Same, same but different: the epistemic norms of assertion, action and practical reasoning

Abstract

What is the relationship between the epistemic norms of assertion and the epistemic norms of action/practical reasoning? Brown argues that the epistemic standards for practical reasoning and assertion are distinct (Brown in Philos Phenomenol Res 84(1):123–157, 2012). In contrast, Montminy argues that practical reasoning and assertion must be governed by the same epistemic norm (Montminy in Pac Philos Quart 93(4):57–68, 2012). Likewise, McKinnon has articulated an argument for a unified account from cases of isolated second-hand knowledge (McKinnon in Logos Episteme 3(4):565–569, 2012). To clarify the issue, I articulate a distinction between Equivalence Commonality and Structural Commonality. I then argue against the former by counterexamples that doubly dissociate the epistemic standards for assertion and action. Furthermore, I argue that such a double dissociation compromises knowledge accounts of both assertion and action/practical reasoning. To provide a more accurate diagnosis, I consider speech act theory and argue that principled differences between the epistemic norms of action and assertion compromise Equivalence Commonality. In contrast, a qualified version of Structural Commonality may be preserved.

1 Introduction

What is the relationship between the epistemic norms of assertion and the epistemic norms of action/practical reasoning?¹ Brown argues that the epistemic standards for practical reasoning and assertion are distinct (Brown 2012). In contrast, Montminy argues that practical reasoning and assertion must be governed by the same epistemic norm (Montminy 2012). Likewise, McKinnon has articulated an argument for a unified account from cases of isolated second-hand knowledge (McKinnon 2012).

In order to clarify the issue, I articulate a distinction between Equivalence Commonality and Structural Commonality. I then argue against the former by counterexamples that doubly dissociate the epistemic standards for assertion and action. Furthermore, I argue that such a double dissociation compromises knowledge norms of both assertion and action/practical reasoning.

To provide a more accurate diagnosis, I consider speech act theory and argue that principled differences between the epistemic norms of action and assertion compromise Equivalence Commonality. In contrast, a qualified version of Structural Commonality may be preserved.

2 Pro et contra - and a distinction

There are several generic rationales for assuming that the epistemic norms for action/practical reasoning and assertion are the same. One such generic rationale involves the assumption that assertions are speech acts and, hence, acts. We do something when we assert something. So, we are epistemically responsible for our assertions in ways that are similar to the ways in which we are epistemically responsible for our actions (Gerken 2012). Moreover, the profiles for ascribing epistemic blame in cases of assertion and action/practical reasoning appear to be very similar. If the stakes are high and further evidence is easily available, someone may be blameworthy for acting on the assumption that the meeting is at 10:00 as well as for asserting that the meeting is at 10:00. Many theorists pursue epistemic norms of action/practical reasoning and assertion by assuming that they must account for such cases of epistemic assessment (Gerken 2013b). So, if the profiles of epistemic assessment for action and assertion are relevantly similar, it is prima facie evidence for the assumption that the relevant epistemic norms are also relevantly similar.²

The issue has substantive consequences for the epistemic norms of action/practical reasoning and assertion. If the epistemic norms for action and assertion are the same, an argument against a given norm for assertion may seem to automatically discredit the analog norm of action—and vice versa. But if the norms are distinct, such arguments will not automatically transpose. So, settling the issue about commonality will both shed light on the overall structure of epistemic norms and bear on specific proposals. As an example of the latter, I will argue that the issue raises a problem for “knowledge first” theorists who accept knowledge norms for both action and assertion.

It will be instructive to approach the issue by considering some strong arguments which have been given for each side of this debate.

2.1 Arguments against commonality

Brown coins the term ‘commonality’ to denote the claim “…that there is a common epistemic standard for assertion and practical reasoning.” (Brown 2012: 123. See p. 126 for an elaboration). I will adopt this terminology although I will make an important qualification below in Sect. 2.3.

Brown’s central aim is to problematize principled rationales for commonality: “My interest in this paper is not primarily in whether commonality is true but, whether, if it is, there is a deeper theoretical motivation for it.” (Brown 2012: 124). Indeed, her central conclusion is “…that there are significant difficulties with three main strategies for attempting to explain commonality.” (Brown 2012: 155). Hence, her arguments are primarily negative. For example, Brown argues against the idea that assertion is an act as a motivation of commonality. There are many different speech acts, and many of them are governed by different epistemic standards than those governing action/practical reasoning. So, the mere fact that an assertion is a speech act does not motivate commonality. More generally, Brown uses this point to argue against the “instance strategy” according to which the epistemic norm of assertion is an instance of the epistemic norm of practical reasoning. Furthermore, she argues against the “inheritance strategy” according to which the relevant norms are communal because they inherit the properties from a more general epistemic norm. Finally, she argues against some versions of the “licensing strategy” according to which “… one of the functions of assertion is to entitle hearers to rely on the asserted proposition in their practical reasoning” (Brown 2012: 144).

