Assertion and Isolated Second‐Hand Knowledge *
Abstract and Keywords
A common view in the recent philosophical literature is that knowledge is sufficient for proper assertion. More precisely, it is frequently said that one is properly epistemically positioned to assert that p if one knows that p. This chapter argues that this thesis is false. In particular, it is shown that there are various kinds of cases in which a speaker asserts that p, clearly knows that p, and yet does not have the proper epistemic authority or credentials to make such an assertion, thereby showing that knowledge is not always sufficient for epistemically proper assertion. A diagnosis is then offered of what is salient in the cases challenging this sufficiency claim and a broad feature is highlighted that needs to be accounted for in any view of the norm governing proper assertion.
A common view in the recent philosophical literature is that knowledge is the norm governing proper assertion. Thus, according to Keith DeRose (2002: 180), “one is positioned well‐enough to assert that P iff one knows that P.” Let us call the thesis expressed here the Knowledge Norm of Assertion, or the KNA, and formulate it as follows:
KNA One is properly positioned to assert that p if and only if one knows that p.
As stated, there are two dimensions to the KNA; one is a necessity claim and the other is a sufficiency claim. More precisely:
KNA‐N One is properly positioned to assert that p only if one knows that p.
KNA‐S One is properly positioned to assert that p if one knows that p.
Much attention has been devoted to the KNA‐N, both in terms of arguments presented on its behalf and in terms of objections offered to challenge it.1 I shall not here contribute to this debate. Instead, I shall restrict my focus to the KNA‐S, which, by comparison, has been the explicit topic of relatively few extended discussions.2
Now, while clear endorsements of the KNA‐S tend to be more implicit or undeveloped than those of the KNA‐N, they are nonetheless quite prevalent. For instance, immediately prior to arguing that knowledge is the norm of assertion, Steven (p.252) Reynolds (2002: 140) poses the question “what epistemic relation to p is good enough to make it permissible to assert that p?” Similarly, shortly after defending the KNA‐N, John Hawthorne (2004: 23 n. 58) adds that it may also be “arguable that knowledge suffices” for the “epistemic correctness” of assertion. And, according to John Hawthorne and Jason Stanley (2008: 578), “Where one's choice is p‐dependent, it is appropriate to treat the proposition that p as a reason for acting iff you know that p.” Thus, if assertion is a species of action, knowing that p is both necessary and sufficient for properly asserting that p. In all these passages, then, we find different characterizations of the KNA‐S.3
Surely, however, even if I unquestionably know that my colleague made a fool of himself while we were all drinking the other night, it may still be improper for me to assert that this is the case on Monday morning. It may, for instance, be imprudent, because it would strain our friendship; or it may be impolite, because he would find it utterly embarrassing; or it may simply be pointless, because everybody in the department already knows that this is the case. So in what sense is knowing that p sufficient for one to be properly positioned to assert that p?
The answer to this question can be found in the quoted passages above. Reynolds, for instance, asks what epistemic relation to p is good enough for the permissibility of assertion, and Hawthorne talks about knowledge being sufficient for the epistemic correctness of assertion. Given this, let us clarify the sufficiency claim as follows:
KNA‐S* One is properly epistemically positioned to assert that p if one knows that p.
According to the KNA‐S*, then, knowledge is sufficient for possessing the epistemic authority for assertion, even if it is insufficient for various other kinds of propriety. For instance, while it may be imprudent, impolite, or pointless for me to assert that my colleague behaved foolishly over the weekend, my knowing that this is the case suffices for my having the epistemic credentials to make such an assertion.
The KNA‐S* has a great deal of intuitive appeal. If I assert that the university is closed because of an impending snowstorm, my knowing that this is the case seems sufficient to render such an assertion permissible. If my assertion is questioned, appealing to my knowledge adequately meets the challenge, while offering anything less—such as my suspecting that the university is closed, or being moderately justified in believing that it is—does not. Moreover, the KNA‐S* has significant theoretical power. According to DeRose (2002: 147):
The knowledge account of assertion provides a powerful argument for contextualism: If the standards for when one is in a position to assert warrantedly that P are the same as those that constitute a truth condition for “I know that P,” then if the former vary with context, so do the (p.253) latter. In short: The knowledge account of assertion together with the context sensitivity of assertability yields contextualism about knowledge.
This link between knowledge as the norm of assertion and contextualism requires both the necessity claim and the sufficiency claim of the KNA, and so the truth of the KNA‐S* provides critical support to a central argument on behalf of contextualism.
Despite the intuitive plausibility and theoretical power of the view that knowledge suffices for epistemically permissible assertion, however, I shall argue in what follows that the KNA‐S* is false. In particular, I shall show that there are various kinds of cases in which a speaker asserts that p, clearly knows that p, and yet does not have the proper epistemic authority or credentials to make such an assertion, thereby showing that knowledge is not always sufficient for epistemically proper assertion. I shall then offer a diagnosis of what is salient in the cases challenging the KNA‐S*, and suggest a broad feature that needs to be accounted for in any view of the norm governing proper assertion.
1 Isolated Secondhand Knowledge: Expert Testimony
To begin, consider the following three cases, all of which involve a phenomenon that I shall call isolated secondhand knowledge:
DOCTOR. Matilda is an oncologist at a teaching hospital who has been diagnosing and treating various kinds of cancers for the past fifteen years. One of her patients, Derek, was recently referred to her office because he has been experiencing intense abdominal pain for a couple of weeks. Matilda requested an ultrasound and MRI, but the results of the tests arrived on her day off; consequently, all the relevant data were reviewed by Nancy, a competent medical student in oncology training at her hospital. Being able to confer for only a very brief period of time prior to Derek's appointment today, Nancy communicated to Matilda simply that her diagnosis is pancreatic cancer, without offering any of the details of the test results or the reasons underlying her conclusion. Shortly thereafter, Matilda had her appointment with Derek, where she truly asserts to him purely on the basis of Nancy's reliable testimony, “I am very sorry to tell you this, but you have pancreatic cancer.”
EXPERT PANELIST. In the wake of the Space Shuttle Challenger disintegration, the United States House Committee on Science and Technology conducted a hearing in an effort to determine the cause of the disaster. One of the experts called to testify at the hearing was John Smith, a manager at NASA. Though it was part of Smith's responsibilities to monitor the details of the shuttle operation, both before and after the accident, he has been preoccupied with personal problems and has thus been negligent in carrying out his official duties. On the morning of the hearing, Smith met very briefly with one of his co‐workers, who told him only that the cause of the shuttle's disintegration was the failure of an O‐ring seal at liftoff. Despite the fact that Smith is not privy to any of the data or reasoning underlying this explanation, and (p.254) has only his co‐worker's reliable testimony to ground his belief, he truly asserts at the House Committee hearing, “The Space Shuttle Challenger disintegrated because of the failure of an O‐ring seal at liftoff.”4
PROFESSOR. Judith is a professor at one of the best law schools in the country, and today's lecture is on US copyright law. While she is generally quite knowledgeable of this topic, she has failed to keep up with some recent developments in this area. Over lunch yesterday, one of her colleagues briefly expressed his belief that it is extremely improbable that the Supreme Court will consider a case challenging the addition of twenty years to the original copyright protection of fifty years after the death of authors. Though Judith does not know any of the reasons or considerations underlying this claim, she asserts to her students in class, “The Supreme Court is unlikely to hear the upcoming challenge to the recent extension of US copyright protections to seventy years after the author's death.” While this assertion is in fact true, it is based purely on the basis of the reliable testimony of Judith's colleague.
Though there are some interesting differences among these cases, they are united in all involving what I earlier called isolated second‐hand knowledge. 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 second hand; and, second, the subject knows nothing (or very little) relevant about the matter other than that p—hence the knowledge is isolated. The combination of these features, by itself, is not necessarily problematic, even when assertion is involved. But when a subject's assertion that p is grounded in such knowledge in contexts where the hearer reasonably has the right to expect the asserter to possess more than merely isolated second‐hand knowledge, there is a problem.
