The Informativeness Norm of Assertion

Abstract

Although assertions are often characterised as essentially informative speech acts, there is a widespread disagreement concerning how the informativeness of assertions should be understood. This paper proposes the informativeness norm of assertion, which posits that assertions are speech acts that essentially deliver new information. As a result, if one asserts something that is already commonly known, one’s assertion is improper. The norm is motivated by appealing to unique conversational patterns associated with informative and uninformative uses of assertions, an analogy between assertions and inquiries, and a distinction between assertions and uninformative speech acts. By focusing on the normative approach to speech acts, the paper discusses how particular norms of assertion deal with the data supporting the norm of informativeness. To be informative, the speaker must consider the epistemic position of the audience. Since the majority of norms proposed in the literature are speaker-centred, they fail to explain the submitted data. Looking more broadly, focusing on the informativeness of assertions underscores the crucial role of the audience in construing adequate speech act accounts.

1 Introduction

What assertions are for? Particular theories of assertion emphasise different aspects of assertoric practice and thus deliver various answers to this question. The most common theories locate what is central in assertion in (i) expressing a particular mental state (like belief or knowledge), (ii) making a move in a language game (like following one of the norms of assertion), (iii) producing something in the audience (e.g., knowledge), (iv) undertaking a commitment (say, by asserting that p, the speaker is committing to the truth of p), or (v) proposing to add the proposition to the common ground.¹ Of course, many of these answers, at least on a general level, go together and complement each other. For instance, by asserting a proposition that the speaker believes is true, she proposes to add this proposition to the common ground and commits to its truth. Sometimes, however, these answers are in tension. For instance, the speaker can produce knowledge in the audience by asserting a proposition that she does not believe, or the speaker can assert a proposition that is already commonly known and so it cannot produce knowledge in the audience.

However, there is one answer to the question “What assertions are for?” that seems to be shared universally, namely, assertions are for spreading or transmitting information. In other words, assertions can be characterised as informative speech acts. Let us call this feature of assertoric practice the informativeness of assertions. This view is assumed by all the above theories of assertion, e.g., the speaker needs to transmit information to express a belief, or to produce knowledge in the audience. It seems, therefore, that there is an inextricable link between assertions and transmission of information.

The informativeness of assertions is both an important and timely topic for several reasons. First of all, assertions are central speech acts to our communication. Since the informativeness of assertions may be something that lies at the core of our assertoric practice, it is crucial to establish what it amounts to. It may, for instance, turn out that it is something that can help us in discriminating assertions from other speech acts. Furthermore, assertions play unique functions—they are the paradigmatic (or exclusive) vehicles for testimony (Goldberg 2015), and lying (Stainton 2016), to mention just two. Thus, it is essential to establish what role we should ascribe to the informativeness of assertions since it may be something that illuminates our thinking about other concepts, such as testimony and lying, that are defined in terms of assertion.

Although the claim that assertions are informative speech acts is widely accepted, there is a deep disagreement concerning how it should be understood. Specifically, there is the question of the centrality of informativeness for the assertoric practice. On the one hand, informativeness can be treated as a core feature of assertions. In one reading of this view, a proper assertion is such that delivers new information. Such a view can be called the Strong Informativeness Hypothesis (STRONG). For instance, in this view, my assertion that I had tacos for breakfast is proper only if I utter it to someone who does not know that, but improper when I utter it to someone who already knows it. Since an assertion that fails to deliver new information is improper, by STRONG, the propriety of assertion is linked to the epistemic position of the audience. On the other hand, informativeness can be thought of in more relaxed terms and thus as an optional feature of assertion. This view can be called the Weak Informativeness Hypothesis (WEAK). According to WEAK, my assertion can be proper even if it does not deliver new information. Thus, the propriety of assertion is independent of the epistemic position of the audience.

The aim of this paper is to propose a normative account of the informativeness of assertions along the lines of STRONG. In the centre of my proposal lies the norm of informativeness.² In this paper, I focus on the informativeness of assertions on the level of their content—this is the most reliable and the main source of information that comes from assertions. Thus, for instance, my assertion “The capital of Barbados is Bridgetown” is informative only if my audience does not know the content of this statement.³

The plan is as follows. Section 2 introduces the theoretical framework of the paper, i.e., the normative account of assertion, and shows how particular instalments of this account deal with the informativeness of assertions. Section 3 defends the informativeness norm of assertion and motivates it by appealing to relevant linguistic data. Section 4 addresses some of the potential counterarguments to my proposal. Section 5 discusses the consequences of the presented view. Particularly, I show that only a specific and rarely discussed type of normative account—one that is directed towards the audience—can naturally accommodate the linguistic data that support the norm of informativeness. Section 6 concludes the paper.