In arguing against the instance strategy, Brown articulates a concern which may be construed as a counterexample to commonality. She points out that when one meets the epistemic requirements on relying on p, one is, relative to the aim of misleading one’s audience, meeting the epistemic requirements on asserting that not-p. More generally, when one meets the epistemic requirements on relying on p, “…one can rely on p in deciding to assert p, or in deciding to assert not-p…” (Brown 2012: 131). I discuss the case of insincerity below in Sect. 3.1.

Thus, Brown not only argues against prominent rationales for commonality, in the process she also provides some positive reasons to suspect it.

2.2 Arguments for commonality

In contrast to Brown, Montminy argues that “…assertion and practical reasoning must be governed by the same rule” because “…the epistemic rule governing assertion derives from the epistemic rule governing practical reasoning together with a very plausible rule of assertion.” (Montminy 2012: 57–58)

Montminy considers the knowledge norm and argues that a number of objections to it are superficial since they can be responded to by minor adjustments in formulation. On the basis of his reformulations, he seeks to derive the knowledge norms of assertion from the knowledge norms of action and the following principle:

(MRA):

One must: assert that p only if one’s assertion manifests one’s belief that p.

Montminy’s adjustments of the knowledge norms exemplify how addressing the question concerning commonality can lead to substantive changes in the articulation of epistemic norms. Moreover, Montminy responds to Brown’s objection that it is insufficient for motivating commonality that assertion is an act.³ According to (MRA), assertion differs from other speech acts in being constitutively associated with belief, and belief is what figures in the premises of practical reasoning. While this response does not fully defend commonality, it may perhaps be said to restore the limited force of the original rationale.

Montminy’s discussion invokes knowledge norms, and his arguments articulate an important commitment for knowledge norms of action and assertion. According to Montminy, knowledge first theorists who uphold knowledge norms for both assertion and action/practical reasoning are committed to assume that the epistemic standard for action and assertion is the same. Montminy’s assumption aligns well with how proponents of knowledge accounts approach the issue. Indeed, many such theorists double down on knowledge first by embracing knowledge norms for both action/practical reasoning and assertion (Hawthorne 2004; Stanley 2005; Hawthorne and Stanley 2008. See also Fantl and McGrath 2009; 2012). Moreover, an important aspect of the knowledge first methodology is that knowledge is taken to be the central explanatory notion in epistemology and, more specifically, in the articulation of epistemic norms (Williamson 2000). So, although knowledge first theorists have not always spoken directly to the issue, Montminy’s arguments clarify a common presupposition among proponents of knowledge accounts. According to Montminy, a commitment to the knowledge norm of action/practical reasoning and (MRA) yields a rational commitment to a knowledge norm of assertion, and this suggests that there is a deep theoretical reason to uphold commonality.

However, like the rest of the debate about commonality, Montminy’s point should be seen in the light of an important distinction between two kinds of commonality.

2.3 A distinction and a corollary

As mentioned, Brown characterizes commonality in terms of a “common epistemic standard.” This claim can be understood in multiple ways. Here’s one way to understand it:

Equivalence Commonality

S is an epistemic position to act on p iff S is in an epistemic position to assert that p.

And here’s another:

Structural Commonality

The epistemic norm of action/practical reasoning and the epistemic norm of assertion have relevantly similar structures.

I think it is most charitable to interpret Brown as arguing against Equivalence Commonality. Structural Commonality can be retained even though Equivalence Commonality fails. Structural Commonality only allows and requires that similar factors can determine, in a similar way, the epistemic position required to assert and act. But rejecting that one could be in an adequate position to act on p without being in an adequate position to assert that p—and vice versa—is compatible with postulating relevantly similar structures between epistemic norms of assertion and action/practical reasoning. The structures are ‘relevantly similar’ only if their similarity is explanatory. An example of this requirement is that the structural similarity should provide a principled contribution to an explanation of the similar patterns of epistemic blame ascription. For example, epistemic norms of assertion and action/practical reasoning have relevantly similar structures insofar as both have it that the relevant epistemic requirements vary with contextual variances. The epistemic norms of assertion and action/practical reasoning that I have developed elsewhere exemplify this kind of structural commonality. According to my epistemic norm of action, the deliberative context that one is in determines the degree of warrant that is epistemically required for action (Gerken 2011a, 2013b). Likewise, my epistemic norm of assertion has it that the conversational context one is in determines the degree of warrant that is epistemically required for assertion (Gerken 2012, MS). Here the common denominator is structural. Contextual factors determine the degree of warrant that is epistemically required for both action and assertion.

This common structural property provides a principled contribution to explanations of patterns of blame ascription: Someone may be epistemically blameworthy for a certain action in one deliberative context but not in another. Likewise, someone may be epistemically blameworthy for a certain assertion in one conversational context but not in another. I return briefly to my own favorite epistemic norms of assertion and action/practical reasoning in Sect. 4.4.

