In “How to Make Invidious Distinctions Amongst Reliable Testifiers,” Elizabeth Fricker's central aim is to defend what we might call the Knowledge View of Testimony against the Reliability View. According to Fricker (Reference Fricker2015a), the latter includes the following key thesis:
Reliability Supply-Side View of Testimony (RSSVT): On the supply side, all that is needed for the Core Mechanism to operate to yield knowledge on an occasion O is that the testifier is a reliable testifier – one who satisfies (RT).
where (RT) is understood as:
Reliable Testifier (RT): The speaker is such that on this occasion not easily would she assert that P unless P.
She writes that for the Reliability View “all that matters [within epistemology] is that there are [reliable testifiers], and being so they are an epistemic source for other people, recipients of their reliable testimony … and all reliable testifiers are equally good qua reliable source; so that further distinctions amongst reliable testifiers are irrelevant to the epistemology of testimony” (Fricker Reference Fricker2015a). In contrast, Fricker favors the Knowledge View of Testimony, according to which there are “grounds to discern epistemically relevant distinctions within the class of reliable testifiers.” In particular, “only those who are reliable because they speak from knowledge of what they assert are sources from which knowledge can be gained via the Core Mechanism—the proprietary way in which speech acts of telling as such are able to spread knowledge” (Fricker Reference Fricker2015a). Thus Fricker argues that (RT) needs to be replaced with:
Knowing Testifier (KT): The speaker is such that on this occasion not easily would she assert that P unless she knew that P.
In this short response, I will raise five problems for Fricker's view and the arguments put forth to support it.
1. THE TARGET
The first problem that I would like to discuss is that it is not at all clear who Fricker's target is. Reliabilists themselves focus not on the modal thesis found in (RT) but on reliable processes of the appropriate level of generality – processes that reliably lead to true beliefs that fit broad, explanatory patterns. Thus, for the reliabilist, it is not the case that “all reliable testifiers are equally good” and that “further distinctions amongst reliable testifiers are irrelevant to the epistemology of testimony.” Rather, some reliable processes have the appropriate level of generality and thereby have greater explanatory power than others.
The view of the epistemology of testimony that I develop in my (Reference Lackey2008) also has a reliabilist component and thus might be seen as a target of Fricker's criticisms. But again, my view fails to fit the description she provides. I argue that a necessary condition for testimonial knowledge is that the speaker's statement needs to be reliable, which can, in turn, be fleshed out in a number of different ways. For instance, it might be that the statements need to be produced by reliable testimony-producing processes, which we have just seen does make epistemologically relevant distinctions among reliable testifiers. It might also be the case that the statements need to be produced via testimony-producing intellectual virtues. This, too, would enable important distinctions to be drawn among reliable testifiers, as within this class of testifiers, only some would be testifying virtuously.
Moreover, Fricker characterizes the (RSSVT) as saying that all that is needed on the supply side for the core mechanism to yield knowledge is a reliable testifier, as understood in terms of (RT). But, again, this fails to apply to Fricker's intended target. Most reliabilists and virtue theorists include, in addition to the reliability in question, that the processes or virtues function in a suitable environment. So more is indeed needed on the supply side than mere reliability.
Relatedly, even if Fricker successfully reveals problems with the Reliabilist View as characterized by the (RSSVT), this does not support her preferred Knowledge View, as there are many views in between these two. As I've already made clear, my view, the standard reliabilist view, and one grounded in terms of intellectual virtues are all candidates.
2. THE CASES
Fricker goes on to support the Knowledge View by challenging the Reliability View via the following two examples:
Would-be Liar: Simonetta for some reason wishes to deceive Paul, by telling him a false story. To do so, she decides to tell, as if true, all and only the details from what she believes to be a fictional story about an incident in a small English village during the second world war. Unbeknown to her, this narrative is in reality not fiction, but an entirely accurate true historical account.