To see this, let us begin by considering DOCTOR. The first point to notice is that Matilda clearly knows that Derek has pancreatic cancer—it is true, she believes it, she has good reason to trust the testimony of her medical student, and Nancy is in fact a reliable source. Given that Matilda herself has not reviewed any of the results of Derek's test, and has no independent information supporting the diagnosis of pancreatic cancer in this case, it is equally clear that Matilda's knowledge is both second hand and isolated. Of course, qua oncologist, Matilda knows a great deal about pancreatic cancer in general, and she has some limited data about Derek's symptoms from meeting him. But this broad information in no way grounds the specific knowledge in question—abdominal pain is, after all, a sign of numerous conditions, ranging from gallstones and food poisoning to intestinal obstructions and appendicitis. The knowledge that Matilda possesses that Derek in particular has pancreatic cancer, then, is grounded entirely in Nancy's testimony, and she has no additional information relevant to this specific diagnosis other than the fact that her student communicated to her. The question we must now consider is whether, under these conditions, (p.255) Matilda is properly epistemically positioned to flat-out assert to Derek that he has pancreatic cancer.
And here the answer is clearly no. For, while Nancy's reliable testimony may be sufficient for Matilda's knowing that Derek has pancreatic cancer, and while its isolated nature may not pose an epistemic obstacle to this being the case, the isolated second‐hand nature of Matilda's knowledge makes it improper for her to flat-out assert this diagnosis to Derek. One reason for this is that Matilda is an expert—she is an oncologist and Derek's physician, and such roles carry with them certain epistemic duties. In DOCTOR, 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. But the overarching epistemic duty here is that, qua oncologist, Matilda should be able (at least partially) to explain or justify the diagnosis of pancreatic cancer that she is offering to her patient. Moreover, as her patient, Derek reasonably has the right to expect his doctor to fulfill such a duty. Suppose, for instance, that he asks Matilda what exactly the ultrasound and MRI revealed, or how large his tumor is, or why she thinks it is pancreatic cancer, and she is unable to answer any of these questions. Indeed, suppose that she reveals to Derek that she had been told that he has pancreatic cancer by her student Nancy, that she had not actually seen any of the test results herself, and that she has no additional information to offer about his particular diagnosis. Would not Derek be entitled to resent Matilda under such circumstances, to feel that he has been epistemically cheated by his doctor, who owes him more than a diagnosis grounded purely in isolated secondhand knowledge? The upshot of these considerations, then, is that, in DOCTOR, we have a case where a speaker knows that p without thereby being epistemically positioned properly to assert that p, thereby falsifying the KNA‐S*.
Similar considerations apply in EXPERT PANELIST and PROFESSOR. In both cases, it is clear not only that the asserter has the knowledge in question, but also that it is second hand and isolated: in the former, John Smith truly believes that the Challenger disintegrated because of the failure of an O‐ring seal at liftoff solely on the basis of the reliable testimony of his co‐worker at NASA, whom he has very good reason to trust, and he has no additional or independent information grounding this explanation of the disaster; in the latter, Judith's true belief that the Supreme Court is unlikely to hear the upcoming challenge to the recent extension of US copyright protections to seventy years after the author's death is grounded entirely in the reliable testimony of her law‐school colleague, whom she knows to be trustworthy, and she has no further evidence justifying this claim. Are these speakers properly positioned epistemically to offer the flat‐out assertions in question? Once again, the answer to this question is no. In both cases, the asserters are experts who are offering assertions in contexts that call for their expertise, and thus they are expected to be able to defend or offer support for the assertions that they make when occupying such roles. In EXPERT PANELIST, John Smith—qua manager of NASA whose responsibilities included the monitoring of the shuttle operation, and qua expert called to testify at the House Committee hearing— (p.256) should be able to explain or justify the conclusion that a failed O‐ring seal is the cause of the Challenger disaster. Were fellow NASA workers and those present at the House Committee hearing to press Smith for additional information supporting this explanation, they would rightly feel epistemically cheated when they hear that he is basing this claim entirely on one co‐worker's testimony and that he is unable to offer anything beyond this assertion. And in PROFESSOR, Judith—qua law‐school professor teaching her students about US copyright law—should be able to provide some support on behalf of her statement that the Supreme Court is unlikely to hear the upcoming challenge to the recent extension of US copyright protections. If her law students raised their hands and asked for an explanation or defense of her assertion, they would be entitled to feel resentful when she informs them that she cannot offer any further data or support for this claim, because it is grounded entirely in an isolated statement made by her colleague.5
Now it is important to note that the epistemic problem with the assertions in the above cases is the result neither entirely of their being second hand, nor entirely of their being isolated, but, rather, of their being both second hand and isolated. This can be seen by modifying the scenarios so that only one of these features is present, and comparing the intuitions elicited from such modified cases with those from the original ones. For instance, suppose that, while all Matilda's knowledge regarding Derek's pancreatic cancer is second hand, it is not isolated. Perhaps Matilda did not see any of the test results herself, though she had an extensive conversation with her student Nancy about both the information discovered in the ultrasound and the MRI and why all the data strongly support a diagnosis of pancreatic cancer. Otherwise put, while Matilda's knowledge of the situation is entirely grounded in the testimony of her student, she acquires what we might call secondhand expertise regarding Derek's diagnosis of pancreatic cancer. In such a case, it is not at all clear to me, as it is in DOCTOR, that Matilda lacks the epistemic authority properly to assert to Derek that he has pancreatic cancer. Perhaps the most significant difference between this scenario and the original is that, were Matilda pressed about the assertion she offered, she would be able to provide additional support and tap into further explanatory resources on its behalf. Indeed, this is precisely what distinguishes an assertion grounded in secondhand expertise from one that is offered by an expert, though grounded merely in isolated secondhand knowledge.
The same holds in cases where there is isolated firsthand knowledge: suppose, for instance, that John Smith's knowledge of the cause of the disintegration of the Space Shuttle Challenger is grounded in his own perception of the failed O‐ring seal, though he lacks any supplementary information about this matter. Perhaps he saw a video of liftoff that clearly reveals the failure of such a seal, but he has no additional evidence about the cause or circumstances surrounding it. Once again, it is not clear, as it is in (p.257) EXPERT PANELIST, that Smith does not have the epistemic credentials properly to assert that the Challenger disaster is the result of the failure of an O‐ring seal. A significant difference between this scenario and the original is that, were Smith pressed about the assertion he offered, he would be able to say that he saw the failed O‐ring seal while reviewing the evidence from the disaster. This response has an authority or finality to it that is missing from “I was told that this is what happened by a colleague,” particularly when expert testimony is at issue.
We have seen, then, that there are cases where the assertion of an expert that is grounded entirely in isolated secondhand knowledge is epistemically improper, thereby showing that the KNA‐S* is false. Let us now turn to a second class of assertions that poses a problem for this view.
2 Isolated Secondhand Knowledge: Judgments
The three cases from the previous section—DOCTOR, EXPERT PANELIST, and PROFESSOR—all include assertions where the speaker in question is, in one sense or another, offering expert testimony. One question that some may have at this point, then, is whether counter‐examples to the KNA‐S* always involve assertions of experts.
By way of response to this question, consider the following two cases:
FOOD. My neighbor Ken is a connoisseur of fine dining. As we were leaving Starbucks this afternoon, he told me that the food at a new local restaurant about which I was previously quite unfamiliar, Quince, is exquisite, though being in a hurry prevented him from offering any details or evidence on behalf of this claim. While talking to my friend Vivienne later in the day, she was fretting over where to take her boyfriend to dinner for Valentine's Day. I promptly relieved her stress by truly asserting, “The food at Quince is exquisite.”
