1. THE PESSIMIST'S CASE

First, how does testimony work? That is, how does a hearer come to know some proposition on the basis of some speaker's say-so? On some anti-reductionist views of testimony, sometimes referred to as credulism or fundamentalism, hearers have a default justification to believe what speakers say.Footnote ² Christopher Insole refers to this as a hearer's presumptive right (PR) to believe what they're told.

On this view, when Jane tells me that the department meeting is at 3:00 p.m., I immediately gain epistemic justification for my belief that the meeting is at 3:00 p.m., and provided that the proposition is true, I thereby come to know.Footnote ⁴ Whether anti-reductionism is a plausible account of non-aesthetic testimony is a question I will take up later.

We can contrast anti-reductionism with reductionism: speakers do not gain presumptive rights to believe what they're told. Instead, hearers need some additional evidence to their merely being told by someone, such as evidence that the speaker is reliable, or that testimony in general is reliable. On some reductionist views, the hearer's justification for an asserted proposition traces back to the speaker's non-testimonial justification, such as perception. For example, when Jane tells me about the meeting time, my justification for believing that the meeting is at 3:00 p.m. traces back to Jane's having seen a department memo stating the meeting time. My justification may include Jane's having told me, but what ultimately justifies my belief – and gives me knowledge – is the memo, which Jane saw, stating the meeting time.

While it's a widespread view that testimony is a source of knowledge for a wide variety of propositions, a number of philosophers have adopted various versions of pessimism about the possibility of aesthetic testimony.Footnote ⁵ Pessimist views that argue that it's impossible for one to gain knowledge of aesthetic propositions via testimony are what Hopkins (Reference Hopkins2011) labels unavailability pessimism. But for some, the debate is over the legitimacy of forming testimony-based beliefs when those beliefs are about aesthetic propositions. Hopkins (Reference Hopkins2011) refers to this class of pessimist views as unusability pessimism.

Many pessimists take on board some version of the asymmetry thesis (AT).

Some argue, for example, that while aesthetic testimony provides the recipient with some reason, and some justification, to believe what they're told, the level of justification conferred is insufficient to constitute knowledge.Footnote ⁷ Moreover, many pessimists either implicitly or explicitly adopt what has become known as the acquaintance principle (AP).

The principle seems clearly to rule-out the possibility of aesthetic testimony. Unfortunately, very few are explicit in their arguments for it: many seem to take it on board as something to be explained, rather than something to be argued for.Footnote ⁹ Whether AP is plausible is a question I will return to later.

Robert Hopkins (Reference Hopkins and O'Hear2000, Reference Hopkins2011) is one prominent defender of a pessimist view of aesthetic testimony. There are two central features to his view. First, he argues that we should simply reject unavailability pessimist positions.Footnote ¹⁰ So all that's left for a pessimist about aesthetic testimony is to be an unusability pessimist. Second, he argues that while speakers may have knowledge of aesthetic propositions made available to them via testimony, it's not the right sort of evidence on which to form one's beliefs. That is, there's a non-epistemic norm rendering such sources of knowledge “unusable” or “illicit.” He thus calls his view of pessimism about aesthetic testimony, unusability pessimism.

Hopkins's view marks an interesting departure from other pessimist views: he abandons arguments that aesthetic testimony doesn't make knowledge available to hearers – that is, he rejects unavailability pessimism. For example, if the recipient of aesthetic testimony knows that the speaker is highly reliable about, say, the quality of certain local coffeehouses, she can come to know that The Roasterie makes the best espresso in the city.Footnote ¹¹ However, Hopkins argues that while testimony about aesthetic propositions makes knowledge available to hearers, it's illicit for hearers to use a speaker's testimony as the (sole) basis for their beliefs.

Why would aesthetic testimony make knowledge available to recipients, but using such sources remain illicit? Hopkins suggests that there's a non-epistemic norm governing what sources of information we may use in coming to know aesthetic propositions, which is absent in coming to know non-aesthetic propositions. He writes:

In order to make his case, Hopkins provides what he takes to be a non-epistemic norm governing the beliefs of experts. He writes, “Although the notion of an expert is multi-faceted, one element that sometimes comes to prominence is the idea that an expert is someone who ought to settle for herself questions in her domain of expertise.”Footnote ¹³ In a sense, then, Hopkins is suggesting that experts in their domain of expertise have a non-epistemic norm governing belief in a modified form of the acquaintance principle. The further point, then, is that Hopkins argues that a similar non-epistemic norm governs the acceptance – the use – of aesthetic testimony.

