The trust-transport model
The project starts from a formal model of how trust should move between domains: how a rational agent ought to revise trust in an expert in one field given gains or losses of trust in that expert elsewhere. The model rests on four weak requirements on rational trust. It bears on epistemic trespassing, on the difference between virtuous trust and either gullibility or excessive scepticism, and on how broken trust can be rebuilt by starting again in a nearby, lower-stakes domain. It also identifies a new failure mode, reverse trespass.
Over the year we’ll test and extend the model at each workshop, against both philosophical and empirical critique, and look for which of its conclusions hold up across variations.
Working papers
Trust Transport
Experts often opine outside their home domain of expertise. A Nobel Prize-winning physicist might give television interviews, or even write a book, about religion. Or skilled actors might try to sell you a diet by citing health benefits. How much should we trust them? Philosophers writing on “epistemic trespassing” – roughly: speaking outside of one’s area of expertise – agree that expertise in one field does not simply “carry over” to another, but nobody has said how much of it carries over, or why. This paper lays out an answer. As we’ll argue, two things settle the matter, though each as we’ll see is difficult to pin down: (i) how closely the two subject matters are related, and (ii) how broad a range of questions the expert has actually shown themselves to be reliable about. We attempt to make both ideas precise. Using David Lewis’s account of subject matters, relatedness becomes a measurable quantity, namely how much settling every question in one field would tell you about the other, and breadth becomes the generality of the subject matter an expert’s track record covers. From these we derive a rule for adjusting trust as an expert moves between fields, and prove that it satisfies four conditions any adequate account must meet, conditions that simpler proposals fail. We close by showing what the rule implies for epistemic trespassing and for how broken trust can be rebuilt.
Trust as Willingness to Gamble
In this paper we defend a novel account of trust according to which to trust someone is to be willing to stake costly consequences on their acting in ways to which they have committed. Equivalently, trust is the willing intent to gamble on the faithfulness of others, and how much you trust someone is how much their commitment moves the odds you will accept. This account is unexplored in the literature but powerful in its simplicity: it implies the existence of both therapeutic and non-therapeutic instances of trust as special cases; it naturally differentiates trust from mere reliance and explains why trust carries a necessary vulnerability to betrayal; it explains both the action-like character of trust (e.g. its voluntariness) and its belief-like character (e.g. its responsiveness to evidence); it provides a principled explanation for the “wrong kind of reason” phenomenon, i.e., why you can’t bribe someone into trusting you; finally, it yields a natural solution to the Meno problem of trust, or why it is that trust is so socially valuable when it looks like mere reliance might provide the same benefits: an unhedged stake on another’s faithfulness shows a confidence in them that isn’t shown by reliance alone.
More to come
Further working papers from the network are in progress and will be posted here as they’re ready. Early drafts are discussed at COGITO’s work-in-progress seminar in Glasgow.
Recordings
Recordings of the public talk series will be collected here.