Societies around the world face change, including economic, ecological, technological, demographic change, or change through migration. Positions on which forms of change should be welcomed, accepted, prevented, or stopped differ widely depending on ideological viewpoints and perceptions of human control. In Western societies, for example, non-Muslim majority groups debate the impact that “Islam” has on “them” (e.g., G. Pickel et al. 2020), giving less attention to their own influence on Islamic communities. In these societal conflicts, which can spiral into “co-radicalization” between opposing groups (S. Pickel et al. 2023),Footnote 1 far more is being thrown into the ring than questions of cultural diversity, religious freedom, or migration law: it is about the fundamental question of identity (“Who Are We?”; Huntington 2004), the question of the nature and location of the shared We. This comprises various sub-questions, ranging from seemingly straightforward inquiries, such as who belongs to our We,Footnote 2 but also very complex ones, for example, what distinguishes Us from Others; the bread and butter questions of social psychology and related disciplines. One level below, we are confronted with philosophical questions: what is a We anyway? Is it the composition of all the I ’s it contains? Does this view imply that removing or adding a single I changed the composition of the whole We, transforming the earlier We into a different We? And how could we, in times of change, refer to the later We and pretend that it still has something to do with the earlier We? In other words, we might question whether identity under social change can be continuous at all, that means, able to remain itself despite change. This article delves into the underlying assumptions when discussing “Western identity” under the (potentially) transformative influence of “Islam.” However, I will not make any statements about the actual relationship between Islam and Western societies, nor speculate on specific changes, but will try to shed light on philosophical and psychological problems on the foundations of identity under change.

1 Enigmatic foundations of human togetherness

People not only speak in the mode of We, they also think, feel, and act in a way that can be described as collective. When someone says “We Germans,” he or she expresses: “I am German. And there are others who are also German, and together with me, we form a unit.” The person thinks of these others in his or her interactions, feels with them, acts with them, and for them. These others, associates, mates, fellows, colleagues, comrades, or companions neither have to be present nor necessarily alive in the situation of conversation. Membership in groups is understood as the social identity of individuals, since for them, group memberships are far more than just cold facts. Group memberships are an essential part of individuals’ self-concept (or self-knowledge),Footnote 3 which is why they are imbued with many different meanings and feelings. Given that identity is shared yet subject to varying perceptions and interests, it is not surprising that there are so many negotiations and conflicts over group identity.

The social psychological Social Identity Approach, developed in the 1970s and 1980s by Henri Tajfel and John Turner, conceptualizes representations of group membership as a source of the self. Serge Moscovici’s Social Representations Theory analyzes how these representations are constructed in complex social processes, while Identity Process Theory by Glynis Breakwell connects social representations to (social) identity, bridging conceptual gaps. The concept of collective identity is closely related to discourses on social identity. Terminological similarities may suggest that both concepts are synonyms, but they are, in fact, separate perspectives rooted in different research traditions. Research on social identity (e.g., Turner et al. 1987) adopts a psychological perspective, describing the intra-psychic process of how an individual’s self-definition is based on perceived social associations (i.e., the individual’s identity as a group member; a personal characteristic). Collective identity (e.g., Melucci 1989), in contrast, adopts a more sociological perspective, describing the inter-psychic process of how several individuals together form a shared sense of association (i.e., the identity of the group as a whole; a shared characteristic). Given these disparities, should we surgically separate the two concepts and focus on either one or the other? In my opinion, this depends on the purpose of the study. This interdisciplinary article aims to reflect on the foundations of identity at the intersection of philosophy and (social) psychology. Pursuing this broad aspiration, solely concentrating on either social or collective identity would limit our scope, considering the wide range of psychological social psychological perspectives (Social Identity Approach), sociological social psychological perspectives (Social Representations Theory), those in between, and philosophical considerations, which are typically more oriented towards collective than social aspects of identity. Therefore, if we want to avoid narrowing the perspective, it is at the expense of some incongruity between the compared concepts. Nevertheless, I am confident that comparisons of these not-so-distant concepts promise more gains than losses. The collective perspective on identity can benefit from examining the processes within individual group members’ minds, just as the social perspective can benefit from understanding how these individual processes interact.

Usually, there is only a small step from We to I and from I to We, sometimes even less. In ordinary language, both terms refer to ourselves and distinguish Us from other people (You and Them). Drawing on the arguments of Les embarras de l’identité (English: Puzzling Identities) by the analytic philosopher Vincent Descombes, the study of identity revisits language and poses the question: what do we mean when we speak of collective identity? Are we referring to a specific composition of group members, that is, the sum of a group’s parts or something else? When everything is in flux, how could any identity persist beyond linguistic conventions, and would it not be much more appropriate to speak of snapshots of compositions, that is, situationally defined sums of parts rather than continuous collective identities? However, for members to be able to join and leave, there has to be something to join or leave. For a process in flux, there must be something in flux unless our language of identity is devoid of meaning and logic.Footnote 4 To introduce this problem, we explore broader social psychological and philosophical perspectives on identity. The purpose of this excursus is to provide readers with the necessary background knowledge to comprehend subsequent discussions on the continuity of social and collective identities.Footnote 5

2 Social psychological perspectives on identity in the we-mode

2.1 Social identity approach

The Social Identity Approach is an umbrella under which two social psychological theories are subsumed: Social Identity Theory (Tajfel and Turner 1979, 1986) and Self-Categorization Theory (Turner et al. 1987). Social Identity Theory understands social identity as the aspect of an individual’s complex self-concept that refers not to his or her private self-representations in idiosyncratic terms (the self as I) but to the representation of the self as a group member (the self as part of a We) “together with the value and emotional significance attached to that membership” (Tajfel 1978, p. 63). According to Tajfel’s and Turner’s approach, and in contrast to other conceptions in sociology or psychology, the collectives or groups to which individuals feel they belong (i.e., with which they identify) are not external entities separate from the self to which they can instrumentally resort when it seems rational to them. Instead, the authors conceive of groups as internalized within the individual’s self. When a social category (i.e., group) is salient, individuals experience other group members and themselves less as unique and idiosyncratic persons but more as interchangeable exemplars of “their” category (depersonalization; Hornsey 2008). In modern social psychology, Social Identity Approach’s emphasis on subjective perceptions marks an overcoming of purely behaviorist and reductionist perspectives on human attitudes, emotions, and behaviors within social contexts. Categorizing and identifying oneself as a group member blurs the boundaries between I and We. As group members, individuals recognize the group in themselves and see themselves reflected in their group. In their perceptions, people who identify with their group are their group. Thus, in the Social Identity Approach, the group is not conceptualized as an internal occupying force that oppresses one’s self but as expression of one’s self that enables agency (see Fritsche 2022).

In their pioneering works, Tajfel and Turner (1979, 1986) argue that people are motivated to see themselves in a positive light (need for self-esteem), which is why they cognitively, evaluatively, affectively, and behaviorally strive to perceive and establish positive distinctiveness from other individuals and groups (Brown 2020b). In intergroup contexts, individuals self-categorized into groups want to perceive and uphold superiority of their own group (ingroup) over relevant (i.e., comparable, proximal) outgroups. This is because they derive self-esteem from favorable intergroup comparisons. Perceived positive distinctiveness of the ingroup on contextually salient comparison dimensions (e.g., income or education in Western societies) fosters a positive identity and increases self-esteem, while perceived negative comparison outcomes result in negative identity and loss of self-esteem.

The theory discusses various strategies for managing negative social identities, including (I) individual mobility: moving to a higher status group when group boundaries are perceived as permeable, (II) social competition: competing with a higher status group with the aim of outperforming it, when status relations are perceived as unstable and illegitimate, and (III) social creativity: re-evaluating comparison dimensions, changing the comparison group or comparison dimension, when status relations are perceived as stable and legitimate (Tajfel and Turner 1986). Thus, the Social Identity Approach integrates perceptions of socio-structural conditions such as the perceived stability or permeability of social systems, positing that individuals’ subjective perceptions determine social reality by guiding their thoughts, feelings, and actions. Above all, the approach is well-known for its explanation of intergroup dynamics such as conflict, prejudice, or discrimination (Brown 2020b). Specifically, experimental research demonstrates that mere awareness of a trivial and arbitrary categorization into ingroup and outgroup (for example, We over-estimators vs. You under-estimators), without any history of interaction or gain, can be sufficient for preferring other ingroup members or discriminating against outgroup members (Tajfel et al. 1971).Footnote 6 Tajfel and Turner (1979) term these biases ethnocentrism in real-world contexts and ingroup bias in laboratory settings. Other theorists explain the emergence of ethnocentrism with evolutionary advantages through a mechanism that facilitates individually costly cooperation with minimal cognitive requirements when alternative cooperation-supportive conditions, such as continuing interactions, institutions, or norms, are absent (Hammond and Axelrod 2006). In contrast, Bizumic and Duckitt (2012) differentiate the ethnocentrism concept into separate, inter-correlated facets of intergroup expressions (i.e., preference, superiority, purity, and exploitativeness), reflecting the idea that one’s ingroup is more important than other groups, and intragroup expressions (i.e., group cohesion and devotion), reflecting the idea that one’s ingroup is more important than its individual members.

