Modeling the Normativity of Joint Commitments

Abstract

This paper discusses current debates about the normativity of joint commitments in analytical social ontology from a methodological point of view. I argue that collective-acceptance-based, critical-emancipatory, and experimental-philosophy-based approaches to social ontology all face challenges in accommodating the dynamic and strategic aspects of joint commitment. To make up for their shortcomings, I draw on evolutionary game theory and mindshaping approaches to social cognition, while distinguishing between the mental models that individuals use for reasoning about joint commitments and the theoreticians’ models that philosophers construct for studying select aspects of social reality.

1. Contentious Commitments

Commitments are made, not born with. But it matters enormously to each of us by whom, for whom, with what content, and in what context commitments are made, especially when they are of the interpersonal kind. This is because there is a strategic aspect to commitment: I have an interest in having you committed to matters that are important to me, and you have an interest in having me committed to matters that are of importance to you, and each of us has an incentive to be judiciously stingy with one’s own commitments, as too many commitments can lead a life astray and result in a failure to live up to any of one’s commitments, including the ones that truly matter. At least this is so for agents with bounded resources, as spreading oneself too thin may lead to diminished success on multiple fronts of one’s personal and professional life (as the briefest reflection on conflicting commitments between one’s career and family life will make clear to many of us).

Shared commitments have become a topic of major interest in analytical social ontology (which is largely based on the notion of collective intentionality—see Sarkia [2022]) during recent years. However, I think that the methodological tools of current social ontology remain inadequate for addressing this complex topic. This is because most social ontologists assume that joint commitments are an outcome of a more or less harmonious process of collective acceptance or recognition, where shared social norms underlying social practices are mutually recognized by each member of society (see e.g., Gilbert 2014; Searle 2010; Tuomela 2013). According to revisionary critical-emancipatory approaches to social ontology, this appearance of harmony is merely a tool that is used by the powerful to keep the oppressed in bay (Àsta 2018; Burman 2023; Haslanger 2000 [2013]; Jenkins 2023; Richardson 2023). However, they often end up repeating the mistakes of their collective-acceptance-based counterparts by relying on monolithic approaches to ideology and social power, which are not able to model the dynamic and strategic aspects of joint commitment.

The nearly universal methodological strategy of contemporary social ontology has been to start with simple dyadic interactions between individual agents and scale up from such basic interactions to larger-scale social groups, norms, and institutions (see esp. Bratman 1999, 2014). For purposes of exposition, I will also follow this strategy, although I will later criticize it for not being a genuinely bottom-up approach. Accordingly, I will begin the next section by discussing the traditional distinction between normativists and non-normativists in the study of joint action, including recent methodologically innovative work in experimental philosophy in defense of the normativist paradigm (Gomez-Lavin and Rachar 2019, 2022; Michael and Butterfill 2024). While granting that experimental philosophers have moved the field forward, I will argue that they fail to settle the debate about the normativity of joint commitments, because individuals can use a multiplicity of different mental models for reasoning about joint activities, and the methods of experimental philosophy tap into only a subset of these cognitive resources. I will also distinguish between such mental models and the theoreticians’ models that philosophers and social scientists construct to study select aspects of social reality. Challenging the tacit premise of many experimental philosophers and Weberian-interpretivist social ontologists (see esp. Gilbert 1990, 2014), I will argue that the latter may, but need not, correspond to the mental models that individuals use when coordinating their joint activities.

In the third section, I will turn my attention to the strategy of scaling up from the joint commitments that are exhibited in basic social interactions to larger-scale social norms and institutional reality. After criticizing traditional approaches to social ontology for imposing descriptively thick and theory-laden frameworks on joint action phenomena, I will argue that collective-acceptance-based accounts reify existing power relations, because they do not provide the tools for formulating counterfactuals related to the content of collective acceptance—in other words, why one social norm rather than another prevails in a society. While critical-emancipatory approaches to social ontology correctly diagnose this problem, they are challenged by the difficulty of justifying the privileged viewpoint of the social theoretician in identifying unjust power relations and negotiating the viewpoints of different groups of marginalized individuals. I will suggest that evolutionary game theory provides a promising set of tools for addressing the dynamic aspects of social power—but because of their prior commitments, critical-emancipatory social ontologists are unlikely to endorse this crucial set of resources.

