Sharing Knowledge: A Functionalist Account of Assertion

In this excellent book, Christoph Kelp and Mona Simion defend an etiological-functionalist account of the normativity of assertion. Specifically, the etiological function of assertion is to generate knowledge in hearers. Kelp and Simion argue that this functionalist thesis has two important implications: a) that epistemically good assertions are those that are disposed to generate knowledge in hearers, and b) that epistemically permissible assertions are those that conform to the Knowledge Rule of Assertion (KRA). One important feature of the book is a sophisticated defense of KRA by means of an etiological-functionalist framework.The book is lucidly written, rigorously argued, informed, and original. All of this is as expected, coming from two authors who have already made significant contributions to the topic. Indeed, the book draws on both their single-authored and coauthored work on related issues, including the normativity of assertion, the epistemology of testimony, epistemic norms, and epistemic normativity. The result is that an etiological-functionalist account of assertion is now front and center among competing views.The book also contains several other interesting and original discussions, all by way of applying the authors’ etiological-functionalist framework to prominent issues in the literature on assertion and beyond. These include a chapter on epistemic injustice (defending a duty to believe on the part of hearers), a chapter on whether there is a constitutive rule of assertion (rejecting Williamson’s strong constitution thesis, but defending a weaker version), a chapter on contextualism (arguing, contra DeRose, that KRA counts against a contextualist semantics of knowledge attributions), and an appendix on the value of knowledge (offering an original defense of our concern with knowledge, as well as a functionalist account of knowledge’s value). In my judgment, the book advances discussions on all of these topics in important and original ways.In the remainder of the review, I clarify what Kelp and Simion mean by a functionalist account of assertion, and I review what I take to be the authors’ most interesting arguments in favor of KRA and their etiological-functionalist account. I end by raising some questions about the view that results.Kelp and Simion tell us that they are arguing for a “function first” account of the normativity of assertion. Three points of clarifications are important here. First, this an account of the epistemic normativity of assertion, that is, an account of what makes an assertion epistemically (as opposed to practically, morally, or all-things-considered) good, proper, or right. Second, by “function,” the authors mean etiological function. This makes their account a competitor with ones that ground the normativity of assertion in human intentions (or what I will call the intended function of assertion). For example, it is a competitor with accounts that ground the normativity of assertion in human intentions embedded in speech acts or in other social practices or norms. Putting these two points together, Kelp and Simion are arguing for an etiological account of the epistemic normativity of assertion.Third, Kelp and Simion think that the normativity of assertion is multidimensional in the following way: there is a distinction between an assertion’s being good (or valuable) and its being permissible. As such, they argue, an adequate account of the normativity of assertion must address both of these dimensions. Specifically, rules of assertion such as KRA address the epistemic permissibility question. KRA says one must (epistemically) assert p only if one knows p. Alternatively, it is (epistemically) permissible to assert p only if one knows p. But this leaves open the question of the goodness of assertion—what makes an assertion epistemically good or valuable. Their function first account primarily answers this second question about goodness, and then derives an answer to the permissibility question from that. In turn, they argue, alternative accounts tend to conflate the distinction between these two dimensions for normativity, either assuming that good assertions must be permissible assertions or missing the distinction altogether.Regarding the epistemic permissibility of assertion, Kelp and Simion defend KRA in two stages. In the first stage, which comprises part 1 of the book, they look at the main arguments for and against KRA in the literature. Here they introduce a piece of theoretical machinery—their “Basic Normative Framework”—that I think moves discussion forward. In particular, a number of objections to KRA take the form of an apparent counterexample: a person S asserts p in a case where a) S does not know p, and yet b) S’s assertion seems permissible nonetheless. A common move by defenders of KRA is to interpret the case as a blameless norm violation. For example, KRA is indeed governing S’s assertion and violated, but S has an excuse. Objectors to KRA complain that this kind of move is ad hoc, inappropriately insulating KRA from apparent counterexamples. In the context of this dialectic, Kelp and Simion introduce an independently plausible framework for understanding criticism and blame, and they then show how the framework predicts patterns of criticism, apology, explanation, and excuse that favor KRA over other accounts. The Basic Normative Framework and the discussion around its application are too complex to present here. I will only register my agreement with Kelp and Simion on two central points: a) that their Basic Normative Framework is independently plausible and b) that it speaks decisively in favor of KRA in terms of how it predicts relevant conversational patterns.The second stage of KRA’s defense takes place in