This isn't a view unique to Hopkins, or even to the aesthetic testimony literature. In fact, a number of people have recently taken this view in the norms of assertion literature.Footnote ¹⁴ Testimony is merely a special case of assertion (on the part of the speaker), so the two debates have strong parallels. Some have recently argued that some assertions epistemically grounded in a kind of knowledge are inappropriate, and the failing isn't clearly epistemic. I think that we can gain some insight into why Hopkins is wrong about a (non-epistemic) norm governing the acceptance of aesthetic testimony from considering the parallel debate in the norms of assertion. One issue will be whether the view that experts should settle for themselves the questions in their domain is really a non-epistemic norm. Thus, the focus of my arguments will be on rejecting unusability pessimism. In the end, I will argue that we readily acquire aesthetic knowledge through testimony. Since this also amounts to rejecting unavailability pessimism, my view constitutes an optimist view of aesthetic testimony.Footnote ¹⁵

2. EXPERTS, ASSERTION, AND ISOLATED SECOND-HAND KNOWLEDGE

Jennifer Lackey (Reference Lackey, Brown and Cappelen2011) argues that assertions epistemically grounded in what she calls isolated second-hand knowledge (ISHK) are unwarranted, even though the speaker knows what they say. An agent has ISHK when she gains knowledge of a proposition based (almost) entirely on the reliable testimony of another agent. The knowledge is isolated because she lacks independent evidence justifying her belief (or assertion), and the knowledge is second-hand because the epistemic justification she has is derived from testimony rather than first-hand experience.

Lackey's case involves an oncologist coming to know about a patient's diagnosis.Footnote ¹⁶ Matilda, the head oncologist of a major hospital has had a patient referred to her, Derek, complaining of abdominal pain. Since Derek's symptoms are consistent with a number of ailments, Matilda orders a number of tests. The tests happen to come back while Matilda was on holiday, so when she returns to work, she seeks out her fifth year oncology resident, Nancy. Part of what it means to be a fifth year resident is that one is also an expert in making cancer diagnoses – she's just, we suppose, less of an expert than Matilda.Footnote ¹⁷ Nancy, who has seen the test results first-hand, and is very reliable at producing cancer diagnoses, tells Matilda that Derek has pancreatic cancer. Since the diagnosis is accurate, and Nancy has very good evidence for what she says, Matilda gains testimony-based knowledge of Derek's diagnosis. Later that day, when Derek is in Matilda's office, she asserts to him, “I'm sorry, but you have pancreatic cancer.”

Lackey argues that while Matilda knows that Derek has pancreatic cancer, she only has isolated second-hand knowledge, and her assertion is therefore not warranted. And the assertion is unwarranted even though it meets the epistemic requirements of various proposed norms of assertion, including the relatively stringent knowledge norm: assert that p only if you know that p.Footnote ¹⁸ Part of her argument is that “Matilda is an expert – she is an oncologist and Derek's physician, and such roles carry with them certain epistemic duties. In [this case], these responsibilities may include having reviewed the test results first hand, possessing reasons for choosing one condition over another, knowing details about the size and nature of the cancer and so on.”Footnote ¹⁹

It's not that Matilda doesn't know that blocks her from warrantedly asserting that Derek has pancreatic cancer; rather, it's that ISHK isn't the right kind of knowledge on which to make that assertion given her role as an oncologist. Carter and Gordon (Reference Carter and Gordon2011) cash this out in epistemic terms of the difference between knowledge and understanding. That is, while Matilda knows that Derek has pancreatic cancer, she needs to understand that he has pancreatic cancer in order to warrantedly assert, to Derek, that he has pancreatic cancer. There are a number of ways of distinguishing between knowledge and understanding. Kvanvig (Reference Kvanvig2003), for example, distinguishes between knowledge, propositional understanding, and objectual understanding.

When one knows a proposition, one can know that the department meeting starts at 3:00 p.m. without knowing a number of related propositions – such as why the meeting is at 3:00 p.m. Understanding “requires, and knowledge does not, an internal grasping or appreciation of how the various elements in a body of information are related to each other in terms of explanatory, logical, probabilistic, and other kinds of relations that coherentists have thought constitutive of justification.”Footnote ²⁰ One has propositional understanding when one, for example, understands that the meeting is at 3:00 p.m. This may require that one know that department meetings tend to be held at 3:00 p.m., and that one knows the evidence for the meeting being held at 3:00 p.m. (such as having seen a memo stating the time). One has objectual understanding when one, for example, has an understanding of a subject such as cancer, which we readily ascribe to oncologists such as Matilda.

Carter and Gordon (Reference Carter and Gordon2011) argue for a view whereby one must have the right sort of understanding in order to be properly epistemically positioned to warrantedly assert.