MOVIE. My colleague Richard is a movie buff whose assessments of films have proven to be quite reliable. As we quickly passed each other in the hall on our way to our respective classes this morning, he told me that The Diving Bell and the Butterfly, which is a new movie I had not previously heard of, is extremely moving. This evening, while my sister was deliberating about what to plan to celebrate her wedding anniversary, I truly said, “Well, The Diving Bell and the Butterfly is an extremely moving film.”
Now, notice that in both FOOD and MOVIE, not only do I know the relevant propositions that I am asserting; my knowledge is also clearly isolated and second hand. In FOOD, I truly believe that the food at Quince is exquisite solely on the basis of Ken's reliable testimony and, while I have very good reason to trust him on such a topic, I am not privy to any further evidence or reasons in support of this claim. Similarly, in MOVIE, while I correctly believe that The Diving Bell and the Butterfly is extremely moving entirely on the basis of a report offered by Richard, whom I know to be (p.258) trustworthy on such an issue, I do not possess any additional information or data on behalf of this statement. The question we should now ask is whether my possessing the relevant knowledge in these cases is sufficient for being properly epistemically positioned to offer the flat‐out assertions in question.
And here the answer is, once again, clearly in the negative. In both FOOD and MOVIE, the assertions that I make to my interlocutors involve judgments of different kinds: I am providing a judgment about the quality of the food at Quince to my friend in the first case, and I am offering a judgment about the emotional depth of The Diving Bell and the Butterfly to my sister in the second case. Assertions involving such judgments elicit various quite reasonable expectations in one's hearers. For instance, natural follow‐up questions from Vivienne to my assertion in FOOD are “Well, what is your favorite dish at Quince?” or “What kind of food do they serve?” And likely reactions from my sister in MOVIE are “Is this film too somber for an anniversary celebration?” or “What is this movie about?” Such questions betray the expectation on the part of my interlocutors that I am offering assertions that are not grounded in purely isolated secondhand knowledge. Indeed, imagine my friend's reaction upon hearing that not only have I myself never stepped foot in Quince nor tasted a morsel of its food; I also cannot offer a single piece of evidence or data to back my assertion. I cannot name an entrée on its menu, identify one palate‐tantalizing taste, or provide any particular recommendation. Despite my possessing the relevant knowledge in such a case, my friend would rightly feel cheated or resentful that I asserted that the food at Quince is exquisite under such circumstances. Otherwise put, if I am going to offer a flat‐out assertion about the quality of food at a given restaurant, there is a reasonable expectation on the part of my hearer that I can defend this claim—perhaps with some firsthand experiences, a fair bit more secondhand knowledge, or a combination thereof. Similar remarks apply to my sister's reaction upon hearing that not only have I never seen The Diving Bell and the Butterfly; I am also unable to identify a single scene, plot line, character, or feature of the film that renders it moving. She would quite reasonably feel that I had no right to offer such an assertion to her, that my epistemic credentials are inadequate to ground this judgment. The upshot of these considerations, then, is that in FOOD and MOVIE, we find another class of cases—one involving judgments grounded in isolated secondhand knowledge—where a speaker knows that p without thereby being epistemically positioned properly to assert that p. Thus, there is further support for rejecting the KNA‐S*.
Moreover, it should be noted that, as we saw with respect to the cases discussed in the previous section, the epistemic problem with the assertions in both FOOD and MOVIE is the result neither solely of their being second hand, nor solely of their being isolated, but of their being both second hand and isolated. Once again, this can be demonstrated by comparing the intuitions generated by the original cases with those elicited from modified ones that possess only one of the relevant features. For instance, suppose that, while all of my knowledge regarding the emotional depth of The Diving Bell and the Butterfly is second hand, it is not isolated. Perhaps Richard went into great detail about (p.259) the plot, the characters, the acting, and the directing of this film. Here, it is not as clear to me that I lack the epistemic credentials to flat-out assert to my sister that the film is moving. Were she to ask some natural follow‐up question, I would be able to defend my assertion by appealing to specific aspects of the movie. The same holds in cases where there is isolated firsthand knowledge: suppose, for instance, that my knowledge of the food at Quince being exquisite is grounded in my own firsthand experience, though I do not possess any additional information on behalf of this judgment. Perhaps I have a clear memory of having a lovely dining experience at this restaurant, but I cannot recall any details about the entrée I had. Again, it is not clear that I do not have the epistemic credentials properly to assert that the food at Quince is exquisite. Were my sister to question me about the proffered judgment, I would be able to respond that I remember enjoying my meal very much at this restaurant, despite forgetting the details of my experience.
Thus, assertions of judgments grounded entirely in isolated secondhand knowledge provide another class of cases where knowledge is not sufficient for epistemically proper assertion, thereby undermining the KNA‐S*.
3 Isolated Secondhand Knowledge: Presumed‐Witness and High‐Practical‐Stakes Contexts
The cases involving expert testimony and judgments discussed thus far are enough to show that the KNA‐S* is false: it is clear that a subject can know that p despite not being properly epistemically positioned to assert that p. But, once this is appreciated, it can be seen that there is a broad range of phenomena that fall under this description. For the sake of completeness, then, I shall here briefly discuss two additional classes of assertions that challenge the KNA‐S*.
The first class involves assertions grounded entirely in isolated secondhand knowledge in what we might call presumed‐witness contexts. The following provides an example of this type of assertion:
RECOMMENDATION. Josie, who was asked to support a philosophy student applying to Ph.D. programs, wrote in her letter of recommendation for his applications, “Mitchell has very polished writing skills.” While Josie does indeed know this about the student, her knowledge is grounded purely in the isolated, reliable testimony of her trustworthy colleague. Josie herself has had Mitchell in class for only a few weeks, and has yet to see any of his writing.
Now, in RECOMMENDATION, Josie clearly knows that Mitchell has polished writing skills—after all, the reliable testimony of another professor who has taught the student in question can be as a good a basis as just about any for forming beliefs about the student's abilities. Moreover, the knowledge in question is both isolated and secondhand: Josie knows this fact about Mitchell's writing abilities solely on the basis of a (p.260) single remark from a colleague. Does she have the epistemic authority to offer the above flat‐out assertion in her letter of recommendation?
Once again, the answer to this question is no. The assertion that Josie offers in her letter is grounded entirely in isolated secondhand knowledge in what I earlier called a presumed‐witness context. A context of this sort is one in which the parties to the discussion reasonably presume that the asserter has the status of being a witness—which involves the possession of either firsthand or some other kind of privileged information—to the issue or event in question. For instance, given that Josie is offering a flat‐out assertion about the abilities of a student in a letter of recommendation, it is utterly reasonable for readers of her letter to presume that she experienced Mitchell's writing first hand. Indeed, upon hearing the details of the grounding of her assertion, a reasonable question to ask is, “Well why did not your colleague, rather than you, write the letter of recommendation on behalf of Mitchell?” If Josie were able to offer details or particular features about the student's writing that her colleague had shared with her, this might go some way toward rendering her assertion legitimate. But possessing isolated secondhand knowledge here is clearly not sufficient for Josie's being properly epistemically positioned to make such a flat‐out assertion in a letter of recommendation.
There is one further class of assertions that poses a problem for the KNA‐S* that I shall briefly mention: those assertions grounded in isolated secondhand knowledge offered in high‐practical‐stakes contexts. Consider, for instance, the following:
CHEATING. During my office hours today, Jamie—a student from my introduction to philosophy course—came to see me. While we were talking about the recent assignments for the class, she truly asserted, “Sam Smith cheated on the midterm exam.” It turns out that Jamie did not herself see Sam cheat on the exam, but acquired this information via an isolated, though trustworthy and reliable, remark from her friend, Colin.
