It is clearly difficult to define a “social group,” as there are several aspects that must be explained for a comprehensive understanding of the concept. Psychological definitions converge on the categorization of self and others into groups, which produces perceptions of groups or group behaviour (e.g., mutual support). Sociology and anthropology characterize groups by shared norms, standards, and institutions regulating the behaviour of group members.
Adding an evolutionary perspective to the existing literature, the present article proposes a computational theory of social groups that identifies four triadic primitives, defining specific group-constitutive roles, which represent “social groups” in the human mind. These primitives include three actors and the interaction between them. In all these triads, two of the interactions are assumed to be negative (e.g., attack and threat), whereas one is assumed to be positive (e.g., support). In this commentary, we focus on the author's suggestion that the negative relations are the group-constitutive factor within these triads.
Our first argument is that conflict is not a necessary component (or marker) for the constitution of groups. The question whether conflict is indeed a necessary component of groups can be traced back to the debate between the Darwinian and Kropotkinian perspectives (Todes, Reference Todes1987). Although Darwin focused on the “law of mutual struggle” as the driving force in evolution, Kropotkin believed that this law is complemented by a “law of mutual aid” (e.g., Skyrms, Reference Skyrms2014). In line with Kropotkin, we argue that not the “attack” relations but the “mutually-supporting” relations are the actual primitives that are traced by mental representations of a “social group” (Brown, Reference Brown1988; Kessler & Cohrs, Reference Kessler and Cohrs2008; Tajfel & Turner, Reference Tajfel, Turner, Worchel and Austin1979; Turner, Reference Turner and Tajfel1982). The existence of two or more interactants that are in mutually supporting relations (compared to no relations or antagonistic relations) would be sufficient to constitute a group and represent it.
According to this perspective, group members would reliably support each other when one of them (or all as a collective) face a challenge, even if such a challenge is not posed by an antagonistic individual or group (Elad-Strenger, Reference Elad-Strenger2016). As an example, team members working to achieve a common goal constitute a group that is detectable and represented as such, even without being attacked from the outside. Thus, although we agree that the notion of conflict with outsiders may enhance mutual support and commitment within the group (Elad-Strenger, Reference Elad-Strenger2013; Elad-Strenger & Shahar, Reference Elad-Strenger and Shahar2017; Sherif, Reference Sherif1966), we argue that conflict is not necessary for the psychological representation and the actual existence of a group. Group boundaries can just as well be located where the mutual support ends and neutral (e.g., indifference) or negative relations begin (e.g., attack).
Our second argument is that the existence of conflict within a triad does not necessarily create or signal the existence of antagonistic groups (those who are inside the group and those who are outside), but may simply represent an internal dynamic within an ingroup, which may even strengthen the agents' belongingness to the group. An example for such a dynamic is ingroup deviance and group members' response to the deviance. The specific norms of the group define what is considered an aggressive or deviant act towards the group (Ben-Shitrit, Elad-Strenger, & Hirsch-Hoefler, Reference Ben-Shitrit, Elad-Strenger and Hirsch-Hoefler2021; Elad-Strenger, Hall, Hobfoll, & Canetti, Reference Elad-Strenger, Hall, Hobfoll and Canetti2021). Accordingly, some deviances will be met with the exclusion of the deviant from the group, whereas many others will trigger attempts to reform the deviant (Kessler & Cohrs, Reference Kessler and Cohrs2008). Ingroup deviants are remembered better (Hechler, Neyer, & Kessler, Reference Hechler, Neyer and Kessler2016) and punished harsher (e.g., Marques and Yzerbyt, Reference Marques and Yzerbyt1988) than out-group deviants, precisely because the deviants and respondents are members of the same group. Thus, it is not only that groups can exist despite occasional internal deviance or conflict (Elad-Strenger, Fireman, Schiller, Besser, & Shahar, Reference Elad-Strenger, Fireman, Schiller, Besser and Shahar2013; Elad-Strenger, Halperin, & Saguy, Reference Elad-Strenger, Halperin and Saguy2019), but also that the existence of deviance may even strengthen ingroup members' identification with it (Pinto, Marques, Levine, & Abrams, Reference Pinto, Marques, Levine and Abrams2010).
