1 For help with drafts of this material I'm grateful to audiences at Bowdoin College, the Claremont Colleges, Oberlin College, Rice University, Southern Methodist University, the University of Kansas, the University of Michigan, and the Pacific APA (where Tamar Schapiro commented). Thanks to Steve Darwall, David Hills, David Velleman, and Andrea Westlund for help with early drafts. Special thanks to Karen Jones for three full rounds of challenging and encouraging subsequent commentary; the paper would be significantly weaker without her provocations. (A companion paper in the epistemology of testimony, ‘Telling as Inviting to Trust,’ appeared in Philosophy and Phenomenological Research 70 (May 2005); and a companion paper on the nature of judgment, ‘Judging as Inviting Self-Trust,’ is in preparation.)

2 What I want to highlight, as we'll see, is the contrast between (mere) testimony and advice (i.e. testimony that is also advice), not any particular resonance from the verb 'to tell.’ My use of that verb will mislead if it suggests the very different idiom ‘to tell A to ϕ.’ If Stephen tells Anna to be kinder, he is not testifying at all but commanding, and in a way that may or may not amount to advice. (One reason I'm using ‘tell’ here is to make my account of advice in this paper continuous with the account I give of testimony in ‘Telling as Inviting to Trust,’ Philosophy and Phenomenological Research 70 [2005].)

3 I do not deny that there are other interpretations of the melodrama, thus briefly described. I claim only that my Interpretation is a coherent one. If you find it coherent, I ask you to imagine the scenario along lines that would bear it out. If you do not immediately find it coherent, I ask you to treat my fuller characterization of it in this and the next two sections as an argument for its coherence. I ask also that you not be distracted by the somewhat unsavory power dynamic between Stephen and Anna. I have tried to make Stephen's predicament vivid by imagining a realistic — that is, a humanly flawed — context of motives.

4 Though I try not to stereotype gender relations in my examples, when speaking generally I'll follow the Convention of making advisers female and advisees male.

5 Of course, advisers are typically far less than perfectly confident that their advice is sound, a lack of confidence reflected in their unwillingness to ‘impose’ their perspective on the advisee. For simplicity, unless otherwise specified I'll assume that the advisers I'll discuss are quite confident that their advice is sound.

6 I set aside the comparatively rare case in which S tells A that ϕ-ing is in A's self-interest, not because S has judged for herself that it is, but because she has taken someone eise's word for it — call this other Speaker S*. In this case, S* teils S that ϕ-ing is in A's self-interest, then S tells A. Does S count as advising A to ϕ here, or only as passing along S*'s advice? Or is there no advice in play here, merely testimony? Perhaps in such a case S may count as advising A without adopting a perspective of care for A. Still, S is not inviting A to trust her on the matter, so this is not the species of advice I am discussing in this paper. (S may invite A to take her word that S* believes he should ϕ, without inviting A to take her word that he should ϕ.)

7 For an account of the relationship between this perspective and well-being, see Darwall, Stephen Welfare and Rational Care (Princeton: Princeton University Press 2002).CrossRefGoogle Scholar

8 Austin uses the term ‘abuse’ to characterize infelicities that are not misfires. If a speech act ‘misfires’ it does not come off, but when there is an ‘abuse’ the speech act comes off but is ‘not implemented or not consummated, rather than void or without effect’ (Austin, J.L. How to Do Things with Words, 2nd ed. (Cambridge: Harvard University Press 1975), 16CrossRefGoogle Scholar)

9 If you hear it differently, note that nothing pertinent to my principal argument in this paper depends on the assumption. (I could simply move ‘A recognizes that S intends that…’ inside the scope of 's intends that…,’ as on the original Gricean formulation.) The assumption has received most attention in the debate over whether a violation of free speech could occur by virtue of an intended audience's inability to recognize the illocutionary intention. (Say, a woman tries to refuse sex, but the man cannot recognize what she's doing as manifesting an intention to get him to stop. One side of the debate views her as subject to ‘illocutionary disablement.’) For objections to the assumption, see Jacobson, Daniel ‘Freedom of Speech Acts? A Response to Langton,’ Philosophy and Public Affairs 24 (1995) 72–8.CrossRefGoogle Scholar For a reply, see Hornsby, Jennifer and Langton, Rae ‘Free Speech and Illocution,’ Legal Theory 4 (1998), 28–31;CrossRefGoogle Scholar then see rejoinder, Jacobson's ‘Speech and Action: Replies to Hornsby and Langton,’ Legal Theory 7 (2001) 179–201Google Scholar