Now, while the true testimony of a reliable friend is adequate for Jamie to know that Sam cheated on the midterm exam, is she properly epistemically positioned to flat-out assert that this is the case? Once again, the answer to this question is no. For CHEATING involves an assertion being offered in a context with high practical stakes—Jamie is calling into question Sam's academic honesty and integrity and, given that I am the professor of the class, such an accusation brings with it the possibility of serious consequences. Because of this, there is the expectation that Jamie's relationship to the proffered assertion is such that she will be able to defend or otherwise vouch for its truth. I would, for instance, assume that Jamie herself saw Sam cheating upon hearing her flat-out assert that this is the case. At the very least, I would presume that she could offer some support on its behalf, such as providing details about the nature or the extent of the cheating at issue. Upon learning that she neither saw nor can answer any of my reasonable questions about the incident, I would think she had no right to make such (p.261) an assertion to me—that she lacked the epistemic authority to do so. This point can be further illustrated by considering whether it would be appropriate for me, in turn, to flat-out assert that Sam cheated on my midterm exam to my colleague in the Office of Judicial Affairs when my sole basis for possessing this information is Jamie's isolated, reliable testimony. Again, while a student's true, reliable testimony may be sufficient for my knowing that this is the case, asserting that Sam cheated to my colleague in Judicial Affairs requires a different kind of grounding. With so much at stake in this student's life, I had better make sure that my assertion is supported by testimony that is either first hand or detailed enough to justify.6
These reactions stand in contrast to those elicited by an altered version of CHEATING that does not involve an assertion in a high‐practical‐stakes context. Suppose that everything is the same as in the original case, except the secondhand knowledge and corresponding assertion at issue are about Sam having the flu. Were Jamie to flat-out assert this to me purely on the basis of Colin's isolated testimony, it would seem entirely appropriate for her to do so. The fact that so much is at stake in CHEATING—such as Sam's reputation, grade in my class, and standing in the university—is precisely what makes Jamie's epistemic credentials inadequate to ground her assertion.
Thus, we have seen that there are at least four broad categories of assertions grounded in isolated secondhand knowledge that undermine the KNA‐S*—those made by experts, those expressing judgments, those made in presumed‐witness contexts, and those involving high practical stakes.
4 Two General Features
In this section, I shall briefly mention two general features of all the cases considered in the previous sections that provide further support for the rejection of the KNA‐S*.
First, in all the scenarios, the assertions in question clearly should not be flat out but, rather, should be prefaced with “I heard that…” “I've been told that…” or some similar qualification. In PROFESSOR, for instance, Judith instead asserting to her students “I've been told that the Supreme Court is unlikely to hear the upcoming challenge to the recent extension of US copyright protections to seventy years after the author's death” sounds both appropriate and natural. Upon being challenged by her students to defend this assertion, Judith appealing solely to the isolated, secondhand support provided by her colleague's testimony seems perfectly sufficient to justify such a qualified assertion. The same is true of the other assertions in question—my asserting in MOVIE that I heard that The Diving Bell and the Butterfly is a moving film not only sounds more appropriate than my original assertion; it also seems called for in this context. This is evidenced by the fact that my appealing to Richard's testimony to defend this qualified assertion in the face of being pressed by my sister is clearly (p.262) adequate to meet her challenge, which stands in contrast to its inadequacy to support my unqualified assertion in MOVIE.
Second, a natural reaction to the above cases that supports a rejection of the KNA‐S* is that, upon being challenged, the asserters would rightly feel embarrassed or apologetic for offering the flat‐out assertions in question. If Derek requested that his cancer diagnosis be defended or explained in DOCTOR, Matilda would naturally feel as though her flat‐out assertion was epistemically unjustified. She would quite likely regret not having reviewed the test results herself, or at least not insisting on additional support from Nancy, and thus she would find it extremely embarrassing to admit to Derek that she lacks any information about his diagnosis other than the relevant isolated secondhand knowledge. Similarly, were Vivienne to press me on the kind of food served at Quince or the name of my favorite entrée, I would be uncomfortable about the assertion that I had offered on behalf of this restaurant. I would be inclined to retract my earlier flat‐out assertion and replace it with an appropriately qualified one, such as, “Well, what I should have said is that I heard that the food is exquisite because I actually haven't eaten at Quince myself, nor do I know anything about the menu or the entrées.”
These general considerations provide further reason to reject the thesis that knowledge is sufficient for epistemically proper assertion. Let us now turn to some responses that may be offered on behalf of the KNA‐S*.
5 Objections and Replies
One response that the proponent of the KNA‐S* may offer to the counter‐examples discussed in the previous sections is to grant that there is something problematic about the proffered assertions, but to explain the wrongness in question via Gricean conversational implicature.7 In particular, it may be argued that in all the conversational contexts at issue, the speakers falsely implicate that their assertions are not grounded purely in isolated secondhand reasons. For instance, Matilda flat-out asserting in DOCTOR that Derek has pancreatic cancer falsely implicates that she has reviewed the test results herself, or at least that she has support that can be offered on behalf of this diagnosis. Similarly, Josie asserting in RECOMMENDATION that Mitchell has polished writing skills in a letter for a Ph.D. application falsely implicates that she has seen his papers first hand, or at least that she has some other direct familiarity with his writing. Thus, while the assertions themselves satisfy the norm of assertion and are thereby epistemically proper, their false implicatures enable a Gricean explanation of any impropriety found in such cases.
There are, however, at least three good reasons to reject this response to my counter‐examples. First, this sort of move, if it is legitimate, can be extended to account for the linguistic data that proponents of the KNA rely on in support of their own (p.263) view. Thus, if these Gricean considerations weigh against my objections to the KNA, they also weigh just as heavily against the KNA itself. For instance, in support of the KNA, Hawthorne and Stanley ask us to consider the apparent fact that we would regard a sous‐chef's asserting that a cake is done in the absence of knowing that it is so done as intuitively defective.8 But notice: if the impropriety in my counter‐examples can be explained via the false Gricean implicature that the epistemic basis of such assertions is not merely isolated secondhand knowledge, why could not cases such as that of the sous‐chef be similarly explained via the false Gricean implicature that the epistemic basis of such assertions is at least knowledge? That is to say, why could we not account for the intuitive defectiveness of what the sous‐chef asserts by taking his assertion to implicate, falsely, that he knows that the cake is done? This, of course, then opens the door to offering a general explanation of the impropriety of assertions that fall short of knowledge via Gricean implicatures—someone who does not know but asserts anyway that the store is open, or that the neighbor's dog is safe, or that the ice cream does not have peanuts in it is subject to criticism because she is falsely implicating that she has knowledge in such cases when in fact she does not. The upshot of these considerations, then, is that proponents of the KNA will have a hard time explaining why Gricean considerations apply to my assertions involving isolated secondhand knowledge, but not to those that they use in support of the KNA. Thus, this sort of move turns out ultimately to be more harmful than helpful for the KNA.
There is a second problem with attempting to explain the wrongness of the assertions in my counter‐examples via Gricean conversational implicature. For, even if this move did not turn out to be equally undermining to the KNA, there are important asymmetries between the cases involving isolated secondhand knowledge and classic examples of Gricean implicature. To see this, consider the following familiar instances of the latter:
(1) A professor writing on behalf of a student applying to graduate school ends her letter by saying merely, “Ian is very hard working.” However, the professor believes that Ian is also very intelligent.
(2) While talking to a friend, Peter says, “Claire got pregnant and got married.” However, he knows that these events did not occur in the order in which he mentioned them.