What this view seems to imply is that in order for an agent to warrantedly assert, Matilda must have direct (cognitive) access to the evidence, at least dispositionally, that epistemically supports the assertion.Footnote ²¹

What one takes to be the relevant norm of assertion isn't my interest, here.Footnote ²² My point is to show that the same intuition under-riding Hopkins's pessimism about aesthetic testimony is also present in discussions of the norms of assertion. The idea is that experts have a duty to be personally acquainted with the evidence for their knowledge in their domain of expertise. However, this duty is not clearly a non-epistemic duty. If that's cashed out in terms of understanding, as Carter and Gordon do, this is an epistemic norm of expert belief. But I don't think much hangs on whether Hopkins's usability norm of aesthetic testimony is epistemic or non-epistemic. What is more important is whether Hopkins, Lackey, and Carter and Gordon are right that experts have this sort of duty. And if it's false that experts have this duty, this can shed some light on how one can be an optimist about aesthetic testimony. Simply put, my view is that recipients of aesthetic testimony, just like experts, do not need to be personally acquainted with the evidence for their beliefs in order to know.

3. EXPERTS AND TRANSMITTING OR RECEIVING KNOWLEDGE

Experts do not need to have first-hand experience with the evidence that justifies their belief – producing knowledge – in propositions within their domain of expertise. In understanding why, we first need to define what we mean by “expert.” There are a number of philosophical treatments of what makes someone an expert.Footnote ²³ Goldman (Reference Goldman2001: 91) argues “that cognitive expertise be defined in ‘veritistic’ (truth-linked) terms. As a first pass, experts in a given domain (the E-domain) have more beliefs (or high degrees of belief) in true propositions and/or fewer beliefs in false propositions within that domain than most people do (or better: than the vast majority of people do).” This is a reliabilist view of expertise: someone is an expert if they have a good track record of true beliefs in a given domain.Footnote ²⁴ Driver (Reference Driver2006, Reference Driver2013) argues that experts not only have the relevant knowledge of a subject area, but they have a kind of meta-knowledge. They're not just skilled at forming true beliefs in their domain of expertise, they're also skilled at, for example, knowing which knowledge to access and to apply to various questions or problems. I think that this roughly gets things right, but we can make things even more precise.

Ernest Sosa (Reference Sosa2011) offers a useful analysis of what he calls “knowing full well.” On Sosa's virtue epistemic view, knowledge is apt belief: it's accurate (true), adroit (skillfully believed), and accurate because adroit. His favorite analogy is an archer taking a shot at a target. Knowledge is like an archer skillfully shooting (adroit) at her target and hitting her target (accurate), where her hitting the target is appropriately because of her shooting skill. That is, it's not due to luck that she hits her target.Footnote ²⁵ In Turri's (Reference Turri2011b) terms, the archer's success manifests her skill.Footnote ²⁶ So apt beliefs constitute knowledge. However, beliefs may also be meta-apt (and they may be meta-apt without being apt, and may be apt without being meta-apt).

A belief is meta-apt when the agent's exercising the relevant skill is itself the result of exercising a skill. Returning to the archer, a skilled huntress must be skilled not only at shooting and hitting her target, but also at deciding when to shoot. One huntress may be foolish in choosing to shoot, but skillfully hit her target nonetheless. Her shot is apt but not meta-apt. Contrast this with another huntress who wisely chooses to shoot at her target, and hits her target by exercising her shooting skill. Her shot is both apt and meta-apt.

Finally, a performance is fully apt when the aptness of the performance manifests the performance's meta-aptness. When an agent's belief is apt, meta-apt, and apt because meta-apt, she knows “full well.” Experts, I suggest, are agents whose epistemic actions – believing, testifying, and so on – tend to be cases of knowing full well. Experts aren't perfect, though: they make mistakes. However, mistakes tend to be less common with experts than non-experts. And, following Driver (Reference Driver2013), expertise is contrastive: an expert is an expert compared to a contrast class of other people. Experts are better at coming to know, on average, than other people in a given domain.

With this in place, what basis is there for the intuition that experts should have first-hand experience with the evidence for their beliefs? Hopkins (Reference Hopkins2011: 147) writes, “The reason experts are required to settle for themselves matters in their field is that that way we minimize the risk of error across the interested community as a whole.” One worry is that if, as an expert, I merely take a colleague's say-so, then if my colleague makes an error, my accepting her testimony would spread the error. This could then result in further spreading of the error as the misinformation gains support as more experts come to accept it. The worry here, then, seems to be one of shared false testimony of groups. Hopkins seems to suggest that if each expert were to consider the evidence for themself, the likelihood of all experts settling on the wrong answer would be minimized.

There's something to this worry. If, in considering the evidence, each expert is 99% likely to arrive at the right answer, then the probability of any given expert arriving at the wrong answer is 1%. If multiple experts’ judgments are arrived at independently, then the probability that, say, ten experts arriving at the wrong answer is extremely remote.Footnote ²⁷ However, if our judgments are not independent, because one expert formed her own judgment about p, told the rest of us, and we all came to believe that p on the basis of her testimony, then the probability that we all arrive at the wrong answer isn't so remote. So Hopkins's worry is a real one.

