From objections to KNA to a case for a hearer-centered norm of assertion
DAS OBJEÇÕES À KNA À DEFESA DE UMA NORMA DE ASSERÇÃO CENTRADA NO OUVINTE
João B. Ferreira Filho
UFBA – Universidade Federal da Bahia
Recebido: 30/06/2025
Received: 30/06/2025
Aprovado: 30/09/2025
Approved: 30/09/2025
Publicado: 06/10/2025
Published: 06/10/2025
Abstract
An assertion is generally understood as a speech act in the form of a statement meant to impart something of epistemic significance: we assert to clarify, inform, teach, and improve a hearer’s epistemic status. Timothy Williamson (1996) claims that the nature of assertion is intrinsically normative: assertion has a constitutive norm. That norm, according to Williamson, is that the asserter ought to know the content of the assertion (the “knowledge norm of assertion” or “KNA”). In this paper, I examine (a) Williamson’s argument for a constitutive norm of assertion and (b) whether that norm must be knowledge. I will argue, particularly through Jennifer Lackey’s famous case of the creationist teacher, that an asserter can – and often does – consistently impart something of epistemic significance to a hearer without fully satisfying the KNA. I conclude with a recommendation that KNA, albeit an intuitive expectation about asserters, is not a necessary condition for assertion in epistemically successful assertoric practices.
Keywords: assertion; KNA; Timothy Williamson; Jennifer Lackey; testimony.
Resumo
Uma asserção é geralmente entendida como um ato de fala na forma de uma sentença que intenta transmitir algo de significância epistêmica: nós asserimos para esclarecer, informar, ensinar e melhorar o estado epistêmico de um ouvinte. Timothy Williamson (1996) afirma que a natureza da asserção é intrinsicamente normativa: asserções têm uma norma constitutiva. Esta norma, de acordo com Williamson, é que o asseridor deve saber o conteúdo da asserção (a “norma de conhecimento da asserção” ou “KNA”). Neste artigo, eu examino (a) o argumento de Williamson para uma norma constitutiva da asserção e (b) se tal norma deve ser o conhecimento. Argumentarei, em particular através do famoso caso de Jennifer Lackey da professora criacionista, que um asseridor pode – e frequentemente o faz – transmitir algo de significância epistêmica a um ouvinte consistentemente sem satisfazer plenamente a KNA. Concluo com uma recomendação de que a KNA, apesar de ser uma expectativa intuitiva sobre asseridores, não é condição necessária da asserção em práticas assertóricas epistemicamente bem-sucedidas.
Palavras-chave: asserção; KNA; Timothy Williamson; Jennifer Lackey; testemunho.
1. Introduction
Asserting is a fundamental practice in our social lives. Without assertion, knowing and navigating the world would be much harder tasks. Assertion, broadly speaking, is a speech act in the form of a statement meant to impart something of epistemic significance to a hearer. We assert to clarify, to inform, to teach, to improve a hearer’s epistemic status. Thus, assertion is of special concern for social epistemologists: since assertion is so important, there must be requirements to make assertion proper, or right, or truthful, or whatever epistemic good we should expect from assertions.
Assertion differs from other speech acts for it is propositional: asserters do not inquire, command, or express a desire; they state that it is the case that [p]. As a speech act, one concern regarding assertion is what an assertion expresses from its asserter, that is, what kind of act is performed. When asserting that [p], is the asserter expressing their belief that [p]? That has been a common take on what assertion is as a speech act. Thus, in those accounts, some descriptive and others normative in nature, belief is a necessary condition of assertion. Now, whether or not assertion is tied to belief will be examined later.
Also, if the asserter states that it is the case that [p], should it be the case that [p]? In other words, is the truth of [p] another necessary condition of assertion? In the truth account, since we expect assertions to have a positive epistemic status for the hearer, being truthful should be a normative condition for assertion. Even though asserters often fail to satisfy the truth condition – we could state falsehoods by mistake or plain deceitful intentions –, this violation of the rule would not threaten the normativity of the truth condition. The assumption here is that our common expectation about the content of assertions should inform us of its regulating norm.
