2. Trust

Let me begin by describing the nature and value of testimonial trust. Testifying is a particular way of presenting a proposition as true. When one testifies, one extends an offer to one's audience. The (intended) recipients of one's testimony are invited to believe in the truth of the proposition on a particular basis. Here are two ways of identifying that basis, which I take to be interchangeable: when one testifies that p, one invites one's audience to trust one that p, or to take one's word for it that p.

Accepting the testifier's invitation involves taking oneself to be relieved of one's usual epistemic responsibility to review the evidence before making up one's mind about whether p.Footnote ⁶ It also involves abstaining from reviewing the evidence for the purpose of making up one's mind about whether p. A policy of ‘trusting but verifying’ is self-undermining: when one trusts in testimony, one hands over responsibility for judging the sufficiency of the evidence to the testifier, and one cannot wrest it back without ceasing to trust. Finally, when one trusts testimony that p, one commits to dismissing or discounting at least some evidence that ~p. Absolute trust in testimony that p would involve a commitment to believing that p in the face of any and all counterevidence. One can count as trusting testimony that p even if one would be willing to abandon one's belief that p in the face of some indications that ~p, but if I am willing to abandon or downgrade my confidence that p in the face of even weak counterevidence, then I can only be said to trust in an etiolated sense.

One cannot trust another's testimony unless one is willing to attribute two qualities to the testifier. The truster must attribute to the testifier a sufficient degree of honesty and also a sufficient degree of epistemic competence. Epistemic competence encompasses a number of different proficiencies, most centrally the abilities to recognize, gather, and properly evaluate evidence. This second trust-enabling condition is the one we will focus on here, because it bears a particularly significant relation to empathy.

Testimonial trust can be based on evidence. If I believe a testifier T wholly or partly on the basis of evidence that he is epistemically competent and honest, then my belief can count as an instance of testimonial trust.Footnote ⁷ But testimonial trust that p must be distinguished from a different sort of evidentially-based belief that p, one which we can call ‘belief occasioned by testimony’. Such belief arises out of testimony in the sense that the testifier's communication is at least partially causally responsible for the audience's coming to believe that p. Unlike testimonial trust, however, this belief is not based in evidence of the epistemic competence or honesty of the testifier. For example, if T’s testimony that p inspires me to reflect on whether p, but I judge that p based on my own intuitions about the matter, and not because I trust T, then my belief that p is an instance of belief occasioned by T's testimony but not an instance of trust in T's testimony.

I am going to concentrate on testimony about features of circumstances encountered experientially, because (as we will see) trust about this kind of testimony bears the most obviously fraught relationship to empathy. We might think of testimony about experience as testimony about ‘what it's like for’ the experiencer. The locution ‘what it's like for’, though, risks obscuring the difference between two forms of experientially-based testimony. On the one hand, there is testimony like ‘I felt threatened by the man's attentions’, or ‘I experienced the stare as hostile’. On the surface, at least, those statements merely amount to testimony about the presence of a psychological state within the testifier. On the other hand, there is experientially-based testimony like ‘The man's attentions are threatening’, or ‘That stare is hostile’. This latter kind of testimony is about the qualities of things in the world outside the subject, but it is (at least in the kind of cases I have in mind) testimony born of an experience of the world. When I register eyes on me, and shudder at the cold nature of the gaze, I perceive a hostile stare, and I can report it as such. This is the type of testimony that interests me. Of the two kinds of experientially-based testimony, it is also the one that is harder to trust. It is one thing to believe that a person who reports a threatening encounter feels threatened, and it is another to believe her that there actually is something threatening about her situation. The former attitude does not require that we accept that the testifier is a competent detector of threat, the latter does.

3. Empathy: What It Is, and How It Supports Testimonial Trust

With this picture of testimonial trust in hand, we can now turn to the question of the role that empathy can play in supporting it. It is criterial for empathy in the sense that interests me that it involve using one's imagination to ‘transport’ oneself, and more particularly that it involve considering the other's situation as though one were occupying the other's position.