However, this isn't a problem unique to testimony. This is merely a problem for fallible knowledge in general.Footnote ²⁸ Fallibilism, roughly speaking, is the view that when an agent knows that p, their evidence is consistent with ~p. However, since they know, p must be true, because knowledge is factive. So while the agent knows, the risk that they're mistaken is non-zero. Evidence can be misleading, too. Williamson (Reference Williamson2000) argues that evidence is factive: there's no such thing as false evidence. However, evidence can be taken to support false propositions. In these cases, the evidence, while itself true (and on Williamson's view, known) can lead agents to form false beliefs. And this isn't a problem unique to testimony, let alone aesthetic testimony. This is a problem common to all forms of knowing.Footnote ²⁹

When knowledge is transmitted, further risks of being wrong are introduced, particularly when knowledge is transmitted verbally through testimony. Many readers might be familiar with a children's game sometimes called “Telephone.” In the game, the starting person decides on a phrase and secretly tells the person on their left. That person then relays what they heard, verbatim, to the person on their left. The final person hears what they were told, and tells the group. If you've ever played this, you know that the message is almost invariably radically changed from the beginning to the end of the game. “The walls are white,” may be transformed into “Jenny is cute.”Footnote ³⁰ So I think we can reject the possibility of error worry about experts using knowledge made available to them via testimony.

Recall that Hopkins (Reference Hopkins2011) grants that speakers can come to know non-aesthetic propositions via testimony. However, his view is that it's illicit for experts to use non-aesthetic knowledge that's made available to them through the (reliable) testimony of others, even other experts, just as it's illicit for anyone to use knowledge of aesthetic propositions that they come to know through testimony.Footnote ³¹ One way of thinking about this is that experts should make up their own minds because experts need to be autonomous in their expert judgments. So when they rely on the say-so of others, they're not doing any of the epistemic work themselves. Lackey (Reference Lackey2007a) seems to take a view close to this: recipients of testimony don't deserve “credit” for their knowledge because they didn't do anything in coming to know via testimony. Driver (Reference Driver2006) discusses a similar worry in the context of testimony of moral propositions (e.g., <Jane murdered Sally>).

I think that this is a mistake. Consider the following case from Driver (Reference Driver2006: 623).

Driver notes that “the amorphous feeling at this point is that what is odd in the [Kathy] case is that there is nothing of Kathy in the judgments that x is beautiful, which in turn led up to the purchase and display of those items in her house.”Footnote ³² However, even accepting the case as described, is it really true that Kathy plays no active role in coming to believe that the Braque print is beautiful? Some recent work in the epistemology of testimony should have us doubt that this is the case. Moreover, what I have to say later about Lackey's Matilda/Nancy case also applies to Driver's Kathy case: Kathy can indeed come to know that the Braque print is beautiful merely on the say-so of her interior decorator.

Take a standard case from the epistemology of testimony literature.Footnote ³³ We don't know each other, but I need to know the time. I don't have a watch or access to something that can tell me the time, other than hopefully you. Audi (Reference Audi, Lackey and Sosa2006: 26) writes: “I ask you the time; you tell me it is nine o'clock; and straightaway I believe this on the basis of your saying it.” Intuitively, I can justifiably believe that the time is nine o'clock merely on the basis of your telling me. This sort of case is often raised as an objection to reductionism, called the Scarcity of Information Objection: the claim that “[o]rdinary epistemic agents simply do not have information to acquire positive reasons strong enough to justify accepting most of the testimony that is intuitively justified or warranted” (Lackey 1990: 180; quoted in Kenyon Reference Kenyon2013: 59).

However, Kenyon (Reference Kenyon2013: 66) convincingly argues that this case “is oversimplified in every particular. There is excellent reason to think that I don't just ask, you don't just tell me, I don't just believe you, and if this all really did happen ‘straightaway’, there should at least be serious questions about whether I would be justified in my acceptance.” On this case, Kenyon asks for more details about exactly what happened: when I ask you for the time, are you even wearing a watch? Do you answer after looking at your watch, or do you immediately blurt out “9:17!” without looking at your watch or without breaking eye contact with me? How are you dressed? Does your answer accord with my background information – if I know it's between eight and ten in morning, do you answer with a time in between that range? The fact is, who we ask is also part of the picture: we're likely to ask someone who we think is likely to be able to answer us, and, moreover, that if they answer, they're likely to tell us the truth. For example, we're more likely to ask sufficiently well-dressed people rather than someone who looks disheveled. So the truth is, there's a great deal of information available, if only implicitly, in addition to the speaker's assertion for us to justifiably believe what's said.

As Kenyon writes, “lots of things that we thought implicitly or explicitly in the past quite reasonably ought to count as evidentially informing our acceptance or rejection of testimony in the present” (p. 60). Moreover, “testimonial contexts are very often epistemically and socio-linguistically engineered in order to generate testimony that can be rationally accepted quickly and unreflectively” (p. 60). Furthermore, “engineered or not, even our immediate surroundings in a testimonial context express or encode a lot of information” such that “many of the literature's purported examples of justified [beliefs from testimony] that are informationally impoverished turn out to be informationally very rich” (p. 60).