Timothy Williamson (1996) claims that the nature of assertion is intrinsically normative: it has a constitutive norm. By addressing the truth account of assertion, he claims that it is not sufficient to explain assertion. The governing rule of assertion, according to Williamson, is that the asserter should know the content of the assertion. This account is known in the literature as the “Knowledge Norm of Assertion” or “KNA”. That is to say that an assertion [p] is warranted when its asserter satisfies the necessary conditions of knowledge that it is the case that [p].
In section 2, I try to reconstruct Williamson’s arguments that (a) assertion has a constitutive norm and that (b) the rule that governs the norm of assertion is knowledge, with brief commentary; I examine his replies to the main criticisms to KNA and whether KNA can successfully solve the concerns raised by its critics the way Williamson claims it does. In section 3, I offer Jennifer Lackey’s (2008) paradigmatic case of the creationist teacher as a case against KNA and expand on how asserters can consistently state propositions that improve their hearer’s epistemic status without satisfying the belief condition implied in the KNA. In section 4, I argue that KNA mistakes demands for asserter with demands for assertion, and that a constitutive norm for assertion should be centered on the epistemic needs of the hearer.
2. The Knowledge Norm of Assertion
In Knowing and Asserting (1996), Timothy Williamson argues that assertions, much like games, are governed by a constitutive rule. To clarify the notion of constitutive rule, Williamson makes a distinction between those and conventions:
Constitutive rules are not conventions. If it is a convention that one must φ, then it is contingent that one must φ; conventions are arbitrary and can be replaced by alternative conventions. In contrast, if it is a constitutive rule that one must φ, then it is necessary that one must φ. More precisely, a rule will count as constitutive of an act only if it is essential to that act: necessarily, the rule governs every performance of the act (Williamson, 1996, p. 490).
Simply put, constitutive rules differ from conventions in that the conventions that regulate a practice can be altered and replaced without mischaracterizing the practice. Constitutive rules, on the other hand, are constitutive because they make up the practice itself, they are the backbone of an activity and without which the activity itself cannot be.
Expanding on the analogy with games, one of the few conventions in chess is that king pieces touch at the center of the board to signal the outcome of a chess game that took place on that board (with different positioning of the kings meaning one of three different outcomes): failing to do so do not affect the game itself. However, it is a constitutive rule that kings must not touch each other during the game: failing to follow this rule would break the logic of the game itself (because since kings are not capturable, they cannot move themselves into check, which would be the case for a king attack to the enemy king).
Now, Williamson tells us that assertions are also governed by a rule that is constitutive, without which assertion cannot be. He formulates it as follows:
The C(P) rule One must: assert that P only if C(P).
According to Williamson, the constitutive rule of assertion is that one must assert that [p] only if conditions regarding [p] – that is, C(P) – are met. Now, rules in games, roughly speaking, are meant to set conditions of fairness for achieving a pre-determined goal. Likewise, the goal of asserting, that is, its purpose or function, should inform us about its constitutive rule. Speech acts with different functions should have different constitutive rules.
Since asserting is a propositional speech act, its constitutive rule is tied to conditions regarding the relevant proposition; and since assertion is functional, its constitutive rule should be a function of assertion regarding the relevant proposition. However, Williamson claims that the nature of the constitutive rule of assertion is not teleological (Williamson, 1996, p. 491), neither does he give us a more detailed account of what the constitutive rule of assertion is grounded on than the following:
Nevertheless, some sensitivity to the difference – both in oneself and in others – between conforming to the rule and breaking it presumably is a necessary condition of playing the game, speaking the language, or performing the speech act. The important task of elucidating the nature of this sensitivity will not be undertaken here (Williamson, 1996, p. 491).
It is clear from the excerpt above that Williamson acknowledges that expectations about assertions by the actors involved in the assertoric practice inform us of its constitutive rule, for the “sensitivity” he mentions is but an expression of expectations of the actors involved. What is not clear is why Williamson thinks this is not teleological. If by “teleological” we mean that the ends of an act are the foundational regulatory condition for the success of such an act, it is reasonable to think that in order to understand the nature of the constitutive rule of assertion we must look into what assertions aim to do. The role of expectations about assertions seems paramount for the norm of assertion and will be addressed throughout this paper.