In the fourth section, I will return to the strategic aspects of commitment, and argue that both collective-acceptance-based and critical-emancipatory approaches to social ontology overintellectualize the formation of joint commitments. Consequently, they fail to pay attention to the partly subconscious processes of mutual alignment and persuasion, which bear on the structure and contents of what individuals become committed to. Drawing on the mindshaping approach to social cognition, I will argue that individuals are often inclined to form commitments strategically, even when they are not consciously aware of it, and their interests may be in tension with one another, even when they end up settling on shared practical goals. Thus I may have reason to make demands of you that you do not have reason to acquiesce to, and individuals may rationally disagree about what they are jointly committed to. In this sense, it seems to be in the nature of commitments to be contested and fought over. This speaks to the continued relevance of social criticism, as well as the importance of providing it with appropriate, descriptively and explanatorily adequate methodological tools.

2. Intuitions About Commitment

Discussions of the normativity of shared commitments in traditional social ontology have often revolved around the distinction between normativism and non-normativism in the study of joint action (Löhr 2022). Normativists, such as Margaret Gilbert (1990, 2014), have argued that joint actions are intrinsically normative, and give rise to mutual rights and obligations between the participants to joint action. Non-normativists, such as Michael Bratman (1999, 2014), have viewed the types of rights and obligations that arise in the context of joint action as an outcome of extrinsic and circumstantial considerations, which are only contingently involved in joint action, such as mutual assurances and promise making. Traditionally, the debate between normativists and non-normativists has been carried out primarily by means of philosophical intuitions about imaginary cases, such as two people walking alongside one another (Gilbert 1990) or painting a house together (Bratman 1999), although they have sometimes also been informed by (a somewhat selective reading of) ¹ empirical results from fields such as developmental psychology and evolutionary anthropology (e.g., Tomasello 2019).

For a long time, non-normativists seemed to have the upper hand in this debate. The non-normativist position was justified by conceptual intuitions and domain-general considerations of explanatory parsimony and ontological simplicity, which are taken to apply as general rules of investigation across a variety of domains of philosophical investigation (Bratman 1999). Given that the normativist position involved the introduction of new and irreducible notions, it was considered preferable to explain shared commitments with the explanatory resources of standard philosophical action theory, such as individual intentions and beliefs, together with relevant features of the decision-making context. The normativists, on the other hand, justified their non-reductive position by a Weberian-interpretivist approach to social science, where consistency with the intuitions of ordinary people (especially the participants to joint action) was seen as an indispensable adequacy criterion for a successful philosophical account of joint commitment (see esp. Gilbert 1990, 2014). However, until recently, there was little attempt to consult those unrefined folk intuitions directly—rather, the (supposedly) expert intuitions of philosophers were taken as a proxy for what is tacit in folk intuitions.

This situation was overturned by methodologically innovative work in experimental philosophy on the intuitions of ordinary people regarding the structure of joint commitments (Gomez-Lavin and Rachar 2019, 2022, 2023; Michael and Butterfill 2024). In a landmark study, Gomez-Lavin and Rachar (2019) introduced a number of naïve participants recruited from Amazon’s Mechanical Turk platform to a series of vignettes describing the types of cases that normativists have taken to support the existence of irreducible joint commitments. Their test group shared Gilbert’s intuition that participants to joint action should notify the other person of their decision to disengage from a joint action, but not that they are obliged to seek their permission to do so (in the absence of explicit promising). In a separate study, John Michael and Stephen Butterfill obtained results that were largely consistent with this interpretation, finding that naïve participants’ responses to questions about shared intention predicted their responses to questions about joint commitment, but not about mutual obligations (Michael and Butterfill 2024). These studies have often been interpreted as indicating that a more tempered type of normativism may be needed in order to capture folk intuitions about joint action—giving up Gilbert’s emphasis on mutual rights and obligations, while still holding on to the basic conceptual connection between joint action and joint commitment. ²

The broader methodological question that arises as a result of these studies is the extent to which he conceptual intuitions of expert philosophers or ordinary people can be used to settle disagreements about the normativity of joint actions. Following this line of inquiry, Löhr (2022) suggests that intuitions about joint commitment may not be stable across alternative philosophical scenarios, and that more research is needed on the influence of tacit cultural factors on perceptions of normativity. Indeed, cross-cultural variation has been identified in other domains of philosophical investigation (e.g., Machery et al. 2004), and it seems reasonable to expect the same degree of cross-cultural variation to occur with respect to intuitions about the paradigmatically social phenomenon of joint commitment. Indeed, even within the same cultural context, a number of situational cues seem to influence perceptions of normativity (Michael 2022), such as the degree of effort expended by one’s cooperative partner (Székely and Michael 2018), the degree of mutual reliance (Michael 2022, 51-53), whether the joint action is part of a repeated pattern (Guala and Mittone 2010), and the degree of coordination between the participants (Michael et al. 2016). The diversity of such modulating factors suggests that a wholly unified account of the normativity of joint commitments may not be a reasonable goal to begin with.