The Social Identity Approach raised conceptual (and philosophical) questions that were only gradually addressed. For example, in his initial works on relations between established and rather static groups, Tajfel did not delve into their origin or development. Whereas Social Identity Theory does not explain where social identities actually come from, Self-Categorization Theory proposes the emergence of the social self as a situational, rather than constant, cognitive process, positing that group formation occurs when in the comparative context, on average, differences within one collection of stimuli are perceived as smaller than their difference from another collection of stimuli (meta-contrast principle; Turner et al. 1987). Accordingly, social groups are derivatives of cognitive construction based on perceived similarity and distinctiveness. Depending on situationally salient comparison contexts and accessible comparison categories (i.e., groups), individuals self-categorize and socially identify with groups across various levels of inclusiveness, making them susceptible to thinking, feeling, and acting collectively. Consequently, group phenomena, such as collective action, are created and coordinated by a social psychological mechanism wherein individuals represent themselves as exemplars of social categories, presumed to be an individual yet universal cognition rooted in (contrast) perceptions of the social world. This is not only assumed for social identity but also for individuated personal identity: “Rather than personal identity reflecting the relatively stable and enduring features of an individual, the nature of individuality is forged through categorization and social comparison” (Turner and Reynolds 2012, p. 409).

In the late 20th-century research landscape, the Social Identity Approach (and especially Self-Categorization Theory) emerged as a counter-proposal to the prevailing identity theories of the time, which focused on the stability and continuity of identities (including subdivisions into stable domain-specific identities, for example, “academic identity”). In contrast to these theories, the Social Identity Approach emphasizes the situational dependence and fluidity of identity, which result from functionally antagonistic personal and social self-definitions depending on the accessible categorization in meta-contrast comparisons. Nevertheless, it is about the potential of fluidity, rather than positing human identity as inherently fluid. Later, we will delve into reasons provided by the Social Identity Approach why in reality, identity tends to be rather stable and continuous.

2.2 Social representations theory

Another social psychological framework concerned with the questions of identity and how membership in the socio-cultural sphere constitutes individual consciousness is Social Representations Theory (Moscovici 1981, 1988). The theory posits that collective entities (e.g., groups), but also events or history, are not phenomena with objective reality beyond human perceptions (they would not affect, for example, stars or birds) but social representations, that means, complex networks of ever-evolving collective cognitions produced, practiced, and spread like rumors through the interactive discourse of a “thinking society” (Moscovici 1988). By linking “a cognitive form with a content widely accepted by the group” (ibid., p. 221), social representations make reality by (re-)negotiating what is socially accepted (or common sense), while this created reality is represented vice versa in new representations: “Contents that are shared by a whole society lead each mind to draw its categories from them and these categories impose themselves on everyone.” (ibid.) As purposeful products of human thought, social representations are expressions of human agency but simultaneously social influence imposed on human interaction, prescriptively limiting our socio-cognitive capacities (Voelklein and Howarth 2005).

Social representations establish an order for individuals to navigate and act within the social world, enabling communication within groups through common codes for social exchange (Moscovici 1973). This points to the primary function of representations: making the unfamiliar familiar by anchoring the strange within the familiar and by objectifying representations into social realities (Sammut and Howarth 2014). Importantly, representation is not meant as imitation of some external “facts and things that have a meaning outside the communication that expresses them.” (Moscovici 1988, p. 230) Instead, through representation, individuals basically represent themselves, but in a sense that the underlying social meaning system manifests itself through their representations. Indeed, Moscovici counters the idea of representational constructions (exclusively) inside individual minds with the observation that all of these representations are very much social. As social emergence phenomena, representations are inter-subjectively constructed, contested in the political space, and validated within dynamic semiotic triangles of representing subject, represented object, and “the social groups towards whom the subject is positioning him- or herself in undertaking this representation” (Voelklein and Howarth 2005, p. 434). These representations are referred to as social rather than collective to reflect the noisy conflictual and cooperative “plurality of representations and their diversity within a group.” (Moscovici 1988, p. 219) Social Representations Theory, in tandem with the Social Identity Approach, helps to comprehend the social construction processes of “how people come to interpret and make their world meaningful” (Breakwell 1993, p. 199). Meaningful social identities could be formed by social representations if they serve important psychological functions or align with significant goals. Conversely, the centrality of social identities for one’s self-concept should determine exposure to, acceptance, and use of social representations (Breakwell 1993). Everyday endorsement and rejection of social representations “reveals one’s affiliation, loyalty and identity as experienced within particular encounters and contexts” (Howarth 2006, p. 70).

2.3 Identity process theory

Building on both the Social Identity Approach and Social Representations Theory, Identity Process Theory (Breakwell 1986, 1993, 2014) investigates the multi-layered dynamic processes of constructing, maintaining, and modifying identity. Aiming to explain the total identity of individuals, Identity Process Theory integrates biological characteristics (memory capacities, consciousness of the organism, etc.) with physical and societal constituents of the social context into a complex interaction that provides the “material for identity construction” (Breakwell 2014, p. 26, 1993). Constantly (re-)producing and changing, individuals’ thoughts, affects, and actions reflect the reciprocal interplay of social representations and an identity in change. This change is delineated into two distinct processes, each comprising of different components. The first identity process, assimilation-accommodation, absorbs informational content (the abundance of social representations, for example, group memberships) into the identity structure and adjusts this structure for the information to fit in; the second identity process, evaluation, allocates value and meaning to the (new) identity elements (Breakwell 1993). These process algorithms are biased towards self-interest rather than accuracy, therefore, information-processing works selectively to achieve and uphold a sense of self-esteem and uniqueness (positive distinctiveness from other people) as well as self-efficacy, that is, an identity structure characterized by competence and control (ibid.). Moreover, the algorithms strive to ensure self-continuity, maintaining a sense of “persistent sameness with oneself” (see Erikson 1980, p. 109). Importantly, this does not imply objective consistency across time but means the subjective feeling that one’s growth and change is congruent with the development of one’s own identity (Breakwell 1993). In Identity Process Theory, continuous change is therefore not seen as an obstacle to identity, but on the contrary as identity’s core characteristic. But the problem how exactly identity’s variation over the lifespan is knotted into a whole has not yet been solved. Breakwell (2014, p. 28) emphasizes: “A developmental theory of identity processes is in my opinion the Holy Grail.”

2.4 Interim conclusion and specification of the research question

Drawing from the social psychological theories presented above, it is evident that social groups are not extra-psychological entities (Turner et al. 1987). Although earlier social psychologies of the fin de siècle (for example, Gustave Le Bon with his 1895 work The Crowd: A Study of the Popular Mind) might have suggested otherwise with terms such as group mind and collective soul, in which one’s self gets lost (cf. Le Bon 2006): in the absence of a collective brain, groups do not think, feel, or act for themselves over their individual members’ heads. When we speak of a nation “rejoicing over a gold medal,” or “Ukraine defending itself,” we are using personifying and psychologizing conventions of our language which hypostasize. Psychologically, thinking, feeling, and acting are “performed” by individuals, albeit not detached, each on their own, but in interaction. When self-defined as members of a collective, individuals can feel collectively proud or offended, they can collectively conquer or defend something they call “their country.” Methodologically focused on the macro-context level, the more sociological Social Representations Theory emphasizes the role of interactions and contexts in the emergence of collective thinking, feeling, and acting. Although the more psychological Social Identity Approach, due to its methodological micro-level focus, studies perceptions (rather than interactions and context effects), it nonetheless recognizes that interactions and social contexts influence collective thinking, feeling, and acting (Turner et al. 1994). Stimulus processing thus takes place in singular brains. If a particular individual has no perception of a specific category at all, even if this category may be virulent in the broader discourse, it is difficult to imagine how this specific category should (directly) influence his or her thinking, feeling, and acting. At the same time, however, we know that humans are not solitary stimulus processors. In most cases, individuals discuss and test their thoughts with other people, share their feelings, coordinate their actions, and are (at least indirectly) affected by known and unknown categories that shape the actions of others, that is, influence others and are influenced by others. We can therefore conclude that collective thinking, feeling, and acting as processes based on perception and interaction have both individual and social aspects. Following this comparison of social psychological concepts, the philosophical reflection on the nature of collectives is now specified.

The central question of this article: is a We sum of its parts, or is it something else? Both possible answers are puzzling as they raise a new set of problems. If a group is indeed the sum of its parts, it is a mere label for a collection of people devoid of independent existence. The group itself would be identical to the material sum of its members. Beyond its members, nothing would remain of the common We (as a synonym for the group), and the disintegration of its members would necessarily lead to the disappearance of the group. So, is the last surviving member of a group only an individual (an ex-member) who has lost her group—or is she still a member of a persisting, albeit currently almost empty group? If the We is more than just the sum of its parts, groups must have an independent existence that transcends that of their individual members. However, what form of independent existence could this be? Physical or metaphysical existence? And to what extent are these questions of existence questions of identity?