There are candidate explanations of what the conditions entailed in C(P) must be. For now, let’s focus on Williamson’s proposal. Williamson is committed to the warrant rule of assertion, which roughly states that one must assert that [p] only if one has warrant that [p]. He goes on further to claim that what grants warrant to an assertion is knowledge, that is, that the condition that must be satisfied in C(P) for a warranted assertion is that the asserter knows that [p]. He formulates the knowledge rule as follows:
The knowledge rule One must: assert that P only if one knows that P.
Now, the knowledge rule – or knowledge norm of assertion (or KNA), as it has come to be known – entails that, for an assertion to be warranted, the asserter must satisfy the necessary conditions of knowledge, that is, a justified true belief that [p], and their assertion must be in conformity with their epistemic state regarding [p]. It follows that failing to satisfy any of the necessary conditions of knowledge would make an assertion unwarranted. However, Williamson is aware that the KNA is often violated and unwarranted assertions are common. He offers three main arguments for the KNA, as summarized by Goldberg (2015), all of which rely on KNA’s explanatory power:
First, he argued that KNA appears to best explain the Moore-paradoxicality of assertions of sentences of the form “p, but I do not know that p”; second, that KNA appears to be better positioned than any epistemically weaker standard to explain why lottery propositions – propositions representing one’s ticket as a loser – are not properly assertable even when the odds of one’s winning are arbitrarily small; and third, that KNA appears to best explain why “How do you know?” is a proper response to an assertion (even when the assertion’s explicit content has nothing to do with the speaker’s knowledge) (Goldberg, 2015, p. 367).
Let us now dissect Williamson’s three arguments for KNA. First, Williamson examines the wrongness in his version of Moore’s paradox, phrased as “[p], but I do not know that [p]”. He argues that this wrongness can only be explained in virtue of knowledge being a necessary condition of assertion – that is, if one does not know that [p], one is not warranted to assert that [p]. Williamson indicates that the paradoxicality in Moore’s statement lies in a contradiction by suggesting that in order to assert that “[p], but I do not know that [p]”, one must know (a) that one does not know that [p] and (b) that it is the case that [p]. Thus, in order for one to know that “[p], but I do not know that [p]” one would have to know that [p]. It follows that such a contradiction is present because Moore’s paradox violates the knowledge norm of assertion.
Originally, Moore’s paradox takes the form of “[p], but I do not believe that [p]”. Belief is but one necessary condition of propositional knowledge. Equating knowledge and belief would be a fallacy of composition. There are, however, alternative explanations of the norm of assertion that would make Moore’s paradox (as originally phrased) a warranted assertion, explanations that do not take belief to be a necessary condition for properly asserting that [p].[1] Consequently, those accounts could reveal that there is in fact no contradiction in Moore’s paradox; the apparent wrongness claimed by Williamson could be explained as simply an expression of psychological expectations regarding assertions. What we could infer from Williamson’s argumentative tactic is that he advanced a conveniently favorable mischaracterization of Moore’s paradox, that is, a mischaracterization that creates a problem for which KNA is the best solution.
The second argument is raised by Williamson as an objection to the truth account of the norm of assertion, and in favor of KNA. The truth account, as the name suggests, holds that the truth of [p] is sufficient to warrant an assertion that [p]. Williamson presents a lottery case to illustrate how a true assertion would not be warranted by itself if evidence were insufficient. The case goes as follows: consider H has one lottery ticket in a lottery of millions of tickets. There is only one winning ticket, which has already been drawn but not announced. Neither S nor H has privileged information regarding H’s ticket or any other ticket: at t1, the winning ticket has already been determined but neither S nor H knows which one it is. Nonetheless, based purely on the long odds of it being H’s ticket against millions of other tickets, S asserts at t1 to H that his ticket did not win. S does not assert that H’s ticket probably did not win, but simply that it did not win. And it happens to be the case that, at t1, H’s ticket was not the one drawn.