Why might intuitions about joint commitment vary across different circumstances? One possibility is that individuals have a fine-grained, if implicit theory of interpersonal normativity (cf. Gopnik and Meltzoff 1997), which they rely on to assess which features of a joint action bring about normative ties between the participants—e.g., the number of repetitions that are needed in order to turn a recurrent joint action into an element of a stable pattern (Michael 2022, 51-53). However, it seems more plausible that individuals use different mental models for reasoning about joint action in different circumstances, and these models cue different perceptions of normativity (cf. Christensen and Michael 2016; Michael 2022). For example, physically co-present individuals acting simultaneously with their cooperative partners will partly rely on the embodied mental models of action simulation and perception-action matching, while individuals who are acting independently may need to rely on some form of theoretical inference grounded in the explicit ascription of propositional attitudes (see Knoblich et al. 2011; Vesper et al. 2016). Thus the degree to which individuals perceive a joint action as normative may depend in part on the mental models that they bring to bear on the situation. Importantly, such mental models need not be the same for those engaging in a joint action and those observing it from the outside, nor need they be the same for all participants in joint action, as actor-observer asymmetries in other areas of social cognition indicate (Malle 2011 [2022]).

To illustrate this challenge to the experimental philosophy program, consider the influence of physical entrainment (acting in synchrony with a cooperative partner) on perceptions of normativity (Cross et al. 2019; Vesper et al. 2016). From a theoretical perspective, entrainment does not seem like the type of feature that should influence the normativity of joint actions, as it does not change either the pursuit-worthiness of the participants’ goal nor the strength of their interdependence (or any other feature of obvious normative importance). However, Michael et al. (2016) found that individuals estimated the degree of mutual commitment to be higher between joint action partners, who acted in physical synchrony with one another. This effect of entrainment on normative cognition cannot be easily captured by experimental philosophy vignettes, which rely on textual descriptions of imaginary cases, and therefore seem to tap into only the more theory-driven and linguistically embedded mental models that individuals use for reasoning about social normativity. Consequently, the perspective that they provide on the normativity of joint commitments is relatively detached and impersonal, overlooking the significance of actor-observer asymmetries, physical entrainment, and other contextual influences on perceptions of normativity.

The experimental philosophy approach to joint commitment is also constrained by the “psychologistic” assumption that philosophical accounts of joint commitment ultimately stand or fall in line with their correspondence to individuals’ psychological states. This assumption seems to tip the scales prematurely in favor of the normativists, who have often emphasized their allegiance to a Weberian-interpretivist approach to social science, which considers agents’ perspectives on social reality as crucial to social scientific explanation (see esp. Gilbert 1990, 2014). By contrast, the non-normativists have often argued that individuals may be mistaken about what they ought to do, even if they feel some psychological pull toward acting cooperatively with their partners in joint action. This controversy about the degree to which social explanation should be grounded in agents’ perspectives echoes traditional social scientific debates regarding explanation and understanding (e.g., Weber 1922 [2019]; cf. Mantzavinos 2005; Turner 2018), although the philosophical debate is also informed by disciplinary controversies regarding analysis of the notions of “reason” and “justification,” as well as the scope and limits of the BDI-model of intentional agency (Sarkia 2021).

