In their paper, Gilead et al. make an important contribution to the field of predictive cognition by leveraging construal level theory (CLT; Trope & Liberman Reference Trope and Liberman2010) as a unifying framework underlying various forms of abstract representation. In this commentary, we situate their contribution in terms of the interpersonal context. We propose that predictive models are not generated and used solely within the confines of the individual mind – instead, these are fundamentally socially constructed. We examine abstraction as an interpersonal process and argue that it serves a social function.
As Gilead and colleagues point out, the minds of other people may be highly relevant to the predictive process in that people “wish to align themselves with the beliefs of others” (sect. 3, para. 7), even for non-social stimuli, and that “contrary to theories that suggest that prediction is purely an epistemic process, we highlight that in humans, mental travel is often guided by the motivation to arrive at a state of shared belief with other humans – regardless of the truthfulness of those beliefs” (sect. 3.3, para. 1). We strongly agree with this position (Echterhoff et al. Reference Echterhoff, Higgins and Levine2009; Hardin & Higgins Reference Hardin, Higgins, Sorrentino and Higgins1996; Higgins Reference Higgins2019; Rossignac-Milon & Higgins Reference Rossignac-Milon and Higgins2018a). But further, we propose that predictive models are socially constructed more often than not. Evidence from the field of shared reality supports this idea.
Shared reality theory proposes that humans are fundamentally motivated to turn to each other in order to validate their interpretations of the world and understand what is real and true (Echterhoff & Higgins Reference Echterhoff and Higgins2017; Hardin & Higgins Reference Hardin, Higgins, Sorrentino and Higgins1996; Higgins Reference Higgins2019). Research on the Sharing-Is-Believing effect (Higgins Reference Higgins1992; for reviews, see Echterhoff & Higgins, Reference Echterhoff, Higgins, Van Lange, Kruglanski and Higginsin press; Higgins Reference Higgins2019) has shown that in communication, people tune what they say to fit with their communication partner's perspective on the object of communication (i.e., the target referent). For example, when describing a target person's behaviors to a colleague, people will categorize (i.e., abstract) the behaviors more positively (vs. negatively) if they know that their colleague likes (vs. dislikes) this person. More importantly, when speakers feel connected to their audience, they subsequently align their cognitive representations of the target to fit with the evaluative tone of what they said. Thus, they integrate their communication partner's attitude into their abstract mental representation of the target.
As described by CLT, attitudes are judgments about the “essence” of an object, which are more abstract than its concrete details (Trope & Liberman Reference Trope and Liberman2010). Critically, the new evaluatively-tuned mental representation is a distortion of the original given information about the target's behaviors. Thus, the attitudes of others can significantly shape abstraction, even at the cost of accuracy. Indeed, Gilead and colleagues touch on the uniquely human ability to disregard different dimensions in order to create intangible abstracta. These intangible abstracta are created not only to summarize information, but also (and above all) in the service of creating a common representation – a shared reality. As demonstrated in the Sharing-Is-Believing paradigm, the communicator produces an evaluatively biased message based on the audience's attitude, in the service of creating a shared abstractum that reflects their common understanding of the object. Moreover, the communicator's message can also bias the listener's understanding (Hirst & Coman Reference Hirst and Coman2018), thereby further aligning their understanding.
As suggested by Gilead and colleagues, categorical representations are primarily instantiated through social interaction. Categorization and multimodal abstracta serve a communicative purpose: when determining which level of abstractness to use (and specifically, which lemma; Roelofs Reference Roelofs1992), people take into account the perspective of their conversation partner and select the one most useful for the conversation at hand; for example, people refer to spare change as “coins” and not “round metal objects” because their monetary utility is most relevant to their conversation partner (cf. Brown Reference Brown1958). Communication goals determine the selected level of abstractness.