S’s assertion that H’s ticket did not win has been true since t1. However, Williamson argues that S never had sufficient evidence to assert that H’s ticket did not win (even though it was true): on probabilistic grounds short of 1, argues Williamson, only probabilistic assertions are warranted. By not fulfilling the necessary conditions to know that H’s ticket had not won, S presented himself as having an evidential authority to assert (that H’s ticket did not win) that he did not actually have. In fact, both S and H were symmetrical regarding the evidence on H’s ticket winning or losing. Thus, S asserts a true proposition for which both of them are equally lacking in evidence: he presents an asymmetry that is not there.[2]
Lastly, Williamson argues that KNA is best suited to explain this common response to an assertion: “How do you know?”. Though it is not always the case that the hearer believes the asserter to satisfy the knowledge condition, the mere form of the question reveals the nature of assertion’s warrantability: it is a challenge to the asserter as to whether his assertion is warranted. Williamson’s argument is that since that question is often an appropriate response to an assertion, it can only be so because there is an expectation from the hearer that the asserter should know the content of their assertion.
However, drawing from common linguistic habits as evidence for KNA – or any other hypothesis, for that matter – seems dubious. In fact, that is the position of Kvanvig (2009), as noted by Goldberg (2015): Kvanvig argues that “Are you certain?” is an equally proper response to an assertion.[3] Likewise, one could say that there is a variety of common and proper responses to assertion, some of which would, following Williamson’s reasoning, point to different norms of assertion: “Are you sure?”; “Is that true?”; “Do you believe that?”; “Can you prove it?”; and so on. Moreover, cultural differences at a global scale could present us with even more diverse examples of how hearers challenge an assertion and, consequently, of what those hearers consider to warrant an assertion. Thus, claiming that the question “How do you know?” is the standard challenge to an assertion – from which we can derive that knowledge is the norm of assertion – is misguided at best. I will not develop this objection further.
Let’s now turn, in the next section, to Jennifer Lackey’s case of the creationist teacher as a major objection to the assumption that belief is a necessary condition of assertion and, hence, as an objection to KNA as a condition for the warrantability of assertion altogether.
3. Objection from the “selfless assertion” account
There is a common assumption – which KNA is rooted in – that assertion is an expression of belief. KNA is a more qualified characterization of such assumption turned into a norm: assertion should be an expression of a belief that is true and justified. This, indeed, puts KNA in a good position to explain Moore’s paradox as Williamson intended. In this section, I will challenge the assumptions that (a) assertion is defined as an expression of belief and, consequently, that (b) there is a contradiction underlying Moore’s paradox. I will advance this objection from Jennifer Lackey’s “selfless assertion” account, famously exemplified by her case of the creationist teacher. It reads as follows:
CREATIONIST TEACHER: Stella is a devoutly Christian fourth-grade teacher, and her religious beliefs are grounded in a deep faith that she has had since she was a very young child. Part of this faith includes a belief in the truth of creationism and, accordingly, a belief in the falsity of evolutionary theory. Despite this, she fully recognizes that there is an overwhelming amount of scientific evidence against both of these beliefs. Indeed, she readily admits that she is not basing her own commitment to creationism on evidence at all but, rather, on the personal faith that she has in an all-powerful Creator. Because of this, Stella does not think that religion is something that she should impose on those around her, and this is especially true with respect to her fourth-grade students. Instead, she regards her duty as a teacher to involve presenting material that is best supported by the available evidence, which clearly includes the truth of evolutionary theory. As a result, after consulting reliable sources in the library and developing reliable lecture notes, Stella asserts to her students, ‘‘Modern-day Homo sapiens evolved from Homo erectus,’’ while presenting her biology lesson today. Though Stella herself neither believes nor knows this proposition, she never shares her own personal faith-based views with her students, and so they form the corresponding true belief solely on the basis of her reliable testimony (Lackey, 2008, p. 48).
Let us break down Stella’s epistemic status. Stella does not believe that [p] (in her case, that “Modern-day Homo sapiens evolved from Homo erectus”). In fact, she believes that ¬[p]. Stella is aware that there is an overwhelming amount of evidence in favor of [p] and also that believing [p] is a consensus among the epistemic authorities within the relevant community (researchers of the natural history of humans). Stella is also aware that her belief that ¬[p] is not evidentially sound, that it is rather a matter of spiritual preference. Thus, she does not know that [p] since she does not satisfy the belief condition that [p], and she is in an epistemically inferior position regarding the evidence that [p]. Yet, Stella proceeds in asserting to her students that it is the case that [p].