2 Informativeness and the Normative Account of Assertion

Particular theories of assertion position themselves differently towards the informativeness of assertions. I discuss this issue focusing on the normative account of assertion, one of the dominant accounts nowadays.⁴ It has been recently popularised by Williamson’s (2000) approach to assertion, according to which the speech act of assertion is governed by the norm constitutive for its performance. The norm is applicable only to a specific speech act type, as Williamson puts it, “flat-out assertions” (2000, p. 246), or our “default use” of the declarative mood (2000, p. 258). Williamson proposes that assertions are governed by the knowledge norm:

KNA One must: assert p only if one knows p.

Many other norms, which have been proposed, follow the same formula. Consider, for instance, the truth norm (Weiner 2005), and the reasonableness norm (Lackey 2007):

TNA One must: assert p only if p.

RBNA One must: assert p only if it is reasonable for one to believe that p.⁵

It is worth clarifying what the claim about the constitutiveness of the norm of assertion amounts to.⁶ First of all, in the Williamsonian framework, the norm (whatever it may be) is not intended to give the correctness condition for assertion; rather, it characterises what is essential for assertion. As García-Carpintero (2020, p. 271) puts it, “assertion is an act essentially constituted by its being beholden to the relevant norm.” The norm is unique, i.e., it is the only constitutive norm that governs assertion. Even though Williamson (2000, p. 240) maintains that constitutive norms “do not lay down necessary conditions for performing the constituted act,” they are individuating, i.e., assertion is the only speech act governed by this norm, which allows us to differentiate assertions from other speech acts, especially from other speech acts made in the declarative mood.

One of the central features of the Williamsonian view is that the constitutive norms can be broken without ceasing to perform the speech act (Williamson 2000, p. 240). In Austinian terms, a violation of the norm amounts to abuse, not a misfire (Austin 1962, pp. 167-8). Consider an analogy with games. Just as one can cheat without ceasing to play a game, one can break the norm of assertion (for instance, by saying something false) without ceasing to make an assertion. Finally, the norms are defeasible—they can be overridden or overruled by obligations imposed by other, say, moral or prudential norms; for instance, by saying something false to save someone’s life.

What is common for all the above norms is that they impose conditions only for the speaker, i.e., characterise what is essential for an assertion in terms of the speaker’s obligations. For instance, following KNA, a proper assertion is such that the speaker knows its content. Let us call such norms speaker-centred. Looking at the informativeness of assertions, the speaker-centred norms are consistent with WEAK since, following these norms, nothing prevents the speakers from making uninformative assertions—there is nothing improper in asserting that p, and later on, reasserting p.

Consider how such norms classify reminders.⁷ Typically, they are made in a performative way, by saying, for instance, “I remind you that p.” They are standardly theorised as utterances whose content was already known (and may still be known) by the audience. Reminders are interesting because they are utterances which may be uninformative (on the level of content), but which are still intuitively appropriate to make. Different speech act theories categorise them either as a kind of assertions or as distinct speech act types. Whether a particular norm of assertion favours WEAK or STRONG plays a significant role in classifying reminders either as a kind of assertion or distinct speech act type. The speaker-centred norms seem to classify reminders as cases of assertions. In fact, Weiner (2005, p. 239) lists reminders as subsuming to assertions.⁸

There are more types of norms that the normative account offers. Consider the audience-centred norms, which bring the audience into the characteristic of assertion. Here is one example of such a norm, i.e., the transfer of knowledge norm of assertion (García-Carpintero 2004, p. 134; cf. 2020, p. 270):

TKNA One must: assert p only if one’s audience comes thereby to be in a position to know p.

Although both KNA and TKNA are knowledge-based norms, i.e., for both, knowledge is the norm of assertion, they differ significantly from each other. Imagine that I assert that p to someone who already knows that p. Such an assertion satisfies KNA, but it violates TKNA—since my audience already knows that p, they cannot come to be in a position to know that p. Thus, by the TKNA standard, such an assertion is improper (more on this in the next section). For this reason, TKNA commits to STRONG. Consider again reminders, which García-Carpintero (2004, p. 156) lists as distinct speech acts from assertions. Thus, for instance, when I preface my utterance with “Let me remind you of the following” I am not performing an assertion. Since this is not a flat-out assertion, it is not subject to TKNA. Moreover, by definition, it does not transfer new knowledge to my audience.

The general picture that emerges is the following. Particular norms of assertion are implicitly or explicitly committing either to STRONG or WEAK.⁹ For an account to be consistent with STRONG, to a first approximation, we need a requirement such that the speaker is sensitive to the epistemic position of the audience. For instance, for TKNA it is not enough that the speaker knows that p—she must transfer this knowledge to the audience. Following this observation, in the next section, I argue for an informativeness norm of assertion along the lines of STRONG.

