1 Introduction

Several people think knowledge is the norm of assertion: one should only assert that p if one knows that p (KNA).Footnote 1 Others disagree and impose weaker conditions (justification, truth), or stronger conditions (certainty), or even contextually variant conditions on permissible assertion.Footnote 2

Whatever their preferred norm might be, however, most people in the debate agree upon one broadly Gricean thought: it should be able to account for assertion’s epistemic aim, or function.Footnote 3 Some think that the relevant function is expressing one’s beliefs (e.g. Bach (2008), Hindriks (2007)). Others take a pragmatic stance: assertion and belief alike should provide/store actionable information (e.g. McKenna Forthcoming).

Last but not least, many people believe the main epistemic function of assertion is generating some epistemic stance in the hearer.Footnote 4 When this is the case, people draw upon results in the epistemology of testimony in order to justify their preferred account of assertion, or the other way around: they look into what the normativity of assertion implies for the normativity of testimonial belief.

Now, one very interesting territory, lying at the intersection of the epistemology and ethics of testimony, is the normativity of belief formed via moral testimony. Can one come to know via moral testimony? Is it permissible to rely on it in action? Should one trust moral testimony to begin with? Given this, it is rather strange that the literature does not seem to care that much about the normativity of moral assertion.Footnote 5 After all, one would think that, insofar as discussions pertaining to the normativity of testimonial belief are relevant to the normativity of assertion in general, mutatis mutandis, the normativity of moral assertion in particular will be sensitive to the correct account of permissible belief from moral testimony.

As far as I can see, the reason why this might be is that people have generally assumed that whatever the norm for assertion in general turns out to be, it will also govern moral assertions. This, however, of course, need not be the case: just because waltzing is a species of dancing, it need not follow that it is merely governed by norms pertaining to dancing in general. To the contrary, what makes waltzing into a particular species of dancing to begin with is the fact that is governed by extra-norms, on top of the ones governing general dancing. Similarly, the norm for moral assertions can be stronger than the norm for assertion in general: more can be required in terms of epistemic support for proper moral assertion.

This paper aims to supply this lack: it looks into the prospects of a function-first account of the normativity of moral assertion, where the norm at stake is taken to drop right out of the function of moral assertion in testimonial exchanges. On the view defended here, moral assertion has the function of reliably generating moral understanding in hearers. I argue that the norm of moral assertion borne out by this framework is one that requires speakers to (1) know the contents of their assertion and (2) back their assertion with a contextually appropriate explanation.

To this aim, I first look at the broad lines employed in the literature to derive the norm of assertion from its epistemic function; in particular, I look at the case for the knowledge norm (#2). Further on, I offer a brief overview of the extant case for understanding as the central epistemic goal regarding moral affairs (#3). In Section #4 I outline my proposed account of the normativity of moral assertion, which, I argue, drops right out of the epistemic function thereof. Finally (#5), I consider some possible objections to my account and respond. In the last section I conclude.

2 The Norm and Function of Assertion

Independently of whether one likes KNA or not, the view definitely has a lot going for it: friends and foes of KNA alike generally agree that, if KNA is true, we have a very straightforward and elegant way of explaining quite a few otherwise puzzling linguistic data, such as the paradoxical soundingness of Moorean statements of the form ‘p but I don’t know that p’, the unassertability of lottery propositions and ‘How do you know?’ challenges.Footnote 6

Last but not least, KNA affords a very straightforward answer to the rationale question, i.e., the question as to why assertion is governed by an epistemic norm to begin with: according to several people in the literature, the main epistemic function of assertion is generating testimonial knowledge in the hearer.Footnote 7 In virtue of our physical and psychological limitations, we rely a lot on other people’s testimony. Arguably, we learn most of the things we know about the world from our parents, from the testimony of our teachers and, later on, from media channels; thus, assertion is one of our main epistemic vehicles.

