Abstract
According to doxastic accounts of trust, trusting a person to \(\varPhi \) involves, among other things, holding a belief about the trusted person: either the belief that the trusted person is trustworthy or the belief that she actually will \(\varPhi \). In recent years, several philosophers have argued against doxastic accounts of trust. They have claimed that the phenomenology of trust suggests that rather than such a belief, trust involves some kind of non-doxastic mental attitude towards the trusted person, or a non-doxastic disposition to rely upon her. This paper offers a new account of reasons for trust and employs the account to defend a doxastic account of trust. The paper argues that reasons for trust are preemptive reasons for action or belief. Thus the Razian concept of preemptive reasons, which arguably plays a key role in our understanding of relations of authority, is also central to our understanding of relations of trust. Furthermore, the paper argues that acceptance of a preemptive account of reasons for trust supports the adoption of a doxastic account of trust, for acceptance of such an account both neutralizes central objections to doxastic accounts of trust and provides independent reasons supporting a doxastic account.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Non-doxastic accounts need not deny that trusting often involves a belief about the trusted person or even that it is sometimes in virtue of having such a belief that one can be described as trusting. For example, Frost-Arnold (forthcoming) suggests an account according to which \(A\) trusts \(B\) to \(\varPhi \) if she either believes or accepts the proposition that \(B\) will \(\varPhi \). For our purposes, this is nevertheless a non-doxastic account, because it denies that trust entails belief.
Even some supporters of doxastic accounts concede that the kind of trusting beliefs that are central to trust are not based upon evidence (Hieronymi 2008), and that accounting for some of trust’s distinctive features requires severing the relation between trust and evidence. According to the account defended here, this is not the case.
Both Keren (2007) and Zagzebski (2012) explore the idea that forms of epistemic trust involve responding to preemptive reasons for belief. However, they only address epistemic trust; neither suggests that preemptive reasons play a role in our understanding of trust in general nor that this notion can shed light on the debate between supporters of doxastic and non-doxastic accounts of trust. For further discussion of the difference between the view presented here and Zagzebski’s, see footnote 15.
Note that I make the claim that trust entails a belief, not the stronger claim that trust is a belief. Thus, the account defended here is doxastic, but not purely doxastic. Trust, I claim, requires not merely believing that the trustee is trustworthy but also responding to preemptive reasons (that one sees oneself as having in virtue of what one believes about the trustee) against taking certain precautions. Moreover, my (partial) account of trust does not rule out the possibility that trust also involves having some non-doxastic attitude in addition to belief.
Trustworthiness can also be understood as having a three-place structure (Jones 2012): person \(B\) is trustworthy with respect to person \(A\) and action \(\varPhi \) (or domain of interaction \(D\)). Whether a three-place relation of trust is the most basic trusting relation is a question I will not address here; on this see Jones (2004).
Jones, while committing herself to the claim that trust does not entail a belief (1996, p. 22), also presents a weaker claim—that “trust is not primarily a belief” (1996, p. 5; emphasis added). This seems to be consistent with a mixed doxastic account, according to which trust involves some non-doxastic attitude as a central component alongside a non-central doxastic component, where the centrality of the non-doxastic component is to be understood, at least partially, in terms of its role in explaining features of trust. Without arguing here against mixed accounts of trust, I do want to insist on the centrality of the belief component: as I shall argue, features of trust which Jones attempts to explain by appealing to non-doxastic components are better explained by appealing to the belief component and to the preemptive reasons which an agent has by virtue of her belief.
Kappel does not presuppose that trust is a distinct or uniform type of attitude (forthcoming). Accordingly, unlike other accounts, his does not attempt to draw a line distinguishing trust from other attitudes. Nonetheless, a dispositional account of the kind he proposes may suggest a way of doing so.
Baker (1987), while noting that trust is resistant to evidence, does not argue for a non-doxastic account of trust. Instead, she suggests that this feature should make us reconsider the standards of rationality that apply to beliefs.