To navigate these controversies, I would like to introduce a distinction between the mental models, which individuals use for reasoning about joint commitments, and the theoreticians’ models, which philosophers and social scientists use to study various aspects of social normativity. ³ Regarding the structure of mental models (which most experimental philosophers have been concerned with), it seems that philosophical analysis should be closely associated with cognitive science, and a successful account should in principle represent individuals’ cognitive perspectives as completely and truthfully as possible (I will leave it open here whether the textual vignettes used by experimental philosophers are a successful means of tapping into these resources). By contrast, what I call theoreticians’ models may be responsive to a broader range of desiderata, such as conceptual or ontological parsimony, generality, morality, and precision. Satisfying such desiderata may require abstracting away from select aspects of individual psychology. For example, there is no obvious reason to expect all of the mental models that individuals bring to bear on the structure of joint commitments to be descriptively parsimonious or reducible to well-articulated conceptual primitives, even if this is a valid goal for theoreticians’ models that aim to satisfy the reductionistic criterion of Ockham’s razor (e.g., Bratman 1999). Similarly, it would be naive to reject the view that individuals’ mental models may cue a sense of commitment toward their felonious partners when jointly performing evil deeds (e.g., ethnic cleansing or classroom bullying), even if moral philosophers would readily agree that such wicked forms of interpersonal normativity cannot give rise to full-blown moral “oughts” and should therefore play no role in a theoretician’s model of social normativity that prioritizes moral considerations. These examples illustrate how requiring that all philosophical models of joint commitment must be aligned with individuals’ cognitive perspectives would exclude many worthwhile forms of philosophical inquiry while holding on to an outdated, Weberian-interpretivist approach to social understanding. From a broader social scientific perspective, it would also be at odds with currently popular multi-methods approaches that emphasize the complementarity of different (qualitative as well as quantitative) research methods in social science (e.g., Goertz and Mahoney 2013).

3. Commitments, Norms, and Society

Most philosophical research on joint commitment has focused on small-scale cases of joint action, such as walking together or singing a duet with another individual. This approach, which is common to both normativist and non-normativist, as well as traditional and experimental-philosophy-based approaches to joint commitment, may be viewed as expressing a common philosophical preference for basic cases: insofar as an analysis can be shown to hold across the most basic forms of joint action, one might expect it to scale up to larger-scale cases of collective, corporate, and institutional action as well—e.g., an orchestra performing in a music festival or an army marching to lay siege to a foreign city. Following this line of reasoning, John Searle (1995, 2010) famously applied his accounts of collective intentionality and status functions to the metaphysics of social institutions, such as money and language, with Gilbert (2014) and Tuomela (2007, 2013) following close behind with their respective accounts of “plural subjects” and “we-intentionality.” ⁴

The analytical strategy of scaling up from basic cases to the macroscopic scale of institutional reality repeats the mistake of considering only a small subset of the mental models that individuals use for reasoning about social normativity. Are two civil servants jointly emptying the dishwasher in the coffee room (one dries the dishes, the other puts them in the cupboard) really engaged in the same type of joint action as two managers sealing the deal for a corporate merger while acting as representatives of the corporation? Echoing points made in the previous section, it seems more likely that two physically co-present individuals simultaneously performing complementary parts of a joint action are likely to be relying on a rather different set of mental models for coordinating their activities than two individuals signing off on an agreement with explicit knowledge of their rights and responsibilities as representatives of an organization (“we-moders” in Tuomela’s [2007, 2013] terminology). In general terms, the preference for scaling up seems to be a symptom of the persistent legacy of the classical approach to philosophical analysis, which was built around the goal of identifying conceptual primitives that give rise to more complex concepts and phenomena (see Strawson 1992). However, although this point is often misrepresented in the literature, a focus on small-scale cases does not amount to adopting a bottom-up methodology (Sarkia 2021). A central requirement for a methodologically bottom-up approach is that one should make minimal precommitments about the best way to describe or explain a set of phenomena. ⁵ However, most philosophical approaches to social ontology draw on rich background theories of intentional agency, which greatly constrain the set of possible explanations for various aspects of social reality (e.g., Sarkia 2022; Tuomela 2007, 2013).

These limitations to the strategy of scaling up have not held analytical social ontologists back from grand aspirations to explain larger-scale institutional reality. For Searle (1995, 2010), this goes by way of the distinction between collectively accepted constitutive rules, which make possible entirely new forms of social action, such as financial exchange, and regulative rules, which merely regulate existing forms of activity, such as driving on the left-hand side of the road (see also Rawls 1955; cf. Guala 2016b). Constitutive rules are taken to depend on collective intentions and the imposition of status functions, which bring about deontologically binding rights and obligations between the participants to a social practice. However, although Searle’s approach allows for unintended “systemic fallouts” from sets of constitutive rules (e.g., increasing inequality in financial capitalist societies) his approach ultimately seems to reify existing social structures (Mallon 2017). This is because it provides no tools for analyzing the power asymmetries that result in the imposition of one set of constitutive rules over another—i.e., the content of what is collectively accepted—while it still insists that the deontic norms governing social interaction must be collectively accepted or recognized by all (or most) members of society. Insofar as the formulation of counterfactuals is one of the central success conditions of social scientific theorizing (Woodward 2003; Ylikoski 2012), Searle’s account accordingly seems to provide very limited explanatory leverage on the structure of the social world.