But, if that is the case, it is argued, KNA readily follows. Here is why: on most if not all accounts of testimony in the literature,Footnote 8 in the vast majority of cases, the speaker needs to know in order to be able to generate knowledge in the hearer. Exceptions are few, and they describe highly unusual scenarios.Footnote 9 Given their oddity, these exceptional cases are highly unlikely to affect the argument from testimony to the knowledge norm in any way. After all, if the characteristic purpose of assertion is generating testimonial knowledge, and in the vast majority of cases knowledge on the part of the speaker is needed for generating testimonial knowledge in the hearer, it makes sense to have a knowledge norm governing assertion. To see this, consider driving: norms regulating speed limit within city bounds are presumably there to make it so that we arrive safely at our destination. Surely, though, driving 50 km/h within city bounds is not always the ideal speed; there are instances when, for instance, overtaking at 80 km/h will avoid a major accident. However, presumably, the reason why the norm says ‘Drive at most 50 km/h within city bounds!’ is because, most of the time, that is the ideal speed for safety purposes.

Note, also, that in the friendly epistemic environment we inhabit, in a wide range of areas, knowledge is widely and readily available. All I have to do in order to come to know that there’s a computer in front of me is take a look; all I have to do in order to come to know whether you are dropping by on Sunday is give you a call. Long story short, happily, our epistemic predicament is such that knowledge is ubiquitous and belief that falls short of knowledge is the exception: all we have to do to acquire knowledge is open our eyes, listen to what other people tell us, attend to our feelings, etc.

In sum, things stand as follows: on one hand, in the vast majority of cases, assertions need be knowledgeable in order to deliver the epistemic goods they are meant to deliver; also, this does not constitute much in the way of epistemic burden on the shoulders of the speaker: knowledge is readily available. If that is the case, we have a nicely motivated picture in support of KNA.

Now, as stated, KNA makes a necessity claim. Importantly, though, that is not all there is to it: the ambition is to offer the strongest necessity claim. To see this, note that, although knowledge is factive, it would be weird to say that defenders of KNA and defenders of a truth norm of assertion do not disagree: they definitely take each other to make importantly different claims. The debate, thus, is over the most plausible strongest necessary condition for assertion.

Note, also, that what is at stake is the strongest necessary condition for assertion in general. As such, KNA does not exclude there being particular types of assertion that are governed by stronger norms. Assertions in the court of law or, more generally, in high stakes scenarios are often taken by many to be cases in point.Footnote 10 Again, this also affords a nice explanation: given how much hinges on getting it right, in these cases, more than knowledge is required for proper assertion.

Here is, then, the question that will concern us next: what about the norm for proper moral assertion? What is the functionalist story to be told there? This paper’s hypothesis is that the answer is to be found in the literature discussing moral testimony.

3 Testimony and Moral Understanding

Consider the following case from Alison Hills (2009, 94):

ELEANOR: Eleanor has always enjoyed eating meat, but has recently realized that it raises some moral issues. Rather than thinking further about these, however, she asks a friend who tells her that eating meat is wrong. Eleanor knows that her friend is normally trustworthy and reliable, so she believes her and accepts that eating meat is wrong.

Many people in the literature have the intuition that there is something wrong in the way in which Eleanor proceeds; that rather than taking her friend’s testimony for granted, she should reflect on the issue herself. Call these people ‘pessimists’Footnote 11 about moral testimony. Now, importantly, there are two ways in which one can be a pessimist: a weak and a strong way.Footnote 12 According to the strong pessimist, one should not form moral beliefs on the mere say-so of others, because the latter cannot generate knowledge. This fellow takes it that knowledge of moral matters is too highbrow an affair to be gained second hand: more work than that is needed. Furthermore, he generally takes it that the function of testimonial exchanges is to generate the corresponding knowledge in the audience. Since moral testimony cannot do that, according to the pessimist, one should not form beliefs about moral issues from mere testimonial sources.

Now, a lot of ink has been spilled on disputing the claim that one cannot gain knowledge from moral testimony, and with some degree of success too. It is fair to say that, in recent years, the state of the art is such that most people would agree that moral testimony can generate knowledge ((Driver 2006), (Hills 2009), (Jones 1999)). As such, for the purposes of this paper, I will set strong pessimism aside. At the same time, a new variety of pessimism is born: weak pessimism. What these philosophers want to say is that, even though one can get knowledge of moral facts from testimony, this is still not good enough; there is still something wrong about forming moral beliefs on mere say-so, even though they are knowledgeable beliefs.