10 See the quotation from Austin in n. 8 above.

11 The formulations in this paragraph and the next parallel those given in section III of ‘Telling as Inviting to Trust’ for the illocutionary act of telling A that p. Though the illocutionary and perlocutionary intentions differ, each act — telling and advising — is consummated in trust.

12 Or that she succeed in getting A to ϕ. I'm following Grice in assuming that we can intend that others do things, but nothing hangs on that assumption.

13 One may object that he must be exercising some species of self-care, or he wouldn't have managed to get dressed for the reunion, etc. But we can easily imagine cases — someone who really cannot manage to leave the house or even get out of bed, someone needing to be talked down from threatening a suicidal plunge, someone with a gun in his mouth — where the deficit in self-care seems closer to complete. My point is not that we must view Adam as approximating these cases but that an adviser needn't tap into his capacity for more fundamental self-care in order to reach him. Insufficiently self-caring without the melodrama, Adam contains the seed of our common predicament.

14 It is what Thomas Nagel calls a ‘motivated’ desire whose basis does not lie in any ‘unmotivated’ desire, a desire ‘motivated exactly as the action is… by reasons stemming from certain external factors’ (The Possibility of Altruism (Princeton: Princeton University Press 1970), 29-30).

15 I'll give this claim more explication and defense in sections VII and VIII.

16 My account of the reasonability of trust constitutes the second contribution of this paper, in principle separable from my quasi-Gricean account of advising. The reader is therefore free to accept one but not the other, though I of course think the best view combines them.

17 By ‘evidence’ here and in similar contexts I mean good evidence on balance. This needn't be conclusive evidence.

18 A number of philosophers have argued for this view; for one such argument, see Burge, Tyler ‘Content Preservation,’ Philosophical Review 102 (1993) 457–88;CrossRefGoogle Scholar ‘Interlocution, Perception, and Memory, Philosophical Studies 86 (1997) 21-47; and ‘Computer Proof, Apriori Knowledge, and Other Minds,’ Philosophical Perspectives 12 (1998) 1-37. I give a very different argument for the view in ‘Telling as Inviting to Trust.’

By forming a judgment (whether about what to believe or about what to do) ‘on the basis of evidence’ I'll mean forming it on the basis of antecedently possessed evidence of the Speaker's trustworthiness. I don't claim that when you form a judgment on some basis other than such evidence you do not form it on the basis of evidence at all. It may be that forming the judgment itself gives you evidence of its truth — if, for example, it's a practical judgment of the form ‘I'll ϕ.’ But in that case, (a) the evidence was not available until you formed the judgment, and (b) the evidence will not typically concern the speaker's trustworthiness (although it could, since trusting someone may cause her to become more worthy of trust). For discussion of (a), see Velleman, J. David Practical Reflection (Princeton: Princeton University Press 1989), 61–4;Google Scholar and, more fully, ‘Epistemic Freedom,’ in his The Possibility of Practical Reason (Oxford: Oxford University Press 2000).

19 I don't claim that it is always unreasonable (see below, three paragraphs hence).

20 In a fuller treatment we might characterize a Standard for assessing care-relations as such by examining empirical literature on the many complex ways in which a care-relation can break down. See, for example, Sameroff, Arnold and Emde, Robert eds., Relationship Disturbances in Early Childhood (New York: Basic Books 1989).Google Scholar

21 Of course, she may be advising partly from self-interested motives, if she believes that what you should do coincides with what she independently wants. I'll tacitly assume this qualification when I write ‘non-self-interested motives’ (and cognate locutions): I mean motives that are not exclusively self-interested.