3 Normative Account of the Informativeness of Assertions

This section consists of two parts. In the first one, I propose the informativeness norm of assertion, and in the second one, I motivate my proposal by appealing to relevant data.

3.1 The Informativeness Norm of Assertion

I propose the following informativeness norm of assertion:

INA One must: assert that p only if p is not common knowledge.

I will argue that INA is in part constitutive of assertion. The recently dominant accounts of assertion either do not subscribe to INA or deny its correctness. For this reason, they give wrong predictions for what utterances should be counted as proper and improper assertions.

Before discussing the nature of INA, it is worth looking at its predecessors since some classical accounts of assertion have proposed a similar condition for assertion. Here are two examples. Searle (1969) proposes a set of rules that, taken together, deliver a full characterization of assertion. Apart from a constitutive rule whose satisfaction is necessary for counting a particular utterance as an assertion (differently than in the Williamsonian account where the constitutive rule can be violated), there are additional, preparatory rules that fulfilment results in a successful (or non-defective) assertion. Crucially, preparatory rules can be broken. One of them states the following:

INA-Searle “It is not obvious to both S and H that H knows (does not need to be reminded of, etc.) p.” (Searle 1969, p. 66).

For Searle, if I reassert that p knowing that p is already known by my audience (leaving the obscurity of what “being not obvious” exactly means here), my assertion is improper since I violate the preparatory rule of assertion. Thus, a proper assertion must take into account the epistemic position of the audience. Searle also mentions reminders as cases that are not subject to this rule—I will get back to them in the next subsection.

Consider Stalnaker’s (1974, 1978) account.¹⁰ According to Stalnaker, an assertion that p amounts to proposing to make it common ground that p. When my audience accepts p, its content is added to the common ground of the conversation. As a result, the truth of this proposition will be presupposed in our future interactions. Stalnaker proposes three rules (or principles) for assertion. Commenting on the first rule, he formulates something that can be considered as the informativeness condition for assertion:

INA-Stalnaker “to assert something which [is] already presupposed is to attempt to do something that is already done.” (Stalnaker 1978, p. 89).

In Stalnaker’s view, an uninformative assertion cannot update the common ground with p, since p is already part of the common ground.¹¹ Stalnaker explicitly says that failing to conform to the above rule can be classified as “unreasonable, inefficient, disorderly, or uncooperative” (1978, p. 89).

Although in distinct ways, both Searle and Stalnaker address the impropriety of asserting a proposition that already belongs to the common ground. Following these observations, I propose INA as in part constitutive of assertions. Above, I listed three features of the Williamsonian framework, i.e., the constitutive norms are supposed to be essential, unique, and individuating. In this section, I argue that taken on its own, INA satisfies only the first feature; nevertheless, it is enough for taking the informativeness of assertions as constitutive for the assertoric practice. In Sect. 5, I propose that INA should be theorised as an integral part of the constitutive norm of assertion.

The norm is essential for assertions not because all assertions conform to the norm, but because they are subject to it. It is the same with INA. An assertion is a speech act type essentially governed by being subject to INA. Thus, one who asserts something that is already commonly known violates INA but does not cease to make an assertion. Such assertions are judged as improper and, because of that, can be legitimately criticised, which I discuss in the next subsection. The situation is different for uniqueness and individuation. INA, taken on its own, is not unique since, apart from it, there are other constitutive norms of assertion, like KNA, or TNA. INA is also not individuating since more speech acts are informative in the way it proposes. Specifically, the informativeness of speech acts has been recognised in some speech act theories (Searle 1969; cf. Jary 2010). Such speech acts as guaranteeing, warning, or admitting, can be seen as informative, possibly in a similar way as assertions. The fact that INA is neither unique nor individuating, I argue, does not lower its significance.¹²

Since INA does not say what should be the outcome of assertion, it does not make a commitment regarding what the norm of assertion amounts to. Nevertheless, since knowledge-based norms of assertion are the most commonly used and discussed types of norms, formulating INA in terms of common knowledge gives a better perspective on how the informativeness of assertions fits into the overall picture of norms of assertion and assertoric normativity.¹³ Moreover, INA does not commit to any specific mode of how knowledge in assertion comes from the speaker to the audience, i.e., whether assertion transfers, expresses, or generates, knowledge in the audience.¹⁴ This is what the preferred account of assertion can specify. The informativeness norm ensures that this transfer (expression, or generation) of knowledge considers the epistemic position of the audience.