Empathy that by definition involves perspective taking is sometimes called ‘cognitive empathy’.Footnote ⁸ This is a potentially misleading term, especially since it is sometimes implied that empathy of this sort has nothing to do with feeling. Empathy as I understand it does draw upon the empathizer's emotional resources. A rough and ready way of thinking about emotions’ role in empathy is to conceive of the empathizer as encountering the other's situation through an appropriate emotional lens. When we try to empathetically imagine how things are for a recent widower, for instance, we attempt to look at his situation through the lens of grief. It does not matter for my purposes here whether we ought to describe the empathizer's emotional experience as one of really feeling, or as feeling-in-imagination, or as pseudo-feeling. What matters is that empathy in my sense is not bloodless or coldly cognitive.

To see how empathy supports trust, we can return to one of the earliest descriptions of empathy's operation: Adam Smith's The Theory of Moral Sentiments. According to Smith, empathy (which he calls ‘sympathy’) provides the unique means by which we judge the propriety of others’ attitudes.Footnote ⁹

According to Smith, after we come to know (through testimony or observation) how another person emotionally registers their circumstances, we assess their emotional apprehension of their situation by trying to imaginatively enter into their position, such that we imaginatively duplicate their experience. If (and only if) we succeed in matching our imagined emotional response to the situation to the testifier's own reported response, we approve of the other's attitude as a fitting one. Our own empathetic emotion furnishes our standard of assessment. I am now going to argue that while this Smith's account needs to be qualified, the core insight is correct: there is a deep connection between empathy and approval. I'll then make it clear how this connection underlies empathy's positive contribution to testimonial trust.

Commentators have been sceptical of this proposal of Smith's for a number of reasons. Some have protested there are surely other ways of judging others’ emotional reactions to be appropriate, ones that have nothing to do with empathy.Footnote ¹¹ I will not deal with that objection here. Rather, I will be concerned with the kind of worries expressed in Thomas Reid's commentary on The Theory of Moral Sentiments: ‘When, [Smith] says, I judge your resentment by my resentment, your love by my love & so on, this is a way of speaking altogether new–this should be confined solely to our judging powers… to say that my resentment is the faculty by which I judge of your resentment is incommutable to all of our ideas of Resentment; we may as well say, I judge of my hunger by your hunger’.Footnote ¹²

There are two possible ways of understanding Reid's complaint. His objection might be that our hunger response is simply an impulse that assails us, one that cannot even get things wrong or right. Far from being a power of judgment, the thought might go, the sense of hunger does not even offer a representation of the world that could in turn be the object of judgment. Reid may be recommending that we think of instances of emotion, whether empathetic or normal, in the same way: as mere sensations that do not represent the world at all.

In response to this first reading of the complaint, we can note that while some forms of hunger may shade into emotional experience, the physical sensation of an empty belly is importantly different from emotion, because emotions present the world as being a certain way. Suppose I am angry at some slight. What does that involve? There are the familiar bodily symptoms: a pounding heart, sweaty palms, etc. There are also the tendencies to act and to express myself in particular ways: to speak harshly, to refuse attempts at reconciliation. Importantly, these actions, symptoms, and expressions are not arbitrary. The many physiological and behavioural dispositions associated with the emotion are unified by their connection to my recognition of certain features of the situation as inviting or demanding my ire. Emotions focus our attention on some features of our situation, push others into the background, and present certain objects or events as calling out for particular kinds of treatment. So, I attend to the intolerable smugness of the offender's smirk, the outrageousness of their audacity. The relative triviality of the situation and the good qualities of the offender recede into the background. It is helpful to think of emotions as being, at heart, ways of seeing in virtue of which things take on a particular evaluative sheen. Since emotions present situations as laden with evaluative properties, they (unlike hunger sensations) can represent reality.

Even if that point is granted, though, one might suspect that the fact that we are able to imaginatively match another's emotional representation of her circumstances is itself evaluatively neutral: it is still equally open to us to accept her representation as accurate or reject it as illusory. And this is the other thought we might draw out of the hunger analogy. Whether or not our sense of hunger issues representations, one might insist, it is not itself a judging faculty. The fact that we are hungry for some delicious cake does not entail that we judge our desire for the cake to be appropriate; that judgment must be issued by a distinct faculty, whose verdicts are not in any way constrained by our hunger. Similarly, the fact that we experience an emotional presentation of the world does not at all constrain our judgment about the presentation's propriety.