22 One can imagine bizarre cases in which your evidence that her motives are not self-interested is also evidence that she is motivated from concern for someone who hates you and moreover just insofar as this person hates you. So the very weak formulation of the necessary condition needs to be tweaked in order to rule out these cases. The necessary condition is not met, then, unless the evidence that you have that her motives are non-self-interested is compatible with the proposition that she is advising from a perspective of care for you. Since that's too complicated to keep repeating, I'll simply take it for granted.

23 One might try to formulate (b) in terms of evidence that the addressee has rather than evidence that there is. But it is not reasonable to trust when there is good evidence of untrustworthiness that you lack through epistemic sloth. I am applying here a point that is by now Standard in debates about externalist theories of epistemic warrant. For perhaps its first clear articulation, see Bonjour, Laurence The Structure of Empirical Knowledge (Cambridge: Harvard University Press 1985), ch. 3.Google Scholar

24 By that ‘robust’ let me mark the fact that we're no longer discussing the very minimal sort of trustworthiness involved in condition (i), i.e. the possession merely of non-self-interested motives.

25 The ‘ought’-clause leaves it open that in a given case, though there is evidence that S is untrustworthy, we don't think that A ought to possess that evidence — an agent is by no means responsible for all his evidentiary failures: not all arise through epistemic sloth (see note 23).

26 Remember the acknowledgment of complexity in note 22.

27 Nor need we assume a moralized conception of care. That is, we need not assume it irrational to advise someone to do what's in his self-interest while conceiving his self-interest to conflict with what morality requires of him.

28 I'll henceforth treat the further clause on (b) as understood.

29 I leave it open what counts as moral competence and incompetence. I intend my account to be compatible with any normative theory or conception of morality. I do not, in particular, assume a cognitivist conception of moral competence. Moral competence needn't be like competence in seeking the truth.

30 This is often the case with parents of adult children: they care enormously about their children, but are incapable of assessing accurately what's in their children's self-interest. Such a parent can care for (as opposed to merely about) his or her child only where the child has regressed (perhaps through illness) to the State of having most saliently only those interests that the parent can assess (perhaps mere recovery of health).

31 This is true even when the incompetent one is you yourself. Competence in assessing your own self-interest figures in a more robust form of self-care than the form discussed in section IV; there I had only caring-about in mind.

32 At the beginning of the next section, I'll note how Standards of trustworthiness are relative to context and subject matter. It is important also to note that some advisings serve, given the Standard appropriate to their context and subject matter, to give immediate evidence that the adviser is untrustworthy. If I hardly know you yet give you advice that presumes an intimate familiarity with your daily rhythms, you need seek no further for evidence that I am untrustworthy. And perhaps there are some matters that are so intimate — or so complex? — that no one could give advice without thereby giving evidence of untrustworthiness. Here is another respect in which it is important to join Fricker in ‘disaggregating’ (see the next note).

33 Here I agree with the spirit of Elizabeth Fricker's polemic on behalf of what she calls ‘disaggregation’ in the epistemology of testimony (‘Telling and Trusting: Reductionism and Anti-Reductionism in the Epistemology of Testimony,’ Mind 104 (1995) 393-411, section VI). But I do not agree with the letter of that polemic, whose upshot would, if applied to advice, amount to the thesis that trusting advice is presumptively reasonable only given certain contexts and subject matters. I've argued that trusting advice is always presumptively reasonable (given satisfaction of the weak evidential condition). As I'll argue presently, differences in Standards of reliability mark differences in what it would take to defeat this presumption.

34 Jones, Karen ‘Trust as an Affective Attitude,’ Ethics 107 (1996), 20–5;CrossRefGoogle Scholar and ‘Second-Hand Moral Knowledge,’ Journal of Philosophy 96 (1999), 70-7

35 ‘Second-Hand Moral Knowledge,’ 72. Jones is disagreeing here with Coady's, C.A.J. claim that our default stance should be one of trust in testimony (Testimony: A Philosophical Study (Oxford: Oxford University Press 1992), 46–7).Google Scholar But Coady articu lates nothing like my defeasibility account of the rational presumption in favor of trust, nor does Jones consider anything like my account.

36 It may seem that seriousness about morality requires a commitment not to take moral advice without assessing the adviser as morally competent. But that, I believe, is too strong. Moral seriousness requires that you be very sensitive to evidence that you should not take the advice. Yet it doesn't require that you manifest that sensitivity by actively searching for evidence. Actively searching for evidence of X is only one way of manifesting sensitivity to evidence of X.

