What's so testimonial about testimonial injustice? A reply to Richard Pettigrew

I am interested in the paper "What is the characteristic wrong of testimonial injustice?" by Richard Pettigrew. I have quibbles both about its first-order discussion and its methodological maneuvers. At the first order, Pettigrew objects to one account of the "characteristic wrong" on grounds that I don't think come off and makes a positive proposal that I deem peculiar. (In fact, he makes two positive proposals, of which the second, which he seems to favor, strikes me as both unmotivated and unworkable.) This first order and his methodology intersect when he argues that while his positive proposal doesn't cover one example in the literature, that's fine, because the provision of an analysis and the determination of what the analysand even is are part of a hermeneutic circle in which both affect the other, and we learn, through his analysis, that that example actually is not an example of testimonial injustice, though it is an injustice.

Three examples are treated as the canonical motivations for the notion of testimonial injustice:

1. Cal, a visitor to Glasgow, receives advice from Elspeth, about how to get to Glasgow Central Station, but doesn't believe her as much as he should (or rather he grants her testimony insufficient credence, to use the parlance of the times). In a variation he does grant her sufficient credence, but only because his prejudice about the abilities of women to give directions is offset by his observation that she's a mountaineer and hence presumably reliable—even for a woman—about orientation.

2. Marge Sherwood is disbelieved by Dickie Greenleaf's father about Tom Ripley, in The Talented Mr Ripley.

3. Tom Robinson is disbelieved by the jury in claiming that he was Mayella Ewell's friend, in To Kill a Mockingbird.

The third is the example that Pettigrew will claim is shown not to be testimonial injustice on the grounds that his analysis does not apply to it. The first and its variants are his starting point. The second he more or less leaves aside.

Pettigrew's methodological interlude

Observing that the concept was introduced by Miranda Fricker apparently quite thoroughly, and yet "others have argued that she circumscribes its extension too narrowly" (1431), Pettigrew finds that "something strange" is afoot: "Fricker introduces a technical term of art partly by ostension and then by definition, and yet there is debate over whether it's the correct definition. How can this be?" (ibid.) His answer is that "her interlocutors are trying to pick out the most natural or unified or useful concept in the vicinity of the one that she specifies with her definition (ibid.; emphasis added). We want to include in the concept "all genuinely similar instances" (1432), where "genuinely similar" means that "the same wrong is done by hearer to testifier … and so the question of the distinctive wrong of testimonial injustice is entangled with the question of the concept's extension: we settle both together" (ibid.).

Now, I do not think that anything strange is afoot. Imagine that Fricker had, under the name "testimonial injustice", ostended the exact same examples and given the exact same definition, except she had also specified "the testimony is given and received in New Jersey". No one, I think, would be tempted by the thought that that was just what testimonial injustice was henceforth to be, as if Fricker had a proprietary interest in the term or must, by having given a definition, have penetrated into the nature of what she was discussing, even if all her examples really did take place in New Jersey. Any other observer is entitled to say: "you're clearly on to something—you have rendered a service by pointing this phenomenon out under this rubric—but your analysis is too restrictive". Or if we insist on calling it a definition, we can say: "you can introduce a term with a definition (but no one is required to think it refers to anything interesting), and you can introduce a term by ostension, but if you ostend and define we are all allowed to look at what is ostended and deny that the definition fits it, or to define the same good term differently".

Partly this is just because Fricker names the phenomenon with a descriptive name. If we know that the phenomenon in question is "testimonial injustice", then what possible relevance could its taking place in New Jersey have? If Fricker had ostended the same examples, given the same definition, and said "I dub this phenomenon 'qwxzlp'", then a reasonable reaction might be to simply let her have it. How would I know what qwxzlp is supposed to be in the first place? On the other hand, I could still look at the examples and think that something else is going on, which is more interesting.

Really, the issue is that (on Pettigrew's presentation!) Fricker is not giving a definition at all. She is pointing to a phenomenon, hitherto neglected, and offering an analysis of it. The name still guides us toward finding something specifically testimonial in the offing, and her analysis is oriented to explaining what, testimonially, is unjust, but we can all look at the (fictional) cases and judge for ourselves if that's really what's going on. We might conjecture that the order of priority is noticing that there's something interesting about such-and-such a case, surmising that it's a self-standing phenomenon in its own right and has something to do with testimony and justice, giving it the name "testimonial injustice", and then offering an analysis, guided by the surmise the name embodies.

And, given the name, we others can cast about for what other things might also merit the name. If notation is productive, how much more productive is a name? This need not be in service of the idea that some one wrong is involved in all cases; it could simply be widening the field of investigation. This is why I used "merit the name" rather than "be testimonial injustice". What other cases does that phrase aptly describe? There need be no suggestion that we are looking for "genuine similarity" at all, rather than investigating how best to use this evocative phrase.