However, it is important to note that Structural Commonality does not entail or presuppose my own norms. I merely invoke a key aspect of them here in order to illustrate one way in which norms of assertion and action may exhibit an explanatory structural commonality. However, there are other ways in which the norms in question may have relevantly similar structures. Structural Commonality is articulated as a fairly broad and theoretically uncommitted thesis that allows for a number of non-equivalent developments. Indeed, the thesis is compatible with a number of very different epistemic norms.

In contrast to Brown, Montminy argues in favor of Equivalence Commonality. Montminy’s arguments may be understood as articulating a rationale for the view that those who uphold knowledge norms of both action/practical reasoning and assertion have a rational commitment to Equivalence Commonality. Indeed, this claim appears to be little but an explicit expression of the view that knowledge is claimed to be necessary and sufficient for practical reasoning/action and assertion alike.⁴ We may articulate this “knowledge first commitment” (‘KFC’ for short) in the following manner:

KFC

A proponent of knowledge norms for both action/practical reasoning and assertion is rationally committed to Equivalence Commonality.

Knowledge first theorists of both assertion and action rarely advocate a commitment to Equivalence Commonality in an explicit manner. So, KFC is a diagnostic, rather than an exegetical, thesis. However, Montminy’s line of reasoning suggests that acceptance of Equivalence Commonality is a rational commitment of a knowledge-first account of both assertion and action. If knowledge is required for action as well as for assertion, it is reasonable to assume that one is in an epistemic position to assert that p if and only if one is in an epistemic position to act on p. While the thesis appears congenial to the knowledge first theorists ambition of unifying epistemic norms in terms of knowledge, it is important to note that knowledge first theories do not entail Equivalence Commonality. There are ways for a knowledge first theorist of both assertion and action to reject Equivalence Commonality. However, I will argue that the available alternative positions are very problematic (Sects. 3.3, 4.3). This is the reason for using the phrase “rational commitment” in the articulation of KFC.

A commitment to Equivalence Commonality is not unique to knowledge first theorists, however. Some but not all non-knowledge norms have a similar commitment. Smithies’ epistemic norm for both assertion and action in terms of (propositional) justification to believe that one knows that p is an example:

JK Rule

One has justification to assert and to act on P if and only if one has justification to believe that one is in a position to know that P (Smithies 2012: 270).

The JK Rule sets forth a single epistemic standard as both necessary and sufficient for assertion as well as for action.⁵ So, it appears to involve a commitment to Equivalence Commonality. But not all epistemic norms share this commitment.

3 Counterexamples to Equivalence Commonality

Counterexamples to Equivalence Commonality may serve to guide a principled diagnosis of the relationship between the epistemic norms of action/practical reasoning and assertion. So, I approach the issue via some candidate counterexamples.

3.1 Insincere assertions

As mentioned, Brown considers a case in which someone, who is in an excellent position for believing that p, asserts that not-p in order to deceive her audience (Brown 2012: 131). In such an insincerity case, the speaker may appear to meet the epistemic constraint on action/practical reasoning although she violates the epistemic norm of assertion. Relative to her aim of deceiving the audience, she is in an epistemic position to rely on p in a line of practical reasoning concluding that she should assert that not-p. But the assertion appears to violate the epistemic norm of assertion.

Are such insincerity cases counterexamples to Equivalence Commonality? A direct defense of Equivalence Commonality consists in restricting the epistemic norms of assertion. According to one such restriction, only sincere assertions are governed by the default epistemic norm. According to such a view, assertion is governed by a norm of sincerity and when this norm is violated, the epistemic norm of assertion is no longer operative (Gerken 2012, fn. 8). So, the case of insincerity would not show that the epistemic norms of action/practical reasoning are met although the epistemic norm of assertion is violated.

Could the insincerity case instead compromise Montminy’s arguments by compromising (MRA) since it is a case in which violating the principle furthers the speaker’s aims? This is not clear because the speaker’s aim may override the normal communicative aim expressed by (MRA) (see also Montminy 2012, Objection 2 and reply). In addition to the insincerity case, Brown argues contra the idea that assertion licenses the hearer’s reliance on the asserted content in practical reasoning. The assumption that the speaker should not be required to put the hearer in certain epistemic positions is central to Brown’s arguments. But since (MRA) is a norm governing the relationship between the speaker’s own belief and assertion, it is not clear that it is compromised.⁶ Since proponents of Equivalence Commonality may argue that insincerity violates other norms of assertion than epistemic norms, insincerity cases raise complex substantive and dialectical issues (Saul 2012). So, the case of insincere assertion is not a clear-cut counterexample to Equivalence Commonality or to (MRA). I therefore turn to other candidate counterexamples.

3.2 Doubly dissociating counterexamples

I will set forth two novel counterexamples to Equivalence Commonality. If successful, they doubly dissociate the epistemic position one needs to be in to act on p from the epistemic position that one needs to be in to assert that p.