3 Identity as a philosophical problem

In Puzzling Identities (2016), Descombes delves into the historical linguistic usage of the term identity, aiming to unravel its conceptual meanings and meaning shifts from an analytic-philosophical point of view. Semantically, the identity term has undergone fundamental change. For what follows, it is crucial to differentiate between the classical and modern understandings of identity.

3.1 The classical understanding: identity as statement of congruency

From ancient times, the term identity has been used to express complete congruency and indiscernibility between two or more things (from Latin identitas, meaning sameness). Answering an identity question, according to this classical understanding, means indicating whether the things in question are identical with each other. Expressed in logical terminology, this involves statements of the form a ≡ a. Identity questions thus refer to the non-distinctness and indistinguishability of things because things are identical only if they have all properties in common (Rosa 2008). Identity as congruency can apply to relations between the thing and itself or to relations between the thing and a second thing that is indistinguishable from the first (ibid.). However, when two things are identical, there is, in fact, only one single thing. With few controversial exceptions, most classical identity questions are trivial in nature because it is obvious that Socrates’ identity is to be Socrates, and not Plato. “Roughly speaking, to say of two things that they are identical is nonsense, and to say of one thing that it is identical with itself is to say nothing at all.” (Wittgenstein 1922, 5.5303) Accordingly, Descombes observes that until a few decades ago, identity questions in Western societies were almost exclusively posed in legal and criminological contexts to determine “whether an individual is who he claims to be” or “whether a corpse is that of a particular person, ‘the presumed victim of a crime,’ identified by the elements of his civil status.” (Descombes 2016, p. 4) The insignificance of identity questions in everyday language is underlined by an entry in the 18th-century Dictionnaire de l’Académie Française, which defines identity as “a scholarly term […] only used in Didactics” (ibid.) that answers the question: “which makes it the case that two or more things are but a single thing.” (ibid.) Non-trivial, in fact, is the concept of identity when it communicates that we are dealing with only one “selfsame thing that we sometimes call A and sometimes call B” (ibid.; for example, “Bruce Wayne is Batman”). Rare exceptions where the identity question emerges concerning oneself include conditions like amnesia or delirium—that is, when one’s own life story is forgotten or no longer retrievable in consciousness (cf. ibid., and Locke 1847). In such cases, individuals may learn new information about themselves when finding an answer to first-person questions. However, why do we commonly ask “Who am I?” and “Who are We?”, although we do exactly know what our name is, what alternative names we bear, and whether we lead a double life with a second identity, like Bruce Wayne?

3.2 The modern understanding: identity as statement of quality

Descombes uses a tourist guidebook for the city of Rome to illustrate a shift in the usage and meaning of the identity term. In the said guidebook, the urban zone of San Lorenzo is described as “one of the working-class neighborhoods whose identity has been best preserved.” (Descombes 2016, p. 5) This sentence precisely does not (only) mean that San Lorenzo is identical to itself, but that according to the modern concept of identity, San Lorenzo has a unique quality distinguishing it from other districts of Rome. In contrast to the surrounding neighborhoods, San Lorenzo still possesses its own unique and original character or essence, that is, what makes San Lorenzo distinctive and special. “Identity is now a quality that can be preserved, which means that it is also a quality that one can lose or that one can seek to defend against whatever threatens to destroy it.” (ibid.) Identity is now understood as the ability to remain oneself and remain distinctive, without merging with one’s surroundings to the point of indistinguishability. We feel compelled to demonstrate our authenticity to others, and in the face of groups losing their binding powers, previously taken for granted, we feel compelled to prove our belonging. In contrast to the classical understanding, the modern identity question can no longer be assessed objectively from outside, as it is based on subjective self-assessments and motivated ideas. Thus, identity questions have become a complicated endeavor. Unlike the classical understanding, the modern concept of identity does not describe an equation or statement of congruent interchangeability (“I am who I am”), but rather an attempt to absorb certain qualities through identification (“I am German,” in the sense of: “What concerns the Germans also applies to me.”).Footnote 7 Our modern identity is often not what we truly are but aspire to be. Unlike the classical concept, the modern concept of identity has built-in escape doors. Just as we can identify ourselves with a quality or group, we can dis-identify and deny that this quality or group belongs to ourselves.

However, the modern concept of identity has not simply replaced the classical view. We still understand the classical concept of “two things being identical.” Although the classical usage (identity in the sense of the objectively identical) seems to take a back seat to the dominant modern usage (identity in the sense of subjective identification), people have a need to remain identical with themselves (i.e., authentic, recognizable, genuine, not a copy). However, even if the use of the term may have changed, striving for authenticity (psychologically: distinctiveness) is not a modern invention but has always been rooted in the conflicting fundamental needs for group membership (thus, conformity, synchronization, depersonalization) and preserving or achieving distinctiveness. Therefore, it is not surprising that the motivation to remain identical to oneself (maintaining one’s distinctiveness) may even intensify with increasing opportunities for identification: the more volatile, multifaceted, growing, and normatively influential the marketplace of social identities, the more people want to remain identical to themselves, that is, defending what they perceive to be their distinctive identity (which, of course, has always been a socially identified identity). This could be the source of constant pressure to prove that one’s identified identity is genuinely one’s own, will forever be one’s own, and always has been one’s own (i.e., identity continuity). Individuals may perceive themselves as typical of their identified group, yet it is inevitable that they become (or will become) aware of incongruencies between themselves and what they identify with. Group members have to validate their identification through normality (i.e., conformity with group norms), and, particularly in contexts where individuality is normative as well, they have to emphasize their originality while doing so (which is, of course, impossible). In social psychology, Optimal Distinctiveness Theory (Brewer 2012) describes the pursuit of optimally balancing the opposing needs for uniqueness (recognizability) vs. homogeneity, i.e., differentiation from vs. immersion in groups. The theory assumes that the more one need is satisfied, the more the other gets activated, as both social motives hold the other in check (ibid.). For Brewer, this reciprocal regulation of social identification poses the dilemma of how a universal need for distinctive identities, rooted in our evolutionary history, can be met in a modern world “where interdependence transcends group boundaries at a global level” (ibid., p. 95).

For Descombes, Erik Erikson is the key figure in establishing the modern concept of identity. The psychoanalyst of Jewish descent emigrated to the U.S. in the 1930s, where he treated traumatized WW2 soldiers. Many of these young war returnees had great difficulty reintegrating into the U.S. society. Erikson observed in the traumatized soldiers a loss “of what he calls ‘a sense of identity.’” (Descombes 2016, p. 17) This loss of identity, however, cannot be attributed to disorders or disturbances in one’s personal identity. None of the soldiers undergoing therapy forgot their own names. Instead, Erikson noted, “The patient of today suffers most under the problem of what he should believe in and who he should—or, indeed, might—be or become.” (Erikson 1986, p. 279) The patients are “unable to find the mooring they needed within the ideals and models of the group.” (Descombes 2016, p. 19) Erikson’s notion of identity crisis, which found its way into everyday language via psychology, thus refers, in the case of his treated soldiers, not to a loss of memory or delirium (a crisis of the identical), but to the breakdown of their normative and idealized self-image as Americans (that is, a crisis of identification; cf. ibid., p. 18). In the post-war period, identification with the attributed (and contested) qualities of one’s group became thematizable.

By shifting the focus to collective identity as a social environment that operates in and through the individual (pivotal tenet of the later social psychological theories), Erikson distances himself from orthodox Freudianism. For the latter, the external world beyond the individual is a separate objective reality or object world that cannot penetrate the self (cf. ibid., p. 20). However, in response to the cultural anthropological positions of his time, Erikson recognizes that the Freudian conception “cannot take account for the environment as a pervasive actuality. The German ethologists introduced the word ‘Umwelt’ to denote not merely an environment which surrounds you, but which is also in you. And indeed, from the point of view of development, ‘former’ environments are forever in us; and since we live in a continuous process of making the present ‘former’, we never—not even as a new-born—meet any environment as a person who never had an environment.” (Erikson 1968, p. 24) Therefore, the quality of being distinguished from others through identity not only demarcates oneself from one’s environment, but signifies the expression of a social environment that operates within and through oneself.