Moreover, as the above quotation from Baier suggests, it can undermine trust even in cases where initially we had strong evidence supporting the trustee’s trustworthiness.
Moreover, if the parents’ trust in their daughter does not involve any belief in her trustworthiness, then such trust may appear to be rational: lack of evidence for the daughter’s trustworthiness need not tell against the rationality of their trust.
Frost-Arnold (forthcoming), while agreeing that merely acting as if a speaker’s testimony is true does not amount to trusting her, argues that this is no objection to accounts of trust suggesting that trusting speakers sometimes involves accepting their testimony without believing it, because merely acting as if a proposition is true does not amount to accepting it. However, this does not avoid the above objection. Since acceptance, unlike belief, is context-dependent, we can accept a speaker’s testimony in one context while asserting, in other contexts, that we don’t know whether her testimony is true. Surely such acceptance would not satisfy speakers’ expectations that we trust them.
Raz coins the term “preemptive reason” to describe a reasons that “displaces” other reasons (1990a, p. 10, 1986, p. 42). At other places Raz describes them as reasons that “exclude” or “preempt” other reasons, and in some writings refers to them as “exclusionary reasons”. The core idea of all these expressions is that such reasons are second-order reasons against acting on certain first-order reasons. It is in this sense that I shall use the term here.
Note however that the term “preemptive reasons” is sometimes used, for instance, by Darwall (2010), as interchangeable with what Raz calls “protected reasons”. In this sense, a preemptive or protected reason is both a first-order reason to act in a certain way and a second-order exclusionary reason not to act on conflicting first-order reasons (Raz 1990b). To avoid any confusion, I shall use the term “protected reasons” to refer to this kind of combination of first- and second-order reasons. As I explain below (footnote 20), reasons for trust are also protected reasons.
For a discussion of the concept of preemptive reasons for belief, see Zagzebski (2012).
Both Zagzebski (2012) and McMyler (2011) propose accounts of epistemic trust, and both employ in this context ideas about the preemptive nature of beliefs on authority. Zagzebski explicitly adopts the notion of “preemption” derived from Raz (1986), and the notion plays a central role in her account of believing on the epistemic authority of others. However, unlike the account of trust suggested here, the notion of preemption does not figure in her understanding of trust per se and of epistemic-trust in particular. While on her account, epistemic authority has everything to do with preemption, trust per se, whether epistemic or non-epistemic, does not. Instead, preemption figures within her discussion of epistemic trust, because “among those we are committed to trusting are some whom we ought to treat as epistemic authorities” (2012, p. 3). In contrast, on the account suggested here, the notion of preemption is central to our very understanding of trust. For in trusting we see ourselves as having preemptive reasons against taking precautions. What distinguishes those whom we treat as epistemic authorities is not the fact that we see their judgment as issuing preemptive reasons for belief, but the strength and scope of these reasons. When a trusted speaker tells me that \(p\), it might be epistemically responsible for me to treat her saying so as issuing a preemptive reason to believe that \(p\), but trusting her need not involve thinking that it would be epistemically irresponsible for me to form an opinion on my own consideration of the evidence. In contrast, deference to an epistemic authority involves not merely thinking that I may allow her weighing of the evidence to replace my own, but that it would be irresponsible for me not to treat her judgment in this way (on the difference between speaker-trust and deference to an authority, see also footnote 31).