Stella’s assertion is not in conformity with her belief. In fact, for the purpose of her lecture, her belief – that ¬[p] – has no effect. According to KNA, Stella’s assertion on the evolution of humans is not warranted (neither would it be asserting that ¬[p]). However, Lackey is not convinced that Stella fails as an asserter. Lackey’s claim is that
[…] an unreliable believer may nonetheless be a reliable testifier, and so may reliably convey knowledge (justified/warranted belief) to a hearer despite the fact that she fails to possess it herself (Lackey, 2008, p. 49).
Therefore, even though Stella fails as an epistemic agent for disregarding the influence of the relevant defeaters to her doxastic state, she still asserts consistently in an epistemically successful way to her students, that is, she offers reliable testimony that [p]. It should be clear now how “the creationist teacher” fares against (a) and (b): Stella does not assert from belief and yet her speech act is not contradictory.
Now, the form of the “selfless assertion” – as it is called the cases in which the asserter does not believe the proposition being asserted – can express many instances of the case. Moreover, the “selfless assertion” account is not a claim to the norm of assertion, but rather a challenge to the hypothesis that knowledge is: it remarks that sometimes it is epistemically preferable to assert from something other than belief, thus making a norm of assertion that does not require belief a more reasonable candidate. However, Turri (2015) presents us with a few empirically-informed cases to object to the selfless assertion as a threat to KNA, which I will summarize next.[4]
The premise of Turri’s experiments is to use laypeople’s judgment of the epistemic status of a selfless asserter as evidence for KNA: that is, whether they think that the selfless asserter either (1) believes or (2) knows what he is asserting and whether they think that the asserter is warranted to assert that [p]. Participants were divided into two groups: the “Believes” and “Knows” groups. They were then presented with a story that takes the form of Stella’s, edited to replace any terms that could skew the participants’ judgment (such as “belief” and “know”) with more ordinary ones (such as “conviction” and “doubt”). After reading the story, the participants in the “Believes” group were asked to rate the degree to which they agreed with the statement of the form “at some level, the asserter believes that [p]” – while the participants in the “Knows” group were asked to rate the degree to which they agreed with the statement “at some level, the asserter knows that [p]”. Lastly, both groups were asked to rate whether they agreed to the statement of the form “the asserter should assert that [p]”.
Unsurprisingly, the results of the experiments revealed that the overwhelming majority of participants in either group (1) deemed the assertion by the asserter warranted and (2) attributed positive doxastic states to the asserter – that is, that the asserter either believed or knew the assertion in question.[5]
Turri’s hypothesis was that there is a connection between laypeople’s expectations regarding asserters and laypeople’s judgment of the warrantability of assertions. The approach of the experiments was what he called a “natural interpretation” of the case of the selfless asserter. However, some reasonable concerns can be raised regarding the connection between the results of the “natural interpretation” approach and an analytic norm of assertion.
First, Turri’s conclusion, that is, that his experiments support KNA, is based on a specialized interpretation of epistemically charged terms: “believe”, “knows”, “recognize”, and so forth. It is fair to assume that should the participants have a qualified (epistemological) understanding of these concepts and their implications, the results could have been more controversial (or even outright contradicted the hypothesis). In other words, the language of the researcher might not be consistent with the participants’, which would not necessarily entail the conclusion of the experiments.[6]
Additionally, Turri’s experiments seem to have overlooked confirmation bias: they fail to test the alternative. Are the participants’ expectations that the asserter should believe/know that [p] or merely that the asserter should be a reliable testifier regarding [p]? There is no way to tell from the experiments. Again, the language employed seems to preemptively exclude the alternative. If Turri is assuming that expectations regarding assertion should inform us of its norm, it should have been made unequivocal what the participants’ expectations are.
Third and foremost, even though the overwhelming majority of participants perceive the asserter as (to some degree) believing/knowing what is being asserted, it is a fact of the story that the asserter does not. Therefore, the participants’ belief regarding the asserter’s mental state is false, and yet it warranted the asserter’s speech act. That a false belief regarding an asserter epistemic status should warrant assertion seems wholly controversial.