Fricker, after all, is not guaranteed to have found the most illustrative, or all the illustrative, cases that seem to merit the name "testimonial injustice" merely by dint of having been first. Indeed, given the above sketch of how testimonial injustice may have come to be a topic at all, it seems that "testimonial injustice" ought to be regarded as a label pro tem for a set of phenomena of which we ought not assume that the "best" concept under which to array them makes essential reference to testimony at all (indeed we may also wish to hold open the possibility that the wrong involved is not a form of injustice but another form). Lichtenberg thought that "parabola" and "hyperbola" were names of rare quality, "for they express qualities of these lines from which all the others can be derived"; we oughtn't assume that "testimonial injustice" expresses any quality of the cases under consideration, even the involvement of testimony, as far as getting at what's really interesting in them goes. Guidance by the name may be very productive in scaring up other testimony-involving examples, but it may also prevent one from noticing other examples which bid fair to be, at root, the same kind of thing. In fact I think it's not too hard to work up such an example. Consider again Cal and Elspeth. In the example as given, Elspeth approaches a a visibly confused Cal, offers her aid, and is insufficiently believed for her troubles. But what if a confused Cal, realizing he could consult a local, looks about him for someone whose aid he will affirmatively seek? Elspeth is in his field of vision and he considers asking her before plumping instead for a man ten feet further off, solely because she's a woman and he a man. It seems to me as if, whatever is going on in the case of Cal's disvaluing of Elspeth's actual act of giving information, the same thing is going on in Cal's act of ignoring Elspeth as a potential source of information; that the two should be given the same analysis; that no difference of kind is in question at all. And yet since there is no real presence of testimony it would not count, for Pettigrew, as testimonial injustice.

The admission is now likely long overdue that I am in fact quite unfamiliar with the literature on testimonial injustice; I'm sure that the question of the unity and the testimoniality of testimonial injustice has been raised before. It is unreasonable in the extreme to expect each paper to include, just in case n00bs to the topic should chance upon them, a justification of the legitimacy of the topic at all before getting underway; one ought to be able to address a community that basically agrees on such things. But in Pettigrew's case in particular I think it is a fault that he doesn't say more about the presupposition that the "genuine similarity" depends on the same wrong occurring, or that the proliferation of examples of putative testimonial injustice is concerned with getting at what the most useful concept in the neighborhood is, since he brings the points of methodology up himself and attempts to use these presuppositions to argue that the Tom Robinson example isn't one of testimonial injustice. This he can only maintain if "testimonial injustice" does name one single unified phenomenon, rather than being a façade (in something like Mark Wilson's sense),¹ or being otherwise multifarious. Moreover, given the way testimony actually gets into his account, the question of testimoniality does seem worth addressing explicitly. The jurors, in Pettigrew's analysis of the Tom Robinson example, meet every single criterion of committing testimonial injustice against Robinson, except that they aren't responding wrongly specifically to his testimony. (Actually, I think that even on Pettigrew's own description of the case, they are, but he doesn't think so; I'll return to this below.) Pettigrew concludes from this that it isn't a case of testimonial injustice, and that this is a positive step, because we've learned something new. It is reasonable, I think, to ask why we don't learn from this that "doesn't happen to involve testimony" is more like "doesn't happen to take place in New Jersey" than previously suspected—that the category is less distinctive than we thought.

Pettigrew is aware that on his account "the characteristic wrong of testimonial injustice … doesn't seem distinctive to testimonial injustices", but deflects this with the remark that his "aim has been to identify the wrong that is present in all cases of testimonial injustice", not "a wrong that is present only in those cases", and after all "the alternative accounts [he's] considered [also] don't identify a wrong that occurs only in those cases" (1444). Fair enough. But if the ambition is to find "the most natural or unified or useful concept" where the unification comes in virtue of "the same wrong [being] done by hearer to testifier" (1432), then either one must give up on the idea that testimony has much essential to do with testimonial injustice—because one will be unifying with "too much"—or one must find a wrong that does essential involve testimony. (But wait: didn't I just say that Pettigrew excludes the Tom Robinson case on testimonial grounds? Yes, because his story is basically bipartite: testimonial injustice is an injustice of such-and-such a type, that occurs in the right sort of testimonial setting. The suspicious newjerseyness of this structure is apparent, I take it.) At a minimum, I think, a paper titled "what is the characteristic wrong of testimonial injustice?" should have more to say on its own behalf when it turns out that the characteristic wrong of testimonial injustice is not characteristic of testimonial injustice, the way most readers would construe that phrase, than "the goal was never to describe the characteristic wrong of testimonial injustice that way".

Pettigrew's first-order account

In what follows I wish to address the first-order story Pettigrew tells, starting with the objection he raises to the view that the wrong involved in testimonial injustice is that of failing to treat one's interlocutor with due respect, and then moving on to his actual proposed account what the wrong is, and his treatment of the Tom Robinson example.

Respect

"The hearer wrongs the testifier by not treating her with the epistemic respect his evidence tells him she is owed" (1437) sure seems like a plausible start. Pettigrew asserts that it falls to the following example. First, he says that "often our evidence … comes from many different sources and there are many permissible ways to weight these against each other; it pulls in different directions, and there are many permissible ways to resolve these tensions. And so, in such cases, there is a range of credences that rationality permits as a response to that evidence, rather than a single credence it demands" (ibid.).