Now notice that, in (1) and (2), the impropriety at issue is with the way in which the asserters offer their assertions. In (1), for instance, concluding a letter of recommendation by merely asserting that a student is hard working implicates that he is not particularly smart or talented. It is not that the professor in (1) should not have expressed her belief that Ian is hard working in her letter, nor that she lacked the epistemic authority to make such a flat‐out assertion. Rather, she should not have put (p.264) her point about Ian the way that she put it—for example, in isolation, at the conclusion of her letter of recommendation, and so on. Similarly, in (2), saying of a friend that she got pregnant and got married implicates that this is the order in which these events took place. We would not criticize Peter for expressing his belief that such events occurred in Claire's life, nor say of him that he lacked the epistemic credentials to offer this flat‐out assertion. Instead, we would say that he should not have put his point about Claire the way in which he put it—for example, in the reverse order from which the events in fact occurred. Indeed, Grice (1989: 39) himself says that “implicature is not carried by what is said, but only by the saying of what is said, or by ‘putting it that way’” (emphasis added).
In contrast, the impropriety relevant in the cases involving isolated secondhand knowledge is with the epistemic grounding of the assertions in question and thus with their having been made at all. In DOCTOR, for instance, the problem with Matilda's assertion that Derek has pancreatic cancer is not that she should have inserted it later in the conversation, or expressed its content in an alternative order, or put her point in a slightly different way. Rather, Matilda lacked the epistemic authority to flat-out assert this diagnosis to Derek in any way at all. This is evidenced by the fact that the natural criticism of her assertion is that it has grounds that are simply not good enough. Similar considerations apply to the other cases of isolated secondhand knowledge: in RECOMMENDATION, for instance, it would not alter the wrongness of Josie's assertion that Mitchell has polished writing skills for her to express her thought differently. For, no matter how Josie puts her point, her epistemic credentials are inadequate to ground such a flat‐out assertion.9
This contrast between my cases of isolated secondhand knowledge and classic examples of Gricean implicature is further evidenced by an asymmetry involving cancelability. Specifically, in most instances of the latter, there is no residual impropriety once the implicature has been canceled, but canceling the supposed implicature in my counter‐examples does not thereby remove the impropriety in question. In (1), for example, if the professor were to follow up her remark about Ian's being hard working with, “but I do not mean to suggest that he is not also very intelligent,” there would be no lingering impropriety with respect to her assertion. Members of graduate admissions committees would not feel angry or resentful that the professor said this, but, rather, would take her assertion as highlighting two different virtues of a student that are not mutually exclusive. Similarly, if Peter were to add after saying that Claire got pregnant and got married, “but of course not in this order,” there would no longer be impropriety left to explain. His friends would not feel cheated or outraged by his having said this, and they surely would not regard Peter as lacking the right to have made such an assertion. In contrast, suppose that immediately after asserting to Derek (p.265) that he has pancreatic cancer in DOCTOR, Matilda adds, “but I myself have not seen any of your test results, nor do I have any specific reasons to offer to defend this diagnosis.” Even if Matilda succeeds in canceling the supposed implicature of her assertion, she does not thereby eliminate the wrongness or epistemic impropriety of it. Derek would still rightly feel resentful, even incensed, that his oncologist had flat-out asserted a cancer diagnosis to him without being able to offer any direct support on its behalf. The same is true of Josie's assertion in RECOMMENDATION—even if she were to follow her assertion about Mitchell's writing skills with, “but I myself have never seen any of his papers, nor can I identify any particular virtues of his writing,” the members of graduate admissions committees reading this letter would reasonably be outraged by this flat‐out assertion. They would quite likely agree that Josie had no right to offer such a claim about Mitchell's writing in a letter of recommendation, regardless of her attempt to cancel any implicature about possessing knowledge that is not both isolated and second hand.
Of course, there are some classic examples of Gricean implicature where the impropriety of the assertion remains even after cancellation. Consider, for instance, the following:
(3) In response to a passerby asking where the nearest gas station is, Susan says, “It's around the corner.” However, Susan knows that the gas station in question is in fact closed.
Now, if Susan were to add to her assertion in (3), “but the gas station is closed,” there is still the sense that the assertion should not have been made at all. But notice: the reason why Susan should not have offered her assertion in (3) at all is because it is pointless or irrelevant.10 It is standardly assumed that, if a passerby is asking about the whereabouts of the nearest gas station, she is looking for one that is currently open. So, even if Susan cancels the implicature of her assertion in (3)—that is, that the gas station around the corner is open—her assertion is still irrelevant or pointless given the current purposes of the exchange. It is only in this sense that Susan should not have offered the assertion in question. From an epistemic point of view, there is nothing whatsoever wrong with what she said. In my cases of isolated secondhand knowledge, however, the remaining impropriety involves the asserter's epistemic relation to that which she asserted. For, even if the asserter cancels the supposed implicature that she possesses a firsthand or non‐isolated basis for what she says, there is the clear sense that she lacks the epistemic authority or credentials to offer the assertion in question. Thus, this difference casts serious doubt on the attempt to explain the wrongness of the assertions in my counter‐examples via Gricean conversational implicature.
KNPR‐S One is properly epistemically positioned to use the proposition that p in practical reasoning if one knows that p.12
Thus, just as it is said to be intuitively permissible to flat-out assert that p when one knows that p, so, too, it is said to be intuitively acceptable for similar reasons to act on the proposition that p when one knows that p. For instance, if I decide to leave for the airport an hour later than was expected, my knowing that the relevant flight was delayed seems sufficient to render such a conclusion permissible. If my choice is questioned, appealing to my knowledge adequately meets the challenge, while offering anything less—such as my suspecting that the flight is delayed, or being justified in believing that it is—does not. It is not difficult to see, however, that modified versions of some of the counter‐examples discussed thus far can be used equally to undermine the KNPR‐S. Suppose, for example, that, instead of flat-out assert to Derek that he has pancreatic cancer in DOCTOR, Matilda schedules surgery and begins operating on him purely on the basis of Nancy's isolated testimony. It seems clear that it is just as problematic for Matilda to act on the basis of isolated secondhand knowledge as it is for her to assert flat out to Derek on the basis of this knowledge.13 Yet notice: there is simply no plausible sense in which Gricean considerations can be appealed to here to explain the intuitive impropriety in Matilda's action. Given this, combined with the fact that similar reasoning often underlies both the KNA‐S* and the KNPR‐S, there is good reason to doubt the ability of Gricean considerations adequately to account for the data on the assertion side as well.
A second response that may be offered to the above counter‐examples on behalf of the KNA‐S* is to grant the impropriety of the relevant assertions, but to explain the wrongness at issue through various institutional norms.14 For instance, it may be argued that the medical profession itself requires that doctors, particularly specialists, offer diagnoses that are not grounded entirely in isolated secondhand knowledge. Thus,qua asserter, Matilda is not subject to any criticism—she possesses all the necessary epistemic credentials properly to assert that Derek has pancreatic cancer. But,qua doctor, she has violated norms imposed by the medical profession, and the institution at issue here can adequately account for the wrongness of her assertion. Similarly, it is part of our institution of letter of recommendation writing that assertions offered in such contexts not be based purely on isolated secondhand reasons. Qua asserter, then, Josie has not violated any norms—she possesses the requisite epistemic basis properly to assert that (p.267) Mitchell has polished writing skills. But qua letter‐of‐recommendation writer, she is subject to criticism for not satisfying the norms imposed by this institution. This enables a satisfying explanation of the impropriety of the assertions from the previous section, without calling into question the KNA‐S*.
This response on behalf of the sufficiency of knowledge for assertion fails for two central reasons. First, the assertions involving isolated secondhand knowledge are not epistemically problematic because various institutions say that they are wrong; rather, the institutions say that they are wrong because such assertions are epistemically problematic. To see this, consider what would happen if the institutions changed such that it was no longer improper to offer assertions grounded purely in isolated secondhand knowledge. If, for instance, the medical profession changed so that diagnoses from specialists could be grounded entirely in a single instance of reliable testimony, this institution would no longer serve the epistemic purpose for which it was created. Patients would no longer regard the medical verdict of an expert as having a certain kind of epistemic authority, and thus they would cease to consult specialists to obtain precisely the specialized information that the medical profession intended these doctors to provide. Similar considerations apply when we consider the institution of letter‐writing changing so that it is permissible to offer assertions about candidates that are grounded entirely in isolated secondhand knowledge. If letter‐writers routinely offered information about applicants based purely on such knowledge, this institution would be utterly ineffective in serving the epistemic role it was meant to serve.15 Members of admissions committees would no longer regard the assessments of candidates found in these recommendations as having an epistemic authority backed by the recommender's reputation and acquaintance with the applicant, and thus these letters would cease to be precisely the crucial factors in admissions decisions that they were intended to be. This gives us reason to conclude that the assertions grounded entirely in isolated secondhand knowledge at issue here are deemed wrong by the relevant institutions because they are epistemically improper, rather than the other way around.