3.3 Which group affiliations does social identity refer to?

To peel themselves out of their social environment, people resort to their social identities. The We’s reference point is one’s group (in Social Identity Approach terminology, the ingroup), to which individuals self-categorize and feel a sense of belonging. The Social Identity Theory makes only sparse statements about the nature of these groups. Whereas in ordinary language, groups often refer to gatherings of a countable number of people, in the Social Identity Approach, all kinds of social categories are conceivable as groups, from very small task groups to billion-sized categories such as humanity. Only an interaction of at least two persons is needed for the transition from personal to social relationships; groups are formed when two or more individuals share the perception of being members of the same social category (Turner 1982). Interactive identity formation takes place in political, discoursive opinion spaces, whereby beyond deductive categorization based on membership in long-established groups, categorization can also emerge inductively through perception of current attitudinal congruence in opinions (O’Reilly et al. 2024). Social identification with spontaneously arising opinion-based groups suggests that there is no fixed set of “real” identities, but that the (un-)reality of identities can fluctuate and be updated depending on, for example, perceived cumulative attitude congruence (ibid.). According to Tajfel and Turner, group identification depends on cognitive but also motivational factors, particularly the extent to which a social identity satisfies the individual’s fundamental need for self-esteem. Meanwhile, social psychology theorists discuss whether other needs could be universally motivating factors, such as the need for belonging (Baumeister and Leary 1995), certainty (Hogg 2000), existence (Castano et al. 2002), coherence (Jaspal and Cinnirella 2010), significance (Kruglanski et al. 2022), or control (Fritsche 2022). These motives build on the perception that group membership allows individuals to be part of something bigger. Therefore, we are more likely to identify with tradition-rich nations or successful sports clubs rather than with categories based on trivial attributes like shoe size, as the former fulfill our basic needs more effectively (in most cases). However, if it became salient in a particular social context that wearers of size eight shoes are considered to be positively distinct from wearers of all other shoe sizes on a relevant comparative dimension (for example, everyone knows that no one can run as fast as wearers of size eight shoes), people with this particular characteristic would presumably identify more frequently with this social category. For the Social Identity Approach, all of these groups (nations, sports clubs, wearers of a specific shoe size) are ontologically on the same level, because, from its constructivist perspective, groups are for us in our subjective consciousness. The Social Identity Approach does not address what groups actually are in themselves. Psychologically, all forms of group identification are assumed to be possible and functionally equivalent, relative to their personal value and emotional significance.

Descombes, on the other hand, warns against lumping all possible social affiliations together. He argues that there are fundamental differences between logical and sociological affiliations that should not be confused (cf. Descombes 2016, p. 12). Membership in nominal groups refers to the logical affiliation to an infinite number of possible categories, i.e., taxonomic games with groups that are nothing more than a mere name, such as the aforementioned affiliation to the wearers of size eight shoes. Real group memberships, like belonging to nations, are different from nominal memberships, as real, historically grown groups hold factual significance for individuals’ self-identification and create social bonds through the collective consciousness of a shared fate (cf. ibid., p. 13). The Social Identity Approach may challenge Descombes’ thesis by referring to findings on inductive, instead of deductive, group formation (cf. O’Reilly et al. 2024). However, Descombes is primarily concerned with logical considerations, not with empirical findings. According to his reasoning, sharing nominal attributes with other group members is logically contingent, and it is even possible, though unlikely, that one does not share specific nominal attributes (for example, “I like licorice”) with any other person in the world. In contrast to nominal attributes, the attributes of real group affiliations are necessarily shared. One may be the last member of a real group but never the only one that ever existed, as this would negate the existence of the group itself: “By definition, to have a given nationality is to be from the same country as the other citizens of that country. Or, to be from the same family assumes that said family exists” (Descombes 2016, pp. 12–13). For Descombes, the attributes of real group memberships are necessarily shared with someone. However, we should critically question: with whom are they necessarily shared? With anyone or with specific group members? Descombes is vague at this point, but seems to lean toward the latter with his example: “someone who speaks her own language is not someone who just happens to have a linguistic resemblance with other individuals […] She speaks the same language that her mother does not because she bears a resemblance to her mother […] but because she learned the language from her mother.” (ibid., p. 13) He seems to argue that one shares one’s mother tongue not contingently or coincidently with one’s mother but necessarily. However, this is not the case. The social construct of mother tongue is often understood as one’s first language, and this can also be a language that one’s mother does not speak. One might also happen to speak the same language as one’s mother without having been taught it by her, but because one learns it later in life from someone else (e.g., at school). One’s mother may have even forgotten the language by then. These considerations show that Descombes’ example is not convincing: one’s language is not necessarily shared with a specific group member. Unfortunately, Descombes does not provide further examples of non-contingently shared attributes. Hence, the question remains unanswered as to whether there are memberships in real groups that are necessarily shared with specific group members, not anyone. It is at least questionable whether the (non-)contingency of shared attributes is a criterion for distinguishing between nominal and real group memberships.

Sociological nominalism, which Descombes may also attribute to the Social Identity Approach, sees no fundamental differences between real and nominal group memberships. “Of course, even for a nominalist sociologist, a human group is real in one way: it is made up of real people. But that does not make the group itself real as a group. For such a sociologist, there is nothing more in a human group than a plurality of human individuals […] he will refuse to say that certain groups are endowed with a group identity.” (ibid., p. 141). However, what implications follow if any kind of group is nothing but the sum of its parts? Subsequently, I will try to explain the challenge that arises when thinking of continuous collective identity, given that in their composition, groups are ever-changing in the succession of members rather than being a solid block. I will draw on a popular philosophical thought experiment to illustrate this problem.

4 Twice in the same river: can groups have a history?

The ancient Greek writer Plutarch told about the mythological hero Theseus, who had slain the Minotaur in Crete (Plutarch 1914, p. 49):

“The ship on which Theseus sailed with the youths and returned in safety, a thirty-oared galley, was preserved by the Athenians down to the time of Demetrius Phalereus. They took away the old timbers from time to time, and put new and sound ones in their places, so that the vessel became a standing illustration for the philosophers in the mooted question of growth, some declaring that it remained the same, others that it was not the same vessel.”

The Ship of Theseus is a philosophical paradox that brings the identity question to a head: can an entity remain identical with itself if it is subject to a process of permanent change? Does Theseus’ ship lose its identity and become a different ship once one of its wooden planks is replaced by a new one? If not, is it a different ship as soon as all planks are exchanged? And would the Ship of Theseus be duplicated if the old, discarded wooden planks were picked up and reassembled into their original configuration? Regarding ourselves, how can we assume that we are still the same person as we were last year, even though we have different skin cells, different hair, different opinions, and different memories today? Applying this analogy to social groups, the vessel’s wooden planks that are replaced from time to time are the members who leave or die, join, or are born. In this context, Aristotle compares cities to rivers: “the city is also the same, although the citizens are always dying and being born, as we call rivers and fountains the same, although the water is always flowing away and coming again” (Aristotle 1999, p. 55). Does Aristotle’s argument truly imply that collective identity persists amid permanent change, or does a group alter into another group with changes in its composition? According to Descombes, a continuously persisting, i.e., diachronic identity presupposes an independent substance that does not change despite the constant change in the group’s composition of members. “If a totality is nothing more than a collection of the elements of which it is composed at a given instant, this totality will disappear each time its composition changes.” (Descombes 2016, p. 142) Nominal groups, which are nothing but the sum of their parts, consequently cannot have a history of their own, “since the identity of a purely taxonomical class is a function of the identity of its members, such as can be set forth in an itemized list or inventory of the collection’s elements.” (ibid.) They are replaced in the stream of time by a new group with each change, since there is nothing permanent about them that could legitimize a diachronic identity. Similarly, Self-Categorization Theory’s meta-contrast principle indicates that social identity is not rigid, static, or monolithic but in a state of flux, contingent upon perceptions of the social environment. As Descombes points out with regard to modern identity, we nonetheless want to maintain a sense of authentic continuity. In our social identity as group member, we are motivated to perceive our groups as continuous. To bolster this perception, we rely on intersubjective agreement about the continuity of our group, confirming its existence to others, and receiving this confirmation from others (according to the Social Identity Approach, groups differ from non-groups in that a common identity is shared by at least two people; Brown 2020a). However, can collectives fulfill this demand for continuity objectively, beyond (inter-)subjective perceptions?

Historians have cast doubt on the thesis proposing seamless objective continuity in collective regional or national identitiesFootnote 8 over decades and centuries. In the past, historiography has taken for granted that our collective traditions and customs are passed on from generation to generation, suggesting continuity as if a nation were an unchanging or inalterable substance (cf. ibid., p. 136). “‘We have a national dish, one that our ancestors also ate’ […] or ‘We have a national language, which was also that of our ancestors.’ Similar things are claimed for national literatures, national musical forms, etc.” (ibid., p. 137) Historical research unearths disillusioning findings, revealing that supposed traditions actually date back much less than believed; thus, one’s ancestors ate different dishes, communicated in different languages, and believed in different values than their descendants (cf. ibid.). Reconstructing the past from a present-day perspective tends towards nationalization, that is, homogenization of life realities of earlier people in a given territory, which were actually much more heterogeneous than in the idealizations of contemporary nationalists: “Consequently, according to some historians, collective identities (national identities, say) are not the reliable products of collective memory that they seem to be but are, rather, fabrications—i.e., inauthentic; for the historical process of their fabrication can be described” (ibid.).

Consistent constructivism does not distinguish between nominal and real groups, and insists that no group can have a history. Due to fluctuations and transformations, the glorious traditions of (supposedly diachronic) collective identities are merely pleasing myths and phantasms. Imagined collective identities might even be misleading and harmful insofar as they distract people from the actually relevant lines of conflict. From a Marxist point of view, this is class struggle, from a libertarian perspective, this could be the individual struggle for survival (cf. ibid., p. 140). Descombes, however, considers it premature to reject the possibility of collective historicity. The fact that the history of a collective was not as glorious or impressive as it is claimed does not prove that this collective had no history at all (cf. ibid., p. 138). As elaborated and discussed in the following sections, it is indeed conceivable that some non-nominal groups can have a diachronic identity. For this, it is crucial to abstract a group from its members.