Before even finishing Pettigrew's narrative, we can see where one objection to his objection will surface. "There is a range of credences that rationality permits" is true if it means "one can rationally arrive at any credence in this range by some particular route in each case"; it is not true if it means "rationality permits any credence in this range, no matter how arrived at". Given the many permissible weighings and resolvings of evidentiary sources and tensions, it may be the case that one can permissibly end up with credences in the 80—90% range, but only by a permissible weighing and resolving. This drops out from Pettigrew's presentation, in which he simply says that "Cal's background evidence is such that … his credence … should lie between 80 and 90 percent" (ibid.); as it happens, whenever a woman helps him, he believes her to the tune of 81%, whereas a man he believes to the tune of 89%. Therefore, per Pettigrew, Cal accords Elspeth due epistemic respect.

How might we think of such a range of permissible credences in this case? Perhaps Cal has reason to believe that those who approach one to offer help are generally knowledgeable (why else would they be offering help?) but sometimes, as he also has reason to believe, they're instead ignorant but overconfident (that's why else!), and sometimes too they have the competence but botch the performance; perhaps also he's heard that the Scottish are a friendly people (so that they may be more inclined to try to help even though they don't know), but Elspeth seemed more neutral than friendly (so maybe that's not what's going on) and as nothing in the scenario compels a specific weight or resolution to the different directions these point, there's some wiggle room for where he ends up. When Pettigrew says that (epistemic) disrespect in this case involves A treating B "as less reliable than A's evidence says B is" (1437; slightly modified), we must be cautious: it's not that A's evidence says B's reliability is "anywhere between 80 and 90%, take your pick in that range", it's that A's evidence doesn't assign any reliability to B, until A resolves it, and all permissible resolutions lie between 80 and 90%.

Suppose that Cal, having awoken in an optimistic mood, full of faith in human kindness, minimizes considerations that would lead him to a lower credence, and would accord Elspeth 87% credence, but for the post-resolution operation of a blanket "prejudice against women" tax of six percentage points, leading to a final assignment of 81%. Here surely he has accorded her less credence than his evidence-as-finally-resolved demanded, since it demanded 87%. Does Pettigrew wish to add that his prejudice against women is another source of evidence to be weighed, which it is permissible for him to take into account? Presumably not, because the construction of the original case was such that Cal assigns a credence too low given his evidence, which he would not be doing if his prejudice were part of his evidence. Does he wish to say that resolving the conflicts by appeal to a prejudice is legitimate? Well, yes, actually (1438f)! Since "everything that is relevant to reliability is already included in the evidence", and thus arriving at some specific credence will always involve an "appeal to something that is arbitrary in some way", there is no basis for complaint. Strictly this is not true: Cal's buoyant mood, which silences certain considerations, is not a consideration he appeals to, and frankly I find the idea that in this situation, in general, we make any kind of appeal to a consideration to "pick our credence" suspicious. Cal's prejudice need not be a consideration he appeals to, either; it could simply be operative in his prejudicially weighing certain parts of his evidence too much or too little (that is, Cal does not resolve his evidence by appeal to a prejudice, he resolves his evidence in a prejudiced way), and it certainly seems plausible that even if he ends up with a "respectable" level of credence, prejudicially playing sources of evidence off each other is disrespectful to Elspeth. Respect is not an output of this process, but a mode in which it is conducted. But even if it were, what of it? The idea that we'd have to go beyond our evidence to something not evidential doesn't mean that we can permissibly go beyond it to anything.

All that to the side: even if we ignore the idea of permissibly arriving at a given credence, so that the mere fact of 81% credence would be fine no matter how arrived at, I still wish to say: of course Cal disrespects Elspeth (epistemically, even). There's a sort of fallacy of composition at play here: what may be properly respectful in any one case is not necessarily respectful when it takes place systematically. If you do the minimum required of you as a friend vis-à-vis one person on one occasion, that may be compatible with your respecting them; perhaps you have a more than sufficient reason for not having done more, and after all you did do the required. If you always do the minimum with respect to that friend, while showing that you're perfectly capable of doing more because you do do more with respect to another friend, then you are indeed being disrespectful, even though on each occasion you exceed the threshold. Systematically differentially doing the bare minimum is disrespectful.² (Perhaps this is not "epistemic disrespect".) Even aside from this compositional argument, the slide from "rationality permits me to grant you 81% credence" to "in granting you 81% credence I am not being disrespectful" is far too quick. Perhaps the reason Cal grants Elspeth merely 81% credence is not his specific prejudice about women's ability to give directions but precisely his disrespect for women. The fact that he doesn't disrespect women enough to dip below the 80% threshold in this case does not mean he is being adequately respectful; at best, it means that his disrespect hasn't led him to make an additional epistemic error.

The positive story

Here is Pettigrew's "account of the wrong that is present in cases of testimonial injustices"; it is subject to a few adjustments and then a fairly radical modification, which will be discussed below:

B gives testimony to A; A's response is guided by [prejudices about relevant competences]; so A treats B's testimony as less reliable than A's evidence tells them it is. This is then expressive of a social hierarchy with respect to which A is advantaged relative to B, and so A thereby treats B as his [sic] social inferior; and in virtue of that, A commits a relational injustice against B—A's treatment of B prevents them from meeting as equals. According to the account built on Fraser's suggestion, this is the distinctive wrong of testimonial injustice. (1442)

As we've seen, "distinctive" does not mean that it is a wrong distinct to testimonial injustices—you cannot distinguish a testimonial injustice from another by its presence. You can (according to Pettigrew) only distinguish a non–testimonial injustice by its absence. An adjustment (1443): one might be met not as equal because one isn't an equal, in some relevant social hierarchy (eg one person may simply be a scoundrel and unworthy of esteem). Thus we append "without warrant" as appropriate. It will be seen that the role testimony plays in this account is that it occasions the injustice; A could express the same hierarchy and commit the same injustice with other promptings.