Second, even if the appeal to institutional norms adequately accounts for the problematic nature of some of the assertions in question, it surely is not an adequate explanation of all of them. For, unlike the role of the medical profession in DOCTOR and that of the institution of letter writing in RECOMMENDATION, there is no plausible institution to shoulder the burden of wrongness in FOOD, MOVIE, or CHEATING. What institution, for instance, has norms governing judgments about good food, assertions about moving films, and claims about fellow students cheating? Moreover, assertions involving judgments and those offered in presumed‐witness and high‐stakes contexts (p.268) contain so much variety of content within each category that it is unlikely that there even are particular institutions that can plausibly be appealed to. High‐stakes contexts, for instance, can involve an endless variety of risky behavior, including cheating, lying, betraying, murdering, plotting, gambling, accident prevention, and so on. To suppose that there is an institutional norm, or even a multitude of such norms, that can explain the wrongness of such wildly different assertions seems implausible. Of course, despite how multifarious such assertions are with respect to content, they are united in all being assertions offered in high‐stakes contexts. But then the most likely candidate to subsume this sort of unity is precisely the norm governing assertion that is at issue.
However, what if the proponent of the KNA‐S* were to respond here that there are norms, beyond the strict institutional ones considered above, that can be invoked to explain the impropriety of the assertions involving isolated secondhand knowledge? Perhaps there are social norms, or broader institutional norms—such as those governing friendship and other kinds of relationships—that can be appealed to in order to explain the intuitive defectiveness of such assertions.16 For instance, perhaps there are values guiding our social interactions that prohibit asserting aesthetic judgments grounded purely in isolated secondhand knowledge. Or perhaps norms of friendship preclude offering recommendations to our friends without possessing either non‐isolated or firsthand knowledge of the information in question. Moreover, these norms could presumably be epistemic in nature by requiring that asserters have a particular kind of epistemic grounding in order to offer proper assertions. Given this, the mere fact that there are not strict institutional norms that can account for the impropriety of the assertions in my counter‐examples does not rule out there being other norms that can shoulder this explanatory burden.
Even if this move could be rendered clear and substantive enough to provide a genuine response to my cases, there are compelling reasons to reject it as a defense of the KNA‐S*. To see this, notice that, in order to take the KNA‐S* seriously, we need first to be able to understand what it means to say that knowledge is sufficient for proper assertion. For, as was noted at the beginning of this chapter, there are all sorts of senses in which this norm, without appropriate qualification, is obviously false—for example, if I know that my colleague made a fool of himself while we were all drinking the other night, it may still be imprudent, impolite, or pointless for me to assert that this is the case at a department meeting on Monday morning. This is, of course, why proponents of the KNA‐S* emphasize that knowing that p is sufficient for possessing the requisite epistemic credentials properly to assert that p. And we can certainly understand the KNA‐S* when this distinction is made, for epistemic authority is one thing, and prudence, politeness, and relevance are quite another. But, if the proponent of the KNA‐S* were to adopt the line of response under consideration here, and argue that there are a host of other norms that can explain the impropriety in my counter‐ (p.269) examples—some of which directly govern whether an asserter has the proper epistemic credentials to assert that p—then the KNA‐S* is no longer comprehensible. For now, whenever evidence is adduced that concerns the epistemic authority requisite for proper assertion, it may bear on the norm of assertion or it may bear on these other social or broader institutional norms. For instance, cases where criticism does not seem appropriate when an assertion is made in the presence of knowledge may support the KNA‐S* or may support the existence of these other norms. Given that these other norms explicitly govern matters of epistemic authority and assertion, it will be extremely difficult, if not impossible, to tell which is being defended. Moreover, if some of these social and institutional norms govern assertion itself, this may be a good reason to doubt that there even is a single norm of assertion, as proponents of the KNA‐S* maintain. Thus, this strategy for responding to my counter‐examples ultimately turns out to be undermining to the KNA‐S*, in terms of both understanding and motivating such a view.
A third response that may be offered in defense of the KNA‐S* is to exploit the high stakes in my counter‐examples to argue that the asserters in question do not possess the relevant knowledge.17 For instance, subject‐sensitive invariantists, contextualists, and those who endorse pragmatic encroachment18 all maintain that high practical stakes can make a difference to whether a subject knows a given proposition. Given this, it may be argued that my cases fail to pose a problem for the KNA‐S* because assertions concerning, for example, cancer diagnoses and space‐shuttle disintegrations involve high practical stakes, which thereby prevent the asserters from possessing the knowledge in question.19
There are three central problems with this response. First, even if this strategy can plausibly be applied to some of my counter‐examples to the KNA‐S*, other cases clearly do not involve high practical stakes in any sense whatsoever. For instance, offering a judgment about the entrées at a restaurant to a friend in FOOD or recommending a film to my sister in MOVIE may have some negative consequences if they turn out to be misleading, but surely Vivienne and her boyfriend eating less than exquisite food or my sister and her husband watching a less than moving film do not involve high stakes in any reasonable sense.20 Second, even if all my counter‐examples could be construed as involving high stakes, the response offered by the subject‐sensitive invariantist is to require more justification or more warrant for knowledge in such cases. Yet surely even if the testimony in the relevant cases is extraordinarily truth (p.270) conducive, the fact that it is isolated and secondhand renders it nonetheless an inadequate ground for the assertions in question.21 Nancy's testimony in DOCTOR, for example, may be far more reliable than it is in the original scenario, yet it still seems like the wrong kind of basis to ground Matilda's flat‐out assertion that Derek has pancreatic cancer. Third, those cases that do involve high practical stakes can easily be altered so that the stakes in question are no longer high, and yet impropriety nonetheless remains with respect to the assertions.22 For instance, the diagnosis in DOCTOR may involve bunions or strep throat rather than pancreatic cancer. Still, if the diagnosis is grounded entirely in isolated secondhand knowledge, the intuition remains that Matilda's assertion to Derek is improper. Similarly, Smith may testify at a House Committee hearing in EXPERT PANELIST about the discovery of a minor comet or recent promotions in the astronaut program. If his testimony is grounded only in isolated secondhand knowledge, however, there is still the intuition that Smith's assertion is epistemically improper. And, in RECOMMENDATION, Josie may be writing on behalf of Mitchell's application for volunteer work at a local library, rather than for a Ph.D. program in philosophy. Once again, if her assertion is based only on knowledge that is isolated and secondhand, it is intuitively improper. Thus this strategy for defending the KNA‐S* fails to provide an adequate response to the full range of my counter‐examples.
Finally, it may be argued that the intuitive impropriety in my counter‐examples can be accounted for in terms of general, social expectations about tact and politeness. Consider, for instance, DOCTOR. Here it may be argued that it is rude or shocking for a doctor to assert a diagnosis to her patient that is grounded purely in isolated secondhand knowledge. Thus, even though knowledge is sufficient for proper assertion, the proponent of the KNA‐S* may claim that the impropriety involved in Matilda's telling Derek that he has pancreatic cancer can be explained by its rude and appalling nature.23 Similar considerations apply to the other cases—Smith's asserting in EXPERT PANELIST that the Space Shuttle Challenger disintegrated because of the failure of an O‐ring seal at liftoff, for example, is impolite and shocking, because we do not expect expert testimony to be grounded purely in isolated secondhand knowledge.