Psychometric experimentation as evidence in philosophical arguments is famously a risky approach. There are, however, two important aspects to Turri’s approach: first, and in agreement with the “selfless assertion” account itself, it acknowledges that a norm of assertion should be derived from the hearer’s perspective which, in turn, goes against KNA’s asserter-centered account; second, it indicates a disconnection between how laypeople rationalize their assertoric expectations and reliable testimony itself.[7]
Circling back to Lackey’s cases of selfless assertion, there is a fundamentally important aspect in the stories whose effects seem to have been conveniently ignored by Turri: the asserter acts “out of duty”. As Green (2006) explains, “On this sort of view, T [testifier] may assert that [p] if T has enough certainty for his audience’s needs, but which might not be enough for T’s own” (Green, 2006, p. 142).
In many dimensions of social life, a sense of commitment to professional duty outweighs asserting from belief. It would be ideal that, in such contexts, one’s own doxastic states were in conformity with one’s assertion in a professional capacity, but when it is not, one’s own beliefs should often be ignored. It is preferable that a scientist report (assert) her findings as they are even if they contradict her beliefs; that a journalist report (assert) the facts of public interest regardless of her opinions on the matter; that a science teacher teach (assert) about the evolutionary origins of man based on scientific consensus even if she does not agree with it herself.
It seems that whenever reliability is an epistemic desideratum, conformity between assertion and belief becomes an epistemic liability. If it is our expectations about assertions that inform us of its constitutive norm, there are dimensions of social life in which we expect the asserter to be a reliable testifier, which would not necessarily entail asserting from belief – which, in turn, falls short of propositional knowledge. The selfless asserter, then, does not act for herself, but in the hearer’s best interest.
This fact of social life – that commitment to professional duty supersedes one’s own doxastic states – should point us towards a paradigm shift in the conception of a constitutive norm of assertion. In the next section, I will briefly expand on this shift: from an asserter-centered norm of assertion, chiefly advanced by KNA, to a hearer-centered norm of assertion.
4. From norms for asserter to norms of assertion
At this point, what the selfless assertion account provides us with that KNA does not should be evident: it not only makes many of our epistemically successful assertoric practices that fall short of knowledge warranted, but it also makes it desirable to be so. By “epistemically successful” I mean that the hearer’s epistemic status is improved by the assertion: they become better informed, their beliefs come closer to the truth, they amass better evidence, and so on. Thus, if assertion is a speech act that aims to improve the hearer’s epistemic status, a constitutive norm of assertion should have the hearer’s epistemic needs in mind.
If KNA takes the hearer’s epistemic needs as a foundation for the constitutive norm of assertion, then its approach is misguided: it mistakes demands on asserter for demands on assertion. Ironically, that is a criticism raised by Williamson himself while criticizing the “believing that one knows” account: “Defenders of the BK account cannot deny that we distinguish faults in the assertion from faults in the asserter” (Williamson, 1996, p. 513). The selfless asserter account makes such a distinction and, unlike KNA, demonstrates that faults in the asserter do not entail defective assertions.
A potential solution to the issue would be discarding our expectations about assertions as a foundation for a norm of assertion. That would, in principle, make it possible to sustain KNA on a different basis. However, that would still require demands on the asserter to satisfy conditions of knowledge. Assertions by public asserters (such as journalists and teachers) would only be warranted if the latter had a justified true belief that [p] and they asserted correspondingly, as Williamson intended it to be. It would also beg the question as to what a more adequate foundation for a norm of assertion should be.
One might inquire as to what alternatives to KNA would look like. There are a few candidate norms of assertion found in the literature. I would like to propose an operational, general norm drawn from my conclusions so far, then compare it with two of those candidate norms. It reads as follows:
General Norm One must: assert that [p] only if one is in a position to reliably assert that [p] and [p] is appropriate to the target hearer’s epistemic needs.[8]
This operational GN has two conditions that should conform with the selfless assertion account: (a) the assertion is reliably grounded on relevant evidence and (b) the assertion must meet the epistemic needs of the target hearer. These two conditions are meant to ensure that proper assertions are neither accidental nor irrelevant or redundant. It is my understanding that most theories of the norm of assertion have, to some extent, been attempts at satisfying (a).[9] The “appropriateness” in (b) is intentionally left sufficiently vague: as we will see below, those aforementioned theories fail at satisfying (b) in different ways. In addition, this GN does not require a positive doxastic state from the asserter.
Contra KNA, Lackey (2007) proposes the “reasonable to believe norm of assertion” (RTBNA):
RTBNA One should assert that p if (i) it is reasonable for one to believe that p, and (ii) if one asserted that p, one would assert that p at least in part because it is reasonable for one to believe that p.