What, then, defines a “group” as such, despite internal conflicts? Our third argument is that a mental representation of a group does not only reflect its existence, but also produces and perpetuates it (Turner & Giles, Reference Turner, Giles, Turner and Giles1981), by means of a “self-fulfilling prophecy.” The self-categorization of two agents as ingroup members tends to produce coordination and cooperation between these agents, and antagonistic behaviours towards agents who are categorized as outsiders, thus forming a group via behavioural confirmation (Sassenberg, Kessler, & Mummendey, Reference Sassenberg, Kessler and Mummendey2003; Snyder & Swann, Reference Snyder and Swann1978; Tajfel, Billig, Bundy, & Flament, Reference Tajfel, Billig, Bundy and Flament1971). This mental representation of a group can even define ingroup and outgroup members despite the behavioural primitives would point the other way. For example, when police officers are attacked, they defend one another (signalling the existence of a group) but also defend (and are defended by) their police dogs. Nonetheless, when talking about the police, people rarely think about the dogs as belonging to this group, despite defending policewomen against attack, simply because they are not included in the mental representation of the group. In short, the mental representation of a group can determine the behaviours of agents and thus the existence of groups, as well as the extent to which these behaviours are interpreted as signalling the existence of a group. Considering this role of mental representations (i.e., cognitive categorization) in group constitution, attack or defence in response to conflict are, therefore, also not sufficient to constitute (or mark) groups.
To summarize, we argue that conflict within triadic relations is neither necessary nor sufficient to constitute groups. Rather, we suggest that mutual support is the relevant ecological invariance to trace the existence of groups. In addition, ingroup norms define which behaviours categorize the actors as ingroup or outgroup members. Finally, the mental representations of a social group determine, at least partially, the behavioural primitives and their interpretation. Therefore, we propose that the combination of different approaches to social groups (psychological: cognitive categorization; sociological and anthropological: shared norms; evolutionary: ecological invariances) can paint a more dynamic picture of groups than is represented in the computational theory of groups.
It is clearly difficult to define a “social group,” as there are several aspects that must be explained for a comprehensive understanding of the concept. Psychological definitions converge on the categorization of self and others into groups, which produces perceptions of groups or group behaviour (e.g., mutual support). Sociology and anthropology characterize groups by shared norms, standards, and institutions regulating the behaviour of group members.
Adding an evolutionary perspective to the existing literature, the present article proposes a computational theory of social groups that identifies four triadic primitives, defining specific group-constitutive roles, which represent “social groups” in the human mind. These primitives include three actors and the interaction between them. In all these triads, two of the interactions are assumed to be negative (e.g., attack and threat), whereas one is assumed to be positive (e.g., support). In this commentary, we focus on the author's suggestion that the negative relations are the group-constitutive factor within these triads.
Our first argument is that conflict is not a necessary component (or marker) for the constitution of groups. The question whether conflict is indeed a necessary component of groups can be traced back to the debate between the Darwinian and Kropotkinian perspectives (Todes, Reference Todes1987). Although Darwin focused on the “law of mutual struggle” as the driving force in evolution, Kropotkin believed that this law is complemented by a “law of mutual aid” (e.g., Skyrms, Reference Skyrms2014). In line with Kropotkin, we argue that not the “attack” relations but the “mutually-supporting” relations are the actual primitives that are traced by mental representations of a “social group” (Brown, Reference Brown1988; Kessler & Cohrs, Reference Kessler and Cohrs2008; Tajfel & Turner, Reference Tajfel, Turner, Worchel and Austin1979; Turner, Reference Turner and Tajfel1982). The existence of two or more interactants that are in mutually supporting relations (compared to no relations or antagonistic relations) would be sufficient to constitute a group and represent it.