Looking more broadly, INA is the first example of the norm of assertion that focuses exclusively on the informativeness of assertions. This aspect of the assertoric practice has been systematically omitted in the debate and it is independent of the favoured norm of assertion. Thus, in principle, a norm analogical to INA can be formulated in terms of any norm of assertion. (Although, as I discuss in Sect. 5, it does not sit well with every norm.) For instance, the advocates of the belief norm of assertion (e.g. Bach 2008) could say something like “one must: assert that p only if p is not commonly believed,” kee** in mind that such a norm would make different predictions than INA. More generally, the informativeness norm of assertion could be formulated without committing to any specific norm of assertion, and say, roughly, “one must: assert that p only if p is new information to the audience.”¹⁵ Nevertheless, I take INA to sit better with the current debate, and, importantly, explain the data, to which I now turn.

3.2 Data

The crux of my argument lies in the following three linguistic observations, each of which supports INA in a distinct way.¹⁶

The first datum is based on the observation that there are unique conversational patterns associated with the informativeness of assertions. Specifically, we can legitimately criticise and challenge uninformative assertions (cf. García-Carpintero 2004, p. 159). There is a wide range of possible responses to such assertions. Specifically, they can vary in the strength of the speaker’s critique upon hearing something that they already know. The critique may be weak and just inform that the asserted information is already known; thus, one may respond with, for instance, “I already know that,” or “I already knew that, thanks.” But it may also be much stronger; thus, one may respond by saying “You know that I know that!,” “Why do you say that? I already know that!,” or “Why are you repeating it? You know that I know that!” The answer may depend on the context and on how blatant the violation of INA is. Additionally, there are specific hedges that indicate that our statements may be uninformative. Thus, for instance, we can say “I’m not sure you already know, but p,” or “Maybe you have heard it already, but p.” We preface our statements in this way to shield ourselves from the above critique and challenge. Notice that such hedged statements are not subject to the constitutive norm of assertion since these are not cases of flat-out assertions. Nevertheless, using such hedges shows that the speakers are aware of the norms and try to follow them.¹⁷

How conclusive is this datum? An examination of conversational patterns has become a widely accepted practice when it comes to analysing and individuating various speech acts. One of the most discussed patterns concerns how we criticise or challenge particular speech acts. Consider this sample of cases. Firstly, there is a unique way in which we criticise assertions qua the primary norm of assertion, like KNA. Specifically, assertions can be challenged by the “How do you know?” question.¹⁸ When I assert that p, you can legitimately challenge my assertion by asking me “How do you know?” Just as above, the strength of this challenge can be modified—one can issue a stronger challenge by saying, for instance, “You don’t know that!” Secondly, just as we can challenge speakers who do not know, we can challenge inquirers who know the answer to the asked question.¹⁹ Thus, when you ask me a question that you already know the answer to, I can legitimately criticise you, by saying, for instance, “You already know this!” Thirdly, we can also challenge presuppositions, specifically with the so-called “Hey, wait a minute!” test (von Fintel 2004). When I assert “Sam quit smoking,” by saying “Hey, wait a minute! I didn’t know that Sam used to smoke!” I directly challenge the presupposition that Sam used to smoke. The patterns associated with challenging uninformative assertions can be added to this list.²⁰

The second datum concerns an analogy between assertions and inquiries, where the latter can be seen as a reverse of the former. This analogy goes to the core of these concepts. Consider the constitutive norms of these speech acts. While knowledge is widely assumed to be the norm of assertion, lack of knowledge is the presumed norm of inquiry:

INQ One must: inquire whether p only if one does not know the answer to p.²¹

Just as assertions are default uses of declarative mood, inquiries are considered to be default uses of interrogative mood. Furthermore, just as KNA is supposed to individuate assertions from other speech acts, INQ does the same for inquiries.

Consider further that INQ, just as KNA, is the speaker-centred norm. However, there is also a recent proposal concerning the norm of inquiry resembling audience-centred norms. Haziza (2023, p. 1) proposes adding to INQ the following audience-centred norm (hence INQ-A):

INQ-A Ask addressee A the question Q only if A knows Q.

Haziza (2023, p. 6) emphasises that this norm, just as other audience-centred norms of assertion, applies to the speaker, “it is the speaker, not the addressee, who complies with or violates the norm. [INQ-A] tells you which addressees to ask given a question. But it can also tell you which questions to ask given an addressee.” It makes sense to ask questions only to those addresses that we have reason to believe that they know the answer to our query. Only in this way, we can acquire information we want to know.