Contra the right-to-left direction of Equivalence Commonality, there may be cases in which S is in an epistemic position to assert that p although she is not in an epistemic position to act on p. Here is a candidate case:

QUIZ SHOW

You and three friends are on a quiz show. Your team is given a question and four possible answers and may deliberate for up to two minutes or until the captain indicates the team’s answer by pushing one of four buttons. You are the captain in the final, all-important round in which you can lose everything or win a million dollars. Winning is crucial since the team is playing to finance a necessary million-dollar operation of a teammate’s son. You are given the question “What is the capital of Trinidad and Tobago?” Upon seeing the four options, you recall, and form the outright belief, that the answer is “C – Port of Spain.”

In this case, you may assert to your team that the capital of Trinidad and Tobago is Port of Spain. Given that you are very confident, it is appropriate to make this assertion outright rather than by qualifying it in a way that might raise doubts, lead to second-guessing etc. Note three features of the case. First, the stakes are extremely high. Second, acting is relatively non-urgent. Third, further evidence might be available from your peers (Gerken 2012, 2013b, 2011b). Given this context, it would be epistemically irresponsible to simply act and press the C button. Even though you have a very good memory, it is not infallible, and double-checking with your team is cost-free. So, QUIZ SHOW appears to illustrate that someone can be in an epistemic position that is strong enough to meet the epistemic requirements on asserting that p without being in an epistemic position to rely on p in some extra-linguistic action.⁷ I return to QUIZ SHOW in connection with the issue of isolated second-hand knowledge in Sect. 4.2.

Contra the left-to-right direction of Equivalence Commonality, here is a candidate case in which S is in an epistemic position to act on p although she is not in a position to assert that p:

SPY

You have encountered strong evidence on the basis of which you believe that your immediate superior, Ortcutt, is a spy from a competing company. However, the evidence may not be strong enough to get Ortcutt convicted. Moreover, if you assert that Ortcutt is a spy during the board meeting, there will, according to company regulations, be a formal process before further evidence may be gathered. So, you decide to act independently by setting a trap in which Ortcutt will, if things go well, expose himself in a manner that may lead directly to his arrest.

In SPY, it seems that you are not in a sufficiently strong epistemic position vis-à-vis the proposition that Ortcutt is a spy in order to assert so given that the aim of your assertion is that Ortcutt be arrested. (Below I return to the issue of aims.) Not only does your evidence for believing that Ortcutt is a spy not “hold up in court” as it stands, it will moreover be impossible to gather stronger evidence if you assert so. However, given that your evidence for believing that Ortcutt is a spy really is quite strong and underwrites your outright belief, it seems epistemically reasonable to act on this assumption by setting up the trap in question.⁸ So, I take the SPY case to illustrate that someone can be sufficiently warranted to meet the epistemic requirements on relying on p in some extra-linguistic action without being sufficiently warranted to assert that p.

QUIZ SHOW and SPY are, like most case-based counterexamples, inconclusive. But they suggest that an individual may simultaneously be in a strong enough epistemic position vis-à-vis p to assert that p without being in a strong enough epistemic position to act on p—and vice versa. Hence, the cases are candidate counterexamples to Equivalence Commonality. Below I will supplement the cases in order to cast doubt upon Equivalence Commonality in a variety of ways. But it may be worthwhile to pause to consider the consequences of rejecting Equivalence Commonality.

3.3 A corollary contra the knowledge norms

Defenders of Equivalence Commonality may object to the cases above. However, they may do so in a number of ways. Consequently, I doubt that it will be productive to try to preemptively consider the responses to the cases that are most likely to be invoked.⁹ So, until concrete responses to the cases have emerged, I will illustrate the significance of the issue by assuming that the cases are genuine counterexamples and investigate their consequences.

Given the counterexamples to Equivalence Communality and (KFC), it is problematic to uphold knowledge norms of both action/practical reasoning and assertion. If knowledge is the epistemic standard for assertion and action alike, the standards for assertion and action should not come apart in a single case. So, given such cases, knowledge first theorists are under pressure to abandon either the knowledge norms of action/practical reasoning or the knowledge norm of assertion. The ‘or’ is inclusive. In fact, SPY and QUIZ SHOW may be specified in ways that make them direct counterexamples to the knowledge norms.¹⁰ But here I only argue for the more modest disjunctive point (for direct counterexamples, Brown 2010; Reed 2010; Gerken 2011a).

Let me anticipate a response to this corollary: Some proponents of knowledge norms uphold various pragmatic encroachment theories according to which whether one knows depends on practical factors (Hawthorne 2004; Stanley 2005; Fantl and McGrath 2009, 2012). If pragmatic encroachment provides a good response to the present objection, proponents of this view would have an advantage over knowledge first theorists who reject it (notably Williamson 2000, 2005).