According to this perspective, group members would reliably support each other when one of them (or all as a collective) face a challenge, even if such a challenge is not posed by an antagonistic individual or group (Elad-Strenger, Reference Elad-Strenger2016). As an example, team members working to achieve a common goal constitute a group that is detectable and represented as such, even without being attacked from the outside. Thus, although we agree that the notion of conflict with outsiders may enhance mutual support and commitment within the group (Elad-Strenger, Reference Elad-Strenger2013; Elad-Strenger & Shahar, Reference Elad-Strenger and Shahar2017; Sherif, Reference Sherif1966), we argue that conflict is not necessary for the psychological representation and the actual existence of a group. Group boundaries can just as well be located where the mutual support ends and neutral (e.g., indifference) or negative relations begin (e.g., attack).
Our second argument is that the existence of conflict within a triad does not necessarily create or signal the existence of antagonistic groups (those who are inside the group and those who are outside), but may simply represent an internal dynamic within an ingroup, which may even strengthen the agents' belongingness to the group. An example for such a dynamic is ingroup deviance and group members' response to the deviance. The specific norms of the group define what is considered an aggressive or deviant act towards the group (Ben-Shitrit, Elad-Strenger, & Hirsch-Hoefler, Reference Ben-Shitrit, Elad-Strenger and Hirsch-Hoefler2021; Elad-Strenger, Hall, Hobfoll, & Canetti, Reference Elad-Strenger, Hall, Hobfoll and Canetti2021). Accordingly, some deviances will be met with the exclusion of the deviant from the group, whereas many others will trigger attempts to reform the deviant (Kessler & Cohrs, Reference Kessler and Cohrs2008). Ingroup deviants are remembered better (Hechler, Neyer, & Kessler, Reference Hechler, Neyer and Kessler2016) and punished harsher (e.g., Marques and Yzerbyt, Reference Marques and Yzerbyt1988) than out-group deviants, precisely because the deviants and respondents are members of the same group. Thus, it is not only that groups can exist despite occasional internal deviance or conflict (Elad-Strenger, Fireman, Schiller, Besser, & Shahar, Reference Elad-Strenger, Fireman, Schiller, Besser and Shahar2013; Elad-Strenger, Halperin, & Saguy, Reference Elad-Strenger, Halperin and Saguy2019), but also that the existence of deviance may even strengthen ingroup members' identification with it (Pinto, Marques, Levine, & Abrams, Reference Pinto, Marques, Levine and Abrams2010).
What, then, defines a “group” as such, despite internal conflicts? Our third argument is that a mental representation of a group does not only reflect its existence, but also produces and perpetuates it (Turner & Giles, Reference Turner, Giles, Turner and Giles1981), by means of a “self-fulfilling prophecy.” The self-categorization of two agents as ingroup members tends to produce coordination and cooperation between these agents, and antagonistic behaviours towards agents who are categorized as outsiders, thus forming a group via behavioural confirmation (Sassenberg, Kessler, & Mummendey, Reference Sassenberg, Kessler and Mummendey2003; Snyder & Swann, Reference Snyder and Swann1978; Tajfel, Billig, Bundy, & Flament, Reference Tajfel, Billig, Bundy and Flament1971). This mental representation of a group can even define ingroup and outgroup members despite the behavioural primitives would point the other way. For example, when police officers are attacked, they defend one another (signalling the existence of a group) but also defend (and are defended by) their police dogs. Nonetheless, when talking about the police, people rarely think about the dogs as belonging to this group, despite defending policewomen against attack, simply because they are not included in the mental representation of the group. In short, the mental representation of a group can determine the behaviours of agents and thus the existence of groups, as well as the extent to which these behaviours are interpreted as signalling the existence of a group. Considering this role of mental representations (i.e., cognitive categorization) in group constitution, attack or defence in response to conflict are, therefore, also not sufficient to constitute (or mark) groups.
To summarize, we argue that conflict within triadic relations is neither necessary nor sufficient to constitute groups. Rather, we suggest that mutual support is the relevant ecological invariance to trace the existence of groups. In addition, ingroup norms define which behaviours categorize the actors as ingroup or outgroup members. Finally, the mental representations of a social group determine, at least partially, the behavioural primitives and their interpretation. Therefore, we propose that the combination of different approaches to social groups (psychological: cognitive categorization; sociological and anthropological: shared norms; evolutionary: ecological invariances) can paint a more dynamic picture of groups than is represented in the computational theory of groups.
Financial support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Conflict of interest
None.