To see what that suspicion misses, consider how our non-empathetic emotional experiences relate to approval. It is certainly true that we can have an emotional response without fully endorsing it. We might register the prospect of crossing the chasm as utterly terrifying, but still think that all in all we oughtn't really be afraid. Or we might continue to resent a friend for missing our party, even though we judge that she had a reasonable excuse. But this does not mean that there is no significant connection between having an emotional experience and approving of it.

Two observations will bring out the nature of this connection. First, although it is true that there is a gap between having an emotional response to some phenomenon and fully endorsing that response, we are in general powerfully inclined to treat our emotionally-coloured perceptions of the world as accurate. Confidence in our own emotional representations is our default attitude. If this latter claim is surprising, it is only because we have a tendency, when thinking about emotions in the abstract, to focus on the cases in which they differ most sharply from more ‘rational’, ‘objective’ mental states. So, we call to mind burning rage or acute, all-consuming terror. But it is misleading to think of these as paradigmatic emotional experiences. Our ordinary experience of the world is valenced – things show up as boring, enticing, regrettable, and so on, and we are accordingly attracted or repulsed by these appearances. That is, ordinary experience is emotional, and we usually take those experiences at face value.

It is possible for us to shy away from our own emotional evaluative representations: I might suspend my acceptance of an emotional representation of the party as delightful until I've had an opportunity to reassess whether the party meets standards for delightfulness that I can wholeheartedly endorse. But we do not typically downgrade our confidence in emotional representations until we have done this kind of reflective work. And this is just as well: if we were to refuse to accept what is right before our eyes, emotionally speaking, we would have an awful lot of trouble making our way through the world.Footnote ¹³

Second, although we undoubtedly do sometimes depart from this default, it is very difficult for us to sustain an emotional representation without regarding it as to some degree appropriate. To bring out the point: imagine that I am at first angry with my friend for spilling my secret, but after reflection I conclude that her passing the secret on was not a genuine betrayal. Importantly, I don't just try to convince myself of this. I genuinely judge that there is nothing in what she did that calls out for my ire. Can I remain angry with my friend? Certainly the residual ‘heat’ of anger may be retained; I can still feel pumped up with adrenaline. Still, if I wholeheartedly believe that there is no offense in any way worthy of ire, it will be difficult, if not impossible, for me to continue to really represent her conduct as outrageous. And, as I've already suggested, such representations are a critically important part of emotions. Without them, emotions dissolve, leaving behind mere sensations.

Now let us apply these observations about the connection between emotion and approval to empathy. Suppose an acquaintance tells me that her colleagues’ admiration of her ‘beautiful accent’ is irritating, and I regard her testimony as honest. If I succeed in imaginatively seconding her exasperation, and am confident that this is what I have done, then it will be very difficult if not impossible for me to believe that her irritation is entirely uncalled-for. And what is more, if I have what I've suggested is the normal attitude concerning my emotional experience, if I take my sense of the irksome to be accurately attuned, my empathy with my acquaintance's frustration will incline me to think that she is totally right about her colleagues’ compliments.

This all means that Smith's claim about the connection between empathy and approval is not grossly overstated. He held that ‘to approve of the passions of another, therefore, as suitable to their objects, is the same thing as to observe that we entirely sympathize with them’. I suggest we amend that to: when we empathize with the passions of another, and recognize that this is what we are doing, we are in general powerfully inclined to regard those passions as fully reflective of the evaluative properties her situation really manifests, and it is extremely difficult if not impossible to dismiss them as wholly inappropriate.

What are the implications of this observation for testimonial trust? First, I will briefly articulate the primary implication, and then I will point to a complication that requires us to render that initial articulation more precise.