McMyler does not explicitly employ or address the notion of preemption; nonetheless, his account of epistemic trust is in some ways closer to the one defended here. The main concept underlying McMyler’s account is that of a second-personal reason for belief: a reason that, unlike monadic reasons (Thompson 2004), has a fundamental element of address, and that is made available to audiences addressed by a speaker, but not to overhearers. McMyler claims that such second-personal reasons for belief provide the former, but not the latter, with a right of deferral: a right to defer challenges to her testimonially-based beliefs to the speaker on which her belief is based (2011, 2013). While I agree that in trusting speakers we see ourselves as having such a right of deferral, and while this right of deferral fits well within the preemptive-reasons account of trust suggested here, the understanding of this right suggested by my account is quite different than that suggested by McMyler’s. First, the account defended here suggests that in trusting speakers we see ourselves as having not just this right of deferral, but also other rights not to take other kinds of precautions,which are not discussed by McMyler. Thus, this feature of epistemic-trust is an instance of a more general feature of epistemic-trust and of trust more generally. Second, the account suggested here explains why we see ourselves as having this right not by pointing at some second-personal reason for belief, but by arguing that reasons for trust, quite generally, involve second-order reasons against acting for precautionary reasons. One problem with tying this right of deferral (and possibly other rights against taking precautions) with a second-personal reason for belief available only to addressees is that this, pace McMyler, makes it difficult to see how a general theory of trust could include such rights in accounting for both testimonial and non-testimonial trust. For it is difficult to see how a general theory of trust can admit that a trustor can have a right of deferral in cases of non-testimonial trust where there is no act of testimonial address, but insist that in the testimonial case, only a trustor who is an addressee can have such a right.
Holton maintains that while trusting \(B\) to \(\varPhi \) does not require believing that \(B\) will \(\varPhi \), it is nonetheless incompatible with believing that \(B\) will not \(\varPhi \) (Holton 1994, p. 71).
An objector might suggest that all that is needed to explain the difference between epistemic reliance and speaker-trust is the contrast between trusting someone for the truth and using someone as a source of information. Any account of trusting-for-the-truth, she may suggest, must say that \(A\) cannot trust \(B\) for the truth knowing that \(B\) is likewise trusting \(A\) for the truth.
Now I of course agree that any account of trusting-for-the-truth must say this. But why must trusting-for-the-truth satisfy this constraint? Take an account of trusting-for-the-truth that says that \(A\) trusts \(B\) for the truth of \(p\) only if \(A\) believes that \(B\) knows that \(p\) (for a similar account of trust, see Fricker 2006). Such an account can explain the validity of the constraint only if it can explain why symmetrical trusting-for-the-truth cannot generate knowledge. But why can it not generate knowledge, if trust does not involve disregarding evidence in the way suggested by the preemptive-reasons account? Indeed, Fricker’s explanation of why knowledge cannot be generated through (asymmetric) trust does not work for symmetrical trust. In the asymmetric case, the very idea of knowledge generation requires trust to be based on a false belief that the speaker knows; according to Fricker, this explains why such trust does not generate knowledge. But in symmetrical trust, the idea that trust generates knowledge does not entail that one knows on the basis of a false belief. Here, then, the claim that symmetrical trust does not generate knowledge remains unexplained. After all, it has been suggested that testimony is sometimes a generative source of knowledge (Lackey 1999, 2008; Graham 2000). If, as some have suggested, trust, unlike other ways of relying on testimony, does not generate knowledge (Fricker 2006; Keren 2007), this is something that an account of trust must explain if it is to explain the difference between symmetrical trust and symmetrical reliance.
Finally, our objector may try to explain why A cannot trust B for the truth knowing that B is likewise trusting A by noting that trust-for-the-truth involves taking the trustee as an authority. But this explanation, even if successful, does not help our objector. For it is of the essence of authority, practical or epistemic, that it issues preemptive reasons (Zagzebski 2012; Raz 1990a), so this is an explanation that supporters of the preemptive reasons account will be happy to endorse.
I do not claim that other accounts cannot accommodate the idea that one does not trust \(B\) to \(\varPhi \) if one takes precautions against \(B\) not \(\varPhi \)’ing, but that they do not entail it. Therefore, unless they incorporate the claim that reasons for trust are preemptive reasons, they are incomplete.
As I note in Sect. 7, trust comes in degrees for at least two reasons: First, like other beliefs, the belief that a person is trustworthy which is necessary for trust, comes in degrees. Second, a person can be believed to be more or less trustworthy. Both of these are reflected in the kind of precautions we see ourselves as having reason not to take, and in the strength of these reasons. See also footnote 31.