This criticism might be countered by claiming that Searle’s account of social ontology was never intended to provide an explanatory theory of the social world, as contrasted with a metaphysical account of its constitution. While I agree that this may well have been originally his intention, I think that philosophers in the Searlean tradition have often been strategically ambiguous about what the goals of their theorizing are, and have not clearly distinguished explanatory goals from metaphysical ones (see also Guala 2007, 2016a; Sarkia 2022). ⁶ For example, Searle (2010, ix) describes as his aim to “explain the exact role of language in the creation, constitution, and maintenance of social reality.” However, it is not clear how the creation and maintenance of social facts can be addressed without formulating the kinds of counterfactuals that are characteristic of an explanatory theory. This type of “explanatory slippage” (as we might call it) from metaphysical to explanatory theorizing may in part be diagnosed as a reflection of the growing popularity of methodological naturalism (Guala 2016a; Kincaid 2012a, 2012b; Ross 2023) as a general thesis about the in-principle inseparability of philosophical and scientific investigations of the (social) world (Sarkia 2022; Sarkia and Kaidesoja 2023).

The overlooking of asymmetric power relations is one of the central aspects of collective-acceptance-based theories that critical-emancipatory approaches to social ontology have taken issue with (see Richardson 2023). They share with traditional social ontology the constructivist idea that our concepts and theories in part create the social world that we live in. However, collective acceptance theorists have emphasized the reflexive and performative aspects of folk concepts as the foundation of social practices and institutions (see Guala 2007). By contrast, critical-emancipatory approaches often draw attention to the performative aspects of social theories framed in terms of ideology and power (see e.g., Burman 2023; Haslanger 2000 [2013]; Mills 2005; Pateman 1988). Consequently, collective acceptance theorists typically view their models and theories as morally and politically neutral, albeit ranging over a subject matter with moral and political implications, while critical-emancipatory approaches view such theories as already tainted by political interests and ideologies. Judging such bias to be a bad thing, they aspire to put forward alternative social ontologies, which give a voice to underprivileged groups, such as racial or sexual minorities (e.g., Cull 2024; Jenkins 2023; Àsta 2018). The dialectic between these two perspectives echoes interactions between earlier social scientific theories that have emphasized stable structures and those focused on social critique, such as Parsonsian structural-functionalist approaches and the critical theories of Habermas (1981 [1985]) and others in the latter half of the twentieth century (see Joas and Knöbl 2004 [2009]).

The most important challenge that critical-emancipatory approaches to social ontology face is accounting for the special capacity of the critical theorist adopt the viewpoint of the oppressed. How can (s)he accommodate the viewpoints of different groups of oppressed individuals—such as academics in precarious jobs, racial and sexual minorities, and illegal migrants with hardly any political rights to speak of at all? These groups are diverse in many ways, and to assume that they can speak with one voice seems to overlook the complexity of their differences. For example, first-generation migrants with established political status and recent cross-border arrivals in the United States may have conflicting interests, even if both groups can be regarded as marginalized from the viewpoint of middle-class American society. While theories of intersectionality draw attention to this challenge (e.g., Cull 2024; Àsta 2018), they fail to provide the conceptual tools for modeling the complex social dynamics that amplify the voice of some groups of individuals over others (however, see O’Connor et al. 2019).

This challenge stands even if the critical role of philosophy is limited to the negotiation of diverse epistemological and political perspectives, instead of taking on an explicitly political agenda of its own (cf. Fraser 1985 [1989]). For example, popular ameliorative approaches to conceptual analysis are often framed as replacing essentialist questions about social categories—such as “What is gender?” or “What is race?”—with pragmatic questions promoting active participation in the definition of the relevant categories, such as “What should gender be?” (e.g., Haslanger 2000 [2013]). Prior to answering these questions, one cannot avoid being confronted by thorny questions about political participation, such as “Who are we?” and “Who is allowed to participate in the debate?” An answer that is simultaneously extraordinarily optimistic and mildly pessimistic to these questions is presupposed by epistemological standpoint theories, which suggest that social critique can be effectively conveyed by individuals who are representatives of oppressed groups (e.g., Harding 2009; Wylie 2013). By assuming that such diverse viewpoints can and will be taken into account, they presuppose that we are already living in a democratic, liberal, and tolerant society, providing limited emancipatory leverage for individuals living in genuinely oppressive regimes, such as religious theocracies. In this respect, the original vision of Marx and his followers in the Frankfurt school (e.g., Arendt 1951), who were aware of the all-encompassing character of totalitarian power, seems to have been more successful in identifying the essential role of social contestation in shaping social institutions, even if their epistemological and methodological views were less developed in other respects.