37 Thanks to Ann Cudd for pressing me on this point. Alasdair Norcross reminds me that the point may hold for prudential advice as well. Imagine you live in a Community in which it is evident that few people are willing to advise, or are competent in advising, from the perspective of care for you. In that case, since there's this evidence of general untrustworthiness, you'll need positive evidence that any given adviser is advising from the perspective of care for you. I can accommodate the case along the lines that I'm about to use to accommodate the moral case. Against this unfortunate background, sensitivity to evidence of untrustworthiness requires a search for evidence of trustworthiness.

38 What would mark this difference in communities? Doug Ehring informs me of a local eccentric who had the habit of spending weekend afternoons on a parkbench under a sign proclaiming ‘Free Advice.’ Taking separately the cases of moral and prudential advice, and assuming the availability of evidence that the eccentric's motives were not exclusively self-interested, the question ‘Would advice thus dispensed directly give a pro tanto reason to take it?’ could serve as a barometer of manifest trustworthiness in the cultural atmosphere. It is not obvious to me that there are not such reasons, both prudential and moral, available where I live. (Of course, manifest eccentricities and an overwillingness to advise typically count as defeaters of reasonable trust, so we'd have to imagine the Free Adviser free from these defects, which doesn't seem impossible.)

39 I do not offer the reflections in this paragraph as a positive argument for my thesis that trust is presumptively reasonable. (My positive argument was in the previous sections; I am here merely deflecting objections.) Considered as a positive argument, the paragraph would run afoul of Fricker's distinction between global and local reductionisms ('Telling and Trusting/ sections IV and V). Fricker argues that it is no argument against a local reductionist requirement in the epistemology of testimony — for example, that you cannot here-and-now acquire a reason to believe that p from having been told that p unless you have good evidence of the teller's reliability on the matter — to note, contra global reductionism, that it cannot be the case that our entitlement to all of what we take ourselves to know through prior testimony should meet that requirement. It may be that the critical faculties whose employment is required by the local reductionism cannot develop except against a background of entitlement that presupposes the falsity of global reductionism. And the point seems to hold not only for epistemic but for practical reasons: that we could not build a humanly livable set of practical commitments from the ground up does not show that, having done so with the help of advice, we are not now responsible for assessing new advisers for positive competence.

Fricker's argument depends on a sharp distinction between immature and mature deployments of practical reason. Although the point is by no means decisive, I regard it as an attraction of my anti-reductionist approach that it views the maturity of an agent's practical reason as a relative matter — relative both to the capacities of other agents and to past and potential future capacities of the agent himself.

40 This is a principal theme of ‘Trust and Reasons,’ where I defend the claim more fully and argue that this resolution of the traditional (practical) internalism/externalism issue generates a new internalism/externalism issue, now about the nature of fiduciary reason. (I don't mean the epistemic issue that I discussed in sections IV and V.)

41 Darwall, Stephen ‘Reasons, Motives, and the Demands of Morality: An Introduction,’ in Darwall, S. Gibbard, A. and Railton, P. eds., Moral Discourse and Practice (Oxford: Oxford University Press 1997)Google Scholar

42 By ‘reasons/motives internalism’ I throughout mean what Darwall calls ‘reasons/ motives existence internalism’ (not what he calls ‘reasons/motives judgment internalism’).

43 For a Humean defense of reasons/motives internalism, see Williams, Bernard ‘Internal and External Reasons,’ in his Moral Luck (Cambridge: Cambridge University Press 1981);CrossRefGoogle Scholar ‘Internal Reasons and the Obscurity of Blame,’ in his Making Sense Of Humanity (Cambridge: Cambridge University Press 1995); ‘Internal and External Reasons (A Reply to McDowell),’ in World, Mind, and Ethics: Essays on the Ethical Philosophy of Bernard Williams, Altham, J.E.J. and Harrison, Ross eds. (Cambridge: Cambridge University Press 1995);CrossRefGoogle Scholar and ‘Values, Reasons, and the Theory of Persuasion,’ Ethics, Rationality, and Economic Behaviour, Farina, F. Hahn, F. and Vannucci, S. eds. (Oxford: Oxford University Press 1996).Google Scholar For a Nietzschean rejection of morality/ reasons internalism, see Williams, Bernard Ethics and the Limits of Philosophy (Cambridge: Harvard University Press 1985);Google Scholar and ‘Nietzsche's Minimal Moral Psychology,’ in his Making Sense of Humanity.

