by Fabio Tollon
One of the cornerstones good of science is that its results furnish us with an objective understanding of the world. That is, science, when done correctly, tells us how the world is, independently of how we might feel the world to be (based, for example, on our values or commitments). It is thus central to science, and its claims to objectivity, that values do not override facts. An important feature of this view of science is the distinction between epistemic and non-epistemic values. Simply, epistemic values are those which would seem to make for good science: external coherence, explanatory power, parsimony, etc. Non-epistemic values, on the other hand, concern things like our value judgements, biases, and preferences. In order for science to work well, so the story goes, it should only be epistemic values that come to matter when we assess the legitimacy of a given scientific theory (this is often termed the “value-free ideal”). Thus, a central presupposition underpinning this value-free ideal is that we can in fact mark a distinction between epistemic and non-epistemic values Unfortunately, as with most things in philosophy, things are not that simple.
The first thing to note are the various ways that the value-free ideal plays out in the context of discovery, justification, and application. With respect to the context of discovery, it doesn’t seem to matter if we find that non-epistemic values are operative. While decisions about funding lines, the significance we attach to various theories, and the choice of questions we might want to investigate are all important insofar as they influence where we might choose to look for evidence, they do not determine whether the theories we come up with are valid or not.
Similarly, in the context of application, we could invoke the age-old is-ought distinction: scientific theories cannot justify value-laden beliefs. For example, even if research shows that taller people are more intelligent, it would not follow that taller people are more valuable than shorter people. Such a claim would depend on the value that one ascribes to intelligence beforehand. Therefore, how we go about applying scientific theories is influenced by non-epistemic values, and this is not necessarily problematic.
Thus, in both the context of validation and the context of discovery, we find non-epistemic values to be operative. This, however, is not seen as much of a problem, so long as these values do not “leak” into the context of justification, as it is here that science’s claims to objectivity are preserved. Is this really possible in practice though?
It seems both natural and acceptable that scientists choose certain research topics because they find them interesting, and that institutions fund and promote research in fields that are deemed important for political or social reasons. The value-laden nature of these contexts is all very well, again, so long as they do not come to bear on the context of justification. Justification here refers to how we assess scientific theories, and one way of defining it would be to say that it is based on empirical evidence in conjunction with various epistemic values. Together, these two elements, however they are understood, combine to aid us in assessing the scientific merit of a given theory. One problem with this, and the problem I will focus on, concerns the influence of non-epistemic values in the context of discovery, and the fact that these non-epistemic values amount to more than just the selection of a research topic. In this way there is not such a clear distinction between epistemic and non-epistemic values, which would undermine the value free ideal.
Non-epistemic values not only impact the selection of research topics, but also influence how this research is in fact carried out. How a research question is framed, and which aspects of it are deemed significant, are value-judgements. “Significance” is not something we find out in the world but rather something we attribute to things in the world, and the reasons for such an attribution could be social, financial, religious, etc. The point is, it need not be (in many cases is not) epistemic (although of course it could be). This becomes especially salient in the era of Big Science, where scientists are not limited to working in siloed academic laboratories but sometimes find themselves in large-scale organizations with political backing (and political agendas).
And this feature of science, well, is significant. This is because scientists cannot disinterestedly investigate every aspect of the world and must be by necessity picky. And so, which hypotheses are developed and tested, and which data are generated, are all influenced by these non-epistemic values. This is not to say that scientific theories are determined by these values, but rather to emphasize how these values can come to guide and shape research programs and the eventual support of specific theories.
Going back to the context of justification, which you will recall, was concerned with the evidential support for a theory and various epistemic values. However, if, as I have argued above, the data that are generated and the research directions that are pursued (and hence the available theories) involve non-epistemic values, then the context of justification also seems to involve such values. Essentially, we will always have “blind spots” when deciding on the best theory that explains some phenomena. The value-laden nature of the context of discovery prohibits us from coming up with theories that might have been viable alternatives to the one chosen, and so there is really no “value free” part of scientific investigation. Thus we see how a central assumption underpinning the value free ideal (that of an independence of justification from discovery) is not maintained in practice.
This, however, does not undermine science’s objectivity. I have not dealt with the empirical adequacy claim in the context of justification, and there is much to be said about how this constraint supports claims to objectivity in science. However, what this argument does do is shine light on how the idea of a “pure” science free of values is untenable. Even if you are unconvinced by the argument wholesale, it seems right (at least socially and politically) that we do not constrain our choices of research lines on epistemic grounds alone. Such a constraint seems far from what actually occurs in science, and, moreover, could be seen as ethically indefensible as new social issues arise and demand scientific investigation.