By way of response to this objection, consider the second claim first—namely, that the assertions in question are improper because they are shocking. Now notice that it would be shocking if a passerby on the street announced that the President had just died of a sudden heart attack, but surely such an assertion would not be improper in any relevant sense. So, even if it were granted that the assertions in my counter‐examples are shocking, it is unclear why this would thereby explain the intuitive impropriety present. Regarding the first claim—that the assertions in question are rude—it is not at (p.271) all obvious that this applies to the cases at hand. Why, for instance, would anyone consider a doctor's diagnosis to be rude? Is not receiving such a diagnosis the whole point of visiting the doctor? Indeed, would it not be rude if the doctor did not give the diagnosis to the patient? Still further, we can easily imagine a case in which the doctor in question is extremely kind, gentle, polite, and so on. When she gives the diagnosis of cancer to the patient, she does so in a profoundly empathetic way. On the account under consideration, however, this would still count as an instance of rudeness. But this seems absurd. Finally, the fact that the doctor's assertion strikes us as improper, even when delivered in this compassionate way, lends additional support to the claim being defended in this chapter—that what is wrong with the assertions in question concerns the nature of epistemic support present, not any sort of rudeness.
We have seen, then, that the four most plausible responses to the counter‐examples facing the KNA‐S* fail to support the sufficiency of knowledge for epistemically proper assertion. Let us now turn to some general conclusions that can be drawn from such cases.
A feature common to all the cases discussed above is that the assertions in question carry with them expectations about the kind of grounds underlying them. In this sense, the problem with the knowledge in these cases is not one of quantity, but of quality. In DOCTOR, for instance, the source of Matilda's information about Derek's pancreatic cancer may be extraordinarily truth conducive, so much so that her knowledge of his condition is very high on the spectrum between knowledge and certainty. Regardless of this very high quantity of epistemic support, however, if all of it is in the form of isolated secondhand support, the intuition that Matilda lacks the epistemic authority to flat-out assert to Derek that he has pancreatic cancer remains. Similar considerations apply to the remaining cases: even if Ken is a connoisseur of fine dining, and his testimony provides me with a highly justified belief that the food at Quince is exquisite, Vivienne is still entitled to feel as though I have no epistemic right to flat-out assert that this is the case when my knowledge is entirely isolated and secondhand. And, no matter how well supported Josie's report is about Mitchell's writing skills or Jamie's testimony is about Sam's cheating, if such assertions entirely lack both firsthand and non‐isolated grounding, they are not offered by asserters who possess the appropriate epistemic credentials.
This point can be put in a slightly different way: suppose, for instance, that the degree of epistemic support on behalf of Jamie's belief is lower than that found in CHEATING so that it falls slightly below what is needed for knowledge. Perhaps Colin's testimony is less reliable than in the initial case, or Jamie's cognitive faculties are not functioning as properly as they are in CHEATING, or Jamie has less evidential support for the trustworthiness of her friend than she does in the original scenario. Despite this overall inferiority in the strength of her epistemic position, however, suppose that some of it is in the form of non‐isolated firsthand support; say, Jamie saw Sam suspiciously (p.272) and repeatedly consulting his BlackBerry during the exam, and she can offer details about her experience to defend the proffered assertion. When we compare this case—which we can call CHEATING*—with the original one, it seems clear that Jamie is in a better epistemic position to flat-out assert to me that Sam cheated on the midterm exam under these conditions, despite the fact that she possesses the relevant knowledge in CHEATING and fails to possess it in CHEATING *. Indeed, even if I were to hear that Colin is an extremely reliable testifier in the former case, and that Jamie's epistemic support in the latter case is largely inferential, I would still regard her assertion in CHEATING as less epistemically appropriate than in the modified CHEATING *.
The same is true of the other cases: for instance, suppose that the amount of epistemic support enjoyed by Josie's belief that Mitchell has polished writing skills in RECOMMENDATION is just lower on the quantity spectrum than what is needed for knowledge. The testimonial source of her belief may be slightly less reliable than it is in the initial case, or Josie may lack the amount of evidence she possesses in RECOMMENDATION on behalf of this source's testimony, and so on. Nevertheless, suppose that some of the epistemic support in question is non‐isolated and first hand; perhaps Josie read one, fairly short paper written by Mitchell, and she can offer some details from this single essay defending her positive attitude toward his writing skills. Once again, when we compare the assertion in this case—which we can call RECOMMENDATION*—with the original one, Josie clearly seems better positioned epistemically to assert that Mitchell has polished writing skills here than she does in the initial scenario. In particular, even though she lacks the knowledge in question in RECOMMENDATION*, and possesses it in RECOMMENDATION, her assertion seems epistemically proper in the former, but not the latter, case. This reveals that one's position on the epistemic quality spectrum can trump one's position on the epistemic quantity spectrum with respect to the propriety of one's assertion.
Of course, the fact that one's assertion is grounded entirely in isolated secondhand knowledge is not sufficient, by itself, for one to lack the appropriate epistemic credentials in question. Suppose, for instance, that, instead of Matilda asserting flat out to Derek that he has pancreatic cancer in DOCTOR, she casually asserts this fact to her husband over dinner. In both cases, Matilda's assertion is grounded entirely in isolated secondhand knowledge provided by Nancy's reliable and trustworthy testimony. Yet, intuitively, Matilda's assertion lacks the appropriate epistemic grounding when made to Derek, but not when made to her husband. This reveals that isolated secondhand knowledge cannot be the full explanation of what renders the assertions in question epistemically defective. What else does the explanatory work here? While a complete answer to this question lies beyond the scope of this chapter, some of the factors relevant here may include the broader social expectations governing assertions made in various contexts. For instance, with respect to the epistemic grounds needed for proper assertion, we have higher expectations when our doctors are offering our diagnoses to us than we do when our spouses are offering casual assertions about other peoples' diagnoses over dinner. Thus, a full account of the kinds of assertions that falsify the (p.273) KNA‐S* will include not only isolated secondhand knowledge, but also the nature and scope of these expectations.24
It is of interest to note that these conclusions hold, regardless of whether one endorses externalism or internalism with respect to epistemic justification. For instance, in EXPERT PANELIST, it can be stipulated that the co‐worker from whom John Smith acquired the information about the disintegration of the Space Shuttle Challenger is extraordinarily reliable, has perfectly functioning cognitive faculties, and tracks the truth impeccably. It can also be built into the case that Smith possesses excellent reasons and compelling evidence for trusting his co‐worker's testimony. Even with this powerful externalist and internalist justification for accepting his co‐worker's testimony, however, there is still a clear sense in which Smith lacks the appropriate epistemic credentials to assert at the House Committee meeting that the Space Shuttle Challenger disintegrated because of the failure of an O‐ring seal at liftoff. Qua expert, he is expected to be able to explain or support his view of the disaster, regardless of how close to certainty the knowledge grounding his assertion is.
Given the distinction between the quantity and the quality of epistemic support needed for proper assertion, however, might it be argued that there is a weaker version of the KNA‐S* that is not falsified by the counter‐examples in question? In particular, while knowing that p may not be sufficient for being properly epistemically positioned to assert that p, it may be sufficient for having the requisite epistemic strength to assert that p. Thus, the following sufficiency claim may still be true:
KNA‐S** One has the quantity of epistemic support required properly to assert that p if one knows that p.