So it's simply false to say that there's nothing of Kathy, or recipients of testimony more generally, in her coming to believe what she's told about the beauty of the art print. There's plenty of information in addition to her decorator's say-so that goes into Kathy's belief that the print is beautiful. While she trusts that what her decorator tells her is beautiful is beautiful, that it's her decorator telling her figures into the formation of her belief. She might not have been so inclined to believe the same proposition if it were told to her by her carpenter. Recipients of testimony don't simply believe whatever they're told, by whomever they're told. That would amount to an extreme version of gullibility. As Driver puts it:

She continues:

I think Driver is correct on all counts. It's just not the case that even non-aesthetic (“mundane”) testimony behaves according to an anti-reductionist model. Recipients of testimony use a great deal of available information, including information about who the speaker is, even if only subconsciously, in deciding whether to trust the speaker's say-so. So we can dismiss the autonomy worry about experts coming to know via testimony. But then the question remains: may experts permissibly come to know, and use that knowledge of, something in their domain of expertise merely on the basis of testimony?

Returning to Lackey's discussion of the oncologist Matilda, recall that Lackey grants that Matilda knows that Derek has pancreatic cancer. Her argument is that this isn't the right kind of epistemic support to warrantedly assert that he has cancer. Lackey seems thus to reject an unavailability pessimist position about expert testimony. Like Hopkins, the argument is that while Nancy's testimony makes this knowledge available to Matilda, Matilda isn't permitted to use it to come to know that Derek has cancer (and perhaps to go about prescribing treatment, or asserting the diagnosis to Derek). My argument is that this is incorrect.

Suppose we push on Hopkins's (and Lackey's) view: a consequence of the view is that even if Matilda knows that Nancy knows that Derek has pancreatic cancer, she can't use that information in coming to know that Derek has pancreatic cancer. One way to make this work is to adopt something like DeRose's (Reference DeRose2002) contextualism about knowledge, where Matilda is in one epistemic context, but Nancy is in another. However, Hopkins (Reference Hopkins2011) explicitly disavows this move. So there's a problem for Hopkins: if Matilda knows that Nancy knows that p, then Matilda knows that p. This follows from factivity of the knowledge operator. So assuming that we're rejecting contextualist accounts of knowledge, if S knows that R knows that p, then S knows that p.Footnote ³⁶ And it seems absurd to say that it's inappropriate for Matilda to form her knowledge, even though she's already formed the knowledge (from her knowing that Nancy knows).

But let's dig deeper into the unusability claim. In thinking about whether experts can appropriately come to know merely through receiving testimony in their domain of expertise, we need to recognize that what Lackey says about the medical institution is just false: experts regularly rely solely on the say-so of other people (including those less expert than themselves). In fact, medical establishments are purposefully engineered for this to be the case.Footnote ³⁷ Who Matilda asks matters. She asks Nancy, a trained fifth year resident.Footnote ³⁸ Part of what this means is that Nancy is herself an expert, though perhaps less of an expert than Matilda. Lackey specifies that Nancy is reliable at forming cancer diagnoses, and that Matilda knows this. The truth is, it's common for physicians to rely on the say-so of others in making important decisions: physicians simply don't have the time to see all of the evidence for themselves in treating patients.Footnote ³⁹

Of course, Lackey argues that Matilda should be personally acquainted with Derek's test results. But we should ask the key question: what are the test results themselves if not the testimony of, for example, the radiologist who performed the CT scan of Derek, or the technician who performed Derek's blood tests? In generating test results, someone performs a test, then records them in the final report. And it's this final report that Nancy sees, which is the basis for her forming and communicating Derek's diagnosis to Matilda. Nancy knows, for example, that (we may suppose) Derek's T-cell count is x merely on the basis of the report's say-so. That report was generated by a person (let's suppose it was one person). So Nancy's coming to know that Derek's T-cell count is x is itself ISHK. After all, neither Nancy nor Matilda performed the tests themselves. So even if Matilda were “personally acquainted” with the test results, her assertion to Derek would be epistemically grounded in ISHK yet again: this time it would be the test results rather than Nancy's say-so, though. So why is it sufficient for Matilda to rely on that testimony, but not her well-trained oncology resident Nancy? And why is Nancy able to warrantedly assert on the ISHK from the test results, but Matilda isn't able to warrantedly assert based on her ISHK from Nancy?

The truth is, I think that there's nothing clearly different between Nancy's reliably looking at the test results and coming to know Derek's diagnosis on the basis of the test results (which is itself an instance of ISHK), and Matilda's asserting to Derek on the basis of Nancy's testimony. Moreover, it's a good thing that things work this way: physicians’ days are just too busy to be personally acquainted with the evidence supporting every diagnosis and decision they need to make. And if this is correct, then there's nothing illicit in an expert's coming to know some proposition in their domain of expertise from the say-so of another person, even someone less expert than themself. And if that's the case, then there's no non-epistemic norm governing the use of knowledge through testimony, as supposed by Hopkins. Consequently, it's simply not at all clear that recipients’ use of aesthetic testimony is illicit. Quite the contrary: provided that the speaker is a reliable source of aesthetic testimony, and the recipient is being a responsible recipient of testimony (as outlined above), recipients can come to know aesthetic propositions through testimony.