We can express the primary implication in terms of our case, above. When I empathize with my acquaintance, imaginatively seconding her irritation at her colleagues’ comments, I will be practically bound to conclude that there is something right in my acquaintance's judgment, and I may also be inclined to conclude that her judgment is wholly correct. These conclusions about her judgment, taken separately or together, provide a reason to believe that my acquaintance is appropriately sensitive to irritating phenomena. And, since appropriate sensitivity is part of being epistemically competent, I thereby secure a reason to attribute epistemic competence to my acquaintance. The attribution of epistemic competence constitutes one of testimonial trust's enabling conditions. So, my empathetic effort will furnish a basis for trust in my acquaintance's testimony.

It is important to note that a successful empathetic effort does not provide support for trust in all of the testifier's testimony. In particular, if T testifies that p, and I conclude on the basis my empathetic effort that T is right that p, my conclusion does not provide a basis for my trust in T that p. It fails to do so simply because my undertaking the relevant empathetic effort precludes my trusting T that p. When my acquaintance testifies that her colleagues’ admiration is irritating, and I succeed in empathetically working myself into her position such that I apprehend the irritating quality of the admiration, I thereby secure a reason to believe that the admiration is irritating, namely that I have ‘seen’ it for myself. Notice, though, that my empathetic assessment of the testimony involves considering how the other's situation appears to me. In doing so, I am reviewing the evidence that the admiration is irritating and on that basis making up my mind about whether the admiration is irritating. When I empathize with my acquaintance, I am verifying the truth of her testimony, and as I explained earlier, it is not possible to ‘trust but verify’.

Still, my empathetic engagement with my acquaintance's judgment about her colleague's admiration can provide a basis for my trust in her testimony about other cases and even other evaluative properties. My conclusion that my acquaintance is right about her colleague's admiration provides evidence for the conclusion that she is appropriately sensitive to irritating phenomena, and that attribution of competence constitutes a basis for trust in her other testimony about what is irritating. Furthermore, insofar as I think that accurate emotional attunement in one domain is positively correlated with accurate attunement in other domains, the conclusion that she is right about the admiration gives me reason to attribute to her broader emotional sensitivity to evaluative properties. That broader attribution of competence in turn provides a basis for wider-ranging trust in testimony about what is threatening, heartening, offensive, and so on.

Empathy has the power to dramatically reshape one's willingness to trust others’ testimony. Suppose my acquaintance's testimony about the irritating nature of her colleagues’ remarks surprises me. It doesn't fit with my general understanding of what kind of things are irritating. The colleagues are highlighting one of my acquaintance's positive traits, and being admired for a positive trait is usually a good thing. In this case, it would be natural for me to initially suspect that my acquaintance is just touchy, and the thought of her as touchy will incline me to discount her testimony about what is troublesome or offensive. But if I take the time to imaginatively place myself in her position, and feel the stirrings of annoyance in myself, matters will look very different. Her irritation will no longer look unaccountable. The possibility of dismissing her testimony as the erroneous upshot of an excessively querulous outlook will recede from view. And next time she mentions that some apparently benign treatment actually isn't, I will be more inclined to take her word seriously.

4. Trust without Empathy

So far, the picture of the relation between empathy and trust that I've been developing is generally rosy. Empathy provides a means of redeploying our approval of our own emotional perceptions to support attributions of epistemic competence to others. In so doing, it helps to clear away a barrier to testimonial trust. We might therefore think that if we want to promote trust in oppressed people's testimony, we ought to lean heavily on empathy. However, a policy of relying on empathy to support testimonial trust is more epistemically and morally problematic that it initially seems. To begin to bring the relevant difficulties to light, let me now introduce a few reports from members of oppressed groups about how they want their testimony to be taken.