The tools of game theory developed in the latter half of the twentieth century provide one promising framework for developing such epistemological and methodological views. However, although game theory is often glossed as the study of strategic interaction, traditional game-theoretic approaches (in technical terms, games expressed in standard form) are relatively static in the sense that while they can be used to model situations of interdependence between two or more agents, they still admit of an axiomatic (rather than dynamic) solution in terms of pure or mixed equilibria from a set of solution concepts and the payoff table (von Neumann and Morgenstern 1944 [1953], 4). ⁷ While games in extensive form can be used for modeling sequential play, even they admit of an axiomatic solution via backward induction. Accordingly, although games such as the Hawk-Dove, Matching Pennies, and the Prisoner’s Dilemma are useful for modeling situations of coordination, cooperation, and conflict that are of concern to social ontologists (see e.g., Guala 2016b), they do not yet provide an account of the social dynamics that could be used to overcome the limitations of collective-acceptance-based and critical-emancipatory approaches to social ontology.

The approach of evolutionary game theory overcomes this limitation. In particular, the replicator dynamics approach (Taylor and Jonker 1978) allows us to model learning and cultural evolution in games, where agents learn from past experiences and consciously or sub-consciously choose the strategies that have produced the best payoffs in past iterations of the game. Evolutionary game theorists interested in social norms and social justice have often studied the Nash Demand game, where individuals make demands for a share of a resource that can be divided between the participants in several ways (O’Connor 2019; Skyrms 1996). If the individuals overdemand the resource, bargaining breaks down, and both agents receive what is called their disagreement point (a low allocation of the resource or nothing at all). An equilibrium solution to the game is any distribution that divides the resource without remainder (e.g., 6-4, 5-5, or 1-9), because otherwise at least one player could improve her share by demanding more. However, only some of the equilibrium solutions are fair in the sense of dividing the resource equally between the players.

	Low	Medium	High
Low	3,3	3,5	3,7
Medium	5,3	5,5	0,0
High	7,3	0,0	0,0

Relying on replicator dynamics, it has been shown that homogeneous groups of undifferentiated individuals tend to converge toward an equilibrium with an equal distribution of resources (Skyrms 1996). However, introducing labels that make it possible to distinguish between two or more sub-groups in the population is sufficient for generating unequal equilibrium outcomes (O’Connor 2019). This is because individuals will learn to associate labels with strategies, and condition their own strategies on payoffs received with representatives of specific groups during previous interactions. Unfair social norms are especially likely to evolve if the groups are of unequal size, because minorities will encounter majorities more often than members of their own groups, and evolve different strategies in response. Accordingly, minorities will settle for less than the equal distribution of resources, while majorities will expect more than their fair share, and any individual unilaterally deviating from the equilibrium will perform less well than the individuals, who play the dominant strategy. This problem is aggravated if the groups have different disagreement points, as more powerful groups can then accommodate more risks with unfair strategies, while underprivileged groups are forced to accept unfair offer due to their lower fallback position when bargaining fails (Bruner and O’Connor 2018).

The dynamic analysis of social power and injustice represented by evolutionary game theory should make it a useful complement to collective-acceptance-based accounts of social ontology, as well as a natural ally for critical-emancipatory approaches. However, surprisingly few collective acceptance theorists have engaged with evolutionary game theory in any detail, perhaps because it seems to conflict with their static assumptions regarding the relatively harmonious and unitary nature of collective acceptance, and even methodologically naturalistic critics of traditional social ontology have sometimes dismissed evolutionary game theory because of its lack of empirical foundations and unrealistic modeling assumptions (e.g., Guala 2007). Perhaps more surprisingly still, many critical-emancipatory social ontologists seem to have endorsed methodological commitments that explicitly disavow their natural ally in evolutionary game theory due to the extensive abstractions and idealizations that it involves. For example, Burman (2023) approvingly cites Charles W. Mills’s (2005) discussion of ideal theory in defending what she describes as a nonideal approach to social ontology:

What distinguishes ideal theory is the reliance on idealization to the exclusion, or at least marginalization, of the actual. … But ideal theory either tacitly represents the actual as a simple deviation from the ideal, not worth theorizing in its own right, or claims that starting from the ideal is at least the best way of realizing it (Mills 2005, 168).