RTBNA emphasizes the role of evidence (or reasonability) in warranting assertion: whether or not one believes that [p], asserting that [p] should be result of good “epistemic support” that [p].[10] Thus, it satisfies (a). Also, RTBNA explains the propriety of cases of selfless assertion while also capturing cases in which the asserter reasonably believes what is asserted. However, RTBNA does not provide any requirement to ensure that the target hearer will be benefited. As Gaszczyk (2024) puts it, RTBNA is too speaker-centered, which leads to undesirable consequences.
For instance, RTBNA overlooks the fact that assertions should be relevant for the target hearer: it might be reasonable to believe that [p] and one might assert accordingly to a target hearer to whom knowing whether [p] is epistemically insignificant. Additionally, one might even reasonably believe that [p] and assert accordingly to a target hearer who already knows that [p], thus making asserting that [p] redundant. RTBNA fails (b) for it makes irrelevant and/or redundant assertions warranted/proper. It is curious that Lackey neglects the fact that in the cases of selfless assertion, relevance and informativeness are underlying requirements in effect for the asserter to act in the best interest of the target hearer. On the topic of informativeness, we turn to INA by Gaszczyk (2024) next.
Gaszczyk (2024) advocates for a hearer-centered norm of assertion. He proposes “the informativeness norm of assertion” (INA): according to this norm, an assertion is warranted/proper if the content of the assertion is not “common knowledge” – that is, if the target hearer does not already know that [p]. One of the undesirable consequences of KNA – as well as RTBNA – is that the asserter is warranted to assert that [p] even if the target hearer already knows that [p], an instance in which asserting that [p] would be an epistemically redundant act.[11] INA remediates that flaw. Moreover, under INA, as long as the assertion is informative for the target hearer, a positive doxastic state from the asserter regarding the content of the assertion is not required. Thus, INA fits well within the selfless assertion account for it takes the epistemic needs of the hearer – namely, the need to be informed – as its guiding rule while avoiding committing the speaker to a positive doxastic state regarding the assertion, maintaining the assertoric practices epistemically meaningful.[12]
Now, Gaszczyk does not consider INA and KNA to be competing theories: in fact, he considers INA to be a complement to KNA, both being constitutive norms. Also, he acknowledges that, analogous to KNA, being constitutive does not keep INA from being violated: one can assert without being informative, it is just not proper/warranted. Gaszczyk does not argue to reject KNA, but rather to supplement it with INA. Redundant assertions (assertions that are already known by the target hearer) are proper/warranted under KNA, a flaw that, according to Gaszczyk, INA remedies by imposing that assertions ought to be informative. I maintain that (1) as long as INA is satisfied in assertoric practices, KNA is dispensable and (2) commitment to KNA may damage informativeness given the selfless assertion paradigm.
Just like RTBNA, one property for a norm of assertion that is not addressed by INA (at least not explicitly) is relevance. Though one informatively asserts that [p] to a target hearer that does not know that [p], learning that it is the case that [p] may not be within the target hearer’s epistemic needs. For example, asserting that “Tarawa is the capital of Kiribati”, albeit informative to anyone who does not know the capital of Kiribati in advance, is of no epistemic relevance to, say, someone struggling with an exam on chemical bonding. Informativeness without relevance is not sufficient to ensure that assertion meets a target hearer’s epistemic needs: INA also fails (b).
Additionally, one can inform regarding [p] without asserting that [p]. That is, consider the following: a target hearer that seeks to know whether [p], an asserter who cannot inform whether [p] but can reliably inform that [q], from which the hearer, knowing that [p] entails [q], can infer that [p]. One can hardly object that asserting that [q] in this scenario is not informative regarding [p]. Our hypothetical asserter is a reliable source regarding [p], even though they do not address [p] directly. This sort of informativeness is not captured by INA – which posits only the informativeness of the proposition in question in an assertion –, but it seems relevant for the informativeness of assertions in general whenever they have a target hearer’s epistemic needs as their core. As INA stands, asserting that [q] is proper/warranted if our target hearer is seeking to know whether [q]: if not, asserting that [q] is not informative, thus improper/unwarranted.