Crucially, especially in the linguistic literature, the most emphasised aspect of the relationship between assertions and inquiries lies in their approach to informativeness. Thus, just as assertions are essentially information-giving speech acts, inquiries are information-seeking speech acts. Consider the above conversational patterns. What follows from INA, is that uninformative assertions are improper, and thus can be criticised. Analogically, it is improper to inquire, and thus criticisable, when one already knows the answer. The assumptions and expectations concerning the speaker and the audience in assertions and inquiries reverse (cf. Farkas 2022). In asserting, the speaker is expected to be competent and the audience’s ignorance is assumed. In inquiring, the speaker is ignorant and the audience is assumed to be competent (alternatively, the speaker has a reason to believe that the audience knows the answer, or at least the audience can redirect the speaker to someone who knows the answer). In short, while assertions are non-inquisitive and informative, inquiries are non-informative and inquisitive (e.g. Moyer and Syrett 2021). The upshot is that assertions and inquiries complement each other, which is seen especially well in how they relate to informativeness.

The final datum concerns a distinction between assertions and uninformative speech acts. Following INA, it is improper to reassert the same content. However, we do say things that do not deliver new information into the common ground. Such uninformative content can be naturally reintroduced by means of uninformative speech acts, like reminding, or other types of content, like presuppositions. Let me start with presuppositions, as this is the most common way of reintroducing already-known content. A presupposition is a piece of information that is taken for granted. Many theories of presupposition treat presuppositional content as commonly known by the participants of the conversation.²² Since presuppositions, just as assertions, received a normative treatment, we can compare INA with the norm of presupposition. García-Carpintero (2020, p. 272), treating presuppositions as ancillary speech acts, i.e., as speech acts that can only occur within another speech act, proposes the following norm of presupposition:

PR For one to presuppose p in a context is correct if and only if p is common knowledge in that context.

PR is a constitutive norm in the Williamsonian sense (cf. Keller 2022). Thus, following INA, while assertions are speech acts that introduce content that is not commonly known, presuppositions reintroduce content that is already commonly known. Crucially, García-Carpintero observes that “when it is correct to presuppose p, it is incorrect to assert it” (2020, p. 291). Thus, for instance, if you know that I have a sister, it is incorrect to assert it, but it is perfectly natural to presuppose this information.

Here is an initial case for distinguishing between assertions and reminders.²³ Firstly, we can observe that there is nothing inappropriate in reminding someone about something, i.e., there is a stable linguistic practice of reintroducing the already known content through reminders. Secondly, uninformative assertions can be criticised as improper. The normative account can accommodate both observations by proposing that these are two speech acts governed by distinct norms. Recall that while the constitutive norm of assertion tracks a specific speech act type performed by “flat-out assertions,” reminders are performed in an explicit way. So, I can assert that p, and later on—when, for instance, I have a reason that you might have forgotten, or I want to be sure that you remember what I asserted—remind you of p, by saying, for instance, “Just to remind you, p.” Crucially, we can make a similar observation as in the case of presuppositions, namely, when it is proper to remind someone about something, it is improper to assert it. For instance, if I already told you that you have an appointment, it is improper to assert it again; however, I can remind you about it. Thus, we can distinguish between essentially informative speech acts, like assertions, and uniquely uninformative speech acts and types of content, like reminders and presuppositions.

4 Objections

At this point, one might see the intuitive appeal of INA and accept the data. Nevertheless, one might worry about potential counterexamples. In this section, I discuss three cases.²⁴

4.1 Asserting Something that will not be Accepted by the Audience

Question

Sometimes we assert that p knowing that p will not be accepted by the audience. It seems that such cases sit uneasily with INA. Does INA classify them as genuine assertions? If so, are they proper assertions?

Answer

Consider the following example:²⁵

Stephen Miller puts credence 0 in the proposition that refugees benefit the American economy more than they cost. Jennifer Arangio, a lower-level aide who has looked at the relevant studies, has credence 1 that refugees benefit the American economy more than they cost. Arangio is well aware that, whatever she says, Miller’s credence in this proposition will not be shifted one bit. Nevertheless, she tells Miller the truth and thereby risks her job. (Krstić 2022, p. 8)

Krstić puts forward this case to show that one can reasonably and rationally assert that p without intending to give the audience a reason to believe that p. Particularly, he targets neo-Gricean theories of assertion, but claims that his argument generalises all theories that involve the audience in their accounts. While I agree that such cases are problematic for neo-Gricean theories, I disagree that the theories of assertion should focus only on the speaker. Krstić (2022, p. 9) argues that “whether it is possible for Arangio to sincerely assert something should depend on her and the context, not on her hearer.” By (at least some) one-Gricean theories, Arangio’s assertion is impossible, i.e., one cannot assert that p knowing that p will not be even considered as a reason to believe that p. However, the normative approach does not run into this problem.