According to weak pessimists, first, when it comes to moral matters, our epistemic goals are more highbrow affairs than just merely gathering a fair number of knowledgeable beliefs: what we are after is understanding the relevant matters (e.g. Nickel (2001), Hills (2009) Hopkins (2007), Howell (2014), McGrath (2009)). Alison Hills, for instance, argues that there are at least two reasons why this is so: moral understanding is an essential mean to morally good character and morally worthy action.

The further assumption is that understanding is more than just knowledge of the relevant matter. Hills (2009) suggests that understanding is constituted by a set of abilities that allow one to employ the relevant piece of information beyond the issue at hand: “To understand why p you must have an ability to draw conclusions about similar cases, and to work out when a different conclusion would hold if the reasons why p were no longer the case.” (194). For this, it is argued, mere knowledge is not enough.

One possible reply that easily comes to mind, however, is that maybe what is actually needed is, well, more knowledge. To see this, think again about Eleanor: it looks as though, if on top knowing that eating meat is wrong, she were to also know why this is so, her moral beliefs would be beyond criticism. In reply to this worry, pessimists rely on views taking understanding to be a sui generis epistemic state; according to these views, there is more to understanding that p than merely having knowledge why p. In line with people like Stephen Grimm (2006), for instance, Sarah McGrath suggests that understanding is a matter of possessing a particular skill of grasping the relevant connections between your pieces of knowledge, and, in this, it works a lot like a priori insight. There seems to be a clear difference between knowing that a conclusion is entailed by a set of premises and being able to see the entailment relation. Similarly, according to McGrath, moral understanding requires not only knowing that an action is wrong, or even why it is wrong, nor simply the ability to make inferences about its moral status based on information about its morally relevant properties, but also seeing or grasping the connection between its having those features and its being wrong. As such, McGrath argues, merely relying on moral testimony cannot and will not get you far enough, epistemically; what needs be the case is that you come to grasp the relevant relations yourself.

Now, there is a lot to be said here. First, the sui generis view about understanding that pessimists employ is hardly the sole view on the market: many people in the relevant epistemological literature believe that, to the contrary, understanding that p is, eventually, reducible to one form or another of knowledge that p (Kelp (2015) Riaz (2015)). Furthermore, several optimists about moral testimony, such as, notably, Julia Driver (2006) and Paulina Sliwa (2012), disagree about the limited capacities of testimony to generate permissible belief or built good moral character. According to them, this is a mere contingent feature associated with most cases of moral testimony, but not with all. In particular, according to Sliwa, when we (justifiably) take the speaker to be more reliable than us on the matters at hand – maybe because we are concerned that our judgment is compromised by bias or self interest, or because the other person is generally better at making moral judgments – it looks as though not only is it perfectly permissible to reply on their testimony; to the contrary, it is the responsible thing to do. Similarly, it seems perfectly fine for one to seek and follow moral advice when not able to make up one’s own mind on a particular moral matter. As such, Sliwa argues, it is not the case that there is something essentially wrong about relying on moral testimony.

Furthermore, Sarah McGrath puts forth cases featuring agents who, against their own beliefs on the matter, take good moral advice from others, to show that there are instances where pure moral deference is not only permissible, but quite a good idea. According to McGrath, this faces us with a puzzle: what is it that explains the fact that, in some cases (e.g., ELEANOR), pure moral deference seems criticizable, while in others it seems the right thing to do?

Whether weak pessimists or optimists are right about these issues will fortunately be irrelevant for our purposes here. Note, however, that there are a few things that most actors in the debate - both friends and foes of moral testimony - seem to agree upon: in most cases, moral testimony that p is deficient in generating the right epistemic standing with regard to p in the hearer; the right epistemic standing is moral understanding, which, in turn, is reliably conducive to morally worthy actions and character traits.

Now, recall the rough lines of the derivation of KNA in the previous sections: the function of assertion is to generate testimonial knowledge in hearers; since in the vast majority of cases, knowledge on the side of the speaker is needed for generating knowledge on the side of the hearer, KNA holds.