From Mx Nillin (a nonbinary person): ‘I know that all of this [my experience of my gender] is very hard for many cisgender, and even many binary trans people, to understand. But the truth is that I don't need you to understand….You don't have to get it. Just respect that I know my body and gender better than you do’.Footnote ¹⁴

From an anonymous African American, writing on behalf of the collective Real Talk: ‘Don't question our experiences. You'll never understand them because you don't walk through the world with black or brown skin. Just sit there and listen’.Footnote ¹⁵

From Kirsty Major (a woman): ‘If a woman is saying you are being sexist, you definitely have to listen, because you've never lived life as a woman and don't know what it is like to face sexism every single waking hour of your life…If you still can't see it or understand it, at least acknowledge that the other person did’.Footnote ¹⁶

I'm not interested in whether these particular requests are totally fair or wise. Rather, I highlight them because they can be read as expressing a demand for testimonial trust that is not based on empathy. Nillin's remark can be understood in two different ways, depending upon the tone with which it is read. It could be understood as a declaration that efforts to really feel what it's like for Nillin are supererogatory: nice, but not required. Shift the stresses, though (‘You don't have to get it’) and it is readily understood instead as communicating exasperation with others’ apparent need to ‘see’ the other's point of view for themselves before placing their confidence in Nillin's testimony about their experience. The latter two remarks similarly recommend deferring to testimony rather than pursuing an understanding ‘from the inside’ of oppressed others’ points of view. They also communicate frustration with the emphasis their addressees place on ‘getting’ how women and racial minorities see things. All three claims can be understood as expressing a sense that others’ need to see things from the testifier's point of view, their reluctance to move on to trusting before doing so, is actually symptomatic of a distressing distortion in their patterns of trust.

I will now turn to two tasks. First, I'll show it can be morally important and epistemically responsible to trust others’ testimony even when one finds one cannot empathize with their evaluative outlooks. Second, I'll suggest that in relying on empathy to help us sustain and extend our testimonial trust, we make it harder for ourselves to trust under this kind of condition.

As we saw in the previous section, empathy works to enhance our confidence in others’ epistemic competence because when we manage to (imaginatively) work our way into experiencing things as others do, our own confidence in the accuracy of our emotional representations inclines us to regard their attitudes as appropriate. An important consequence of this dynamic is that the more alien we find the other's purported perspective, the less support empathy can provide for trust. Is it important, good, needful that we trust the other's testimony even when we find we cannot enter into the other's position, despite our best efforts to do so? To answer that question with a full-throated affirmation, I take it that we need to establish both that it could be epistemically responsible to trust in the face of failure to empathetically match the others’ emotional response and that it could be morally and/or practically valuable to trust in this way. If trust in the face of empathetic failure were morally important but epistemically irresponsible, or epistemically responsible but not particularly morally valuable, we would have reason to doubt whether we ever ought to trust testimony under these conditions, all things considered.

Let us begin with the question of epistemic responsibility: would a willingness to so trust necessarily compromise our status as appropriately cautious believers? There is reason to believe that, far from being reckless, this kind of trust could be a critical corrective to distortions in the attribution of competency that would arise from making trust highly dependent upon empathic success. Our success or failure to empathetically match another's emotional perception is often not a good guide to the propriety of others’ emotional expressions, because our ability to empathize is subject to two kinds of impairments that have nothing to do with the accuracy of the point of view that serves as our empathetic target. Interestingly, these impairments will end up making themselves felt particularly strongly in cases where a privileged person is asked to empathize with the perspective of an oppressed person.

First, our imaginative abilities are affected by numerous idiosyncrasies that have a real effect on what we can empathize with, and how. These features are undoubtedly helpful in some respects. Still, they can warp our empathetic abilities, such that the outcome of our empathetic efforts becomes a poor test for whether testimonial trust is warranted. I will offer just a couple of examples here. For one thing, imaginatively engaging with painful experiences is normally itself a painful activity. Imaginatively attending to experiences of rejection and humiliation as though from the perspective of the one demeaned involves intense focus on a bad experience, and that kind of focus may prompt dismay and dread in the empathizer. These are emotions we would rather avoid. All things considered, then, we will be more resistant to empathy with painful experiences than delightful ones, even if there is just as much to approve of in the perspective of the suffering person as there is in the joyful one. Our perspective-taking abilities also vary substantially depending upon whether the target of our empathy is part of the social ‘in-group’ with which we primarily identify.Footnote ¹⁷ Human beings struggle to imaginatively adopt the points of view of individuals who are not part of their social circles, even in the absence of a good reason to think that social outsiders have less coherent or accurate perspectives on the world. These two features of our imagination give us some reason to think that facts about what we are able to empathize with are not perfect guides to facts about whose perspectives are most worthy of our approval. More particularly, they also give us reason to think that suffering individuals who are part of one or more oppressed groups will receive especially little empathy from the relatively privileged (which means that empathy-based trust in their testimony on the part of the relatively privileged will correspondingly be in short supply).