4.1 An interjection: is diachronic collective identity inherently illogical?

The Port-Royal Logic (Arnauld and Nicole 1996), first published in 1662, takes up the ancient Ship of Theseus paradox in modern times to reject the possibility of diachronic identity—even in the context of collective identity. Its authors Arnauld and Nicole attribute the problem of diachronic identity despite constant change to the observation that propositions have a grammatical and a logical form; however, these two forms are not congruent, but follow their own laws (cf. Descombes 2016, p. 146). Due to grammatical conventions, we sometimes speak about what is changing in the singular (e.g., “my rabbit is growing”) “as if it were, from one moment to the next and from its birth until its death, the same body.” (ibid., pp. 146–147) According to Arnauld and Nicole, however, changing animals (or collectives) are logically no single individuals, but the material totality of several successor animals (or collectives), as every change in the material composition would create a new individual (cf. ibid.). Because we can hardly follow these subtle changes sensually, non-logical ordinary language would tend to treat successive entities in time (e.g., my rabbit yesterday, today, and tomorrow as three distinct individuals) as if they were a single entity.Footnote 9 What grammatically seems to be a single selfsame entity would logically be the succession of many entities. Hence, the Port-Royal logicians come to a radical conclusion: material totalities with changing compositions cannot have a history (cf. Descombes 2016, p. 146). Their history is illusionary because, in fact, we are dealing with predecessors and successors. “It is as if, when attending a parade of brass bands at a votive festival in the village, I point out the passing groups one by one while saying in turns ‘This band is from Plougonvelin,’ ‘This band is from Locmaria-Plouzané,’ and so on.” (ibid., p. 147) Unlike successive marching bands, many “material totalities” with changing compositions occupy the same space across time, exemplified by the transformation of Rome: “Augustus said that he found the city of Rome made of brick and left it made of marble.” (ibid., p. 148) Logically, brick Rome and marble Rome cannot be identical because of their fundamentally different compositions. Cities are not simultaneously completely made of brick and completely made of marble. However, as both Romes occupy the same space and human perception tends to continuation, the mind constructs an entity (the diachronic spiritual Rome), which, as an immaterial bearer of temporally successive predicates (being first brick and later marble), can have a history (“never leave the stage;” ibid., p. 145) The immaterial entity remains identical to itself, while its predicates are mutable and multiple (Znepolski 2019). From Arnauld’s and Nicole’s point of view, however, the gain of being able to tell the history of Rome is dearly bought by logical inconsistency in language. Following the duplication, the same word sometimes refers to the (illusionary) body outside of time, sometimes to its material snapshot in time.

Arnauld and Nicole propose that the “self-deception” (Descombes 2016, p. 153) of identifying oneself with more successful group members can also not be understood logically, but rather psychologically: identification (here, putting oneself identical with other group members) is an attempt to elevate oneself by the deeds of others. “This exposes the grounds of vanity people derive from the fine actions of their nation in which they took no part, which is as foolish as for a deaf ear to glory in the eye’s liveliness or in the skill of the hand.” (Arnauld and Nicole 1996, p. 119). Thus, Arnauld and Nicole explain the phenomenon of feathering one’s own nest with someone else’s feathers in terms of misplaced vanity. Three centuries later, Tajfel and Turner (1986) identify the need for self-esteem as driving force behind social identification, including identity management strategies such as engaging in social creativity when identification does not provide positive distinctiveness. However, what insights does the Port-Royal critique offer and is identification really logically inconsistent and morally illegitimate? Descombes dissents, deeming it an error by the Port-Royal logicians to attribute the pride of the deaf ear to the confusion with other body parts, and not to the quality of belonging to the unique, selfsame body as those more talented body parts (cf. Descombes 2016, p. 152). Descombes argues that identification is not about a misguided confusion between self and others; instead, it signifies the “will to be oneself, a will to name oneself in conformity with a collective representation of oneself—one that allows each [..] to envisage a collective history in which her own individual existence participates.” (ibid., p. 153) This alternative interpretation enables social identification to be something different than the appropriation of others’ merits, namely, pride in a common We. When there is no crisis of the identical, identification does not refer to the other group members but to the group itself.

Now, let us look at three possible solutions to the problem of Theseus’ ship that have been offered in the history of philosophy. These three analogies share the commonality of constructing collective entities that are continuously existent despite permanent changes in their membership structure. Specifically, these constructions portray collectives as legal persons, political agents or embodying the general will. We will analyze their theological, historical and philosophical arguments to discuss whether these analogies actually answer the Ship of Theseus problem and whether they can help us to understand the continuity of collective and social identity.

4.2 First analogy: diachronic collective identity as legal person

As the Port-Royal Logic elaborated, diachronic collective identity is conceivable only for spiritual collective bodies and not for their material snapshots in time. Only a potentially immortal body can be the bearer of temporally successive predicates (e.g., qualities of group identity) across generations. Unlike Arnauld and Nicole, Descombes does not want to reject this idea: imaginary constructions can still be significant institution-givers and thus be real (cf. Descombes 2016, pp. 197–198). The first analogy grasps groups through a legal lens, tracing back to ancient Roman inheritance law, which differentiated between the possessions and debts of the city and those of its citizens. From the Roman jurists’ point of view, if all but one inhabitant left the city, the city’s property, assets, and debts are not passed as inheritance to the sole remaining inhabitant; instead, they remain the possessions and debts of the city (cf. ibid., pp. 157–158). Medieval Europe further developed the distinction between individual and collective legal parties from Roman inheritance law. Cities, monasteries, and universities were allowed to engage in contracts, establish juridical trade and business relationships with individuals and each other, lend money, and stand trial independently. In other words, fictitious, human-made collective entities were recognized and treated as legal persons in their own right, capable of legal relationships akin flesh-and-blood individuals (cf. ibid., p. 159). The inclusion of collective bodies of legal persons allowed the medieval repertoire of complex economic, political, and social government to expand enormously compared to the ancient world. The institution of the Church is an obvious example. “Ecclesia nunquam moritur”—the Church never dies (ibid., p. 154). Thomas Aquinas describes the Church as a mystical bodyFootnote 10 whose members act successively in “perpetual existence” (ibid., also Kantorowicz 2016), transcending the temporal limitations and constraints of natural bodies. As a body standing outside of time, the Church represents a claim that is universal and directed into eternity: “all future generations are called to join it.” (Descombes 2016, p. 154). In Aristotelian tradition, it is argued that despite the constant renewal of its matter (i.e., members), the ship of the Church remains identical to itself due to the persistence of its form (cf. ibid., p. 155). Unlike nominal groups, which are reduced to their unambiguously determinable set of members, mystical collective bodies have an indeterminable, uncountable, unidentifiable but infinite set of members, that is, potentially all human beings, not in time but in eternity.

Yet, how can mystical bodies, such as the Church, cities, monasteries, and universities, practically act as legal persons, for example, suing contract partners in conflicts or defending themselves against their charges? Medieval people were well aware that the Church or a university, as mystical immaterial bodies, could not possibly appear in court. They are conceived as “nomen intellectuale” (ibid., p. 156), names for a fiction, but a fiction with social utility, about whose validity there is social consensus. Instead, medieval glossators and jurists resorted to the institution of juridical representation based on Roman inheritance law. Thus, whenever a mystical body engages in legal activities, a natural person is determined as spokesperson to sign documents, pay debts, or attend court hearings on behalf of the collective body (cf. ibid., p. 158). In this process, roles are clearly defined. The representing person does not attend court as an individual, but as representative of the mystical collective body (cf. ibid.). Therefore, awarded profits and penalties do not affect the individual person, but the mystical collective body he or she represents, which is considered present even though it is not physically present.

4.3 Second analogy: diachronic collective identity as political agent

In Politics, Aristotle addresses a delicate matter with astonishing parallels to the medieval legal example (Aristotle 1999, p. 54). Sparta had lent money to the Thirty Tyrants who ruled over Athens. After the Athenians expelled their tyrants in 403 BCE and re-established a democratic form of government, Sparta demanded that the Athenians should repay the old debts. The Athenians refused, arguing that these were the debts of oligarchic Athens, not of democratic Athens. The philosophical content of this political maneuver is as follows: in the event of a fundamental change in the form of government, does a collective (here, the city of Athens) remain identical with itself, or is the collective replaced by a successor collective? Aristotle notes that a city’s identity cannot be determined topographically by its territory, which usually remains the same despite changes in government (cf. Descombes 2016, p. 160). After all, many tribes (ethnè) live together in delimited territory and, nonetheless, would not have the identity of a city (polis). As an example of a mock polis, Aristotle mentions Babylon, which in ancient times had a huge population, but which, when attacked by Cyrus the Great’s army, could not effectively defend itself. According to the legend, it took three days for the news of the enemy’s conquest of the peripheral districts to reach Babylon’s city center (Aristotle 1999). Despite living together behind walls, the Babylonians did not consider it important to report the appearance of the enemy army, continuing with their daily affairs “as if nothing had changed” (Descombes 2016, p. 163).