Hopkins's unusability pessimism depends on his claim that there's a non-epistemic norm governing the acceptability of a hearer coming to know an aesthetic proposition based merely on the testimony of a speaker. He suggests that this is plausible because we find a similar norm governing the acceptability of experts coming to know non-aesthetic propositions based merely on the testimony of a speaker. I've shown that a similar argument is made by Lackey in her discussion of ISHK. However, my argument is that Lackey and Hopkins are wrong about the acceptability of experts forming knowledge – even in their domain of expertise – based merely on the testimony of a relevant speaker (as not just any speaker will do). This gives us purchase in rejecting Hopkins's argument that aesthetic testimony has a parallel non-epistemic norm governing the acceptability of using testimony as a source of aesthetic knowledge. In what follows, I present my positive proposal for when it's acceptable for experts to use testimony in forming knowledge, and by extension, for when it's acceptable for hearers to form aesthetic knowledge via testimony.

4. THE EXPRESS JUDGMENT ACCOUNT OF AESTHETIC TESTIMONY

A number of things remain to be said about how, and under what conditions, recipients of aesthetic testimony come to know. In the remainder of this paper, I put forward a modest proposal, what I call the Express Judgment Account of Aesthetic Testimony.

Much of the foregoing discussion hinges on what we mean by an expert forming a judgment, whether it's appropriate for an expert to form a judgment merely on the basis of another's testimony, and what the connection is between judgment and testimony-based knowledge. In what follows, I'll argue that one can think that experts ought not to form judgments about propositions in their domain of expertise based only on second-hand evidence (i.e., testimony), while still holding the view that experts (and laypersons) can come to know propositions in their domain of expertise based only on second-hand evidence.

To make this argument, we need an important distinction between belief and acceptance. I'll argue that judgment is best understood in terms of acceptance, and that testimony-based knowledge only requires belief, not acceptance (or judgment). Perhaps some will not accept the belief/acceptance distinction. I present the ‘Express Judgment Account of Aesthetic Testimony’ as a proposal for ‘How to be an optimist about aesthetic testimony.’ I don't offer it as a conclusive argument that pessimism about aesthetic testimony is false.

Cohen (Reference Cohen1989) offers what has become the canonical understanding of the distinction between belief and acceptance.Footnote ⁴⁰ On Cohen's view, having the belief that p is the disposition “to feel that it is true that p,” whereas accepting that p is where one chooses “to have or adopt a policy of deeming, positing or postulating that p.”Footnote ⁴¹ Acceptance and belief, then, come apart. One can accept that p without believing that p, and one can believe that p without accepting that p. A distraught mother, for example, might believe that her son is a criminal, without coming to accept it. Or someone persuaded by Pascal's wager may accept that God exists, and act accordingly, without coming to believe that God exists. Acceptance is more voluntary than belief. Cohen writes, acceptance “occurs at will because at bottom it executes a choice – the accepter's choice of which propositions to take as [their] premises.”Footnote ⁴² A person can't always decide what they believe, but they can decide what they accept. The important point here is that belief is relatively passive, whereas acceptance is relatively active. Since judgment connotes an activity, and a particular willingness to act on what one judges to be true, I suggest that judgment is best understood in terms of acceptance rather than belief.

How does this connect to testimony? Testimony is merely a special case of the speech act of assertion. Very roughly speaking, we assert when we claim or represent something as true.Footnote ⁴³ And when a hearer gains the knowledge that p from a speaker's assertion that p, we have a case of testimonial knowledge. Starting with at least Frege (Reference Frege, Geach and Black1892), philosophers have connected assertion and belief to judgment.Footnote ⁴⁴ Some have argued that assertion is merely the external expression of a judgment.Footnote ⁴⁵ However, if that were the case, then lies wouldn't count as assertions, and I take that to be a problem with any view making belief or judgment a necessary condition for asserting. This means that assertion can't be defined as the external expression of belief, either. Whether belief or judgment is necessary for warrantedly asserting is, of course, a separate question. A few recent treatments of the norms of assertion have argued that belief is not necessary for warrantedly asserting.Footnote ⁴⁶

The issue here, though, is whether a speaker needs to believe – and even know – some proposition in order to transmit knowledge via testimony to a recipient. It's intuitive that we should answer in the affirmative: how can I come to know that the meeting is at 3:00 p.m. unless Jane, who told me that it's at 3:00 p.m., herself knows? Treating speaker knowledge as a necessary condition for imparting knowledge through testimony yields the following form of the transmission principle:

The idea is that testimony is a kind of transmission of epistemic properties, and in order for a hearer to gain knowledge (an epistemic property) from a speaker, the speaker must also have that property.