Aside from problems related to limitations on our imaginative abilities, there are also reasons to worry about the reliability of would-be empathizers’ sensitivity to (certain kinds of) evaluative facts. In cases where our failure to empathize is the result of such insensitivity, it would be a mistake to allow the verdicts of our empathetic efforts to shape our testimonial trust.

Some rare individuals suffer from a complete inability to experience anything as frightening: to them, snakes, heights, and knives look exciting rather than terrifying. A person with this kind of impairment, like the real-life ‘fearless woman’ known in the psychology literature as S.M., will not be able to empathize with other's reported experiences of threats as frightening.Footnote ¹⁸ However, it does seem rational for S.M. to trust that others are right when they say that it is frightening to be faced with, say, a gun-wielding stranger: she knows that her abnormal way of looking at the world involves insensitivity to a quality that, were she able to track it, would help her to avoid the dangerous situations she has fallen into over the years. Now, if insensitivity to evaluative facts were confined to unusual cases like S.M.’s, this wouldn't give us much of a reason to think we really need trust not based on empathy. But is such insensitivity limited to cases of brain trauma? Or is it a much more common phenomenon?

Members of oppressed groups often claim that the prospect of real empathy with their position on the part of the relatively privileged is not just difficult but impossible. So, we find David, a homeless person, asserting, ‘You [well-off people] can't understand what it's like to be homeless,’ and Terrell Jermaine Starr, a black racial justice activist, claiming that when it comes to ‘a white supremacist system that oppresses us and excludes us in every area of American life – economic, educational, social and political – [e]ven the most empathic white person is just not going to know what that's like’.Footnote ¹⁹ These sorts of remarks seem to express the thought that occupying a position of privilege is something like being in S.M.’s situation. But it is unclear exactly what we should do with these claims, particularly since what is at issue is the question of when it is epistemically responsible to accept testimony.

Social epistemologists remain deeply divided over whether minorities and members of oppressed groups have unique access to some sorts of knowledge.Footnote ²⁰ I can't hope to adequately address the matter in a programmatic way here. I will suggest, though, that privilege can foster a kind of insensitivity to certain qualities, in analogy to the case of the fearless woman.

Consider, for example, perception of the threat of sexual violence. Men generally do not have a fine-tuned awareness of the threatening nature of casual street harassment.Footnote ²¹ In fact, some men report thinking of their shouted sexual comments and wolf-whistles as not only non-threatening but a boon for their targets: claims that women experience these things as frightening are sometimes met with rejoinders from the harassers that they themselves would enjoy being the target of such attentions, and cannot see them as hostile.Footnote ²² And yet: empirical evidence does point to the existence of a robust connection between the tendency to objectify women (of which street harassment is an expression) and the tendency to engage in sexual violence.Footnote ²³ If there is such a connection, that gives us good reason to believe that women are, in general, sensitive to a real feature of the world that (many) men are much less well attuned to, even when these men attempt to see things from women's point of view.

More generally, it stands to reason that people who are oppressed will be more sensitive than others to those evaluative features of the world that matter to navigating the dangers imposed on them by systems of oppression. Women's safety depends upon a keen ability to detect hostility or threat of a sexual nature in a way that men's safety generally does not. This observation is actually enshrined in some courts’ determinations concerning sexual harassment. When it comes to determining whether a work environment is hostile, for instance, the Ninth Circuit asks not whether any reasonable person would judge the environment to be hostile, but whether a reasonable woman would judge it to be so.Footnote ²⁴

The problems of imagination and sensitivity just surveyed give us strong reason to think that trust even in the face of empathic failure is sometimes epistemically warranted. Let me now turn to the question of such trust's practical or moral value for its would-be recipients, in light of what we have already seen of the value of testimonial trust more generally. Trust under these conditions could be practically valuable in part because it may sometimes be necessary for the uptake of important facts. Some conditions that people urgently need others to believe them about may not be possible objects of trust for those who stick strictly to trust based on empathy. The case of testimony about street harassment discussed above is, I think, one example of a situation where this value could be manifest.