Burman (2023) goes on to argue that there is a paradigm shift in course in contemporary social ontology from a focus on idealized social processes to the recognition of real power relations in the actual world. However, she fails to justify why we should use the tools of critical-emancipatory social ontology to study such power relations, instead of simply carrying out empirical investigations of the social world by means of descriptive statistics, fieldwork, and interviews, or perhaps the kinds of experiments in game theory Guala (2007) advocates as an alternative to evolutionary game theory. For this reason, I think that her approach tracks the wrong kind of paradigm shift. Insofar as we are concerned with questions of social justice, and one of the main virtues of a philosophical (as contrasted with e.g., a sociological) approach is the generality it affords, the answer may be more idealization, not less. By my lights, the problem with traditional collective acceptance accounts of social ontology is not that they involve idealizations, but that they involve the wrong kinds of idealizations: they assume that all members of society are in a position to wholeheartedly endorse the shared norms and values of society, even when they have been influenced by unjust power asymmetries. While critical-emancipatory social ontologists correctly identify this problem, they make the wrong diagnosis of its underlying causes, and thereby overlook a crucial set of resources for overcoming its limitations. In particular, evolutionary approaches to game theory aid in understanding how and why certain forms of structural injustice are likely to arise in the context of a wide range of different social circumstances, and which counterfactual conditions may be effective in alleviating them (and which may not).

4. Shaping Commitments

In this section, I will argue that both collective acceptance and critical-emancipatory approaches to social ontology overintellectualize the processes that lead to the formation of joint commitments, and therefore fail to pay attention to the complex and strategic background processes, which play a role in structuring joint commitments—including their content, persistence, and who they are initiated with. This is because they both rely on a relatively static approach to the mind, which views individuals as engaging in acts of propositional attitude ascription and practical deliberation from their individual or collective viewpoints, without much interaction between them apart from the idealized act of collective acceptance (cf. Di Paolo and De Jaegher 2012). Accordingly, they treat mindreading (Goldman 2006) and theory of mind (Gopnik and Meltzoff 1997) as prior to the formation of shared commitments, although our disposition to ascribe mental states to others arguably depends on mutual ties that we have already established in relation to them through education, prior interactions, and socialization (Mameli 2001; McGeer 2001; Ross and Stirling 2021; Sterelny 2012; Zawidzki 2013). This process can be captured by Tadeusz Zawidzki’s (2013, 2018) notion of mindshaping, understood as a “as a relation between a target mind (the mind being shaped), a cognitive mechanism (the proper function of which involves shaping that mind), and a model that the mindshaping mechanism works to make the target mind match” (Zawidzki 2018, 739). Ross and Stirling (2021, 250) describe the repercussions of the mindshaping perspective on social cognition as follows:

… social interaction is not like doing (prototypical) science: observation normally influences the phenomena ... [Thus] explanation is seldom among the goals of practical coordination; people aren’t generally concerned with verifying or rejecting prior models of one another, but with achieving coordinated behaviour … except in cases of highly asymmetrical power or status, where order-giving is sufficient for coordination, social interaction is mutualistic: each interpreter can influence the other, and each may be willing to accede some normative authority for the sake of consensus.

The mindshaping hypothesis brings to the fore questions about the alignment of one’s own mental models with the mental models that others use. For example, we remarked earlier that mental models relying on the emergent mechanisms of entrainment and action simulation seem to be more conducive to affiliation and pro-social tendencies than mental models relying on a linguistically embedded theory of mind. Accordingly, insofar as I may have an influence on the mental models that you employ in relation to our joint activities, I may mold you as a cooperative partner in a manner that is conducive to my interests (the situation is of course precisely analogous for you). For example, business associates who meet one another to seal a deal in writing, even if the matter could be more conveniently handled by email, are harnessing the power of physical proximity to render their commitments more persistent and tangible (Cross et al. 2019). Similarly, an academic who travels to a foreign land to work with an associate is displaying her commitment with the implicit knowledge that incurring costs in order to form a joint commitment makes it more likely to be reciprocated (Székely and Michael 2018). Such strategic cuing of commitments need not be grounded in intentional processes based on rational deliberation. On the contrary, given our overall cognitive architecture, it may be counterproductive to deliberate about the best means of influencing our cooperative partners, as this may lead to loss of fluency and greater costs of inhibition owing to increased top-down control (see Christensen and Michael 2016; Evans 2008). Consequently, mindshaping seems to work best when it is an effortless process that individuals engage in without giving it much thought at all.