In fact, people may raise the level of abstractness in order to create shared realities. Indeed, Gilead and colleagues assert that the function of beliefs is not always to “accurately represent reality – but rather to facilitate traversing social distance by creating unity of minds” (sect. 3, para. 8). We agree with this point. The purpose of abstraction is not simply to minimize prediction error. Predictive models function to minimize the discrepancy not only between one's internal state and the state of the world, but also between one's internal state and that of another person. Instead of accuracy, the motivation can be to create something that “we” believe in together (Rossignac-Milon & Higgins Reference Rossignac-Milon and Higgins2018b). Although this “we” is an abstract category that subsumes “me” and “you” – thereby increasing psychological distance away from the “me here and now” (Yip-Bannicq Reference Yip-Bannicq2018) – it also decreases social distance. As CLT would predict, this decreased social distance should simultaneously collapse perceived spatial and temporal distance. The social glue created by shared reality was critical to our survival as a species in our ancestral environment (Higgins Reference Higgins2019). Thus, we argue that the evolutionary pressure to raise abstractness stems not only from accurately predicting the environment, but also from the need to create shared realities with others.
Moreover, through this process of creating shared realities, conversation partners jointly align their predictive models (e.g., Hirst & Echterhoff Reference Hirst and Echterhoff2012; for a review, see Higgins Reference Higgins2019). As mentioned by Gilead and colleagues, predictive models can be transmitted from person to person (e.g., parent to child). Another understudied avenue through which predictive models can become shared is through dynamic, dyadic alignment. In this co-construction process, a model is not transmitted from one mind to the other; instead, it is an emergent product of both minds. We believe that these co-constructed predictive models have the greatest psychological weight (see Rossignac-Milon and Higgins Reference Rossignac-Milon and Higgins2018a).
In conclusion, the field of predictive cognition disproportionately emphasizes intra-personal processes. Humans do not build predictive models of the environment in social isolation. People verify their predictive models not only by determining whether they can predict subsequent events, but also by determining whether other people agree with them. Abstraction functions to build not only accurate, but also socially shared models.
In their paper, Gilead et al. make an important contribution to the field of predictive cognition by leveraging construal level theory (CLT; Trope & Liberman Reference Trope and Liberman2010) as a unifying framework underlying various forms of abstract representation. In this commentary, we situate their contribution in terms of the interpersonal context. We propose that predictive models are not generated and used solely within the confines of the individual mind – instead, these are fundamentally socially constructed. We examine abstraction as an interpersonal process and argue that it serves a social function.
As Gilead and colleagues point out, the minds of other people may be highly relevant to the predictive process in that people “wish to align themselves with the beliefs of others” (sect. 3, para. 7), even for non-social stimuli, and that “contrary to theories that suggest that prediction is purely an epistemic process, we highlight that in humans, mental travel is often guided by the motivation to arrive at a state of shared belief with other humans – regardless of the truthfulness of those beliefs” (sect. 3.3, para. 1). We strongly agree with this position (Echterhoff et al. Reference Echterhoff, Higgins and Levine2009; Hardin & Higgins Reference Hardin, Higgins, Sorrentino and Higgins1996; Higgins Reference Higgins2019; Rossignac-Milon & Higgins Reference Rossignac-Milon and Higgins2018a). But further, we propose that predictive models are socially constructed more often than not. Evidence from the field of shared reality supports this idea.
Shared reality theory proposes that humans are fundamentally motivated to turn to each other in order to validate their interpretations of the world and understand what is real and true (Echterhoff & Higgins Reference Echterhoff and Higgins2017; Hardin & Higgins Reference Hardin, Higgins, Sorrentino and Higgins1996; Higgins Reference Higgins2019). Research on the Sharing-Is-Believing effect (Higgins Reference Higgins1992; for reviews, see Echterhoff & Higgins, Reference Echterhoff, Higgins, Van Lange, Kruglanski and Higginsin press; Higgins Reference Higgins2019) has shown that in communication, people tune what they say to fit with their communication partner's perspective on the object of communication (i.e., the target referent). For example, when describing a target person's behaviors to a colleague, people will categorize (i.e., abstract) the behaviors more positively (vs. negatively) if they know that their colleague likes (vs. dislikes) this person. More importantly, when speakers feel connected to their audience, they subsequently align their cognitive representations of the target to fit with the evaluative tone of what they said. Thus, they integrate their communication partner's attitude into their abstract mental representation of the target.
