Grzegorz Gaszczyk
Affiliation: University of Warsaw, Faculty of Philosophy
Category: Philosophy
Keywords: assertion, lying, informativeness, speech acts, common ground
Date: Tuesday 2nd of September
Time: 18:00
Location: GSSR Plenary Hall (268)
View the full session: Communicating the Facts
Traditionally, assertions are considered to be essentially informative speech acts, i.e., speech acts that deliver new information to the common ground. However, this view has recently been heavily criticized. The goal of this paper is to defend the traditional view against one of the critiques, according to which the aim of an assertion is not to update the common ground with “p,” but merely with “S said that p.” I argue that such cases should be analyzed either as defective assertions or as non-assertoric speech acts. In conclusion, I argue why the traditional picture, which places informativeness of assertion at the core of assertoric practice, remains the most viable option.
I. The informativeness of assertions While the informativeness of assertion has been understood in many ways, it can be formulated as a view such that a proper assertion conveys new information. Let us call this the Informativeness Hypothesis (IH). One of the consequences of IH is that an uninformative assertion—i.e., one that merely repeats already known information—is considered improper. Since information exchange is one of the primary aims of our communication, it makes sense that the central speech act in our communication serves to transmit information. The support for IH comes from traditional theories in pragmatics, speech act theory, and linguistics (e.g., Searle 1969; Stalnaker 1978; Farkas 2022; cf. GarcĂa-Carpintero 2004, 2020; Hinchman 2013). Further support stems from experimental data, which indicate a bias toward informativeness (Rohde et al. 2021; Rohde et al. 2022), and conversational data, which show that one can be criticised for asserting something that is commonly known (GarcĂa-Carpintero 2020).
II. A critique of IH While there are strong reasons to accept IH, uninformative communication appears to be ubiquitous. We often remind, repeat, corroborate, or assure others about something they may already know. In certain contexts, corroborating already known information may even be considered obligatory (cf. Brown 2008). Such uninformative linguistic practices have recently been taken as evidence against IH. As a result, it has been argued that the central aspect of assertoric practice is not informativeness but something else; the proposed options include being relevant or cooperative (Rubio-Fernández 2019), coordinating commitments (Geurts 2019), changing the audience’s beliefs or perspectives (Willard-Kyle 2021), directing attention (Keiser 2023), or merely adding to the common ground that the speaker said that p (López 2023; cf. Abbot 2008; Krstić 2022). Experimental literature also suggests that saying something already known may be rational (Degen et al. 2020), or more efficient (Rubio-Fernández 2016). Because IH has been criticized from many directions, I limit my discussion to one type of case, which López (2023; cf. Krstić 2022) calls minimal assertions. They are called minimal because the speaker does not intend to add “p” to the common ground—instead, they merely aim to add “S said that p.” Consequently, the argument goes, minimal assertions are not subject to the standard requirements we impose on assertions, such as norms of assertion (cf. Williamson 2000). The argument rests on the purported cases of minimal assertions. Here are two such examples. The first is bald-faced lies, i.e., lies in which one says something commonly known to be false (Sorensen 2007). Since the audience knows that one is lying, what goes to the common ground is “S said that p.” The second is exam answers, i.e., answers that a student gives during an exam. Since the professor already knows the answer—as in the previous example—what goes to the common ground is “S said that p.” Because such assertions do not update the common ground with a new proposition “p,” regardless of whether they are sincere or not, they are uninformative. Importantly, López (2023) maintains that they are successful assertions.
III. A defence of IH I argue that the category of minimal assertions is ill-conceived. Some of its cases are genuine assertions and thus subject to the standard assertoric requirements, while others are better theorized as non-assertoric speech acts. Consider bald-faced lies. One of the criteria for being an assertion is lie-proneness (Stainton 2016). If so, then because bald-faced lies are lies, they are assertions. Another criterion is criticizability (Williamson 2000). Even though the audience knows that a bald-faced liar says something false, they can still be criticised for it—this is only possible if we classify bald-faced lies as assertions. Such assertions may have some secondary effects (cf. Stalnaker 1978), but since they fail IH, they are defective. The standard requirements for being an assertion concern the default acts performed using the declarative mood (cf. Williamson 2000; GarcĂa-Carpintero 2004). Similar to standard norms of assertions, IH only operates in default contexts. Thus, the above criteria do not apply to speech acts that (i) are subject to weaker norms, such as guesses, and (ii) are performed in non-default contexts, such as acting or teaching. I propose to treat the second case of minimal assertions—exam answers—as an instance of the latter. Consider that it would be unnatural to accuse the student of lying during an exam. The student may still be criticized for giving a false answer, but for reasons different than those applicable to assertions; for instance, the professor might complain that the student did not prepare for the exam. Consequently, because exam answers are performed in non-default contexts, they cannot be criticized for being uninformative. My proposal is corroborated by the analysis of question-answer pairs. Different types of interrogative acts licence different types of declarative acts. In default contexts, we ask knowledge-seeking questions (Whitcomb 2017). Assertions are best suited to answer such questions because they are knowledge-giving speech acts (Kelp and Simion 2021). This partly explains why bald-faced lies are criticisable as improper assertions, i.e., they are uninformative. However, exam answers are not expected to be informative—the professor does not want to know whether p, but rather to verify whether the student knows that p. Thus, exam questions licence a different type of answer than the default way of using the interrogative mood (cf. Farkas 2022). In sum, the category of minimal assertions does not play any explanatory role.