Undue deference

Discussing the case of a white person granting too much credence to a Black man on the topic of illegal drugs (1445), Pettigrew says that his account gets it right, even though A treats B as more reliable than their evidence tells them B is. The reason is that by treating the Black man as more reliable about drugs, the white person still casts him as their social inferior, even if he is cast as an epistemic superior with regard to this unsavory topic. So "less reliable" in the account really should be read as "either less or more reliable", so long as, with respect to the topic in question, lesser or greater reliability is associated with an inferior position in a social hierarchy.

Bear with me as I reiterate this point: on this account, testimonial injustice occurs because the hearer's incorrect degree of belief in the testifier's testimony expresses a hierarchy on which the testifier is inferior. Epistemic inferiority or superiority is material only insofar as it positions the testifier in a socially inferior role.

If we allow ourselves to be guided by the label "testimonial injustice" and the thought that this injustice has to do with not meeting another as an equal, we might think as follows: one can be granted undue deference about anything, not merely a socially disfavored topic. If A treats B as more reliable than is warranted by A's evidence on the topic of, say, the ballet, does A meet B as an equal? B may not be harmed by this deference, but Pettigrew doesn't wish to insist that a harm to B is necessary for testimonial injustice to be in the offing. Insofar as the ballet is a culturally elevated sort of topic, A may be assimilating B to an unequal, but higher, position on a social hierarchy; is this also unjust? It does seem to mean, as can also be said in the hitherto considered cases, that A doesn't really meet B directly. The operation of a prejudice prevents A from encountering the person actually before them.

The examples Pettigrew considers are one-off exchanges (even Tom Robinson's testimony is treated as a single disbelieved utterance). Imagine, though, that this deference plays out not just with respect to a single statement but over the course of a conversation. One might say: in this case the problem is not so much that A doesn't meet B as an equal but that A prevents B from meeting A as an equal, or that the mode of A's not meeting B as an equal is that A withdraws their equality from the meeting. "The yes-man harms/is unjust to his interlocutor" may not be the most sympathetic position, but I think there's something there. Here we are not mediating the one-on-one injustice by a broader social hierarchy; it is something A does directly to B, perhaps because B is a member of some class, but not to B qua member-of-some-class.

Pettigrew doesn't think views that class credibility as a good to be distributed have much promise, and in that I think he's correct; this insight also smooths the way, though, to thinking these cases of too much credibility (even on socially neutral or favored topics) at least can constitute an injustice. There may be nothing wrong with giving someone more than they're owed when it comes to doing them a favor, but granting someone excessive credence isn't doing them a favor at all. I wish to say: denying them your disagreement is disrespectful. It prevents each of you from a genuine encounter with the other—it is a one-on-one injustice.

Idiosyncratic prejudice

As mentioned, Pettigrew actually gives two positive accounts (the second is considered below). Pettigrew thinks that the second one solves the following problem: the original account requires a hierarchy in which A is advantaged relative to B, but what if Cal and Elspeth are both women? Surely a woman can harbor prejudices about women's ability to give directions, and the female Cal should be able to be testimonially unjust to Elspeth too.

Let us stick however with the thought from the preceding section that injustices are often done one-to-one: I do something to you. Why must my harming you by disbelieving you, or believing you insufficiently, be referred to a social hierarchy? I might have quite niche or idiosyncratic prejudices, corresponding to no real social hierarchy. Pettigrew for his part says that "surely we want to identify a wrong that is present only when the person who receives the credibility deficit belongs to a group that is stereotyped as being less credible" (1440),³ where presumably he means "stereotyped by people at large as being less credible"; all I can say is that I'm not at all sure of that. Perhaps I believe that it's men who are bad at understanding and giving directions, and women who are good at it, and discount Alistair's advice about navigating Glasgow central station, simply because he is a man. Surely we want to say that I thereby work an injustice upon him. I mean: how not? I prejudicially disbelieve him, even if not in accord with a socially cognized hierarchy.

Perhaps the problem here is that I don't treat Alistair as my social inferior by dint of prejudicially insufficiently believing him. The wrong on Pettigrew's account is not-meeting-as-equal-because-treating-as-social-inferior. It could be possible to make my treatment of Alistair conform to this story; it could be the case that my behavior is expressive of a "social hierarchy" on which men are inferior (this is connected to their supposed navigational deficiencies), just a social hierarchy that doesn't have any social reality. (It would not be a "social hierarchy" in the sense of being a hierarchy actually expressed in social standings but in that of being a hierarchy that pertains to social standings.) I treat him as my inferior, even though, socially, he isn't. This might enable the account to handle this case, if we wish to handle it. But perhaps my prejudices run like this:

Men on the whole are superior to women, in countless domains—all the ones that matter, really. But you just can't rely on a man to give you good directions! There's nothing wrong with that. They're just not mindful of the details; they're more big-picture types, which is great normally but not if you need to figure out where you're going. Women are the opposite—that's what makes them so good at domestic organization, keeping spaces tidy, stuff like that.

