (1953). “The scientist qua scientist makes value judgements.” Philosophy of Science 20:1-6.
Heather Douglas (2009). “Science, policy, and the value-free ideal.” University of Pittsburgh Press, published July 2009.