So, according to Aristotle, what is the criterion that distinguishes Athens as polis (or political agent) from Babylon? To qualify as a political agent, a collective must share a common political consciousness, with group members perceiving themselves as a unified community of fate with a shared interest rather than scattered atoms. The identity of the political collective is a normative question that does not depend on the composition of group members. The same individuals can associate themselves first with an oligarchic city and later with a democratic city, just as they “sometimes formed a tragic chorus and sometimes a comic chorus” (ibid., p. 160). In contrast, whether they identify with the identical polis is decided by the city’s “form of composition” (ibid.), not its specific material composition. This potentially diachronic form in which the citizens “are positioned or arranged relative to one another” (ibid., p. 161) is articulated in their constitution (politeia), representing the collective’s norms, dispositions, and concept of the public good. Consequently, due to their normatively different constitutions, the two Athens cannot be identical, absolving democratic Athens from the obligation to repay the debts of oligarchic Athens. In the Social Identity Approach’s meta-contrast, comparable to the idea of different (potentially diachronic) constitutions, the perception of two sufficiently different norms in a collection of people (intragroup differences smaller than intergroup differences) causes them to be perceived as two separate groups.

4.4 Third analogy: diachronic collective identity as general will

Finally, Jean-Jacques Rousseau’s concept of general will (volonté générale) can also be understood as an expression of diachronic collective identity. Participation in the general will is considered the precondition for the freedom of the I in the We, as each individual involved becomes “a part—on equal terms with her fellow citizens—of the sovereign and can therefore say of any enacted law ‘this law is nothing but the expression of our will.’” (Descombes 2016, p. 171) By subsuming the individual will into the general will, the I merges into a We, although Rousseau does not demand individuals to give up their self-love: “He is asking them to redirect this love and to love themselves as citizens.” (ibid., p. 172) The process of social identification is thus normative for Rousseau to preserve the freedom (read, self-determination) of all (Rousseau 2002). Comparable to Social Identity Approach’s differentiation between ingroup and outgroup, Rousseau’s We is distinguished into an inside and outside. The We emerges through individuation of a socially shared self from others. Is it, therefore, the sum of its parts? According to Descombes, a We is not a pronoun referring to a plurality of people, but to a “complex moral person” in the singular (Descombes 2016, p. 174). We is not be the plural of I because the word I has no plural form: “‘We’ is not a quantified or multiplied ‘I’; it is an ‘I’ expanded beyond the strict limits of the person, enlarged and at the same time amorphous.” (ibid., p. 175) In pragmatic usage, We is a term to address someone, which can be directed as an inclusive We to address Me, You [singular] and You [plural], the interlocutors who are members of the same ingroup, in contrast to Them, the outgroup (cf. ibid.). This We is inclusive because it semantically includes all the addressed people into a common group. However, there is also an exclusive We addressing Them, the outgroup members in interlocution, to tell Them about You, the members of the ingroup, who may be absent or present but silent during the speech act (cf. ibid.). Whether speaking in inclusive We or exclusive We, the ingroup is represented by Me, the speaker.

Coming back to the general will, its We is inclusive as it represents the consensus resulting from “a negotiation among those present” (ibid., p. 177). As inclusive We, it is a snapshot of the current members’ will and, consequently, the sum of its parts. However, because Rousseau’s social contract is not designed to be renegotiated with each change in membership composition, the general will, as an exclusive We, also addresses You All. You All: those who are not yet part of the general will, like future generations, but who are called upon to join Us. As precondition for a community to be created and to be maintained by future descendants, the exclusive We can thus be an expression of a diachronic collective body (cf. ibid., p. 178).

5 Discussion: social and collective identities from a philosophical-psychological perspective

This interdisciplinary article has directed its spotlight on social psychological theories (Tajfel and Turner 1979; Moscovici 1981; Breakwell 1993) and philosophical-historical semantics of puzzling identities (Descombes 2016). Before delving into discussion, I briefly summarize the previous findings in three steps:

First, to clarify what is meant by identity in principle, we have to distinguish the classical from the modern concept of identity. For the Social Identity Approach, social identity does not refer to objective or logical sameness, but to subjective identification with the qualities of the We. In contrast to classical identity questions (such as, “Is Marcus Tullius identical with Cicero?”—“Yes, because both names refer to one and the same person.”), modern identity questions (“Does Cicero identify himself as a Roman?”) can only be answered subjectively, not objectively. Thus, identity becomes an intra-psychic feeling or experience which is subject to fluctuations, and about which, in Cicero’s case, only Cicero himself, and no one else, can ultimately provide information. The self is understood as self-concept: a subjective view of one’s self. The classical understanding determines externally whether Cicero is identical to a Roman citizen. In contrast, only Cicero himself knows whether he identifies as a Roman citizen. Thus, with the modern understanding of identity, the focus shifts from who people objectively are to how they perceive themselves.

Second, regarding the question of what kinds of groups people actually (and not hypothetically) identify with, the Social Identity Approach answers cognitively with the meta-contrast principle and motivationally with striving for self-esteem through perceived positive distinctiveness. In contrast, Descombes takes the factual historicity of social reality into account. According to him, membership in real groups not (only) results from contingent stimulus processing but follows with a certain logical necessity. Though it must be acknowledged that across time, one cannot possess the attribute of membership in a real group alone (which, however, should be true for nominal groups as well; otherwise, one would not speak of a group), it is not convincing that this attribute is necessarily, and not contingently, shared with specific group members.Footnote 11 Moreover, it is important to emphasize that descriptions of reproducing groups do not imply a moral imperative for the generation of children to live, speak, or believe, as their parents did. Nor does it follow that the generation of children must differ from their parents’ generation.

Third, the central question of this article, whether identified group members’ We is the sum of its parts or something else, requires further differentiation. The question of whether individuals can continuously identify themselves socially is separate yet closely linked to the question of collective identity continuity. Descombes distinguishes between real and nominal groups to highlight that some categories are not actually constituted as groups (although they have the potential to constitute, according to the Social Identity Approach). Categories that are only conceivable are identical to the sum of their parts, and therefore, have no history. However, it could be different for groups that have been constituted, cultivated as collective identities, and proved to have agency. Although Descombes does not view all real groups as unchanging, static, or eternal beings, he posits that they may have the potential to form a diachronic We which, maintaining identity despite the group’s constantly changing composition, allowing people to continuously identify with. According to the aforementioned philosophers, the continuity of the We may result from specific qualities, namely, being a spiritual body with real-world significance (collectives as legal persons), being a political community with a shared set of beliefs and norms (collectives as political agents), or a general will that represents non-particular common interests. These conceptions of a diachronic We enable their members to be subject of their own collective history. These collectives are not substantial entities like individuals, but more than just a mere context in which individuals act (cf. Urfalino 2017). Transitioning from purely nominal history-less groups to real historical groups is possible, as we can imagine from examples of discrimination or dramatic events. If all shoes of size eight are removed from the market, while the other shoe sizes are sold continuously, the consciousness of a community of fate can emerge among the wearers of size eight shoes, motivating them to form a coalition to defend their rights (cf. Descombes 2016, p. 12). Social Identity Theory suggests that perceived disadvantages in social comparison drive processes of identity management, such as increased competition (Tajfel and Turner 1986). In this context, not only perceptions of unequal treatment but also the lack of recognition of social or collective identitiesFootnote 12 can fuel identity struggles when individuals feel that their identity is less valued or less accepted than that of other groups (Dubet 2021; Kruglanski et al. 2022). People seek recognition from others for their experienced lifeworld, which includes their social identities. Likewise, a previously irrelevant affiliation, such as being a member of a random elevator group, can become relevant when the elevator gets stuck, forcing those trapped inside to collaborate as a collective. Of course, knowledge of groups being constituted as “real” entities does not indicate whether they have continuity. After all, it is possible that identities as the We in the elevator are temporarily real during emergencies, but may not persist as the same identity once conditions change.

As we remember, the Ship of Theseus poses a significant obstacle to the possibility of continuous, diachronic identity: the permanent, factual change in group composition makes deriving continuity from its members (i.e., the parts of the sum) impossible. Therefore, collective identity is diachronic only under the condition that the identification does not refer to other group members, as discussed in the Port-Royal Logic, but to its shared We, which is the bearer of ever-changing attributes and qualities. According to Aristotle, the history of a group comes to an end once the norms and goals of the We change drastically, giving the collective a chance to draw a line under its past and become another. However, recent struggles in overcoming one’s problematic or traumatic collective past indicate that developing a new We and taking responsibility for the former We do not have to be mutually exclusive. The interpretability of the collective We is inherent to its conceptualization as social representations made by human beings (Moscovici 1973, 1981), a notion with precursors debated for centuries. According to the medieval glossators, the identified group member can serve as a representative, advocating for the singular image of the collective identity. Representatives speak, feel, and act on behalf of the group, or in the group’s name, striving to portray it positively and communicate its interests or attributes to the outside world. Yet, whether in inclusive or exclusive We, representation always carries the risk of failure. In discourse, other group members may reject representation if they feel that the shared We is not adequately represented by that person, or in that way. Others may speak against and challenge the spokesperson’s claim of representation if they do not recognize themselves in it (cf. Descombes 2016, p. 176).