That is, I endorse the conventional social hierarchy, I just flip who in it is good at giving directions. (It's not as if there is a natural or necessary connection between the social hierarchy and the specific prejudices associated with it.) Now my believing him insufficiently is expressive of a hierarchy on which men are superior, because the valence of navigation is flipped. Even here it seems to me that I'm behaving unjustly toward him.

We also have prejudices which are not obviously connected to a relevant social hierarchy. I happen to be a Pisces, and Pisces apparently are thought, by the people who think this kind of thing, to be overly emotional; some such person might therefore discount my testimony on the grounds that I must be making too big a deal of whatever it is I'm talking about—typical Pisces! I do not believe that, in general, there is a social hierarchy on which Pisces rank low, or high, or at all; I don't even think that the believer in astrology need have a ranking of signs: each sign will have some good, some bad, and some neutral traits to be mindful of in some situations. Unlike being bad at giving directions, which it is not bad to be in itself, being overly emotional is arguably not neutral (being overly anything is bad), but it is, to the astrology-believer, hierarchically neutral; at least, being conceptualized as overly emotional qua Pisces doesn't lead to a hierarchy of astrological signs. And yet it seems to me that being disbelieved on the grounds of my astrological sign is also unjust; I am granted too little credibility because of a prejudice, and this means my hearer doesn't meet me as an equal (this seems a congenial formula, albeit one yet to be filled in): isn't that enough for injustice to be on the scene?

Others' meeting as equals

Pettigrew is much exercised by the thought that A might unjustly believe B insufficiently on a topic on which they acknowledge B to be both skilled in general and more skilled than they are themselves. This leads him to the radical revision foretold above. I do not believe, though, that his motivation really makes sense, or that his revision does.

Here's how he sets up the motivating case, the mountaineer variation on Cal and Elspeth:

As Elspeth approaches Cal, he notices that she's wearing a badge of the Scottish Mountaineering Society. This, together with his background evidence and Elspeth's testimony, requires him to give credence 99 per cent … but he gives only 90 per cent, because, while he recognizes that women mountaineers are more competent at giving directions than the average person, he is still guided sufficiently by his prejudice that they aren't as good as his evidence tells him they are. What's more, Cal thinks of himself as having a pretty average sense of direction … so, at least in a standard colloquial sense, Cal treats Elspeth as his superior, not as his inferior … it doesn't seem that Cal's behavior impairs the abilities of him and Elspeth to meet as equals, which is supposed to be the source of the injustice (1446)

Before going further, I want to address this locution that "Cal's behavior" does or doesn't "impair his ability" to meet Elspeth as an equal. It is new on the scene; previously, we were concerned with whether Cal did or did not meet Elspeth as an equal, and why. This new formula strikes me as confused: that which impairs an ability impairs its exercise; my ankle injury impairs my ability to hop on one foot, but my faulty hopping does not impair that ability: it is the better or worse exercise of that ability. (Of course, my hopping could impair my ability if it exacerbates my injury, and Cal's behavior could impair his ability if it strengthens his prejudice. But such consequentialist construals have been explicitly excluded (1443).) Cal's behavior is the exercise of his ability to meet with Elspeth as an equal or otherwise. His behavior is his meeting her as an equal, or not; it is not aptly described as impairing, or not impairing, the ability to do so. It's not as if he first believes her insufficiently (and this is his "behavior"), then as a second step taken on that basis does some other thing to treat her as his inferior, and finally finishes it all off by not meeting her as an equal, as a third action. If we wish to find something that might or might not impair his ability to meet Elspeth as an equal, his prejudice is a good candidate; in that case, though, it seems clear (insofar as we understand what meeting as equals is at all) that his prejudice does do so. The fact that his prejudice about women is offset by a countervailing belief about mountaineers doesn't mean that his prejudice is not affecting his ability to meet Elspeth as an equal, any more than my wearing a life vest would not mean that my ability to float is not impaired by having a weight strapped to my waist. You can tell that she is still treated as inferior because men don't need to be mountaineers to have their directions believed. Here as in the discussion of respect I suspect that Pettigrew is misled by his use of specific credences expressed as percentages. The important point has become "is the final number correct?", with the question of how that number was arrived at entirely suppressed.

Since, as we saw in the case of supposed expertise on drugs, it is not the mere fact of judging someone to be superior in some epistemic domain, but what that means for their social status, that makes for the wrong on Pettigrew's account, the fact that "in a standard colloquial sense" Cal meets Elspeth here as his (epistemic) superior is irrelevant. We have a clue to this irrelevance in the fact that in the original description of the case no mention was made of Cal's assessment of his own skill at giving directions. There's nothing wrong, after all, about regarding someone as your inferior with regard to giving directions, and that wasn't what Cal was doing (in the original case) anyway: with regard to skill at giving directions, he was regarding Elspeth as inferior to herself (or, at least, inferior to how his evidence told him she was). The harm was that in so doing he expressed an unjust hierarchy on which women in general are inferior socially to men in general.