Undiscerning Mimic: George goes with his art-critic friend Jane to the private view of an exhibition by a newly fashionable conceptual artist. George has absolutely no talent or training in art appreciation, and his judgement on the quality of the art will not depend on whether it is good, but only on whether he enjoys himself at the show –if he does, this will dispose him to believe the sculptures are good, if not, he will believe they are bad. However, his friend Jane is the acme of critical discernment. Moreover, she will be in a good mood if the art is good, and a bad one if it is bad. (She will not say anything to George about whether or not it is good, however.) George will enjoy himself if Jane is in a good mood, and be miserable if she is grumpy. Thus George is such that, on this occasion, not easily would he assert the art is good unless it is so. But this fact does not issue from George's discernment of the quality of the art he sees. His reliably-true belief is not knowledge, since it is formed in response to Jane's mood, not anything about the art he has seen that evening. (We may suppose that it is exceptional for George to get to accompany Jane to an exhibition, he usually goes alone or with less discerning others. So he is not in general a reliable detector of the quality of art, not even by a deviant mechanism.) (Fricker Reference Fricker2015a)
According to Fricker, Would-be-Liar Simonetta satisfies (RT), but not in virtue of satisfying (KT). She intends to deceive but inadvertently reliably speaks the truth in relation to her topic “due to a freak combination of circumstances” (Fricker Reference Fricker2015a). Undiscerning Mimic George satisfies (RT) by a deviant route since he has no sensitivity to the quality of the art. Like Simonetta, George “reliably speaks truth not through having knowledge of his topic, but due to a coincidental one-off combination of circumstances” (Fricker Reference Fricker2015a). That these are “freak,” “one-off” cases is absolutely crucial to Fricker's later argument that reliable testifiers comprise a “rag-bag” category that fails to pick out a genuine, projectible kind. Herein lies the second problem with Fricker's argument: the cases she chooses to focus on – Would-be-Liar and Undiscerning Mimic – appear to be hand-picked to support her conclusion, but there are other instances of testifiers who satisfy (RT) but not (KT) that have been presented in the literature that pose far more serious problems for her. For instance, a case that I have frequently discussed involves a creationist teacher who reliably conveys to her students that Homo sapiens evolved from Homo erectus though she does not believe this fact herself.Footnote 1 This case involves a phenomenon that I have called selfless assertion.Footnote 2 There are three central components to this phenomenon: first, a subject, for purely non-epistemic reasons, does not believe (and hence does not know) that p; second, despite this lack of belief, the subject is aware that p is very well supported by all of the available evidence; and, third, because of this, the subject asserts that p without believing and, hence, without knowing that p. Selfless assertions can occur by many different kinds of speakers in a variety of situations, but we see them most often in professional contexts – e.g., offered by teachers, medical professionals, and jurors. The point that is of importance here, however, is that the category of selfless assertions does not involve freak, one-off cases like those of Simonetta and George. Instead, it is a category unified by the three features identified above, and ones for which epistemologically interesting generalizations and conclusions can be made, e.g., that selfless assertions falsify the Knowledge Norm of Assertion.
This can be further supported by considering the following description that Fricker offers of the problems with speakers like Simonetta:
Simonetta is a reliable testifier about the events in this village; but she does not know what she asserts, since she does not believe it. Someone, say Paul, who is wise to the whole situation has a basis to know of Simonetta that what she asserts on this topic is reliably true (despite her insincerity); and so can and we suppose does reliably and with justification form knowledgeable beliefs in what she states … Paul does not take Simonetta's word about matters in the village. Paul infers from the facts of what Simonetta asserts to truths about the village, due to an empirical basis he has for taking Simonetta's statements as a reliable sign of what in fact happened; but he does not take her word for what she asserts, does not form beliefs on her authority. (Fricker Reference Fricker2015a)
While this might be true of Paul, it does not at all apply to selfless asserters. Consider the creationist teacher again: her students do not infer from facts about her assertions to truths about evolutionary theory due to an empirical basis they have for taking their teacher's statements as a reliable sign of what in fact happened. Instead, they take their teacher's word for it that Homo sapiens evolved from Homo erectus and form their beliefs on the basis of her epistemic authority. That this is clearly the case provides further reason for thinking that Fricker cannot form interesting conclusions against (RT) and on behalf of (KT) by considering cases like Would-be-Liar and Undiscerning Mimic.