Consider first that similar cases have been already discussed by Stalnaker (1978, p. 87), who says that “A person may make an assertion knowing it will be rejected.” Such an assertion may have some secondary effects, however, since it is not accepted, it fails to fulfil what Stalnaker identifies as the essential effect of an assertion, i.e., the content of what is asserted is not added to what is presupposed. Thus, such assertions are defective.²⁶

The normative account delivers similar results. Looking only at KNA, Arangio’s assertion is proper since she knows that p. Still, we can see that something is off with this assertion. We should ask the following questions: “What is the point in her assertion?”, or “What does she try to achieve?” In arguing for KNA, Williamson (2000, p. 267) considers the claim that the main epistemic function of assertion is knowledge transmission. Arguments have since been put forward to support this, or a stronger claim, such that the epistemic function of assertion is knowledge generation (see e.g., Reynolds 2002; Turri 2016; Kelp 2018; Kelp and Simion 2021; cf. Goldberg 2015). If the function of assertion is to generate knowledge, Arangio’s assertion is improper since she knows that she will fail to do that.

In the above story, even though it is not common knowledge that “refugees benefit the American economy more than they cost,” Arangio knows that Miller believes the opposite and that her assertion will not be accepted. Recall that one of the central features of constitutive norms is that they can be broken without ceasing to perform the speech act. This is what happens here. She knows that her assertion cannot satisfy INA, and in this respect, her assertion is improper. Crucially, however, she performs an assertion. This goes against Krstić’s (2022, p. 14) claim that all audience-centred theories, including the normative account, predict that Arangio is not performing an assertion. Finally, notice that Miller can criticise her assertion, for instance, by saying “Why do you say that? You know that I do not believe it!” Thus, showing that her assertion was, in some sense, inappropriate.

4.2 Asserting Something that is Already Commonly Known

Question

Sometimes we assert that p when unbeknownst to us, p is already known by the audience. Is INA not too restrictive in classifying such assertions as improper?

Answer

Assertions that are already known by the audience can be legitimately criticised. However, how strong this critique can be depends on the context. If, for instance, the speaker repeats p knowing that p is already commonly known, she can be criticised directly (e.g., “You know that I already know this!”). However, saying that p while not being aware that p is already known may not generate the same intuitions, i.e., in such cases we usually do not challenge the speaker directly (we can say, for instance, “Thanks, I already know this”).

Such cases can be accounted for by the following well-known distinction between primary and secondary correctness (e.g. Williamson 2000; DeRose 2002). This distinction is used to show that even though the speaker may not obey the norm, she may be excusable and blameless. For instance, assuming KNA, if one asserts something false while being justified in believing that p, one’s assertion is primarily improper (because of its falsity, and hence a violation of KNA), but secondarily proper (because the speaker was justified in believing that p). The same distinction can be applied to cases of asserting something—unbeknownst to the speaker—already commonly known. The norm (i.e., INA) has been violated (hence the assertion is primarily improper), but the speaker is excusable and blameless because she had reasons to believe that she followed the norm (hence it is secondarily proper).²⁷

4.3 Gricean Maxims

Question

It seems that INA gives a general condition applicable to many speech acts. Is INA a pragmatic norm that is already included in Grice’s (1989) Cooperative Principle and maxims?

Answer

Most speech acts are informative regarding the content, and thus most of them will be governed by some sort of norm of informativeness. I favour the formulation of the norm in terms of knowledge. As a result, INA may work well with those speech act types that are governed by a norm at least as strong as knowledge. Thus, such speech acts as swearing, warning, or admitting. Their relation to assertion can be explained in terms of ‘illocutionary entailment’ (Searle and Vanderveken 1985; cf. Marsili 2023b), i.e., we can say that my swearing that p ‘illocutionary entails’ asserting that p. Still, such speech acts can be differentiated from assertions. Consider that the same considerations apply to norms like KNA. Williamson (2000, p. 244) observes that swearing requires more than asserting, and so it is governed by a norm stronger than KNA. However, we do not want to say that KNA is a too general norm of assertion (it is criticised for being too strong). The same concerns the norm of informativeness as specified in INA.

One may still wonder whether INA is a pragmatic norm similar to Grice’s Cooperative Principle and maxims. In a sense, all norms of speech acts are pragmatic, as speech acts are pragmatic concepts. More relevantly, however, constitutive norms are epistemic norms. It is important to note that “The fact that a norm has epistemic content does not imply that it is an epistemic norm” (Kelp and Simion 2021, p. 56). Thus, INA is an epistemic norm not because it is formulated in terms of common knowledge. To a first approximation, norms are epistemic if they “govern what we ought to say, do or think from an epistemic point of view” (Graham 2015, p. 247). If assertions serve to generate knowledge, as I proposed above, the norm of assertion should prohibit performing assertions that are already commonly known because they fail to fulfil the requisite function. Since this is what INA does, it can be characterised as an epistemic norm. As for the Cooperative Principle and maxims, they set general conditions of cooperative communication, applicable to all communicative acts. In turn, a constitutive norm of assertion, such as INA, proposes a stable, context-independent condition of a specific type of informativeness, applicable only to a specific speech act type. Therefore, INA sets a different type of standards than conversational maxims—they are not only much narrower but also independent of the Cooperative Principle.