With regard to the mountaineer case, since Cal still believes Elspeth exemplifies womanly inferiority in her being worse at giving directions than a similarly situated man, ie than a male mountaineer, and expresses this prejudice in discounting her testimony, there seems to be no obstacle to saying that he treats her as a social inferior, while acknowledging her direction-giving superiority to him. It is still the case that he regards her as worse than she is. It would be another matter if her being a mountaineer simply overpowered his prejudice about women, so that it played no role in his uptake of her testimony—but then presumably he would also not commit an injustice.

There is however another way, not explored by Pettigrew, to construe a similar example so that it's at least questionable whether A treats B to occupy an inferior social role. Social hierarchies are multiple and cross-cutting, so there's not necessarily going to be a way, in an arbitrary interaction, to place the participants unambiguously on a single scale with one above the other. In discussing the mountaineer example, Pettigrew considers "two scientists present[ing] their findings, one a man, one a woman", to an "audience [that] knows no more about their competence than that both are scientists". When the audience members "set their credences … they appeal to a sexist distribution in which scientists who are women are, on average, less competent to than scientists are men", so that "they will give less credence to the woman's testimony and more to the man's" (1448). Now we can say everything we said about Cal and Elspeth-the-mountaineer here too: that the audience members treat the woman scientist as a social inferior in that they make use of a prejudice that partly constitutes a hierarchy on which women are lesser than men, and that this is visible in their treating her as less credible than her peer (rather than as less credible than they themselves are). But perhaps "scientist" is a particularly august role, so that while they disvalue her qua woman they value her qua scientist—this is separate from believing her qua scientist. Now their degree of belief is expressive of a social hierarchy on which women are inferior than men, but their acknowledgement that she is after all a scientist is expressive of a different hierarchy on which scientists are superior to the lay community, so that whether they treat her as their "social inferior" overall, thus not meeting her as an equal, may be up in the air. But perhaps all the story needs is her being treated without warrant as an inferior along some axis; perhaps, too, "treats as a social inferior" is just as much a misleading output as "has credence x%" and we should focus more simply on the operation of the prejudice, locating the wrong directly there.

Here's how Pettigrew approaches this mountaineer conundrum. First, he asserts that (1) "to avoid a relational egalitarian injustice, it is not sufficient that A and B regard each other as equals; it is also necessary that the wider society in which they live regard them in that way" (1448); as an example, (2) "partners in a mixed race relationship cannot meet as equals, however fully they treat one another as equals, while society treats one as the superior of the other" (ibid.). In the example of the scientists, while the audience members treat both scientists as their superiors, (3) "it is the two scientists who are less able to meet as equals". Something similar is said to be happening with Cal and the mountaineer: (4) "Cal's behavior does not impair his ability to meet with Elspeth as equals … but it does impair the ability of others to meet as equals, since it expresses a hierarchy that ensures that … [a woman] will be treated as less competent" (ibid.).

(This is the solution that's supposed to solve the problem of a woman being testimonially unjust to a woman (1449).) Now here we really want to know much more explicitly what "meeting as equals" actually does mean in Pettigrew's eyes. If the partners in (2) who always treat each other as equals don't meet as equals, then how is it that in the original statement of the wrong "A's treatment of B prevents them from meeting as equals" (1442)? They never stood a chance of meeting as equals. Perhaps this is why we get the unheralded appearance of the locution "Cal's behavior does/does not impair his ability to meet Elspeth as an equal" rather than simply "Cal does/does not meet Elspeth as an equal": sure, he was never going to meet her as an equal anyway, because of (1), but he could make things worse for himself. As I've said, I don't think this locution makes much sense in itself, and I don't think it helps Pettigrew out here, because if, in (4), Cal's behavior doesn't impair his ability to meet Elspeth, that is presumably because Cal does in fact believe her sufficiently much (because he takes her to be his navigational superior), but if that's what it takes not to derail meeting another as equals—merely getting the credences right, despite the continued operation of personal prejudice—then what about wider society (1) impairs the ability of the partners in (2), since in treating each other equally they presumably get the credences right, too? Then too (reiterating here): Cal is said to treat mountaineer Elspeth as his superior because of his treatment of her navigational capacities in (4), but it isn't navigational but social equality (presumably!) that's in question in (1).