Both alternatives to KNA that I have considered – RTBNA and INA – fail one or more elements from the provisional GN that I presented. As previously stated, I do not mean this GN to be a candidate norm of assertion, but rather simply an operational framework, derived from the selfless assertion account, against which I compare those two candidate norms of assertion. What I propose is that a norm of assertion (a) should not mistake demands on asserter for demands on assertion and that it (b) should be hearer-centered. The selfless assertion account is a competent counter-example to KNA that highlights both issues and deserves a more careful examination. Unfortunately, the present work will not go further in depth in the selfless assertion account, for my main goal was simply to extract its challenges to KNA, then (briefly) examine a few alternatives. Whether or not the selfless assertion account alludes to a better norm of assertion is a task for a different discussion.
5. Conclusion
In this paper, I argued that, as a theory of the constitutive norm of assertion, KNA leaves out many of our epistemically successful assertoric practices. The selfless assertion account, on the other hand, exemplifies those practices and explains why they are successful and epistemically desirable even though they fail the knowledge condition. From those, I also argued that, if what motivates Timothy Williamson’s search for a constitutive norm of assertion is some underlying common expectation regarding what assertions should do, there must be alternatives that satisfy such a criterion without leaving out cases of selfless assertion. These candidate theories should be hearer-centered since it is the hearer’s epistemic status that is at stake in assertoric practices. I examined two of such candidate theories and discovered that they fail defining properties at play in cases of selfless assertion. The project of designing a candidate theory was not carried out here, but I hope the objections raised here are compelling and informative for such a project. As for whether or not the claims made in this paper are warranted: that may not be a matter of knowledge.[13]
References
FERREIRA FILHO, João B. Outlines for a metatheory of expertise from the veritistic approach and its main problems. Peri, Florianópolis, v. 16, n. 2, p. 75-88, 2024.
GASZCZYK, Grzegorz. The informativeness norm of assertion. Review of Philosophy and Psychology, v. 15, p. 1157-1178, 2024.
GASZCZYK, Grzegorz. A critique of Turri’s experimental research on selfless assertions. Diametros, v. 16, n. 59, p. 23-34, 2019.
GOLDBERG, Sanford C. Recent work on assertion. American Philosophical Quarterly, v. 52, n. 4, p. 365-380, 2015.
GOLDMAN, Alvin I. A causal theory of knowing. The Journal of Philosophy, v. 64, n. 12, p. 257-275, 1967.
GRECO, John. What is transmission? Episteme, v. 13, n. 4, p. 481-498, 2016.
GREEN, Christopher R. The epistemic parity of testimony, Memory, and Perception. Ph.D. dissertation, University of Notre Dame. 2006.
KVANVIG, Jonathan. Assertion, knowledge, and lotteries. In: GREENOUGH, P.; PRITCHARD, D. (Eds.). Williamson on Knowledge. Oxford: Oxford University Press, 2009, p. 140-160.
LACKEY, Jennifer. Learning from words: testimony as a source of knowledge. New York: Oxford University Press, 2008.
LACKEY, Jennifer. Norms of assertion. Noûs, v. 41, n. 4, p. 594-626, 2007.
PLANTINGA, Alvin. Epistemic justification. Noûs, v. 20, n. 1, p. 3-18, 1986.
SORENSEN, Roy. Epistemic paradoxes. The Stanford Encyclopedia of Philosophy (Fall 2024 Edition), Edward N. Zalta & Uri Nodelman (eds.). https://plato.stanford.edu/archives/fall2024/entries/epistemic-paradoxes/ .
TURRI, John. Selfless assertions: some empirical evidence. Synthese, v. 192, p. 1221-1233, 2015.
WEINER, Matthew. Must we know what we say? The Philosophical Review, v. 114, p. 227-51, 2005.
WILLIAMSON, Timothy. Knowing and asserting. The Philosophical Review, v. 105, n. 4, p. 489-523, 1996.
João B. Ferreira Filho
Doctorate student in Philosophy at the Federal University of Bahia (UFBA), with a master’s degree in Philosophy and a bachelor’s degree in History from the Federal University of Mato Grosso (UFMT), the author’s main research interests are Analytic Epistemology, in general, and (Analytic) Philosophy of Science, in particular.
The texts in this article were reviewed by third parties and submitted for validation by the author(s) before publication