One question that arises at this point is: how about moral assertion? What is the function of moral assertion? After all, the phenomenon does seem to be fairly ubiquitous, so it would be fairly surprising if it turned out to be a pointless endeavor. If the literature is right, and, in most cases, moral testimony cannot generate moral understanding in hearers, why are we engaging in it to begin with? To what aim?

We have seen above that most people take it that the goal of moral inquiry is understanding. The next section will make an attempt to defend the norm for moral assertion that drops out of this picture.

4 The Function and Norm of Moral Assertion

Recall that, in Section #2, we have seen that, according to many, generally speaking, knowledge is the stronger necessary condition on epistemically permissible assertion, in virtue of knowledge being our most general epistemic goal. Asserting from knowledge is the most reliable way to generate knowledge in one’s hearer.

Now, importantly, note that this does not mean that the speaker’s asserting from knowledge is all there needs to happen for knowledge to be generated on the hearer’s side; on top of this, the speaker needs be decently reliable, the hearer needs to accept the testimony and display minimal non-gullibility when doing so and so on. What is of central concern to us, though, given that we are after the norm for assertion is what epistemic features the assertion involved in the relevant testimonial exchange needs to display for the relevant function to be fulfilled, independently of whatever else needs be present on top of a proper assertion to this effect. As such, the question is, ceteris praesentibus, what should assertion (epistemically) bring to the table in order for the function to be reliably fulfilled? And, importantly, the answer to this question is: knowledge.

We have also seen above, however, that many people think that in some domains knowledge is not enough: the domain-specific central epistemic goal is, rather, to understand the issue at hand. Morality is such a domain: according to many, to be able to engage in morally worthy action, build a morally virtuous character and so on, one needs more than to know the relevant moral matters of fact: one needs moral insight into how they hang together.

Now, importantly, just as in the case of general testimony, it is hardly surprising that, no matter how epistemically endowed one’s moral assertion might be, it might still fail to generate the relevant epistemic standing in the hearer in isolation: other features need be in play to this effect. According to pessimists, on top of the requirements for general testimony, when it comes to moral affairs, the hearer also needs to actively participate in the exchange and think through the relevant matters herself, in order to achieve moral understanding. However, this does not mean that the moral assertion at work does not have its role to play in this affair: we should expect that, just as in the case of testimony in general, moral assertion should meet the strongest epistemic standard needed for generating the target epistemic standing in the hearer, ceteris praesentibus.

Since in the case of moral matters, the widely agreed target epistemic standing is moral understanding, here is, on a first approximation, the proposal made by this paper:

  • The Explanation Proffering Norm of Moral Assertion (EPNMA): One’s moral assertion that p is epistemically permissible only if (1) one knows that p and (2) one’s assertion is accompanied by an appropriate explanation why p.

The intuitive thought behind EPNMA is the following: say that pessimists are right, and no amount of knowledge will get the hearer to understand the moral matter at hand, unless she also ‘runs’ the relevant issue through her own rational capacities in order to ‘grasp’ how things hang together. Now, take a hearer who is perfectly disposed to do the relevant grasping work; what is the most reliable way to generate moral understanding in this particular hearer? It looks fairly intuitive that the answer is: by asserting to the relevant matter of fact (‘p’) and offering the appropriate explanations as to why p is the case. Recall Eleanor, and ask yourself: what would have to be the case on the speaker’s side in order to reliably generate understanding why eating meat is wrong on Eleanor’s side, granted that she would be well disposed to do the needed ‘grasping’ work? Presumably, her friend would need to say something along the lines of: ‘eating meat is wrong because animals are living beings just like you and they deserve to be treated with equal respect …’. In this case, it looks as though, if Eleanor runs the entailment relation through her own rational mechanisms, she comes to understand why eating meat is wrong. Also, it does not look as though there is anything wrong with her corresponding belief.