The KNA‐S** is compatible with knowledge being insufficient for having the quality of epistemic support required for proper assertion, and so it may be concluded that it is not targeted by the counter‐examples from the previous sections.25
The first point to notice about this response is that it concedes the main thesis of this chapter—knowledge is not sufficient for epistemically proper assertion. One may know that p, satisfy the KNA‐S**, and yet still not be properly epistemically positioned to assert that p. Given that the heart of the KNA‐S* is that knowledge is all that is epistemically needed for proper assertion, the KNA‐S** represents a significant weakening of the view in question. Second, at least one of the cases from the previous sections—CHEATING—can be read as showing that knowledge is also not sufficient for possessing the quantity of epistemic support required properly to assert that p. In particular, it is not implausible to think that, when the practical stakes are high in a given situation, more than knowledge, rather than merely a different kind of knowledge, may be needed for epistemically proper assertion. For instance, if Jamie has a (p.274) firsthand experience of Sam behaving suspiciously during the class's midterm exam, but it is barely enough to clear the threshold of support needed for knowledge, it is still questionable whether she is properly epistemically positioned to flat-out assert that he cheated on the exam to me. This intuition is even clearer when the practical stakes in question are even greater. Suppose, for instance, that I am responsible for determining whether a commercial aircraft with 500 passengers on board is safe for travel. Possessing only the minimum quantity of epistemic support needed for knowledge seems intuitively insufficient for my properly asserting to the pilot and crew members that it is safe to travel; something a bit closer to certainty seems necessary here.26 Thus, there is reason to doubt even the substantially weaker KNA‐S**.
Finally, I shall close by noting the extent to which these considerations motivate a significant change in direction in our theorizing about norms of assertion. All the parties to the debate currently agree that the quantity of epistemic support relevant to knowledge is the feature at issue. For instance, questions surrounding this topic typically take the form of whether less than knowledge is sometimes adequate, or more than knowledge is sometimes needed, for proper assertion. As we have seen, however, this is misguided, for even assertions grounded in a very high degree of justification—well above the threshold for knowledge—can fail to be epistemically appropriate. In order to make genuine progress in theorizing about norms of assertion, then, we need to move beyond the exclusive focus on the quantity of epistemic support an assertion enjoys and turn our attention to the quality of that support as well.
Adler, Jonathan E. (2002). Belief's own Ethics. Cambridge, MA: MIT Press.
Brandom, Robert (1983), “Asserting,” Noûs, 17: 637–50.
—— (1994). Making it Explicit. Cambridge, MA: Harvard University Press.
Brown, Jessica (forthcoming a). “Knowledge and Assertion,” Philosophy and Phenomenological Research.
—— (forthcoming b). “Assertion and Practical Reasoning: Common or Divergent Epistemic Standards?” Philosophy and Phenomenological Research.
Cohen, Stewart (2004). “Knowledge, Assertion, and Practical Reasoning,” in Ernest Sosa and Enrique Villanueva (eds), Philosophical Issues, 14: 482–91.
DeRose, Keith (2002). “Assertion, Knowledge, and Context,” Philosophical Review, 111: 167–203.
Douven, Igor (2006). “Assertion, Knowledge and Rational Credibility,” Philosophical Review, 115: 449–85.
Fantl, Jeremy, and McGrath, Matthew (2002). “Evidence, Pragmatics, and Justification.” Philosophical Review, 111: 67–94.
Fricker, Elizabeth (2006). “Second‐Hand Knowledge,” Philosophy and Phenomenological Research, 73: 592–618.
(p.275) Grice, Paul (1989). Studies in the Way of Words. Cambridge, MA: Harvard University Press.
Hawthorne, John (2004). Knowledge and Lotteries. Oxford: Oxford University Press.
—— and Jason Stanley (2008). “Knowledge and Action,” Journal of Philosophy, 105: 571–90.
Kvanvig, Jonathan (2009) “Assertion, Knowledge, and Lotteries,” in Patrick Greenough and Duncan Pritchard (eds), Williamson on Knowledge. Oxford: Oxford University Press, 140–60.
Lackey, Jennifer (2007). “Norms of Assertion,” Noûs, 41: 594–626.
—— forthcoming “Acting on Knowledge” in John Hawthorne and Jason Turner (eds), Philosophical Perspectives.
Reed, Baron (2010). “A Defense of Stable Invariantism,” Noûs, 44: 224–44.
Reynolds, Steven L. (2002). “Testimony, Knowledge, and Epistemic Goals,” Philosophical Studies, 110: 139–61.
Stanley, Jason (2005). Knowledge and Practical Interests. Oxford: Oxford University Press.
Unger, Peter (1975). Ignorance: A Case for Scepticism. Oxford: Oxford University Press.
Weiner, Matthew (2005) “Must we Know what we Say?” Philosophical Review, 114: 227–51.
Williamson, Timothy (1996) “Knowing and Asserting,” Philosophical Review, 105: 489–523.
—— (2000). Knowledge and its Limits. Oxford: Oxford University Press.
—— (2005). “Contextualism, Subject‐Sensitivity Invariantism, and Knowledge of Knowledge,” Philosophical Quarterly, 55: 213-35.
I am grateful to Jessica Brown, Jeremy Fantl, Sandy Goldberg, Allan Hazlett, Ofra Magidor, David Sosa, Jason Stanley, audience members at the 2008 Arché Assertion Conference at St Andrews, the Department of Philosophy at the University of Texas, Austin, the Place of Epistemic Agents Conference at the Universidad Carlos III de Madrid, Syracuse University, and, especially, Baron Reed for helpful comments on earlier drafts of this chapter.
(1) For different arguments supporting the KNA‐N, see Unger (1975), Brandom (1983, 1994), Williamson (1996, 2000), Adler (2002), Reynolds (2002), Hawthorne (2004), Stanley (2005), and Fricker (2006). Cohen (2004) says that he is “not unsympathetic” to the view. For various objections to the KNA‐N, see Weiner (2005), Douven (2006), Lackey (2007), and Kvanvig (2009).
(2) A notable exception is Brown (forthcoming a).
(4) The case of John Smith's testimony is, of course, merely a thought experiment.
(5) It should be noted that, in such cases, it would still be proper for the asserters in question to attribute knowledge to themselves. Thus, a contextualist response to such cases is not available here.
(6) This may be the rationale for disallowing hearsay in courtroom testimony.
(9) It may be the case that some classic examples of Gricean implicature can be read as concerning the epistemic credentials of the asserter. But, as I will argue at the end of this section, this response to my counter‐examples has problems of its own.
(10) Thus, Susan's assertion fails the Maxim of Relation of Grice's Cooperative Principle.
(11) For a detailed discussion of the relationship between the norms governing assertion and those governing practical rationality, see Brown (forthcoming b).
(13) This point is developed in more detail in my (Forthcoming).
(14) Thanks, again, go to Sandy Goldberg for pressing this response.
(15) This is not to say that there cannot be effective letters of recommendation written from, say, the perspective of an entire department, where some of the information conveyed is grounded entirely in isolated second‐hand knowledge. But such an institution is quite different from one whereby individual recommenders write letters on behalf of candidates, for here there is no presumption that the writer of the departmental letter is speaking from personal or direct acquaintance with the applicant.
(16) This objection was raised by several audience members at the 2008 Arché Assertion Conference at St Andrews.
(17) I am grateful to Jessica Brown and Jason Stanley for raising this point.
(19) It is worth noting that, on every existing theory of testimonial knowledge, a story can be told about how the asserters in my cases satisfy the relevant conditions for knowing the propositions in question.
(20) This may also be true of PROFESSOR, for it is not clear that an assertion involving a challenge to current US copyright law involves high practical stakes. If one wishes to argue, however, that such stakes are relevantly high, then, as I argue in my second point below, the details of the case can easily be tweaked without losing the intuition that the assertion is nonetheless improper.
(21) I shall develop this point in much more detail in the next section.
(22) The only possible exception is CHEATING, where high practical stakes are built directly into this type of case.
(23) I am grateful to Jason Stanley for this objection.
(24) I am grateful to comments made by Ben Bradley, Mark Heller, David Sosa, and Robert Van Gulick that led to the inclusion of this point.
(25) I am indebted to Jeremy Fantl for this response.