At any rate, we are certainly entitled to ask how, in (4), Cal's behavior does impair the ability of others to meet as equals. Pettigrew doesn't really elaborate on this, beyond saying again that Cal's behavior is expressive of a hierarchy on which women are inferior, but this doesn't really answer the question: how does Cal's expression of a hierarchy in the privacy of his own assignment of credences do any such thing? Yes, the hierarchy impedes the meeting as equals of others,⁴ but it was going to do that anyway, and we aren't considering a consequentialist interpretation on which Cal's behavior furthers or maintains the hierarchy, or even his own acceptance of it (which, in any given case, it also might not do). Why Cal's behavior on any one occasion, which may be consequentially inert, is relevant is unclear; it strikes me that Pettigrew would be on firmer ground stating simply that Cal's state of being prejudiced impaired the ability of others to meet as equals, but he exists in that state regardless of whether he is currently listening to a woman's testimony or not, and here too the ground is only really firm if we go with the consequentialist interpretation. To the best of my ability to tell, there wasn't a real problem for Pettigrew's original account here, and his solution to this non-problem doesn't make sense.⁵ Cal still seems to be treating Elspeth, even mountaineer Elspeth, as a social inferior, in virtue of his judgment of her navigational abilities being guided by a prejudice which expresses a hierarchy, even though he accords her more skill than he accords himself. That was the whole account to begin with. If we accepted it before, these variations don't seem to be an obstacle to accepting it still.

Respect redux

What's wrong with treating someone as a social inferior, or what conditions must be met for such treatment to be wrong? Pettigrew does observe (with respect to hierarchies of esteem) that there's nothing wrong with treating a dishonorable scoundrel as one's social inferior, because they are one (1443); it is only when such treatment is unwarranted, as it is in the case of Cal and Elspeth, that an injustice is committed. But hang on: just as the wretch who sells out his friend is inferior on some hierarchies to virtuous folks such as you, dear reader, and such as me, isn't Elspeth actually the social inferior of Cal? Recall that Pettigrew even wants to restrict the domain of testimonial injustice so that it can only be done to people who are "stereotyped as being less credible" (1440), which is a way of being a social inferior. Women aren't actually less skilled at giving directions, but they are actually socially disadvantaged. It's not even necessarily bad to treat them as such. Men are sometimes reminded that we can use our privileged positions on women's behalf, for instance by making a point of loudly crediting a woman with having originated an idea in a meeting. Doing something like that is treating a woman as a social inferior—that's why the man does it—albeit not in a way that accepts that she should be an inferior. (Perhaps we might try to distinguish Cal from this present case by saying that Cal treats Elspeth as an inferior whereas here the man treats the woman as treated as an inferior (but doesn't himself treat her as an inferior), baking acceptance of the hierarchy into "treats as an inferior". I'm not really sure about this.)

If that's the case, then perhaps Pettigrew's account shouldn't turn on not treating as an equal but on expressing an unjust hierarchy, since in the second case above the not-treating-as-an-equal comes in response to, not expression of, the hierarchy. (I'm assuming that "expressing" a hierarchy means at least acting in acceptance of it.)

If we simply say that, then the wrong will seem to be a personal failing, not a wrong done to anyone at all; this is actually ok by Pettigrew (1449), though it does to my mind call into question the justice of the label "injustice". If the wrong is the expression of the hierarchy, the agreement with an injustice, that may be a form of vice, but it is not obviously injustice itself. In keeping with the idea that "testimonial injustice" is serving primarily as a label, this is fine with me too; it does, however, seem worth pointing out explicitly.

If we simply say that, then we also won't be able to make sense of my behavior toward Alistair, or the astrologist's behavior toward me, or the over-eager yes-man's behavior, as instances of the same kind of phenomenon. That too might be fine—not because they aren't testimonial injustice, but because "testimonial injustice" is a multifarious thing, which admits of different analyses in different cases. But perhaps we can say something else, something telegraphed by the section heading. When you allow yourself to be guided by a prejudice, whether it's rooted in sexism or astrology, rather than your evidence concerning the person before you, aren't you failing to respect the person before you? To treat them as (as far as your evidence tells you, anyway) they are, not as your prejudice, whether it redounds to their credit or detriment, tells you they must be. This is a less, because not at all, specified form of respect than the one Pettigrew considers, which seems to bottom out in "you respect someone epistemically when you grant them the appropriate credence". But it seems to me to capture something, at least, of what's objectionable about discounting (or overvaluing) a statement on the basis of a prejudice about the speaker—something about its being an affront to the other, even if they never learn of it. It is an affront to mountaineer Elspeth that her credibility is dinged on the basis of her being a woman, even if she is still trusted overall.

The exclusion of Tom Robinson, & methodology redux

Here is the crux of Pettigrew's discussion of this case (1445; emphasis in original):

It is not unreasonable for the jury to believe that, if Robinson is innocent, he'll be telling the truth, and if he's guilty he'll be lying. And that requires no prejudice about the trustworthiness of Black men … the jury's credence in Robinson's guilt ends up determined almost entirely by their prior credence that the version of of events he tells would be true and their prior credence that the version that Ewell tells would be true. … It is here … that their prejudice enters: they give higher credence to the sort of event that Ewell describes than to the sort of event that Robinson describes … So, while they do end up treating Robinson as less trustworthy than he is, this is because of their prior beliefs about the sort of situation he describes, and not because he describes them, and they antecedently think him an untrustworthy person. Testimonially, they treat him like any other defendant: likely to lie if guilty and likely to tell the truth if innocent. Their prejudice toward him is not a prejudice that he's less trustworthy, but a prejudice that he's more likely to rape than he is. … The prejudice is not a testimonial one: it is not a prejudice concerning the standing of the testimony of Black men, but the prejudice that Black men are more likely to rape than they are … a relational injustice occurs in this case, and of course it is grave, but it is not a testimonial relational injustice.

