Trust, Belief, and the Second-Personal

Abstract

Cognitivism about trust says that it requires belief that the trusted is trustworthy; non-cognitivism denies this. At stake is how to make sense of the strong but competing intuitions that trust is an attitude that is evaluable both morally and rationally. In proposing that one's respect for another's agency may ground one's trusting beliefs, second-personal accounts provide a way to endorse both intuitions. They focus attention on the way that, in normal situations, it is the person whom I trust. My task is to develop an account of the latter insight without the controversial theoretical commitments of the former. I propose a functional account for why the second and third-personal ‘systems’ operate not just in parallel, but in tandem, in support of a cognitivist account of trust.

KEYWORDS:

• Trust
• belief
• second-personal
• assurance view
• cognitivism
• functional account

Notes

¹ I defend these claims elsewhere [Simpson 2017].

² As well as the considerations in Interpersonal, non-cognitivists are often impressed by the voluntariness of trust. If one can decide to trust in a way in which one cannot decide to believe, Rationality should be rejected: see, e.g., Holton [1994], Faulkner [2011: 149, 2014a: 1979], and Hawley [2014b: 2030]. The voluntariness of trust is not a fundamental source of tension with Rationality, however, unlike Interpersonal. What those reasons are to which trust is responsive is a separate issue from whether trust is voluntarily responsive to those reasons. For instance, doxastic voluntarists take belief sometimes to be permissibly voluntary. This provides a way to reconcile the voluntariness of trust with Rationality (see Frost-Arnold [2014]). It also does not follow from denying Rationality that trust is voluntary. For one may lack voluntary control over one's practical attitudes, in particular the practical attitude of trust [Jones 1996: 18; McMyler 2017].

³ Darwall denies that testimonial belief is based on second-personal reasons [2006: 9–12, 56–7]. His [2017] account of trust as ‘an attitude of the heart’ is also non-cognitivist.

⁴ Moran is a suggestive and elusive writer, and elements of his [2005] have been developed in distinct ways. Faulkner [2011] shares a claim about reflexive recognition, where the hearer's recognition of the speaker's intention or motivation in part grounds his trust, with Hinchman [2005]. Hieronymi [2008] and Marušić [2015] argue for an interpersonal extension of an intrapersonal claim, that one cannot view one's own commitment as evidence. McMyler [2011] focuses on the transfer of responsibility. Ross [1986] is a precursor. The relation of Moran's [2013] to his [2005] is a further interpretive task.

⁵ Hawley [2014a] argues that all instances of trust involve ascribing a commitment to the trusted. As further support for this extension, proponents of second-personal accounts since Moran—notably Hieronymi, McMyler, and Marušić—apply them to trust (although not Faulkner [2014b]). Linguistic data support the extension. Like ‘believe’, the verb ‘trust’ may take either a person or a proposition as its object [McMyler 2011: 113ff.].

⁶ What follows is not a rebuttal of all second-personal accounts of trusting beliefs. I rebut arguments for the problem of harmony that are based on the right to complain. But there are (or could be) arguments for the problem of harmony based on other premises. Much of Moran [2005] develops an argument that the value of evidence is diminished when it is intentionally produced. It is endorsed by Faulkner [2011: 62–5] and replied to by Keren [2012], in my view soundly. Leonard [2016] targets the reflexive-recognition argument noted earlier (note 4).

⁷ It is an implication of my account that disbelief is normally insulting. See section 4.

⁸ Schmitt [2010: 232] also makes the point.

⁹ Samuel Bowles and Herbert Gintis marshall the impressive empirical evidence for these claims [2011: 19–45]. The findings hold cross-culturally.

¹⁰ Imagine that someone overhears a speaker talking to an addressee. On Hinchman's view [2005: 565], the addressee but not the over-hearer can insult the speaker by not believing her. On my view, both the over-hearer and addressee can insult the speaker by disbelieving what she says. But, to the extent that the speaker commits to its being the case that p only to the addressee, only the addressee can refuse to believe the speaker by refusing to accept her commitment.

¹¹ A full defence of cognitivism about trust must address the substantive considerations in favour of non-cognitivism, derived from scenarios like the ex-convict on the till. In outline, one can deny that is trust, saying that it is only acting as-if you trust [Hardin 2002: 73–5]; or affirm it as a lesser form, compared to the fully-fledged-believing form [Hieronymi 2008]; or be pluralist [Simpson 2012].

¹² For discussion and criticism of earlier versions, I am grateful to Tim Crane, Katherine Hawley, Jane Heal, Alex Heape, Richard Holton, Jeroen van den Hoven, Hallvard Lillehammer, Beri Marušić, Ben McMyler, Alex Oliver, Rik Peels, audiences in Cambridge, Frankfurt, and Notre Dame, and especially to the AJP editor and two anonymous referees.