AAP 2026 CONFERENCE

In this paper I provide a novel argument against the Value-free Ideal (VFI) and explore some of its implications. I begin by arguing that no existing critique of the VFI targets the relations of inductive support between evidence and hypotheses (relations of confirmation). In fact, many critics of the VFI, like Heather Douglas, explicitly state that relations of confirmation remain value-free (Douglas, 2000, p. 656). However, I argue that confirmation is value-laden. After briefly surveying different inductive logics, I claim that the best prospects for a value-free account of confirmation rely on the probability calculus. For these accounts to be value-free, two conditions must be met: (1) probabilities must themselves be value-free, and (2) the choice of confirmation function (a function of probabilities) must also be value-free. Condition (1) fails because all interpretations of probability face the reference class problem (Hájek, 2007), and choosing a reference class requires non-epistemic value judgements. Condition (2) also fails, since the choice of confirmation function is underdetermined by evidence, theory, and epistemic values, requiring further non-epistemic value judgements. Hence, confirmation is value-laden. I conclude by exploring the implications of this conclusion for contemporary defences of the VFI (e.g. Menon and Stegenga, 2023).

REFERENCES

Douglas, H. (2000). Inductive Risk and Values in Science. Philosophy of Science, 67(4):559–579.
Hájek, A. (2007). The reference class problem is your problem too. Synthese, 156(3):563–585.
Menon, T. and Stegenga, J. (2023). Sisyphean science: why value freedom is worth pursuing. European Journal for Philosophy of Science, 13(4):48.