The main point I want to insist on here is that reasons for trust are preemptive reasons in the sense that they exclude or preempt other reasons. That is, they are second-order reasons against acting (or believing) on certain first-order reasons. However, as noted above, the term “preemptive reason” is used by some as equivalent with the Razian notion of “protected reasons,” and on the account suggested here, reasons for trust can also be described as protected reasons. For if we trust someone to \(\varPhi \), we not only rely on her to \(\varPhi \), we rely while being prepared not to take precautions against the possibility that she will not \(\varPhi \). That is, we rely while being prepared not to act on possible reasons for not relying. Thus reasons for trust are a combination of first order reasons to rely in certain ways, and of second-order (exclusionary) reasons against acting for certain conflicting reasons. Thus, they are protected reasons.
More accurately, reasons against taking precautions are reasons against reflecting on such evidence in as much as such reflection is motivated by precautionary reasons or is apt to motivate a precautionary attitude. For neither reflecting on our grounds for trust, nor operating the CCTV cameras amounts to taking precautions, if they have nothing to do with guarding against relevant risks. Thus, for instance, reasons against taking precautions are reasons against reflecting on our past interactions with the trustee, when such reflection is motivated by our desire to guard against the risk of being let down, but not when we know that such reflection will not affect our action, and merely reflect on them while being swept by a wave of nostalgia. In this respect, the kind of preemptive reasons which are central to our understanding of trust aresimilar to the kind of preemptive reasons central to Raz’s account of authority and of authoritative rules. For the latter too, while described as reasons for disregarding certain reasons, do not exclude engaging in thought about the latter reasons, so long as we know that our reflection will not affect our action (Raz 1990b, pp. 183–184). I am grateful to an anonymous reviewer of this journal for pressing me on this point.
It might be objected, that the preemptive-reason account explains why trust is incompatible with reflection, but not why such reflection undermines trust. In cases where we have good evidence for the trustee’s trustworthiness, why would such reflection not result in the restoration of trust, even if it (temporarily) displays lack of trust? The reason for thinking that even in such cases trust would often not be restored has to do with the fact that trustworthiness is partially a matter of responsiveness to the trustor’s trust and dependence on the trustee. Excessive reflection, because it displays lack of trust, may negatively affect the trustee’s willingness to react to the trustor’s reliance on her, even if trust is restored. Moreover, this is something the trustor can anticipate. Accordingly, even if initially trust was based on strong evidence supporting the trustee’s trustworthiness, reflection on this evidence might result not in the restoration of trust, but in the restoration of the belief that the trustee would have been properly responsive to the trustor’s trust, if the trustor had not displayed lack of trust, conjoined with the belief that now that lack of trust has been displayed, the trustee would not similarly respond to restored trust. I am grateful to an anonymous reviewer of this journal for pressing me on this point.
An account of trust as involving the adoption of the participant’s stance is supposed to explain some features of trust, such as the phenomenon of therapeutic trust (Holton 1994), but does not seem to explain these two, while an account of trust as a non-doxastic disposition to rely in certain ways does not appear to be in the business of explaining features of trust. Describing trust as a disposition to act or form beliefs in certain ways does not explain why trust has the features that it has.
Even recalcitrant emotions, which appear to a subject to be in tension with evidence available to her, characteristically survive reflection (Benbaji 2013).
This is not to claim that non-doxastic attitudes cannot provide us with preemptive reasons, but rather that the attitudes that are central to non-doxastic accounts of trust do not provide the kind of preemptive reasons, such as reasons against being attuned to contrary evidence, that are characteristic of trust.
Where \(B\)’s trustworthiness with respect to \(\varPhi \) is to be understood in terms of relevant competence, goodwill, and responsiveness to \(A\)’s dependence on her.