Supporting this view, Williamson (Reference Williamson2000, p. 267) writes, “In normal circumstances, when the hearer knows that the speaker asserted p, the speaker has no reputation for unreliability, and so on, a speaker who asserts p thereby puts a hearer in a position to know p if (and only if) the speaker knows p.” It seems to me that many in the aesthetic testimony literature are implicitly committed to this transmission principle. However, the principle is mistaken: recipients of testimony can come to know a proposition even if the speaker themself doesn't believe or know.

Lackey (Reference Lackey1999, Reference Lackey2007b, Reference Lackey2008) convincingly argues against Transmission_N. Her primary case against it involves what she calls selfless assertions. For example, Stella is a grade school teacher and a creationist. She's been tasked with teaching evolution to her students. Even though she doesn't believe it, she asserts to her class, “Homo sapiens evolved from Homo erectus.” She recognizes that the scientific evidence supporting the proposition is overwhelming, but she doesn't believe the proposition due to a commitment to her faith and the claims of creationism. The proposition is true (let's assume), and she asserts it on the basis of her grasp of the conclusive (let's assume) scientific evidence supporting it. But since she doesn't believe it, she can't know it. However, the students can come to know it on the basis of her testimony. They believe it, have evidence for its truth (it's being asserted by a reliable teacher who is asserting based on the scientific evidence), and it's true: all the conditions for testimony-based knowledge are present. Therefore, Transmission_N is false.

What's important here is that speakers, in offering testimonial knowledge to recipients, don't need to know or even believe what they say. Furthermore, their assertions need not express the speaker's own judgment of the proposition expressed. Moreover, recipients of testimony, even in cases where the speakers don't themselves believe or know what they say, can come to know on the basis of the speakers’ testimony. On my view, all testimony – aesthetic, moral, and “mundane” – behaves according to a reductionist model of testimony.

Here is my Express Judgment Account of Aesthetic Testimony. In order for a speaker to form testimony-based knowledge about aesthetic propositions, the following conditions must be present: the proposition expressed is true, the recipient has good epistemic reasons to believe that it's true (i.e., justification) – and this justification may involve, for example, that the recipient knows that the speaker is an expert – the recipient comes to believe the expressed proposition, and, following Turri (Reference Turri2011a), the proposition asserted expresses a judgment that the proposition is true.

Turri (Reference Turri2011a) proposes as the norm of warranted assertion that one's assertion (that p) express the knowledge that p. However, the speaker themself need not know that p. In special cases where the speaker is aware that p is known by reputable people, but the speaker themself may not believe p for (often) non-epistemic reasons, it's still possible for the speaker to express the knowledge that p via their assertion. Turri argues that cases like Stella are not a counterexample to knowledge being the norm for warrantedly asserting, since Stella's assertion expresses the knowledge (of the scientific community, in this case) that Homo sapiens evolved from Homo erectus. I'm not going that far, though: I don't think that knowledge is required for properly asserting.Footnote ⁴⁸ What's at issue here is whether it's appropriate to believe, accept (i.e., judge), and use knowledge gained through testimony.

Returning to the Express Judgment Account, since Transmission_N is false, it need not be the speaker's own judgment (that p) expressed in their assertion, just as Stella's students can come to know that Homo sapiens evolved from Homo erectus even though Stella hasn't formed the judgment that that's true. What matters is that her assertion expresses a judgment – of an appropriate source – that the proposition is true. In Stella's case, her assertion expresses the judgment of many in the scientific community that the theory of evolution is true.

Some of the unusability pessimist's resistance to this picture seems to be that they think it inappropriate for a recipient of aesthetic testimony to form a judgment that, for example, the Braque print is beautiful, merely on the basis of a speaker's say-so. However, I think that this conflates the important distinction between belief and acceptance, discussed earlier. Acceptance is active and voluntary: it's a decision to treat p as true. And I argued that we should treat judgment in terms of acceptance, rather than belief. Now, one can grant that recipients of aesthetic testimony ought not to judge or accept that p is true merely on a speaker's say-so – in fact, I'm neutral on this issue – but it doesn't follow that recipients ought not to believe that p is true merely on a speaker's say-so. Indeed, I've argued that there's nothing illicit in a hearer's coming to believe a proposition merely on the basis of testimony. In fact, it's entirely appropriate to believe that p, even when p is an aesthetic proposition, provided that the requirements for testimony-based belief are satisfied. I am thus advancing an optimist position with respect to aesthetic testimony. Respecting the belief/acceptance distinction allows us to explain the unusability pessimist's intuition, while allowing us to remain optimists about coming to know aesthetic testimony.Footnote ⁴⁹ However, I am neutral on whether it's inappropriate to judge/accept that p based only on a speaker's say-so. I'm simply offering a way to explain the unusability pessimist's intuition while also arguing that hearers can come to know aesthetic propositions based only on testimony (of reliable sources). That is, we can explain the unusability pessimist's intuition while also holding an optimist position on aesthetic testimony.