However, this response requires a more radical version of pragmatic encroachment than standard versions. To clarify this point, let me again make use of the phrase ‘deliberative context’ to denote the configuration of factors that determines the epistemic position vis-à-vis p one must be in to act on p or rely on p in practical reasoning. Similarly, the phrase ‘conversational context’ denotes the configuration of factors that determines the epistemic position vis-à-vis p one must be in to assert that p (the terminology is from Gerken 2011a, 2012 but does not presuppose the norms promoted there). Arguably, both deliberative and conversational contexts include practical factors such as stakes. That’s why pragmatic encroachment theorists of knowledge conclude, by appeal to knowledge norms, that the truth-values of knowledge ascriptions are partly determined by practical facts (Hawthorne and Stanley 2008; Fantl and McGrath 2009). I do not find such a view plausible. But the point here is that adopting this view does nothing to block the present objection according to which the epistemic requirements set forth by the deliberative and conversational contexts may come apart.

Some pragmatic encroachers accept that the same subject is, in two deliberative contexts that differ only in practical factors, to be regarded as a knower and a non-knower, respectively. Likewise, some pragmatic encroachers accept that the same subject is, in two conversational contexts that differ only in practical factors, to be regarded as a knower and a non-knower, respectively. But the view required to respond to the present objection is that the same subject in the same situation can know(-relative-to-assertion) and not know(-relative-to-). There are, to the best of my knowledge, no pragmatic encroachers who upheld this more radical view.

Perhaps some theorists are prepared to defend this additional view in order to preserve Equivalence Commonality. But it will not be productive to consider such a view before arguments have been advanced in its favor. In the absence of such arguments, knowledge first theorists are under pressure to abandon the knowledge norms of action/practical reasoning or the knowledge norm of assertion. This is so independently of whether they accept pragmatic encroachment or not.

While the failure of Equivalence Communality impacts knowledge accounts, many non-knowledge accounts are impacted as well. As mentioned, Smithies’ JK Rule exemplifies a non-knowledge approach that appears to be committed to Equivalence Commonality. If so, it is equally compromised by the present objection.

3.4 A first step towards a diagnosis

If QUIZ SHOW and SPY are genuine counterexamples to Equivalence Commonality, what may we learn from them? One thing to note is that in these cases, the subject’s conversational context and deliberative context differ. So, the presence of counterexamples to Equivalence Commonality is partly explained by the fact that the deliberative and conversational contexts may, in a single situation, set forth different epistemic requirements on action/practical reasoning and assertion, respectively (for a different use of similar cases, see DeRose 2009: 269ff). Even if epistemic standards for assertion and action/practical reasoning normally go hand-in-hand, the upshot that deliberative and conversational contexts sometimes diverge is significant. It reflects that the aims of action/practical reasoning and assertion may differ.

However, it is not clear that the cases indicate the principled ways in which epistemic norms of action/practical reasoning differ. After all, a single proposition can be relevant to two different actions with different stakes, and this may result in differences in required epistemic positions (Neta 2007; Rysiew 2007; Brown 2010; Reed 2010, Gerken 2011a). The SPY and QUIZ SHOW cases show that one of these acts may be an assertion. This is what makes them candidate counterexamples to Equivalence Commonality. However, if the cases instantiate a general phenomenon, they may constitute genuine counterexamples to Equivalence Commonality but fail to be particularly informative guides to the principled difference between action and assertion.¹¹ Nevertheless, the fact that contextual differences in aims are central to the cases provides an important clue. Consequently, I will pursue a more principled diagnosis of the similarities and differences between the epistemic norms of action/practical reasoning and assertion by adopting a teleological perspective.

4 The communicative aims of assertion

A teleological perspective on the epistemic norms of assertion may be taken by considering speech act theory since speech acts are partly individuated by their communicative aims. As noted, the epistemic position one must be in to assert that p is also partly determined by the assertion’s aim. Thus, the relationship between the communicative aims of assertion and the epistemic norms of assertion may shed light on how the latter may differ from the epistemic norms of action/practical reasoning.

4.1 A principled difference between assertion and action

It is standardly assumed that the communicative aims of an assertion may be quite distinct from what is literally asserted. Gricean implicatures are clear instances of this phenomenon. Consider an instance of Grice’s famous example (Grice 1989: 33)

• (1a) Mr. X is a student of yours. I wonder whether he should be admitted to graduate school.

• (1b) Mr. X has excellent handwriting.