The underlying psychological mechanisms for mindshaping seem to play in part on the basic psychological need for coherence (Thagard 2000): given that I have committed myself to endorsing a certain view or pursuing a certain course of action, I have a tacit interest in accepting further ideas or making plans that are consistent with this prior commitment in the interests of maintaining a coherent sense of self, which is central to normal psychological functioning:

Normally, we want to be (and to be seen) as consistent with our existing commitments—such as the previous statements we’ve made, stands we’ve taken, and actions we’ve performed. Therefore communicators who can get us to take a pre-suasive step, even a small one, in the direction of a particular idea or entity will increase our willingness to take a much larger, congruent step when asked (Cialdini 2016, 159).

This type of psychological “mission creep” can manifest itself in biased attention and selective attention to evidence, while seeking justification to a view that one has already accepted, or failure to consider alternative courses of action, even if the costs of holding on to one’s prior decision vastly outweigh the benefits of reconsideration. Owing to such biased psychological mechanisms, each of us has an interest in influencing the attitudes that the other has endorsed prior to engaging in joint bargaining and deliberation about what to do together (cf. Bratman 1987, 1999). Indeed, what Cialdini (2016) calls pre-suasion seems to be rife in social life: skilled communicators will smuggle pre-suasive prompts into the terms of a debate with seemingly neutral terminology that is suggestive of a certain view or reasoning process, while conflicting ideas are undermined by thick expressions carrying implicit cues about how to respond to or value the issues under consideration. Consider the fluid but ethically charged distinction between undesirable “lifestyle migrants,” who have crossed borders to take Western jobs away from their supposedly rightful owners, and admired “expatriates,” who bring capital and skills into knowledge-hungry post-industrial societies. Or the distinction between heroic “activist investors” and unscrupulous “vulture funds,” which each seek to influence corporate boards in the pursuit of some agenda of corporate restructuring, but where the former positive term is more likely to be used by the agents involved in such attempts, and the latter pejorative term by the targets of such unsolicited corporate interventions. These are just some examples of how, before the outcome of an idealized social ontological process of collective acceptance, bargaining, or negotiation has been initiated, the game has already been rigged by canny individuals and organizations, who know how to play the game to their own interests, or remorselessly choose to ignore the legitimacy of rival points of view.

The mindshaping hypothesis speaks to the concerns of critical-emancipatory approaches to social ontology and evolutionary game theory also by helping us understand the formation of the cultural categories, which they identify as locuses of injustice and oppression. By affiliating with like-minded individuals and shaping their minds to be receptive to prospective joint commitments, individuals can construct shared cognitive-cultural niches, which exclude certain groups of individuals as potential partners for joint activity while including others (Laland and O’Brien 2011; Nagatsu et al. 2023; Sterelny 2012). Accordingly, teenagers build communities around shared tastes in music, while religious communities use symbolic markers, such as the kippah or the hijab, as signs of religious affiliation and symbolic markers of who one is allowed to associate with for certain types of joint activities, such as marriage or the formation of a sports team. While some such practices of reference group formation may be relatively harmless, others arguably serve as a persistent symptom and cause of social injustice insofar as privileged groups of individuals tend to associate with their ilk, and marginalized minorities are excluded from social circles with power and resources. Such processes of cultural niche formation can also lead to the balkanization of the mental models that are shared in a population (cf. Denzau and North 1994; Mantzavinos et al. 2004), and further undermine the potential for harmonious solutions to collective action problems that are presupposed by collective-acceptance-based approaches to social reality.

5. Summary and Conclusions

In this paper, I have argued that the formation of joint commitments is a complex and multifaceted process, where subtle background processes of mindshaping and mutual alignment precede shared deliberation about joint commitments, influencing their content and structure. Both classical analytical and critical-emancipatory approaches to social ontology have downplayed these features by relying on monolithic approaches to collective intentionality and social power, which do not leave much room for an understanding of the dynamic and strategic aspects of joint commitment. They have also presupposed that we may unproblematically scale up from basic social interactions to larger-scale social norms, practices, and institutions. While novel approaches based in experimental philosophy have brought a breath of fresh air to the field, they have failed to consider the multiplicity of mental models that individuals use in coordinating their joint activities, as well as the importance of a distinction between such mental models and the theoreticians’ models that philosophers construct to study select aspects of social reality.