part 2 of the book, in the context of a broader defense of Kelp and Simion’s etiological-functionalist framework. Specifically, chapter 4 articulates and defends a functionalist account of the goodness of assertion. Roughly, one’s assertion is (epistemically) good just in case it has the disposition to generate knowledge in one’s hearers. Chapter 5 then shows how KRA, which speaks to the permissibility of assertion, can be derived from the account in chapter 4. The arguments of both chapters are central to the purposes of the book, and are of independent interest as well.The burden of chapter 4 is to establish the following evaluative norm:The argument begins by endorsing a standard account of etiological function:Kelp and Simon introduce some additional theoretical machinery that complicates the standard account and is employed to argue for their Evaluative Norm of Assertion. Specifically, a “simple economic system (SES)” is one that is constituted by a type of producer, a type of product, a type of consumer, and a type of return. For example, consider the practice of producing and consuming espresso at a local coffee shop (SES-Espresso). In SES-Espresso, baristas (producers) produce cups of espresso (product), which are consumed by customers (consumers) by means of a purchase (return). By E-function, tokens of espresso have the e-function of producing pleasant gustatory experiences in customers in SES-Espresso.Next, Kelp and Simion defend the following general evaluative norm:By this general evaluative norm, cups of espresso that have the disposition to produce pleasant gustatory experiences are good cups of espresso.The next step of the argument is to apply this machinery to assertions. Specifically, Kelp and Simion argue that our practice of making assertions qualifies as a simple economic system. In SES-Assertion, speakers (producers) produce assertions (products), which are consumed by hearers (consumers) by means of belief (return). By E-function, assertions have the e-function of generating knowledge in hearers. Finally, by means of the general evaluative norm for tokens with an e-function, we get The Evaluative Norm of Assertion.The task of chapter 5 is to provide a functionalist rationale for KRA, that is, to explain from a functionalist perspective why KRA should govern the permissibility of assertion. Kelp and Simion begin by describing a kind of situation where it makes sense to govern e-tokens by some rule or other. This is a situation where a) it matters that tokens of the type fulfill their e-function reliably (Reliability), b) there are numerous possible ways of producing tokens of the type, with varying reliability regarding whether tokens so produced fulfill their e-function (Variation), and c) tokens of the type are produced by human agents who are criticism-averse (Humanity). In such a situation, Kelp and Simion argue, it makes sense to govern the production of tokens by means of rules that promote reliability regarding e-function fulfillment. Moreover, they argue, SES-Assertion satisfies these conditions, so it makes sense that assertions should be governed by such a rule. The next step of the argument is to show that assertions should be governed by KRA in particular. That is, to show that governing assertions by KRA promotes the reliability with which assertions generate knowledge in hearers. Kelp and Simion offer several reasons for supposing that it does, thereby vindicating KRA from a functionalist perspective.First, why do Kelp and Simion think that the function of assertion is to generate knowledge rather than transmit it? The question arises because, on their own account, “cases in which testimony transmits knowledge predominate and cases in which testimony generates knowledge are rare exceptions” (114). But if generation cases are rare, it is unclear how assertions can have “the disposition to generate knowledge that p in one’s hearer(s) … by functioning normally in normal conditions.” (See the Evaluative Norm of Assertion, above). It is also unclear how knowledge generation “contributes to the explanation of why” assertions exist in SES-Assertion. (See E-function, above.) By their own reasoning, then, Kelp and Simion should think that the function of assertion is to transmit knowledge and that it is the disposition to fulfill this function that makes a token assertion good.Second, is FFAA supposed to give a full account or a partial account of the epistemic goodness of assertion? If it is supposed to give a full account, then FFAA is implausible. That is because assertions plausibly have an intended function as well as an etiological function, and it is plausible that intended function fulfillment also contributes to assertion’s epistemic goodness. For example, various authors have argued that the speech act of assertion is constituted by the intention to transmit knowledge, to give a reason to believe, or to give license to believe. If so, then it is plausible that the epistemic goodness of assertions is at least partly a matter of fulfilling their intended epistemic function. Suppose, then, that FFAA is supposed to give only a partial account of the epistemic goodness of assertion. In that case, it is unclear why FAA is a competitor to “intended function” accounts, including those that ground the goodness of assertion in the epistemic intentions embedded in constitutive rules, linguistic norms, or other social norms governing the practice of assertion.Suppose Kelp and Simion were to revise their view so as to accommodate these concerns. Many of their arguments and much of their view would be left intact, including their functionalist defense of KRA. And we would still have a compelling (now partial) account of the goodness of assertion, in terms of assertion’s etiological function of transmitting knowledge.