Sometimes a trustor may see herself as having reason for not taking precautions, but not act on this reason because she has no opportunity to take relevant precautions. Nonetheless, she may be responsive to this reason if she is willing not to take certain precautions, should she have the opportunity to take them.
While it is the belief that the trustee is trustworthy that is central to trust on the account suggested here, full-fledged trust also involves the belief that the trustee will do what she is trusted to do. Thus, as suggested in Sect. 2, if \(B\) tells \(A\) that \(p\), and \(A\) fully trusts \(B\) to tell her the truth, then \(A\)’s strong belief that \(B\) is trustworthy (combined with \(A's\) disregard of other evidence that she may have), would lead \(A\) to believe that \(B\)’s testimony is true. Nonetheless, when one does not fully trust the trustee, but only trusts her to a degree, one may believe, to a degree, that the trustee is trustworthy, without believing to the same degree that she will do what she is trusted to do. This is a result of the fact that being trustworthy with respect to\(\varPhi \)’ing is compatible with failing to \(\varPhi \), and the fact that when trust is not complete, one may be responsive, to some extent, to evidence suggesting that the trustee may not do what she is trusted to do.
More accurately, proponents of doxastic accounts are committed to two claims: first, that one does not trust a person at all if one in no way believes her to be trustworthy; and second, that other things being equal, the greater one’s doubt regarding her trustworthiness, the lesser one’s trust (Hieronymi 2008).
There are important differences between the two kinds of cases. An expert speaker typically has advantage over her lay audiences both in terms of the evidence available to her, and in terms of her ability to evaluate this evidence; a non-expert speaker with positional advantage over her audience normally enjoys only the former kind of epistemic advantage. As a result, the preemptive reasons issued by the testimony of these two kinds of speakers differ significantly in terms of both scope and strength. The judgment of a known expert preempts consideration of more evidence than that of a non-expert with positional advantage. Moreover, the preemptive reasons issued by an expert’s testimony often make it epistemically irresponsible for us to base our belief on our own weighing of the evidence, whereas in the case of a non-expert speaker, such reasons often permit, but do not require, not weighing certain evidence.
Zagzebski (2012), following Raz (1986), suggests an argument for treating an expert’s judgment as a preemptive reason for belief (rather than as a reason for belief to be added to all others), which, if successful, should be convincing to the reliabilist. For the argument is supposed to establish the claim that under certain conditions, treating the expert’s judgment as a preemptive reason would improve our “track-record” (Zagzebski 2012, p. 114) or reduce our “rate of mistake” (Raz 1986, p. 68), when compared with other ways of treating her judgment.
See footnote 4.
It might be objected that, where a belief in the trustworthiness of a person is unjustified, it would be better not to hold the belief, and yet to enjoy the possible fruits of trust by acting as if one trusted. However, this is not always possible. Trust involves doing things—such as focusing our attention in certain ways—that we often cannot do unless we believe in the trustworthiness of a person. Moreover, trusting relationships are often valuable in themselves, and we cannot engage in them unless we really trust our partner to the relationship. Even if Mrs. Barnes’s doubts about her husband’s faithfulness are not manifested in any way in her behavior, they still affect the couple’s relationship. As long as she harbors them, she is not engaged in a trusting relationship, even if she acts as if she were. Therefore, our ability to enjoy the fruits of trust without really trusting is limited.
See Zagzebski (2012) for an important discussion of such reasons and of their neglect within much of modern philosophy.
References
Adler, J. E. (1994). Testimony, trust, knowing. Journal of Philosophy, 91(5), 264–275.
Baier, A. (1986). Trust and antitrust. Ethics, 96(2), 231–260.
Baker, J. (1987). Trust and rationality. Pacific Philosophical Quarterly, 68, 1–13.
Benbaji, H. (2013). How is recalcitrant emotion possible? Australasian Journal of Philosophy, 91(3), 577–599.
Darwall, S. (2010). Authority and reasons: Exclusionary and second-personal. Ethics, 120(2), 257–278.