Descombes sees social identities as opportunities for self-definition and individuation, enabling individuals to break free from predetermined roles, re-define, and shape their contingent selves according to their desires and imagining (cf. ibid., p. 199). His arguments are characterized by an optimistic tone regarding autonomy in identity formation.Footnote 13 However, it should be critically questioned how autonomous or emancipatory the choice of identity can actually be. Referring to elaborations on social influence in Social Representations Theory, individuals might be trapped in their social identities. This points to tensions between wanting to identify and having to identify out of historical, material, legal necessities, or normative pressure towards uniformity and definiteness (i.e., forced identification). Nevertheless, the context of these speculative questions is the social world, and there is no way out of this world. As the aforementioned social psychological theories emphasize, identity develops and unfolds in the social sphere, never in a vacuum. Human identity formation requires this social sphere, shared with at least one human counterpart. Since no identity is formed outside the social sphere, it may be appropriate to classify every identity as social.Footnote 14 According to this interpretation, social identity is a pleonasm. Consequently, Robinson Crusoe on a deserted island, without Friday, would not be an individual without a social identity, but a being without any identity. Importantly, this social sphere, or reality, is not (only) given but rather actively accomplished, i.e., a communal scaffolding, influenced by the knowledge we attribute to others (meta-knowledge), enacted communication, world-making (Elcheroth et al. 2011).

Following the previous considerations, we can conclude that the We is neither a plurality of summed-up members nor the plural of I, but a singular image to which everyone who feels like a link in the chain can refer.Footnote 15 The extent to which one’s self-concept overlaps or merges with the singular image of the We could be termed the degree of social identification. Nevertheless, this derived representation does not provide us with a definitive answer to the question of the We’s true nature. In the research literature, representations of collective self (or selves) diverge from singularity to plurality. For example, recent social psychological conceptions of collective agency raise the question of collective singularity anew. Understanding the representation of one’s social self as (I) subject of collective agency to be fundamentally distinct from the representation of one’s social self as (II) object of social identity results in a duality of “We” (oneself as part of an acting and experiencing “plural and self-aware subject,” i.e., collective agent) vs. “Us” (oneself “as an experienced object (defined by its observable self-characteristics),” i.e., exemplar of a social category; Shteynberg et al. 2022, p. 36). In contrast, if agency is understood as expression of individuals’ (observable) social identity, inferred “through the experience of their in-group effectively and autonomously pursuing its collective goals” (group-based control; Fritsche 2022, p. 194), the clear We vs. Us distinction dissolves. This way, one’s own subjectivity as an agent is itself the experienced object (and vice versa) but nothing separate. From this perspective, a singular representations of the collective self seems more plausible than dual or plural representations.

The Social Identity Approach, contrasting Descombes’ perspective, posits that all groups, even nominal ones, are represented singularly. Based on the meta-contrast principle, Turner and colleagues assume that a social category (i.e., group) is cognitively represented as a singular prototype, i.e., the image of the most typical member of this category that is perceived to represent the category as a whole (Turner et al. 1987). On a specific dimension of comparison, the prototype is the target stimulus perceived to have the maximum quotient of difference from members of the comparison category divided by the difference from other ingroup members (Turner et al. 1987). Therefore, the identity to which social identity refers is not the sum of group members but a singular mental concept of the prototype representing the group. If the Self-Categorization Theory’s concepts are valid, the possibility of social identity continuity depends on whether there is (and can be) continuity of the group prototype. Whether individuals maintain consistent and continuous identification and whether their prototypes are continuously mentalized in the same way are empirical questions. However, it is a non-empirical question whether the criteria for assessing and evaluating continuity should be quantitative or qualitative and which threshold values should be used to determine whether there is continuity or not. So far, consensus on these matters seems elusive.

Maybe there is nothing underneath the renewing stock of masks that one perceives as one’s self. In the considerations presented by Self-Categorization Theory, the Social Identity Approach is constructivist, doubting the ideas of an innate true self or a self that, once developed, remains the same forever. According to this perspective, the identification with groups, represented as prototypes, depends on the specific conditions of the situational meta-contrast comparison. Thus, depending on the contrasts which are salient in a given situation, social identities may or may not be robust. Prototypes resist change when the situational conditions of the meta-contrast context do not change, that means, same salience of the same groups and constant ratio of perceived similarities and dissimilarities between their members. This appears to be impossible in absolute terms, considering the factual change in social situations. Consequently, we could get the impression that Self-Categorization Theory advocates the fluidity of social identity: due to the permanent change of the social sphere, people could identify discontinuously, and the only continuity in identification would be its change, as we identify with one group at one moment and with another group at the next. Conversations about the best wine-growing region can evoke a sudden sense of regional identity. When some minutes later, the discussion will turn to the gender pay gap, the discussants will probably see each other less as inhabitants of certain regions, but more as women or men. Over the course of a single day, we can experience ourselves successively in hundreds of social identities, depending on the situational meta-contrast comparisons and the assessment of which salient group membership situationally promises optimal satisfaction of needs. Tying social identification and prototype construction to situational conditions makes identity continuity and mental representation of the group’s prototype contingent on these conditions. Right now, I may attribute the German prototype to be effective in dealing with crises, but this attribution may change in response to any new information. Interpreted in a postmodern way, one’s social self-concept could be fragmentary and fluid, today like this, tomorrow like that, blurring into an infinity of possible identities. Taking these kinds of considerations to the extreme would lead to the idea of a pinball identity, in which an out-of-control ball (as a metaphor for meta-contrast comparisons) activates different flashing fields of identity in rapid succession.

However, it is important to note that these considerations described above are not the conclusions drawn by the Social Identity Approach. On the one hand, the meta-contrast principle emphasizes the situational dependency and fluidity of identity over concepts of identity stability (Hornsey 2008). Contrary to Erikson’s idea of a critical life period of identity formation as core developmental task of adolescence, after which the formed identity consolidates and becomes more or less stable, Self-Categorization Theory emphasizes the (lifelong) “ad-hoc construction” of social identity (Deaux 2014, p. xix). Precisely, “variability in self-categorization is not arbitrary or chaotic but is systematic and is lawfully related to variation in social contexts.” (Turner et al. 1994, p. 458) Nevertheless, the approach underscores the potential for identity fluidity rather than its realization. Turner and colleagues do not imply that due to permanent meta-contrast updates, identity is dissociative, nor do they suggest that individuals are unable to maintain a sense of self. On the contrary, various factors, including contextual, situational, and personal factors, imply that despite meta-contrast comparisons, human identity is stable and continuous to a certain extent, at least in cases that are not pathological: “the fact that aspects of identity have been shown to change in salience from moment to moment does not make the notion of identity stability vanish into thin air. Instead research suggests that such contextual shifts in identity typically occur in a predictable manner, based on features of the context in which individuals find themselves” (Vignoles et al. 2011, p. 11). Among the contextual and situational factors that stand for continuity is the observation that, in our everyday lives, we frequently find ourselves in similar situations. People live their lives in familiar structures and contexts ordered by roles, schedules, conventions, rituals, routines, norms, that repeatedly make similar meta-contrasts salient to them. Regular interaction with one’s children and partner makes one’s parental identity chronically accessible, regular interaction with colleagues, subordinates, and bosses one’s professional identity, activated through structural or comparative fit in meta-contrasts (i.e., perceived [dis-]similarities). People are not only passively exposed to recurring contexts; they actively seek them out repeatedly; for instance, to have their existing views verified, stabilizing their identity (Vignoles et al. 2011). In addition, categorization as a routine is based on normative fit, i.e., the extent to which perceived stimuli are congruent with stored stereotypes or normative expectations about social categories (e.g., Blanz 1999). Just as people’s familiar world (usually) does not fall apart or disintegrate, neither does their identity. Regarding changing contexts, Identity Process Theory describes the assimilation-accommodation process which embeds new information into the identity structure (and vice versa), thereby preventing the disintegration of identity. Importantly, however, neither change nor scaffolding structures nor familiar contexts determine people to identify in a specific way. Instead, these structures and contexts only increase the probability of certain identifications while decreasing the probability of others by making certain similarities and dissimilarities particularly salient. In an office context where collegial relationships are salient, individuals are more likely to identify with a common professional identity. The same people would be less likely to perceive themselves in this professional identity when they meet at a public viewing wearing scarves in different colors but more likely to perceive themselves as fans of rival sports teams. Nonetheless, salient comparison results do not dictate who one can be in a situation, which is why unprofessional sports rhetoric may also occur in the office.