According to Grice, the assertion (1b) implicates a proposition that is very different from its content. The Gricean framework has been superseded in many ways. For example, many theorists argue that pragmatic factors to a large extent determine not only what is implicated by an assertion but also what is asserted in the first place (Bach 1994; Carston 2002; Horn and Ward 2004; Sperber and Wilson 1986/1995; Recanati 2004). Nevertheless, there is wide agreement among contemporary pragmatists that the primary communicative function of an assertion that p may be to convey another proposition, q. So, assume, for the sake of discussion, that (1b) conveys that Mr. X is a bad student.¹²

In some such a cases, it appears that the professor who asserts (1b) does not need to be in any positive epistemic position vis-à-vis the asserted content. Rather, the professor meets the epistemic requirements on asserting (1b) insofar as she is in a sufficiently strong epistemic position vis-à-vis the implicated content. Note that I do not suggest that the professor is in every context not under the epistemic constraint associated with the asserted content. In many contexts, the professor must be in a positive epistemic position vis-à-vis both the asserted and implicated contents. But in some cases, where the professor’s communicative aim is very transparent—i.e., where it is entirely clear to all parties involved that the professor seeks to convey that Mr. X is a bad student without saying so directly—his epistemic position vis-à-vis the asserted content seems irrelevant. Indeed, the asserted content of (1b) generates the desired implicature because it is contextually irrelevant. Consequently, the professor’s epistemic position vis-à-vis the asserted content may be irrelevant as well.

Most contemporary epistemic norms have presupposed that the asserter must stand in a positive epistemic position to the asserted content. So, arguably, most proposed epistemic norms of an assertion require a restriction (for restriction strategies, Gerken MS). Here I focus on the consequences of cases of implicature for Equivalence Commonality, whereas I discuss the consequences for Structural Commonality below in Sect. 4.3.¹³

Equivalence Commonality appears to be compromised by cases of implicature. In the case of (1b), for example, the professor is not in an epistemic position to act on the proposition that Mr. X has excellent handwriting if she is not in any positive standing vis-à-vis this proposition. But given a conversational context in which (1b) clearly implicates that Mr. X is a bad student, she may be in an epistemic position to assert (1b) insofar as she is in an adequately epistemic position vis-à-vis this implicated proposition. To wit: It would be appropriate to blame the professor if she lacked evidence for believing that Mr. X is a bad student. Moreover, given this conversational context, the professor cannot defend asserting (1b) by citing her evidence pertaining to Mr. X’s handwriting.

Importantly, if an asserter blatantly flouts conversational expectations, this may be a conversational cue that the assertion serves to convey something other than its content. The fact that Mr. X’s handwriting is irrelevant can be such a cue. But so can the contextually salient fact that the professor has no evidence about Mr. X’s handwriting. In each case, the speech act may still be an assertion. At least, it is a natural and standard assumption in pragmatics that the speaker’s assertion implicates something other than its content because it flouts conversational expectations.¹⁴ Of course, there will be cases in which an asserter is epistemically responsible for both the asserted content and its implicature. Such cases may be sufficient to mark a distinctive feature of the epistemic norms of assertion. But cases in which an assertion that p implicates that q in virtue of the fact that it is contextually salient that the asserter has no clue whether p may be clearer counterexamples to Equivalence Commonality. In such cases, the asserter is in an epistemic position to assert that p, insofar as she is in an adequate epistemic position vis-à-vis q. But she is not in an epistemic position to act on p or to deploy it in practical reasoning (Gerken MS).

There may also be implicature cases in which someone is in a position to act on p/use p in practical reasoning although she is not in a position to assert that p. Assume, for example, that the professor who asserts (1b) is in an extremely strong epistemic position for believing that Mr. X’s handwriting is excellent but that she is entirely ignorant of his academic abilities. In such a case, the professor is in an epistemic position to act on the proposition that Mr. X’s handwriting is excellent. But she is not in an epistemic position to assert that Mr. X’s handwriting is excellent in a conversational context where this assertion clearly implicates that Mr. X is a bad student. It appears, then, that cases of implicature reinforce the double dissociation of epistemic standards for action/practical reasoning and assertion that compromises Equivalence Commonality.

4.2 Isolated second-hand knowledge

Let us briefly consider the intriguing case of isolated second-hand knowledge which has been discussed in relation to epistemic norms (Lackey 2011; Carter and Gordon 2011; McKinnon 2012). Roughly, isolated second-hand knowledge is knowledge that p which is based on testimony in cases where the recipient has minimal background information about the speaker and p (Lackey 2011; Gerken 2011b). First of all, it is important to note that the present conclusion is consistent with, but does not entail, the claim that isolated second-hand knowledge is sometimes insufficient for assertion (Lackey 2011; Carter and Gordon 2011).¹⁵

However, McKinnon has argued that reflection on isolated second-hand knowledge (ISHK) may provide evidence for a unified epistemic norm of action and practical reasoning: “…insofar as assertions and decisions epistemically grounded in ISHK both seem inappropriate, this is evidence for a unified norm of assertion and practical reasoning.” (McKinnon 2012: 569).¹⁶

McKinnon considers a case in which the protagonist, Jill, has isolated second-hand knowledge that a plastic dish is leak-proof. McKinnon argues that Jill is not in an epistemic position to rely on the premise that the dish is leak-proof in certain lines of practical reasoning. So, if Jill is also not in an epistemic position to assert that the dish is leak-proof, the case is one in which both assertion and decision epistemically grounded in isolated second-hand knowledge both seem inappropriate. McKinnon does not herself affirm the antecedent of this conditional. In fact, she thinks that Jill’s assertion that the dish is leak-proof would be epistemically appropriate.¹⁷ So, McKinnon’s point is only that someone who assumes that Jill’s assertion is epistemically inappropriate could gain some support for a unified norm since Jill’s decision is, according to McKinnon, epistemically inappropriate.