Elster, J. (2007). Explaining social behaviour: More nuts and bolts for the social sciences. Cambridge: Cambridge University Press.
Faulkner, P. (2007). On telling and trusting. Mind, 116(464), 875–902.
Faulkner, P. (2011). Knowledge on trust. Oxford: Oxford University Press.
Faulkner, P. (2012). Replies. Abstracta, 6, 117–137.
Fricker, E. (2006). Second-hand knowledge. Philosophy and Public Affairs, 73(3), 592–618.
Frost-Arnold, K. (forthcoming). The cognitive attitude of rational trust. Synthese, doi:10.1007/s11229-012-0151-6
Goldman, A. I. (1979). What is justified belief? In G. Pappas (Ed.), Justification and knowledge. Dordrecht: Reidel.
Graham, P. J. (2000). Conveying information. Synthese, 123(3), 365–392.
Hieronymi, P. (2008). The reasons of trust. Australasian Journal of Philosophy, 86(2), 213–236.
Holton, R. (1994). Deciding to trust, coming to believe. Australasian Journal of Philosophy, 72(1), 63–76.
Jones, K. (1996). Trust as an affective attitude. Ethics, 107(1), 4–25.
Jones, K. (2004). Trust and terror. In P. DesAutels & M. U. Walker (Eds.), Moral psychology: Feminist ethics and social theory. Lanham, MD: Rowman & Littlefield.
Jones, K. (2012). Trustworthiness. Ethics, 123(1), 61–85.
Kappel, K. (forthcoming). Believing on trust. Synthese, doi:10.1007/s11229-013-0376-z.
Keren, A. (2007). Epistemic authority, testimony and the transmission of knowledge. Episteme, 4(3), 368–381.
Lackey, J. (1999). Testimonial knowledge and transmission. Philosophical Quarterly, 50(197), 471–490.
Lackey, J. (2008). Learning from words: Testimony as a source of knowledge. Oxford: Oxford University Press.
McLeod, C. (2002). Self-trust and reproductive autonomy. Cambridge, MA: MIT Press.
McLeod, C. (2011). Trust. The Stanford encyclopedia of philosophy (Spring 2011 edition), ed. Edward N. Zalta. Retrieved September 3, 2012 from http://plato.stanford.edu/archives/spr2011/entries/trust.
McMyler, B. (2011). Testimony, trust, and authority. Oxford: Oxford University Press.
McMyler, B. (2013). The epistemic significance of address. Synthese 190, 1059-1078.
Moran, R. (2005). Getting told and being believed. Philosophers’ Imprints, 5, 1–29.
Pettit, P. (1995). The cunning of trust. Philosophy & Public Affairs, 24(3), 202–225.
Raz, J. (1986). The morality of freedom. Oxford: Clarendon Press.
Raz, J. (1990a). Authority and justification. In J. Raz (Ed.), Authority. Oxford: Basil Blackwell.
Raz, J. (1990b). Practical reasons and norms (2nd ed.). Princeton: Princeton University Press.
Thompson, M. (2004). What is it to wrong someone? A puzzle about justice. In R. Jay Wallace, P. Pettit, S. Scheffler, & M. Smith (Eds.), Reason and value: Themes from the moral philosophy of Joseph Raz. Oxford: Oxford University Press.
Williams, B. (2002). Truth and truthfulness: An essay in genealogy. Princeton: Princeton University Press.
Zagzebski, L. (2012). Epistemic authority: A theory of trust, authority and autonomy in belief. Oxford: Oxford University Press.
Acknowledgments
Research on this paper was generously funded by the Israel Science Foundation (Grant No. 714/12). An earlier version of this paper was presented at the International Conference on Theoretical and Rational Rationality at the Israel Institute of Advanced Studies. For helpful comments, I am grateful to participants at the conference, to the editors and anonymous reviewers of this journal, and to Iddo Landau, Ariel Meirav, Collin O’neil, and Saul Smilansky. I am especially grateful to Joseph Raz.