44 For an Aristotelian defense of morality/reasons internalism, see McDowell, John ‘Are Moral Requirements Hypothetical Imperatives?’ and ‘Virtue and Reason,’ both in his Mind, Value, and Reality (Cambridge: Harvard University Press 1998);Google Scholar ‘Eudaimonism and Realism in Aristotle's Ethics,’ in Aristotle and Moral Realism, Heinaman, R. ed. (Boulder: Westview 1995);Google Scholar and ‘Deliberation and Moral Development in Aristotle's Ethics,’ in Aristotle, Kant and the Stoics: Rethinking Happiness and Duty, Engstrom, S. and Whiting, J. eds. (Cambridge: Cambridge University Press 1996).Google Scholar For an Aristotelian rejection of reasons/motives internalism, see McDowell, ‘Might There Be External Reasons?’ in his Mind, Value and Reality.

I disagree with Darwall's classification of McDowell as a reasons/motives existence internalist (‘Reasons, Motives, and the Demands of Morality,’ 308). Darwall calls McDowell a ‘perceptual’ reasons/motives existence internalist, since for McDowell no one can know what reasons he has to act without ‘seeing’ his actions in a way that is intrinsically motivating. This, as Darwall notes, is an epistemic thesis concerning, not whether the agent has a given reason, but what is involved in his knowing that he does. Yet McDowell explicitly rejects the internalist thesis that an agent's possession of a given reason entails any fact about his motivational State or susceptibilities. It seems that epistemic internalism should be, not a species of existence internalism, but a genus set alongside existence internalism and judgment internalism. Existence internalism holds that possessing a reason entails being in some appropriate motivational State, judgment internalism holds that self-ascribing a reason entails being in some appropriate motivational State, and epistemic internalism holds that knowing that you have a reason entails being in some appropriate motivational State. McDowell is an epistemic but not an existence internalist (nor, it seems, a judgment internalist).

45 Korsgaard, Christine was the first to point out this difference; see her ‘Skepticism about Practical Reason,’ Journal of Philosophy 83 (1986) 5–25.CrossRefGoogle Scholar

46 One might wonder whether agents whose capacity for reasonable trust is impaired aren't beyond the reach of morality in another respect as well: Are they ultimately morally responsible for their actions? It seems plausible that a capacity for reasonable trust and distrust forms the ultimate basis of an agent's Status as morally responsible for his actions, since it is plausible that to be a fit subject of moral praise and blame an agent must be capable of taking responsibility for himself by rationally adjudicating the interpersonal influences that shape him. We might view this as a candidate for what Wolf, Susan calls ‘deep responsibility’ in Freedom within Reason (Oxford: Oxford University Press 1990),Google Scholar a less flatly externalist candidate than her own view. (I pursue this thought further in ‘Regret and Responsible Agency,’ in preparation.)

47 For simplicity, I assume an externalist reading of the first defeating condition. And we can't, of course, require that you possess evidence that you are judging from non-self-interested motives! There is at least this disanalogy between the intra-and the inter-personal cases.

48 In ‘Trust and Diachronic Agency’ (Noûs 37 [2003] 25-51) I argue that diachronic agency — forming and then later following through on an intention — requires a robust species of self-trust, given that you have to follow through on the deliverance of your earlier self without redeliberating. But I am now making the more radical claim that forming the intention in the first place requires self-trust. I give a full argument for the claim in ‘Regret and Responsible Agency.’

49 For this Interpretation of Hume, see Brown, Charlotte ‘From Spectator to Agent: Hume's Theory of Obligation,’ Hume Studies 20 (1994) 19–35.Google Scholar do not claim that my account supplants Hume's. The capacity to trust trustworthy advice does not exhaust our practical intelligence.