There is something strange about this. In the example of Cal and Elspeth, nothing turned on the idea that Cal thought Elspeth an untrustworthy person, merely that he thought her untrustworthy on this topic, or as one might say about the situation she describes (the situation is that Cal's destination lies to his left, or whatever). The jurors' "prejudice is not a testimonial one": the prejudice that says women are navigationally impaired is also not a testimonial one. Or to make things more obviously about the description of a situation, suppose that Elspeth was telling Cal about the time she and a male friend got lost somewhere, and she says either (a) that she through her navigational acumen quickly realized where they were and had gone wrong and saw them right or (b) that she was totally befuddled and minorly panicked but fortunately her male companion saw them right. If she says (a) Cal will be skeptical that it really happened that way; if she says (b) Cal will accept it without quibble. Can't we say that in this case Cal doesn't antecedently believe she's untrustworthy, but because of his prior beliefs about the situation she describes (women be getting lost!) he believes her insufficiently well? "His prejudice toward her is not a prejudice that she's less trustworthy, but a prejudice that she's more likely to get, and stay, lost than she is." Because of this prejudice, he doesn't trust her. So it really isn't obvious to me why Pettigrew isn't describing by his own lights an example of testimonial injustice.

Perhaps the fault lies with me, though, and indeed his account doesn't cover this case. Nevertheless it seems that he himself is describing a situation in which a Black man's testimony will be disbelieved unless the content of the testimony corresponds to the prejudices of the hearer. Even outside a courtroom setting, if Tom were to say "I'm off to visit my friend, a white woman", he would be disbelieved. "You may testify to whatever you wish to, as long as it's p": perhaps this isn't an example of testimonial injustice sensu Pettigrew. But if this isn't worthy of the name "testimonial injustice", I'm not sure what is. If we had a compelling reason to believe that whatever the extension of "testimonial injustice" is, it must admit of one account, being a single, unified phenomenon, we would be forced to conclude that either Pettigrew's account fails or Tom Robinson's case really isn't worthy of the name. But do we have such a reason? This is a term that was made up recently. It could be a genus name rather than species name; it could be a useful shorthand for a collection of different phenomena united only in that injustice involving testimony is done. There need be nothing wrong with that, insofar as there is some useful purpose being served in any given case, or insofar as, perhaps, no one had hitherto thought that testimony even could be a scene of injustice. (Whereas "New Jersey injustice" would not be useful, insofar as we already knew that New Jersey is home to many and varied injustices.)

Similarly, if we had a compelling account of testimoniality in one case but not the other, that might tell on the one's behalf. Pettigrew as we've seen believes that the wrong done to Tom Robinson is "not a testimonial relational injustice" (1445); he reiterates in his conclusion that "the wrong done [is] ethical, or at least more broadly epistemic, rather than testimonial" (1449). But on his account "the wrong done" is that of meeting not as a social equal, or of impairing the ability others to meet as equals, which is also not "testimonial", as he has acknowledged (1444). Or at least, it is not clear what it means for a wrong to be testimonial, such that Cal's treatment of Elspeth gets in and the jurors' treatment of Robinson does not. In the latter case, it seems that the wrong is not testimonial insofar as Robinson's testimony is merely the occasion for an injustice which according to its kind could also have occurred in other circumstances. But the same seems to be true of the wrong Cal commits with respect to Elspeth's testimony. Indeed, for all that Pettigrew says, allowing that each is a relational injustice involving a status hierarchy, it seems plausible enough to say that the same wrong is committed in each case, so that one wrong's being testimonial and the other not begins to seem rather inessential and the nature of the philosophical benefit we are supposed to receive by learning that one is and the other isn't testimonial begins to seem rather unclear.

In the abstract, we have two cases, A and B, each of which is prima facie plausibly described as "testimonial injustice". Pettigrew presents an analysis of A that (he claims) doesn't include B; if we accept that analysis we accept at the same time that B is not testimonial injustice. (Actually it is not extremely clear to me how much of his argument about Tom Robinson is a working-out of the consequences of his analysis of Cal and Elspeth and how much is just a direct analysis of the case itself and the interplay therein of testimony and prejudice, in which case it wouldn't really be a case of settling the extension and the analysis together but rather a constraint on any analysis that it exclude the case. But he presents it as a consequence of his analysis.) Pettigrew doesn't mind that his account excludes B because of the hermeneutic circle in which anyone analyzing the concept finds themselves: the analysis constructs the cases and the cases influence the analysis. But one must at least acknowledge the influence of starting conditions. Pettigrew takes A as an utterly secure starting place, and relies throughout on his sense of what obviously is or isn't included in the extension; confident in these assessments, he formulates an account which excludes B. But what if he had started with B? Insofar as B is excluded not on internal grounds but on the basis of the account of A, this is presumably legitimate. Perhaps he would have arrived at an account capacious enough to include A as well; perhaps he would have found one that excludes it. In the latter case, what stops us from saying that they are both testimonial injustice, A in its way and B in its? What—given the wan role for testimony actually included in Pettigrew's actual account—stops us from saying that the most informative thing to say about them, as far as injustice is concerned, is simply that they are both injustices, each in its own way?