Consider two maxims more closely. INA may seem to set the standards that are the closest to the maxims of quantity and relevance. Huang (2014, p. 30) labels the maxim of quantity as the maxim of informativeness. Notice, however, that the maxim of quantity tracks a different type of informativeness than INA. The maxim of quantity specifies that one should not say what is under-informative or over-informative for the current purpose of the conversation. In both cases, one can deliver new information to the audience. Such cases could count as violating or exploiting the maxim, but they would not violate INA. Moreover, we can opt out of the maxim by using specific hedges, such as “As you probably already know, p,“ or “I probably don’t need to say this, but p.” Crucially, such utterances are not subject to INA. The maxim of relevance also gives much broader instructions than INA. For instance, in some contexts, it may be highly relevant to remind, repeat, assure, or double-check something that is commonly known. INA, in turn, individuates assertions from speech acts that are used to restate the already known content. When one says, “Just to remind you, p,” “I reassure you that p,” or “I double-checked that p,” one is not making assertions and thus, one’s speech acts are not governed by INA.

5 Consequences for the Normative Account of Assertion

In this final section, I discuss the consequences of accepting INA, focusing on the debate on the norms of assertion. Particularly, I show how INA fits into the distinction between speaker- and audience-centred norms.

Let me start with audience-centred norms. They are on the margin of the debate and can be treated as outliers.²⁸ To my knowledge, there are only three proposals for such norms. Apart from the already introduced TKNA (García-Carpintero 2004), here are two remaining ones—the knowledge provision norm of assertion (Pelling 2013, p. 297) and the norm based on providing testimonial warrant (Hinchman 2020, p. 556; cf. 2013):²⁹

KPNA

One’s assertion that p is proper only if it is fit to give a hearer knowledge that p.

PTWNA

Assert [that p] only when you could provide testimonial warrant [that p] to a potential addressee.

Gathering what is common for these approaches, Hinchman (2020, p. 26, footnote 12) states that

The core thought here is that there is an epistemic norm of assertion set by the addressee’s circumstances: the normative aim of assertion is not simply to express knowledge but to give your addressee knowledge.

All of these norms are knowledge-based, and all situate knowledge as the desired outcome of assertion. Relevantly for the informativeness of assertions, the audience-centred norms maintain that assertions are for spreading new information (in the form of knowledge) to the audience. Such norms are supposed to bring out the social and communicative aspects of assertions. TKNA specifies that an assertion is a speech act that transfers knowledge to the audience (by putting it in a position to know). Since it maintains that one’s assertion must provide new information, we can say that it already presupposes INA. We receive the same result for Hinchman’s (2013, p. 632) PTWNA, which states that assertions are essentially for providing testimonial warrant, i.e., in asserting that p, the speaker aims to put herself in a position to inform the audience that p. Similarly, Pelling (2013, p. 294) argues that assertions fundamentally aim at transmitting information.³⁰ Thus, granting that assertion is governed by one of these norms, we can accommodate the above data. Consider TKNA. Firstly, uninformative assertions are improper since they do not transfer new information. Secondly, just as inquiries are information-seeking speech acts, TKNA regards assertions as essentially information-giving speech acts. Finally, TKNA strictly distinguishes between assertions and uninformative speech acts.

The situation differs for the speaker-centred norms since they fail to accommodate the data. Consider KNA. Firstly, following KNA, nothing prevents the speakers from making uninformative assertions. Thus, KNA cannot explain why we can legitimately criticise uninformative assertions as improper. If KNA is the constitutive norm of assertion, then uninformative assertions are either proper assertions (since the norm is not violated) or each time such assertions are performed, KNA is overridden by other norms. This is not a satisfactory result. One, we do criticise such assertions. Two, if the function of assertion is knowledge generation, then KNA fails to explain it on its own. Secondly, KNA cannot account for the analogy between inquiries and assertions. The speaker who inquires whether p not only does not know whether p but also seeks whether p in the audience. Analogically with assertions. The speaker must not only know that p but also express and transfer this knowledge to the audience. KNA specifies only the former condition. Finally, KNA fails to distinguish between assertions and uninformative speech acts. Again, because of this, it cannot explain the function of assertion. If KNA is the norm of assertion, it is also the norm of reminders and presuppositions—both of them require the speaker’s knowledge. Thus, KNA blends the categories of informative and uninformative contents. As a result, KNA fails to individuate assertions (cf. García-Carpintero 2020).