But such trust isn't valuable just because it makes the transmission of important information possible. It also has a significant symbolic or communicative function. The risk inherent in trust makes the expression of it a powerful way of displaying and affirming one's belief in the testifier's competence and honesty. Still, not all trust has equal communicative power. Trust that is based (albeit indirectly) on one's confidence in one's own imaginative and discriminatory abilities (as trust based on empathetic effort is) comes comparatively cheap: it doesn't involve acknowledging the possibility that the testifier's competence might outstrip one's own. Conversely, if I trust you even though I find that your way of seeing is not empathetically accessible to me, my trust evinces a serious epistemic humility: it amounts to an admission that 1) I take my imaginative abilities to be so flawed as to deliver a seriously distorted picture of your circumstances, 2) I take my sensitivity to (some kind of) evaluative fact to be inferior to yours, or 3) both. For people who are systematically distrusted, an acknowledgement that their discriminatory abilities may outpace the truster's represents a dramatic departure from the status quo, one that is all the more significant because in trusting thusly, the truster makes himself epistemically reliant on the abilities of the one trusted.

I've claimed that empathy does not actually provide support for trust under some circumstances where trust would be both epistemically responsible and practically or morally valuable. At this point, it might seem tempting to conclude that we still have no reason to think that empathy's relationship with testimonial trust is ambivalent: empathy does not play a role in supporting all the testimonial trust one might reasonably demand, but it does nothing to undercut testimonial trust. The problem with this conclusion is that empathy's powerful ability to provide support for the trust it can sustain derives in large measure from the fact that when we ‘get’ some outlook empathetically, we are strongly inclined to regard it as correct. Conversely, when we attempt to empathize, and find we cannot, we are liable to be similarly strongly inclined to think that there can't be anything to what the testifier tells us. Claims that we can't empathize often have a judgmental dimension. Think about what is conveyed when we say: ‘I will never understand why he would want that’, or ‘You say it's worthwhile, but I just can't see it’. Or take this example from a Chicago Tribune Op-Ed: ‘So I will never get this sexting craze. I can't empathize with a teenage girl seeking acceptance by sharing photos of her breasts’.Footnote ²⁵ The literal claim being made is just that the author is not able to make sense of teenage girls’ sexting habits. But the heavy implication is that there is no sense to be gotten.

There is a defeasible but non-accidental connection between our trying and failing to empathize with another's attitude and our disapproving of it. A truly epistemically humble person might well think: whether or not I can imaginatively get myself to a version of the other's attitude has no bearing on whether her attitude is correct, because I myself am (for instance) emotionally and/or imaginatively stunted in some respects. Still, this kind of humility is hard to maintain. It is importantly different from, and more difficult than, deference to testimony about evaluative facts in cases where one has not even attempted to adopt the other's perspective. We may be relatively content, sometimes, to skip the empathizing and let others put in the work when it comes to determining whether something is beautiful, or just, or unsettling. However, once we do try to look at things from the other's point of view, marshalling our emotional sensitivities and applying them to the task of figuring out whether the other is right about how things are, the import of the choice to trust shifts.

Suppose we fail to ‘get’ the other's way of understanding things, despite our best efforts. In that case, trusting the other's testimony means accepting that one is flawed in (at least) one of two ways. First, one might opt to trust in cases like these because one accepts that one's perspective-taking capacities are in some respects deficient. Accepting that condition is no easy thing, because we heavily depend upon this imaginative ability for our everyday understanding of other people. It is a crucial item in our mental tool kit. The ability to take on other perspectives is crucial to our identity as social beings who experience our fellows not as mysterious ‘black boxes’, but rather as creatures with sensible, if not totally rational, preferences and projects.