Now, this being said, one pressing question that arises is: what will count as an appropriate explanation? First, importantly, I take it that it will not be a merely quantitative matter: to the contrary, appropriateness will refer to both amount and kind of reasons brought I support of the target moral claim. Second, importantly, the running hypothesis of this paper is that this will be a contextually variant matter: more/stronger reasons will need to be brought in support of some assertions, and less/weaker for others. One motivation for this is that I take it, together with many people in the relevant literature,Footnote 13 that understanding is a contextually sensitive matter itself: it looks as though it is one thing for my ten year old to count as understanding physics, and quite another in the case of myself: different standards seem to apply.Footnote 14 Also, according to many, explanation itself is a highly context sensitive matter.Footnote 15

This is not all there is to the context-sensitivity of EPNMA, however; even if it turns out that understanding is invariant, some contexts seem to require more than others in the way of explanation for understanding to be generated in hearers, depending on several contextual determiners such as: the previous knowledge and level of general intelligence of the hearer, the general assumptions about morality present at the context, the level of difficulty and controversy of the subject matter, how much (morally)Footnote 16 hinges on the issue, the urgency of the situation etc. As such, we should reformulate our norm as follows:

  • The Explanation Proffering Norm of Moral Assertion* (EPNMA*): At a context C, one’s moral assertion that p is epistemically permissible only if (1) one knows that p and (2) one’s assertion is accompanied by a C-appropriate explanation why p.

EPNMA* does well by the broadly Gricean idea that “the accepted purpose and direction of the talk exchange” determine what counts as adequate warrant for appropriate assertion (1989). Furthermore, it nicely drops out of the widely accepted epistemic goal with regard to moral matters – i.e., moral understanding - and it is, as such, nicely motivated. Note, also, that this result should be fairly uncontroversial, no matter what side of the moral testimony debate one prefers. After all, note that even if the sui generis view about understanding turns out to be false, EPNMA is unaffected: it would only mean that there’s not much ‘grasping’ work lying on the shoulders of the hearers: all they need to do is listen, comprehend and believe what they are being told.

Also, the view nicely explains what is intuitively wrong about the initial Eleanor case: in general, it is impermissible to believe based on epistemically impermissible sources. Unreliable testifiers, misleading evidence, wishful thinking are just a few cases in point. Similarly, one should not believe based on epistemically impermissible assertions. Eleanor should not form a belief based on her friend’s assertion because the latter misses an appropriate explanation.

Furthermore, at a more general level, the account offers a nice solution for Sarah McGrath’s puzzle about moral deference: what is it that explains the fact that, in some cases pure moral deference seems criticizable, while in others it seems the right thing to do? According to the present account, in some cases, contextual determiners are such that little to no explanation is needed for proper assertion and, therefore, for generating the relevant stance in the hearer. This will, for instance, be the case in urgency cases: if I shout ‘Don’t kill!’ and thereby stop a murder from happening, my assertion will be perfectly fine in the absence of much explanation.Footnote 17Similarly, the Sliwa (2012) cases featuring worries about ones own reliability being affected by bias and moral advice are cases in point. To the contrary, in contexts where the hearer, like Eleanor, has never given the subject matter any thought, and there’s nothing keeping the speaker from going more into depth about the subject matter, more and better explanatory support is needed for proper assertion.

5 A Few Worries

In the previous sections, I have been arguing that, if it is plausible to think that knowledge is the norm of assertion in general, in virtue of assertion’s function of generating knowledge in hearers, since in the case of moral assertion, the epistemic function at stake is moral understanding, the norm governing assertion will be something in the vicinity of EPNMA*. That is, epistemically permissible moral assertion will be knowledgeable and accompanied by a contextually appropriate explanation, since this will maximize its reliability in fulfilling its epistemic function: generating moral understanding on the hearer’s side.

The aim of this section is two fold: first and foremost, I will try to answer what I take to be the most obvious objections that come to one’s mind when presented with my proposed account. Secondly, while doing so, I will show whether and how the account can be extended to other domain of interest concerning the normativity of assertion.

  • Objection #1. How about non-sophisticated cognizers? Uncontroversially, small children do not have much in the way of reflective capacities. However, many people believe that they are knowers, and therefore able to perform perfectly fine assertions. Furthermore, it does not look as though they cannot assert when it comes to moral matters: surely, it is fine for a three-year-old to tell his mother that lying is wrong, for instance. Where does EPNMA* leave them?

