Bermúdez presents a persuasive case that framing effects are not as “irrational” as commonly supposed. It is interesting to consider intra-individual variation in preferences over time and across contexts as a kind of framing effect, where complex decision-making is cast as an iterative process of reasoning from different perspectives. Yet by focusing on the reasoning of individual decision-makers, Bermúdez's account neglects the crucial role of the social-communicative context in explaining why (at least some) framing effects arise. Language is the central medium for communicating our beliefs and attitudes and persuading others to adopt them. We argue, as a result, that many framing effects are best explained in terms of basic principles of discourse processing and pragmatic reasoning. This framework highlights a key mechanism by which framing operates: Subtle linguistic cues communicate the speaker's knowledge and perspective on a target problem, and decision-makers rely on those cues to draw reasonable inferences about the problem. Therefore, even seemingly “simple” framing effects are rational.

To differentiate his account from the existing literature, Bermúdez describes certain “classic” framing effects as the consequence of a basic “priming” mechanism, where exposure to a frame “activates” a dimension/attribute of the target problem, driving reasoning. This may be a textbook account of framing – and a useful way to frame the target article – but it paints an oversimplified picture of how people process language. It also fails to capture certain findings in the framing literature. For example, much research has shown that framing social issues using metaphors can shape attitudes in a metaphor-congruent fashion (e.g., Thibodeau & Boroditsky, Reference Thibodeau and Boroditsky2011; Thibodeau, Crow, & Flusberg, Reference Thibodeau, Crow and Flusberg2017). When people read a news story that frames crime as a beast (vs. a virus) ravaging a city, they are more likely to propose enforcement-related solutions to the crime problem that are consistent with how people would address a literal beast problem (Thibodeau & Boroditsky, Reference Thibodeau and Boroditsky2011). In these same studies, however, simply priming participants with the metaphorical source domain (beast or virus) has no effect on their responses. Rather, the metaphor must be used in context to describe the social issue in order to impact reasoning. These findings situate common framing effects under the rubric of basic discourse processing (Graesser, Millis, & Zwaan, Reference Graesser, Millis and Zwaan1997; Thibodeau & Flusberg, Reference Thibodeau, Flusberg, Colston, Matlock and Steenin press; Zwaan & Radvansky, Reference Zwaan and Radvansky1998). Language comprehension involves dynamically integrating linguistic input with prior knowledge to generate a mental representation of the topic of discussion. When the topic is unfamiliar, abstract, or complicated – like crime – metaphors serve as useful scaffolding, structuring the listener's representation of the target domain. While exposure to different metaphors may result in different representations, this is a rational response to (subtle) variation in message content – analogous to the quasi-cycles of iterated reasoning Bermúdez describes for individual decision-makers.

Effective language processing also requires that listeners make certain assumptions about the communicative intentions of the speaker. For example, listeners infer that specific words and phrases were chosen because they are relevant and informative (Goodman & Frank, Reference Goodman and Frank2016; Grice, Reference Grice, Cole and Morgan1975; Sperber & Wilson, Reference Sperber and Wilson1986). Recent evidence suggests that this ability to “read between the lines” and grasp the pragmatic implications of a linguistic frame is critical for many framing effects to obtain (e.g., Flusberg et al., Reference Flusberg, van der Vord, Husney and Holmes2022; Holmes, Doherty, & Flusberg, Reference Holmes, Doherty and Flusberg2021; Leong, McKenzie, Sher, & Müller-Trede, Reference Leong, McKenzie, Sher and Müller-Trede2017). In one set of studies, we examined the impact of “victim framing” on attitudes toward sexual assault. Participants read a news report that described an alleged sexual assault, often in vivid detail. The report also included a quote from a friend, reflected in the headline, that framed either the accuser as the victim (of assault) or the alleged perpetrator as the victim (of false accusations). Relative to a baseline condition, participants expressed more support for the victim-framed character and less support for the other character. However, this was only the case for those who explicitly cited the framing language as influencing their evaluations – suggesting they surmised that the writer chose to cast one individual as a victim for good reason (i.e., to signal who deserves support; Flusberg et al., Reference Flusberg, van der Vord, Husney and Holmes2022).

In another set of studies, we assessed people's ability to pick up on the pragmatic implications of subject–complement statements of equality. Sentences like “girls are just as good as boys at math” appear to express an equivalence between two social groups, yet people tend to infer that the group in the complement position – in this case, “boys” – is superior (Chestnut & Markman, Reference Chestnut and Markman2018). As a result, these sentences can perpetuate, counteract, and even generate new stereotypes in framing studies that manipulate which groups occupy the subject versus complement positions (Chestnut & Markman, Reference Chestnut and Markman2018; Chestnut, Zhang, & Markman, Reference Chestnut, Zhang and Markman2021; Holmes et al., Reference Holmes, Doherty and Flusberg2021). In a recent study, we measured participants' ability to discern the pragmatics of this syntax by asking them, for example, to infer the beliefs of a journalist who uses a particular subject–complement statement of equality (e.g., “Balurians are just as good as Arigans at cooking” implies that the journalist believes Arigans are the superior chefs). Only those who could successfully recognize these subtle pragmatics showed significant framing effects in an experiment that used similar statements to frame the math abilities of various social groups (e.g., “children from Wyoming do just as well as children from Montana at math”; Holmes et al., Reference Holmes, Wu, Elpers, Doherty and Flusbergin prep; Wu, Elpers, Doherty, Flusberg, & Holmes, Reference Wu, Elpers, Doherty, Flusberg and Holmes2021). This is consistent with other work showing that even logically equivalent frames (e.g., a basketball player who “makes 40%” vs. “misses 60%” of his shots) communicate subtly different speaker appraisals, which sensitive listeners readily incorporate into their decision-making (e.g., Leong et al., Reference Leong, McKenzie, Sher and Müller-Trede2017; McKenzie & Nelson, Reference McKenzie and Nelson2003; Sher & McKenzie, Reference Sher and McKenzie2006).

Taken together, such findings suggest a rational basis for seemingly simple framing effects: Decision-makers infer that specific labels or syntactic constructions communicate relevant information about the target issue and – quite sensibly – use this information in the course of their decision-making. Iterative, quasi-cycles of reasoning about complex situations, while fascinating, are not necessary to reveal the rationality of framing.