In addition to situational or contextual accessibility, personal factors play a central role for “intact identity,” referring to inter-individually different chronic (or chronified) patterns of categorization determined by chronic hierarchies of cognitive accessibility (ibid.). Sexists have a tendency to repeatedly perceive gender categories across different situations and contexts, as these categories are extraordinarily accessible to them, whereas the results of meta-contrast comparisons are less tendentious for non-sexists, just as racists tend to perceive race categories, and so on. Individuals typically have hierarchies in how they categorize themselves and the people around them, as well as identify. Generally, individuals do not forget about their former identities and are (often) motivated to continue them. Motives and goals served by one’s identity may be connected to one’s personality traits, biography of socialization, learning, and experience. However, the Social Identity Approach does not delve into the origins of these chronic or chronified identification patterns, leaving these questions to related disciplines, such as personality and developmental psychology (cf. Erikson 1968, 1980, 1986; Marcia 2002), and integrative approaches such as Identity Process Theory (Breakwell 1986, 1993, 2014). As a‑developmental theory, the Social Identity Approach does not make assumptions about how identities stabilize or persist within individuals over time. Especially personality and developmental psychologists examine in depth how people’s chronic identities develop and consolidate, while more sociological (and more qualitative) approaches focus on how contextual patterns of identification develop and consolidate. In conclusion, contextual, situational, and personal factors explain why the Social Identity Approach, despite its foundation in the potentially fluid meta-contrast principle, does not posit a fragmentary or even dissociative pinball identity. Instead, “perceiver readiness and normative fit–reflecting an individual’s motives, desires, memories, habits, and so forth–provide definite internal psychological constraints on self-category variation […] where there are stability and continuity in self-definition, these are produced by the same processes that make possible fluidity and change” (Turner et al. 1994, pp. 459–460), depending on the (in-)stability of contexts, knowledge frameworks, groups, and social influence.

Nevertheless, we have not yet clarified whether it is justified to attribute continuity to this more or less stable sense of identity. Does the notion of continuity of social identification contradict the observation that many people seem to have very similar or even the same identities on Mondays, Tuesdays and Wednesdays, but different ones on weekends? In this example, situational factors seem to outweigh personal ones, as for many people, contexts of social comparison are similar on Mondays and Tuesdays, but change significantly on weekends. It seems to me that there are at least two different ways of understanding the continuity of social identification. One way is to look at a defined period of time and then examine whether identity appears continuous over this period. If the zoom is set to one week, a target person’s identity may appear discontinuous, fragmented, or fluid. After all, we observe how different the target person’s thoughts and behaviors are when comparing him or her on Mondays and Sundays. However, zooming out to the level of a month or year reveals the same target person’s identity as highly continuous, as in repetition, this sequence proves to be a predictable habit or pattern, indicating a pretty solid identity. Therefore, depending on perspective, patterns or calculated mean values of continuity (i.e., identity on average) may be apparent or not. The continuity criterion according to this understanding is comparatively lax because, whether in pattern or on average, what is sought here is approximate continuity.

However, when we remember the argument of the Port-Royal Logic, there is an alternative way of understanding identity continuity that is stricter.Footnote 16 If continuity is defined as a chain that must not break, it is necessary for identity to not deviate (i.e., change) at any point to be continuous. According to this understanding as continuity without exception or continuity until the first change, whether the chain breaks after one minute or twenty years is irrelevant; therefore, perspective does not matter. It is obvious that social identity, as subjective self-concept, can only be continuous according to the first, laxer understanding but not according to the second, stricter understanding. After all, the very idea of developing an identity across biography would contradict the demand of the second understanding. Even if we ignore childhood and adolescence, and even if there is a stable sense of identity in adulthood, the chain of identification would still break whenever an individual’s consciousness is lost, for example, during sleep, delirium, or when fainting. It may be countered that loss of consciousness does not necessarily imply a break in the chain of identification, as this loss could be interpreted as a press on the stop button of identification rather than a real break up. Nonetheless, it seems implausible to me that identification would remain exactly the same (i.e., unchanged) in moments of changed consciousness, for example, during sleep. In their dreams, psychologists-by-day may identify as pirates-by-night, which, in my view, strongly suggests that continuous identity in the strict sense of sameness cannot be sustained for long without disruption.

In their constructivist assumptions, Social Representations Theory and Identity Process Theory can be interpreted consistently with those of the Social Identity Approach, even though the theories’ roots go back to different traditions of thought. Social Representations Theory makes less specific predictions about how social representations determine the formation of social identity than the Social Identity Approach or Identity Process Theory but rather describes the constitution of the social sphere in which representations are associated with memberships and form the framework in which identity evolves. Accordingly, the content that serves as reference in the Social Identity Approach’s meta-contrast comparisons could be formed in negotiations of social representations, which may be constructed upstream before finding their way into meta-contrast comparisons. These considerations demonstrate that there is basically no dichotomy between psychological and philosophical approaches to the question of collective or social identity and their continuity, illustrating that in both disciplines similar questions are subject of debate. Debating the Ship of Theseus, both Social Representations Theory and Identity Process Theory would argue that the diachronic continuity of a collective entity has essentially nothing to do with its material composition. Instead, a collective is represented as the same, ignoring changes in its composition, as long as a socially shared understanding of this entity persists, which is solidified through collective meaning-making practices into a social fact (Elcheroth et al. 2011). However, there is something unsatisfactory in this hypothesis of constructed reality. If there is no truth criterion for representations outside of social exchange, it is difficult to distinguish between true and false representations. In other words, we have to rely on “social facts,” i.e., told stories, without being able to say with certainty whether they are true. With regard to the first interpretation of continuity as an approximate pattern or mean value over a defined period of time, an answer may be less ambiguous: in communicative exchange, people can reach or constitute an (approximately) shared understanding of whether their group has a history, what it looks like, and to what extent the attribute of continuity applies to their group. However, if we understand diachronic continuity to mean that there must be no break in the temporal chain, then the socially shared representation of collective continuity at a certain point in time is, of course, not sufficient to speak of real continuity. To answer this specific question, we cannot rely on memory. In retrospect, every living person can be wrong about the times when none of the living was there. For the representation of collective continuity to be objectively true, the socially shared understanding of collective identity must have existed at all times for as long as the present-day collective claims to exist. Yet, how could such a claim be empirically verified, given the gaps in the sources? Social understandings are never documented every (milli-)second. Concerning the question of continuity, it is also crucial to consider whose understandings are relevant. Does it have to be the shared understanding of self-identified ingroup members (successive members’ ongoing self-perception of “We are one and the same” as necessary condition?) or could the continuity of a group be constructed externally (“They remain the same group”), without any ascribed ingroup member sharing this understanding? However, if the latter were true, how would it be possible to distinguish between real and imaginary made-up groups?

6 Conclusion

The singular image or representation that the group ultimately is could be described through various lenses: in the psychological language of the Social Identity Approach it is a prototype, medieval Christianity would term the singular image a mystical body, Aristotle a polis, insofar as it concerns groups with normative ideals, interests, and intentions, Rousseau the volonté générale that is directed outward and into the future. Crucially, none of these concepts result from the sum of their individual parts but from social emergence. Furthermore, they can deepen our understanding of how imaginary, yet not unreal, groups can endure, resist disintegration, and effectively impact their environment. Probably, the clearest foreshadowing of the Social Identity Approach’s core ideas can be found in Rousseau’s political philosophy, where the demand to love oneself as a citizen, as well as the differentiation between one’s private interests as an individual and one’s common interests as a citizen (Rousseau 2002),Footnote 17 point to the very social self that Tajfel and Turner later emphasized.

However, it is important to clarify that all these presented philosophical and psychological interpretations of collective or social identity do not parallel.Footnote 18 Whereas each concept understands the We as singularity, only the mystical body is characterized as eternal, and whereas the polis and volonté générale are united by their will to shape politics, only the polis is clearly localized and circumscribed. By contrast, the psychological prototype is a mental representation by individual subjects (and Social Representations Theory would specify, not an individual, but socio-cultural representation). Unlike the representatives of the mystical body, the prototype is not a participating yet separate entity but a construct that is in a way identical to the group. Moreover, unlike the other concepts, Self-Categorization Theory’s prototype does not have to be socially shared by necessity. Logically, a prototype is no more or less a prototype, whether it is mentally represented by one, a hundred, or eight billion people. In contrast, for mystical bodies, the polis, volonté générale, or social representations, it is conceptually crucial not to be idiosyncratic images of individuals but images shared by collectives. However, the previous arguments are not meant to imply that prototypes are inherently individual or a‑social as opposed to social representations. Social Identity Approach and Social Representations Theory can very well be thought of together if we transcend the institutionalized separation of paradigms (cf. Breakwell 1993; Elcheroth et al. 2011). Social representations within one’s cultural-historical context may provide the framework for (approximately) stable identity formation. Accordingly, prototypes could be the social emergence phenomena described by Social Representations Theory and Identity Process Theory. Despite surface differences between these theories, we have to keep in mind that essential for the Social Identity Approach is methodological individualism, not epistemic individualism. Even if methodologically, social identity researchers ask individuals about their perceptions, it is not claimed that these measured perceptions emerge individually. Consequently, on an epistemic level, there is no inherent contradiction between the Social Identity Approach, Social Representations Theory, and Identity Process Theory.

Zooming in, the extent of continuity in group prototypes has been discussed previously. However, a satisfactory answer has not (yet) been identified due to the absence of a consensual criterion for defining the threshold or minimum of immutability necessary to speak of prototype continuity. At this point, subjective perceptions set us the limits on what can be determined. The aim of this article, to more holistically elucidate the foundations of identity and the conditions of its continuity, must be considered incomplete. Nonetheless, I hope to have given the readers some impulses for their own reflections and considerations of these questions.