3. KEY MOVES
A third problem with Fricker's view is that several key moves in her arguments appear to beg the question. For instance, she argues that when one takes a speaker's word for it that p, one takes her to be expressing knowledge that p in her assertion, and it is a rational commitment of forming belief on her testimony that this is so. But why is this the case? Fricker argues that when a speaker makes an act in which she presents herself as taking responsibility for the truth of that p, she thereby presents herself as having the epistemic standing required to do this well-groundedly. And, according to Fricker, this is simply for the speaker to know that p.
This, however, simply begs the question against the proponent of the Reliability View. For on such a view, the creationist teacher does take responsibility for the information she is conveying to her students and she certainly presents herself as having the epistemic standing required to do this well-groundedly. Indeed, were her assertions to be challenged by her students, she could present massive amounts of support on their behalf, some of which might even stem from her own research on the topic.Footnote 3 This clearly reveals that she has the epistemic standing to assert that Homo sapiens evolved from Homo erectus well-groundedly. Of course, the teacher doesn't know that this is the case because she doesn't believe it, so for Fricker to assume that presenting oneself as having the epistemic standing required to take responsibility for what one asserts just is to know that p simply begs the question.Footnote 4
4. THE CORE ARGUMENTS
A fourth problem for Fricker is that the core arguments she uses to establish her thesis fail. She argues, first, that one key piece of support on behalf of the Knowledge View over the Reliability one is that finding out that a speaker is not in a position to know that p, or does not know that p, is a defeater for a belief formed on the basis of her testimony. But finding out that a speaker is unreliable in her statement that p also provides a hearer with a defeater for believing that p, so this observation doesn't discriminate between the two views. Moreover, the situation is even worse for the Knowledge View since it is not always the case that finding out that a speaker doesn't know functions as a defeater: suppose that the creationist teacher's students learn that she doesn't believe that Homo sapiens evolved from Homo erectus, but that she is nonetheless a highly reliable source of information about this subject matter. In such a case, awareness of the teacher's lack of knowledge doesn't seem to provide them with a defeater.
Second, Fricker argues that unlike knowers,
speakers who, like Would-Be Liar and Undiscerning Mimic, conform to RT but not in virtue of conforming to KT are not a unitary kind. There is no generally applicable explanation of why they do so, only a variety of serendipitous combinations of facts each of which results in the unlikely phenomenon of someone whose assertions are reliably true though not in virtue of her conscientious conformity to [the norm of assertion] governing tellings. Correlatively, the fact that a testifier on a particular occasion O due to some unlikely combination of circumstances conforms to RT though not KT on O is no reason to predict that on some other occasion she will conform to RT.
These facts mean that categorising speakers in terms of whether or not they satisfy KT is cutting epistemic reality at its joints, while categorising them in terms of RT is not. (Fricker Reference Fricker2015a)
There are problems with this argument from both directions. On the one hand, when the category of reliable testifiers is divided into subgroups of reliable processes of the appropriate level of generality – as the proponent of the Reliability View maintains – then they do constitute genuine kinds. For instance, as mentioned earlier, selfless assertions are a genuine kind about which epistemologically important generalizations, predictions, and conclusions can be made. On the other hand, knowledge as an epistemic category is not as unified as Fricker seems to suggest. for instance, I have elsewhere discussed the phenomenon of isolated secondhand knowledge.Footnote 5 There are two central components to this phenomenon: first, the subject in question knows that p solely on the basis of another speaker's testimony that p – hence the knowledge is secondhand; and, second, the subject knows nothing (or very little) relevant about the matter other than that p – hence the knowledge is isolated. In some cases of this sort, I have argued that it is not epistemically proper for a subject to assert that p or to act on p, despite knowing that p. Given this, it cannot be projected from knowledge to epistemically permissible assertion or action, and thus knowledge does not have the unity as a category that Fricker seems to suggest it has.
5. THE SCOPE
Finally, Fricker might respond to some of my criticisms by arguing that she is interested only in the core mechanism of testimony when it involves firsthand knowledge. But the worry here is that her topic has become so narrow that it is not clear what the interest is in it. In contrast, the proponent of the Reliability View offers a fully general account that purports to explain every instance of learning via testimony.Footnote 6