As described by CLT, attitudes are judgments about the “essence” of an object, which are more abstract than its concrete details (Trope & Liberman Reference Trope and Liberman2010). Critically, the new evaluatively-tuned mental representation is a distortion of the original given information about the target's behaviors. Thus, the attitudes of others can significantly shape abstraction, even at the cost of accuracy. Indeed, Gilead and colleagues touch on the uniquely human ability to disregard different dimensions in order to create intangible abstracta. These intangible abstracta are created not only to summarize information, but also (and above all) in the service of creating a common representation – a shared reality. As demonstrated in the Sharing-Is-Believing paradigm, the communicator produces an evaluatively biased message based on the audience's attitude, in the service of creating a shared abstractum that reflects their common understanding of the object. Moreover, the communicator's message can also bias the listener's understanding (Hirst & Coman Reference Hirst and Coman2018), thereby further aligning their understanding.
As suggested by Gilead and colleagues, categorical representations are primarily instantiated through social interaction. Categorization and multimodal abstracta serve a communicative purpose: when determining which level of abstractness to use (and specifically, which lemma; Roelofs Reference Roelofs1992), people take into account the perspective of their conversation partner and select the one most useful for the conversation at hand; for example, people refer to spare change as “coins” and not “round metal objects” because their monetary utility is most relevant to their conversation partner (cf. Brown Reference Brown1958). Communication goals determine the selected level of abstractness.
In fact, people may raise the level of abstractness in order to create shared realities. Indeed, Gilead and colleagues assert that the function of beliefs is not always to “accurately represent reality – but rather to facilitate traversing social distance by creating unity of minds” (sect. 3, para. 8). We agree with this point. The purpose of abstraction is not simply to minimize prediction error. Predictive models function to minimize the discrepancy not only between one's internal state and the state of the world, but also between one's internal state and that of another person. Instead of accuracy, the motivation can be to create something that “we” believe in together (Rossignac-Milon & Higgins Reference Rossignac-Milon and Higgins2018b). Although this “we” is an abstract category that subsumes “me” and “you” – thereby increasing psychological distance away from the “me here and now” (Yip-Bannicq Reference Yip-Bannicq2018) – it also decreases social distance. As CLT would predict, this decreased social distance should simultaneously collapse perceived spatial and temporal distance. The social glue created by shared reality was critical to our survival as a species in our ancestral environment (Higgins Reference Higgins2019). Thus, we argue that the evolutionary pressure to raise abstractness stems not only from accurately predicting the environment, but also from the need to create shared realities with others.
Moreover, through this process of creating shared realities, conversation partners jointly align their predictive models (e.g., Hirst & Echterhoff Reference Hirst and Echterhoff2012; for a review, see Higgins Reference Higgins2019). As mentioned by Gilead and colleagues, predictive models can be transmitted from person to person (e.g., parent to child). Another understudied avenue through which predictive models can become shared is through dynamic, dyadic alignment. In this co-construction process, a model is not transmitted from one mind to the other; instead, it is an emergent product of both minds. We believe that these co-constructed predictive models have the greatest psychological weight (see Rossignac-Milon and Higgins Reference Rossignac-Milon and Higgins2018a).
In conclusion, the field of predictive cognition disproportionately emphasizes intra-personal processes. Humans do not build predictive models of the environment in social isolation. People verify their predictive models not only by determining whether they can predict subsequent events, but also by determining whether other people agree with them. Abstraction functions to build not only accurate, but also socially shared models.
Acknowledgment
This work was supported by the Anneliese Maier Research Award from the Alexander von Humboldt Foundation awarded to E. T. Higgins.