Note, first, that the child’s assertion in this case does not have much in the way of potential to generate moral understanding on hearer’s side. If that is the case, EPNMA* aside, according to any function-first view, strictly speaking, the assertion is not a permissible moral assertion. There is a different question, however, whether the asserter is blameworthy for making an assertion that has little potential to fulfill its epistemic function. In the case of the child, of course, the answer is no, in virtue of her limited cognitive capacities. This, however, does not make her assertion, as such, a good one. To see the plausibility of this, note that just like Eleanor, the mother should definitely not base her beliefs on the child’s testimony taken in isolation.

Note, too, that treating non-sophisticated cognizers like a special case is hardly uncommon in the epistemology of moral testimony, and therefore not easily imputable to EPNMA*; for instance, both pessimists and optimists will agree that pure moral deference is perfectly fine in the case of unsophisticated hearers, such as small children. This suggests that, when it comes to highbrow moral matters, unsophisticated cognizers just require special treatment both as receivers and as producers of testimony.

  • Objection #2. Many people in the literature take it that there is an interesting analogy between moral and aesthetic testimony. That is, just as in the moral case, it feels equally wrong to form beliefs based on someone’s testimony about matters of taste in isolation. However, it does not look as though EPNMA* will map equally nice to the normativity of aesthetic assertion.

I agree, the analogy breaks here. It looks as though, no mater the amount of explanation brought in support of the corresponding assertion, one should still not form beliefs about whether ice-cream is good based on someone else’s say-so. Two things about this; first, note that the analogy breaks at a different level of abstraction, and such it does not, strictly speaking, directly concern the account of the normativity of assertion at stake here. After all, the problem with matters of taste seems to be that, no matter how many reasons the speaker offers and no matter how much the hearer ‘runs’ the relevant considerations through her own reasoning capacities, she will still not come to have much in the way of epistemic support for believing that ice-cream is good. Here is, for one, Robert Hopkins on the matter: “If you tell me that some film you have seen is excellent, it is far from clear that I can legitimately adopt your view. Your testimony might motivate me to watch the film, but it does not give me the right to the belief that the film is excellent. I won’t have that right until I’ve seen the film myself” (Hopkins 2011). If that is the case, however, it looks as though proper belief about matters of taste is more difficult to come across than proper belief about moral matters. In the case of the former, something like direct acquaintance with the object of approval or disapproval seems to be needed (at least somewhere down the road)Footnote 18; needless to say, one can have moral beliefs about abortion without ever having had one.

It is interesting to note, however, that, in this respect, moral assertion seems to be more analogous to assertion in more descriptive domains. Take, for instance, the sciences: more often than not, assertions in these domains will aim at generating understanding in hearers rather than mere disparate pieces of knowledge. Many people in the philosophy of science think that the epistemic goal of scientific endeavors is understanding. If that is the case, again, plausibly, explanation proffering will be a plausible requirement on proper assertion in scientific contexts.

  • Objection #3. Why not an understanding norm of assertion?

We have seen that, plausibly, if generating testimonial knowledge is the main epistemic function of assertion, and if its norm follows from its function, knowledge is the norm of assertion. One question that arises at this point is: given that we take the relevant function in the case of moral assertion to be generating understanding, why does this not generate a understanding norm of assertion?

The answer is twofold: first, let us grant that whatever the theorist of understanding takes it that needs to be present on the hearer’s side is present; now let us ask ourselves: is it plausible that understanding on the speaker’s side is going to be the most reliable way to generate understanding on the hearer’s side? The answer seems to be ‘no’. Here is why: for one, it is hardly the case that understanding that p reliably leads one to being able to explain why p is the case. Most of us know our domains well, but are hardly good teachers. This, if not immediately intuitively plausible, is evidenced by the amount of resources we invest in training teachers in pedagogical techniques. Second, if the sui generis account of understanding is right, it looks as though understanding is too strong a requirement on the speaker’s side to begin with. That is because she can reliably generate understanding in her hearers while missing the relevant ‘grasping’ capacities herself. All that is needed for epistemic function fulfillment in this case is that the speaker presents the hearer with the relevant explanation and that the hearer does the grasping work herself.

6 Conclusion

This paper has proposed a novel account of the normativity of moral assertion. I have argued that, in view of reliably fulfilling its function of generating moral understanding, moral assertion should be knowledgeable and accompanied by an explanation, where the relevant notion of appropriateness is a contextually determined one.