However, McKinnon cautiously notes that such cases would only provide abductive evidence for a unified account. Even if there are cases in which isolated second-hand knowledge is insufficient for both assertion and practical reasoning, this is only a single case type. Hence, the potential evidence for unification is limited at best.

The point that I wish to add is that even if it is granted that Jill is unable to assert and act in McKinnon’s case or other isolated second-hand cases, such a case type should not be considered in isolation. McKinnon is quite right that the evidence from such case types is abductive. In consequence, we should broaden the range of types of cases of isolated second-hand knowledge. By doing so, I will argue that isolated second-hand knowledge cases can also constitute doubly dissociating counterexamples to Equivalence Commonality although they do not thereby compromise Structural Commonality.

As a candidate case in which one’s isolated second-hand knowledge is epistemically adequate for an assertion that p but inadequate for relying on p in action/practical reasoning, we can recycle QUIZ SHOW discussed above. We just need to add the assumption that you possess isolated second-hand knowledge that Port of Spain is the capital of Trinidad and Tobago. In fact, the isolated second-hand knowledge specification appears to be a quite congenial way to articulate the case in more detail. As mentioned, isolated second-hand knowledge is sometimes invoked to argue that knowledge is insufficient for assertion (see Carter and Gordon 2011; Lackey 2011 as well as the case below). So, it is interesting that isolated second-hand knowledge can also be invoked to argue that it is, in some conversational contexts, epistemically sufficient for assertion but insufficient for action. However, the opposite may also be true.

As an example of a case in which isolated second-hand knowledge is inadequate for assertion but adequate for action/practical reasoning, we may consider a variation of McKinnon’s leak-proof case.

LEAK - PROOF CONTAINER

Gil’s colleague from the biochemistry department, Scott, is about to carry a liquid sample containing a dangerous species of bacteria across campus. He is considering which token of two types of plastic containers to use. Scott tells Gil that their colleagues have told him that both types of container are leak-proof. Since Gil has isolated second-hand testimony that container type 1 is leak-proof from the very same people as Scott, he says nothing. Eventually, Scott makes an on-line search on the basis of which he opts for the plastic container of type 2 and heads out.

Gil himself is a tad late for his physically demanding gym class and realizes that he has forgotten his water bottle. So, he quickly fills the remaining plastic container of type 1 with water, puts it in his gym bag and heads for the gym.

In this case, Gil is not in an epistemic position to assert without qualification that the plastic container of type 1 is leak-proof. Gil is unable to assert so in part because doing so would implicate that he had further evidence than the evidence that Scott already possess. After all, Gil is aware (and Scott is aware that he is aware) that Scott has already received similar testimony from the same source. So, if Gil were to assert outright that type 1 containers are leak-proof, he would, in the conversational context in question, implicate that he possessed stronger evidence or personal evidence. Again the particular feature of an assertion that it can generate implicatures contributes to generating an important asymmetry between the epistemic norms for assertion and action/practical reasoning. This is indicated by the fact that it seems unproblematic for Gil to subsequently rely on the assumption as a premise-belief in his own practical reasoning. The case trades both on the fact that there is a difference in stakes between Gil’s conversational context with Scott and his own deliberative context. But the more principled difference is the above-noted one that assertions can generate conversational implicatures that the speaker must be in a positive epistemic relation to.

So, a broader set of isolated second-hand knowledge cases appear to substantiate the present approach. That is, reflection on isolated second-hand knowledge cases support that the epistemic norms of assertion and action are unified in the sense of Structural Commonality but not in the sense of Equivalence Commonality.

4.3 A reinforced corollary contra the knowledge norms

The pragmatically driven counterexamples to Equivalence Commonality increase the problems for those who uphold knowledge norms of both action/practical reasoning and assertion insofar as they are committed to Equivalence Commonality. To avoid the unattractive view that someone can in the same situation simultaneously know(-relative-to-action) and not know(-relative-to-assertion) and vice versa, they should abandon the knowledge norms of action/practical reasoning or the knowledge norm of assertion or both.

Furthermore, the cases which constitute the counterexamples may also require a restriction of Montminy’s (MRA) principle according to which one must assert that p only if the assertion manifests one’s belief that p. In some instances of (1b), the professor’s assertion that Mr. X has excellent handwriting does not manifest her belief that this is so. This is not to say that (MRA) is entirely misguided but rather that it should be constrained to cases in which a central communicative function of an assertion that p is to convey that p. However, such a restriction would appear to block the argument from a knowledge account of action to a knowledge account of assertion and, thereby, the argument for Equivalence Commonality.