Recall what Hopkins says about the competing intuitions between optimists and pessimists: “What optimists are picking up on is that there is nothing epistemologically wrong with aesthetic testimony. What the intuitions driving pessimism reflect is that there is something wrong with it, nonetheless.”Footnote ⁵⁰ My argument has been that the pessimist is mistaken in their intuition that there's anything wrong with testimony-based knowledge of aesthetic propositions. Instead, the pessimist's intuition seems to be tracking the potential inappropriateness of speakers accepting (or judging) that a proposition is true merely on the basis of someone's testimony. There's nothing wrong with speakers coming to believe (and thus to know) merely on the basis of someone's testimony.

Finally, how does this apply to Hopkins's own cases? His primary case is whether to believe that a particular movie is good based only on the say-so of a speaker. Suppose that you've never seen Gone Girl. And you're faced with the choice between choosing a movie, neither of which you know much about: Gone Girl and Birdman. You ask your friend, whom you know is an authority on movie quality, and you trust their judgment. You also know that you have similar tastes in movies, so that a recommendation from your friend is a good bet. In such a context, can you come to know that Gone Girl is a good movie (and perhaps that it's better than Birdman) merely on your friend's say-so?

Hopkins (Reference Hopkins2011) isn't entirely clear on the answer here. He suggests that the unusability norm lapses, and it's thus rational for one to see Gone Girl rather than Birdman solely on the basis of your friend's recommendation. However, it's not clear whether Hopkins thinks that the unusability norm lapses such that you're now warranted in using your friend's recommendation in coming to know that Gone Girl is good. But let's assume that Hopkins claims that it's appropriate to use the aesthetic testimony in coming to know that Gone Girl is good in this special context, when it's impossible for you to decide which movie to see (without already having seen them to make up your mind). Hopkins also asks us to suppose that after seeing the film, you're left unsure whether to think that it's a good movie (or not). He writes, “[Your] own experience fails to settle the matter. Can [you] then continue to appeal to [your] friend's recommendation, in forming [your] own view?”Footnote ⁵¹ That is, is it appropriate for you to use your friend's testimony as a reason to think that it's a good movie? Hopkins answers in the negative: it's not appropriate. In fact, he argues that you should discount your friend's testimony altogether – even someone you recognize as an authority on aesthetic judgments about movies. This strikes me as mistaken. If you truly regard your friend as an authority on aesthetic judgments of movies, their testimony should still carry weight.

Consider, by analogy, what an expert oncologist, call her Jane, should do if she views a case and isn't sure what diagnosis to make on the basis of a patient's test results. Suppose that her colleague, Mary – whom Jane regards as at least as expert as herself, and perhaps even more expert – has looked at the same evidence and confidently concluded that the patient has pancreatic cancer. Is it really correct that Jane should discount Mary's testimony altogether? It seems clear that Jane ought to regard Mary's testimony with at least some weight. Additionally, if Jane regards Mary as more of an expert than herself (as we suppose is true in the movie case), it seems obvious that Jane should treat Mary's testimony as a reason to believe that her patient has cancer.Footnote ⁵² This is, after all, exactly what we do when we're unsure about a judgment on a matter, even in our domain of expertise: we ask other experts and use their testimony as a basis in forming a belief of our own. And while one may continue to have the intuition that it's inappropriate for Jane to accept that her patient has pancreatic cancer (although I suspect that this would also be permissible), it seems clear that it's appropriate for her to believe that her patient has pancreatic cancer. Thus, even the case to which Hopkins himself appeals seems better explained by the Express Judgment Account than his own unusability pessimist view.

In summary, then, my view is that norms for experts are such that it's appropriate to come to believe – and thus know – some proposition in their domain of expertise, but the norms may be such that it's inappropriate to accept the proposition. However, I am neutral on the latter question: my suspicion is that it's also appropriate for experts to accept that p merely on the say-so of a relevant speaker (e.g., one that the hearer regards as sufficient expert). In any case, the belief/acceptance distinction allows us to make sense of the unusability pessimist's intuition that it's inappropriate for an expert to come to believe based only on the testimony of a speaker, even when that speaker is another expert. The intuition seems well captured by a norm governing the acceptance of testimony rather than the formation of testimony-based belief. Applying this to aesthetic testimony, while it may be inappropriate to accept that a painting is beautiful, or a movie is good, or The Roasterie makes the best espresso in the city, it's appropriate to believe these propositions on the basis of speaker testimony, provided that the speaker is a requisite epistemic authority on the relevant aesthetic domains. Therefore, hearers can come to know aesthetic propositions via testimony – and I suspect that we do this quite readily. Therefore, I've articulated an optimist position about aesthetic testimony.Footnote ⁵³