WP5. Epistemic Risk and Trust

This fifth work package will deal with the notion of trust and its relation to the social production and distribution of knowledge. It will take into account two distinctions: first, the one between ‘mere reliance’ and trust (Baier 1986, 234; Hawley 2014: 1); and second, between the risk of epistemic betrayal risk and the risk of not compliance with one’s epistemic duties.

With respect to the former, a basic hypothesis of this work package is that epistemic rationality alone is enough, in principle, to justify relationships of ‘mere reliance’, whereas epistemic trust implies that they can be justified only if they fall short of epistemic rationality. In other words: agents are only relying on processes, devices or other agents when they depend on them merely on the basis of good evidence that they have that their behavior will be predictable. However, when an agent genuinely trusts others, she depends on them beyond what epistemic rationality alone may justify. Trusting is thus beyond any mere attitude of prediction, and essentially produces a kind of expectation for which we may only find rational explanations in the practical domain, but not in the epistemic one—or at least not in any direct way. Consequently, when some putative attitude of trust is in fact only based on purely epistemic reasons (such as future estimations on the likelihood that the trustee will act in certain ways), is not genuine trust, but just mere reliance—see Goldberg 2018; for related discussion of epistemic risk and trust in the context extreme betting, see González de Prado (forthcoming-a).

Regarding the second distinction, relationships of epistemic trust give rise to specific varieties of epistemic risk, such as the risk of epistemic betrayal and the correlative risk, on the part of the trustee, of not living up to the epistemic expectations of the trustor. Such risks arise as a consequence of the aforementioned feature of epistemic trust, which produces a certain gap between our rational expectations, strictly based on epistemic reasons and predictive attitudes, and the sort of expectations generated by the trustee’s commitment to the trustor. In such cases, agents may place excessive trust on their epistemic partners (testifiers, for sure, but also other partners in different epistemic practices), vis-à-vis their epistemic goals, which are unrealistic or not well calibrated. One key hypothesis to explore is that such varieties of epistemic risk are absent in cases of mere epistemic reliance, where the giver of knowledge, or partner in the epistemic quest, is not involved in the specific kind of moral relationship that trust involves, and which justifies the manifestation of what Peter Strawson famously called ‘reactive attitudes’, such as resentment, indignation, gratitude or forgiveness (see Strawson 1962, and also Jones 1996). A basic task of WP5 will be to connect this issue with Sosa’s discussion of high-order negligence and recklessness in belief and action, which we have labeled second-order veritic risk. Our hypothesis in this respect is that what is specific of cases of second-order

negligence in trusting is that it always occurs in a social context, giving rise to specific forms of social epistemic dependence, such as the ‘should have known’ phenomenon (Goldberg 2015), epistemic partiality—either in friendship (Stroud 2006) or partisanship (Rini 2017)—, and different problematic cases concerning our diachronic epistemic responsibilities (see Axtell 2011).

A proper defense of this hypothesis will require us to take a stance in the context of extant alternative theories of trust and some related moral concepts, such as blame or regret. In particular, we will take sides with positions that highlight the idea that relationships of trust essentially involve personal commitments, in contrast to views based on notions such as good intentions, obligations or virtues—see Hawley 2014 for a representative account of the former and relevant discussion of the latter. Our aim here is to draw some theoretical implications for the idea that the lack of fulfillment of those commitments gives rise to proper reactive reactions such as blame, which, as recently shown by Fricker (2016), has the social function of coordinating our moral perception of given scenarios. When applied to the epistemological domain, such a theoretical approach may yield the result, and we do expect that it will, that the alleged conflict between trust and epistemic rationality is merely apparent. In a slogan: trust pays epistemically back. Thanks to the epistemically unmotivated excess of epistemic dependence we incur in by epistemically trusting others, we contribute to the production of human capital, epistemically scaffolded (Greco 2015) or engineered (Goldberg forthcoming), i.e., environments where the production and flow of knowledge becomes much more efficient.