Some speaker-centred norms can be fixed, however. We can see that INA explains data that norms like KNA fail to account for. On the other hand, what has been a subject of extensive discussion, KNA is well-equipped to explain other data (particularly, the three classical data points for KNA, i.e., the unique way of challenging assertions, intuitive infelicity of Moorean assertions, and impropriety of assertions based on merely probabilistic grounds).³¹ Simultaneously, both norms, by themselves, are insufficient to individuate assertions. However, when taken together, they complement each other. Thus, I propose the following two-part norm of assertion:

KNA&INA One must: assert that p only if:

• (i) one knows that p, and.

• (ii) p is not common knowledge.

Such a norm can explain all the data and remains consistent with the normative account, i.e., it is essential, unique, and individuating for an assertion. Consider why it is individuating. On the one hand, I noticed that INA does not distinguish assertions from, for instance, guarantees. Notice, however, that KNA does. It has been proposed (Turri 2013) that guaranteeing is governed by a stronger norm than KNA (specifically, one may guarantee that p only if one knows that one knows that p). On the other hand, KNA fails to distinguish assertions from uninformative acts, like reminders or presuppositions. However, this is what INA is designed to do. This is merely an illustration of how INA can be conjoined with a specific speaker-centred norm. Finally, consider that both parts of the norm perform unique functions. One specifies the epistemic position of the speaker, while the other accounts for the communicative function of assertion—showing that assertions are for spreading information. A similar intuition can be found in the debate on testimonial transmission, where Greco (2020, p. 99) argues that:

It is true that, in many cases, the transmission of knowledge will be largely due to the competent agency of the speaker, but it is also true that the hearer must do his essential part for transmission to be successful.

To achieve success in knowledge transmission, which happens via assertions, the cooperation between the speaker and the audience is essential. This is the general idea behind the audience-centred norms, which INA articulates explicitly.

The informativeness norm, in its most general form, says that assertions are essentially informative speech acts. Thus, if one finds knowledge a too strong commitment for the norm of assertion, it seems that other norms of assertion can also be complemented by an appropriately modified informativeness condition. However, there are norms of assertion that do not sit well with this prediction. In general, there are two ways of thinking about assertions. The first focuses on individuating a narrow category—a speech act type used by default in the declarative mood. Norms like KNA are a case in point. The second approach, however, treats assertion as a broad category—more like a genus of speech act—that apart from flat-out assertions, encompasses also other speech acts. How broad this category is, depends on the preferred theory; for instance, Weiner’s (2005) TNA includes predictions and reminders, while McKinnon’s (2015) context-sensitive norm extends the assertoric domain to speech acts from guesses to guarantees. INA, or its variants, can work well with narrow accounts, like KNA—here, both norms complement each other. In other words, if we want to individuate a speech act type, then we want to track assertion in the narrow sense. In the case of broad accounts, however, there arises a tension between the norms. Consider TNA. If truth is the norm of assertion, then it includes uninformative speech acts, like reminders, as a species of assertion. Notice that this is not an undesirable side effect (like in the case of KNA), but something purposeful. The central idea behind TNA is that the notion of assertion should be built in such a capacious way to capture all truth-directed speech acts as assertions.³² However, INA’s first and foremost purpose is to distinguish between informative and uninformative speech acts. Thus, INA and norms like TNA do not go hand in hand. Supplementing TNA with INA would restrict the scope of TNA, which stands against the original purpose of the norm.

6 Conclusions

Assertions are first and foremost informative speech acts. To this end, I have made a case for the informativeness norm of assertion. Looking at assertions from the perspective of their informativeness allows us to see that assertions are for the audience, i.e., to be informative, the speaker must consider the epistemic position of the audience. The role of the audience, however, has been neglected and undervalued in the speech act literature and, specifically, in the recent debate on the norms of assertion where most of the proposed norms are speaker-centred. I have shown that the data for the informativeness of assertions can be naturally accommodated by audience-centred norms. Alternatively, at least some of the speaker-centred norms can be complemented with the informativeness norm. Thus, the paper is a plea for underscoring the importance of the audience in construing adequate accounts of speech acts.

If informativeness plays such a crucial role in assertions, it may also play an analogical role in other speech acts. This concerns declarative and non-declarative speech acts alike. Thus, just as my explanations and guesses should be informative for my audience, so should my promises and orders. However, the topic of the informativeness of speech acts has received little attention. It has been even less so in the case of the uninformativeness of speech acts. However, omitting the category of uninformative speech acts from discussion results in proposing accounts, such as speaker-centred norms, that blend informative and uninformative speech acts with each other. Thus, more work must be done to explain the role of the informativeness and uninformativeness of speech acts and to elucidate how these two types of speech acts intersect with each other.