The admission that other's perspectives may (sometimes, in some respects) be beyond our imaginative ken is, of course, not strictly incompatible with a conception of ourselves as social beings of this kind. However, it does pose some threat to it. And the grosser the imaginative deficit we have to admit, the greater the threat. For instance, it will be particularly difficult to accept that our imagination is not up to the task of supporting a trustworthy verdict about others’ attitudes in cases where it seems like we are not taxing our imagination very much. When the other person is not separated from us by obscuring expanses of time or space, and when their circumstances are clearly described (conditions that will frequently apply in cases of contemporary testimony about oppression), we cannot hide behind the defence that too much is being asked of our imaginations.

The concession in the face of empathetic effort and failure that one's imaginative abilities are in some respects inadequate may be paired with a second admission that is perhaps still more difficult. We have already discussed the strength of our inclination to go along with our own emotional representations of the world. The flip side of that inclination, which makes empathy such a powerful enabler of trust, is that it is extremely difficult to accept that one's emotional sensitivities are misaligned. Most of us, most of the time, are practically bound to think of ourselves as well-equipped to see whether a death is worth mourning, whether a habit is irritating, whether a joke is funny…and also whether an encounter is demeaning or threatening. Accepting that we are in some respects like S.M., that our emotional attunement to the world is dull or distorted (even if just with respect to a relatively narrow set of properties or circumstances) is a difficult prospect for reasons that largely parallel those in play in the case of acknowledging imaginative deficiencies.

Some people might seem to be cavalier about their own emotional failings. Certain forms of masculine ideology, for instance, practically make a virtue of insensitivity to the more ‘tender’ evaluative properties, such as the delicate, the adorable, and the kind. But there are forms of emotional sensitivity to evaluative properties that tend to lie near the heart of our self-conceptions, even for those who would not like to be called ‘sensitive’ or ‘emotional’. It would not be that unusual for a person to readily admit insensitivity to what is tactful, and easily opt to simply trust others’ testimony about this property. But accepting that one is not a fully reliable detector of what is cruel, for instance, is a much trickier proposition. An inability to see the cruel for what it is, even in a limited set of cases, cannot be written off as a harmless character quirk. The vast majority of people, I think, would agree that their good sense of what is needlessly hurtful is important to them as people who try to navigate life without being vicious to others.

The same kind of caveat I made in the case of conceding imaginative fault is relevant here. I can admit that my sense of the hurtful is significantly flawed in some respects without thinking that I am totally at sea when it comes to distinguishing what is wounding from what is harmless fun. I might maintain that my sense of this distinction is generally fine, except when it comes to a narrow band of experiences. But especially in cases where others perceive a policy, attitude, or behaviour as a clear-cut and central instance of malignity (as is often the case when we are dealing with testimony about oppression) the admission that one can't ‘get’ it is a costly one.

When one trusts another's testimony even after finding one cannot empathize with the other, it will not always be clear which of the admissions described above one is making. They tend to interact with and shade into one another; our abilities to accurately imagine the relevant background conditions may be shaped by our emotional sensitivities, and our emotional sensitivities may be influenced by our patterns of imaginative perspective-taking. Return to the cat-calling case. Consider the man who can't hear his comments about a stranger's body as hostile, despite his empathetic efforts. If he is nevertheless going to trust a woman's testimony about the issue, an admission like this might be appropriate: I'm having trouble finding my way to her perspective because my social position has both freed me from needing to cultivate sensitivity to the threat of sexual violence and given me permission not to hone that sensitivity through a practice of empathy with women's experiences.

Rather paradoxically, the more confidence we have in the accuracy of our own emotional perceptions, the more support empathy yields for one form of trust, but greater confidence in this accuracy will also tend to diminish our ability to trust in others’ testimony when we find ourselves unable to empathize with them. That latter form of trust is most easily fostered by a policy of humility about our own perceptual and imaginative abilities. When Nillin and Major, whose remarks I cited above, express exasperation at others’ apparent need to empathize in order to believe testimony about experience, part of what they seem to be targeting is a tendency for people to be too taken in by how they imagine things would look to them. At the same time, though, a policy of distrust in the deliverances of empathic efforts on the part of the relatively privileged would not only be very difficult to sustain, but would also deprive people who are in some respect(s) oppressed of an important source of confidence in their word.