In sum, principled differences between action/practical reasoning and assertion put increasing pressure on knowledge first theorists who uphold knowledge norms for both action/practical reasoning and assertion.¹⁸

4.4 Further steps toward a diagnosis

The candidate counterexamples to Equivalence Commonality such as SPY and QUIZ SHOW involve agents with diverging aims of their actions and assertions. The cases of implicature also exemplify how different aims may drive apart the epistemic norms of assertion and action/practical reasoning. But the pragmatic case types considered above may be more principled because it is a distinctive property of the communicative aims that they may go beyond the asserted content: Whereas one does not always convey that p whenever one asserts that p, one is always acting on p whenever one is acting on p.

In consequence, the epistemic requirements on acting on p always pertain to p, whereas this is not the case for asserting that p. This difference is related to a principled difference between thought and talk.

Whereas the content of a sentence may be inexplicit with respect to the content of the thought it expresses, a thought can’t be inexplicit with respect to its own content; there can’t be more—or less—to a thought than there is to its content because a thought just is its content (Fodor 2001, p. 14).

The lesson that many theorists in pragmatics take from this principled difference is articulated by Carston: “…there is no ‘pragmatics of thought’” (Carston 2008, p. 140).

Carston’s dictum contributes to a more principled explanation of the presence of counterexamples to Equivalence Commonality. Given this principled difference between assertion and action/practical reasoning, conversational factors may affect the epistemic requirements for assertion without affecting those for action. That’s why the epistemic positions required for action/practical reasoning and assertion may diverge radically. This is contrary to Equivalence Commonality and, given (KFC), to knowledge norms of both action/practical reasoning and assertion.

The ramifications for the epistemic norms of assertion and for Structural Commonality are more complicated. Cases such as (1b) may suggest that species of the epistemic norm of assertion are individuated in part by the communicative aims of the assertion.¹⁹ Roughly, it is because (1b) aims to convey that Mr. X is a bad student that the epistemic requirement for asserting (1b) pertains to the academic ability, rather than calligraphic ability, of Mr. X. So, assertions with the main communicative function of conveying the asserted content may be governed by a different species of epistemic norm than assertions with the main communicative function of conveying an implicature. If this suggestion is right, the vast majority of contemporary epistemic norms of assertion stand in need of restriction. This is because they have presupposed that the epistemic requirements on assertion always pertain to the asserted content. So, the present suggestion would have consequences that not only concern the knowledge norm but most epistemic norms of assertion (including, I should emphasize, my own proposals. Gerken 2012, 2013b, MS).

Does the suggestion concerning the individuation of species of the epistemic norms of assertion discredit Structural Commonality? On one hand, the suggestion that epistemic norms of assertion should be species-individuated in a different way than the epistemic norms of action/practical reasoning may indicate important structural differences. If the epistemic norms of action/practical reasoning have no clear analogs to the species of epistemic norms of assertion, this suggests a principled limitation of the structural similarities between the epistemic norms of action/practical reasoning and assertion.

On the other hand, the various species of the epistemic norms of assertion are species of a more abstract genus. Importantly, the common denominator of the species of epistemic norms of assertion is plausibly shared by the epistemic norms of action/practical reasoning. This broad common denominator is that the epistemic requirements on action, practical reasoning and assertion vary with context. In some contexts, a very strong epistemic position is required. In others, a weaker epistemic position may do. Even the characteristically linguistic species of the epistemic norms share this abstract but crucial structural property²⁰ (Gerken 2012, MS). So, the present considerations and cases are compatible with Structural Commonality.

However, the assumption that the epistemic norms of assertion and action/practical reasoning are context sensitive yields significant constraints on concrete substantive norms. For example, it supports the view that the epistemic requirements on action/practical reasoning are governed by deliberative context whereas the epistemic requirements on assertion are governed by conversational context. As mentioned, specific instances of such epistemic norms are compatible with those developed in (Gerken 2011a, 2012, 2013b, MS). However, my aim is to diagnose the general issue concerning the relationship between the epistemic norms of assertion and action/practical reasoning, respectively, rather than to directly promote my favorite norms.

5 Concluding remarks

Given the distinction between Equivalence Commonality and Structural Commonality, there appear to be counterexamples to the former thesis that are unproblematic for the latter thesis. Moreover, knowledge first theorists who uphold knowledge norms of both action/practical reasoning and assertion appear to be compromised insofar as they are committed to Equivalence Commonality.

Furthermore, the discussion suggests two things about Structural Commonality. On the one hand, there are important, indeed constitutive, structural similarities between the epistemic norms of action and assertion. On the other hand, reflection on the complex aims of assertion shows that these similarities are very abstract and that a more fine-grained individuation of species of epistemic norms of assertion is, therefore, required.

These conclusions do not directly entail any particular epistemic norms of action/practical reasoning and assertion, respectively. But they compromise prominent proposals and set forth substantive restrictions on any epistemic norm of action/practical reasoning and assertion.