We will now summarize conceptual developments of group psychology relevant for polyadic collusions and then present different types of collusions involving groups.

Mass phenomena were first described in the late nineteenth century by Le Bon (2013), who introduced a psychological perspective on groups and observed an alteration of the individual when joining the group, such as the obliteration of singular particularities, diminished sense of responsibility and increased suggestibility. Freud believed that identification among group members is essential for its functioning (Freud, 1921) and that social groups may be linked to sexual jealousy (Freud, 2013); he thus introduced the notion of unconscious group dynamics.

Bion proposed studying the group as a whole (Bion, 1969) and found that groups function as work groups, collectively executing what is equivalent individual ego functions, and with “basic assumptions”. He divided basic assumptions, which are driven by powerful unconscious dynamics, into assumptions of dependence (between group and leader), pairing (between group members) and action (e.g., to fight against someone or something or to flee). In this last group, especially relevant for collusions involving a scapegoat, Bion assumed that the group focuses on its own preservation: the focus on an enemy diminishes infighting and fosters its cohesion. Basic assumptions can help us understand certain types of polyadic collusions.

Later, other group phenomena, such as group illusion (Anzieu, 1990) or anxieties (Stephan, 2014), were described. Anxieties about real or fantasized dangers, for example “to harm” in the supervision of the palliative care team described above, may lead to collusions. More recently, the concept of valence shed light on how individuals might be attracted to groups. Valences are affective qualities the individual experiences—dichotomized in attractiveness and aversiveness—when facing an event, an object, an environment, or a situation. Valences orient judgment (Shuman et al., 2013) and can thus favor that a person collusively joins a group.

Regarding the wider context, Foulkes and followers view individuals as nodes at the intersection of groups and parts of a larger matrix, subjected to a communicating system, which includes supra-individual elements such as a sociohistorical context (Foulkes, 1975). He thus articulated psychopathology within the larger social context, which also operates in collusions.

We propose to classify polyadic collusions into group collusions, collusions between a group and an individual, and collusions within and between groups.

Collusions can occur at the origin of group formation. Group members share an unresolved issue, constitutive of the group, which motivates a basic assumption (e.g., pairing in a sect, unresolved issue: separation) that exceeds the work group tasks. Members may share characteristics such as weak ego boundaries, dependent personality traits or the use of projective identification to evacuate shared and unwanted self-representations projected on individuals outside the group. The formation of sects is not comprehensible from a mere psychological perspective, but we consider that shared dependency issues may be operant and that the concept of collusion may thus help us understand such formations. In comparison, groups not forged by collusion such as religious movements or football clubs primarily execute the tasks of the work group (charity or to win), and their members are not united by a shared and unresolved issue.

Other examples of whole group collusions, due to prevailing of basic assumptions, are members who forget their working tasks due to the joy of being together (basic assumption of pairing), which can sometimes be observed among health care professionals working in a hospital who momentarily forget that patients are present, or political protesters who become violent when intergroup divisions appear (basic assumption of fighting, fostering cohesion).

In suicide research, shared projective identification—a superglue that welds members of a group (Young, 1992)—is related to suicide clustering among adolescents (Goldblatt et al., 2015). Suicidal wishes in adolescents can be understood as a fantasized means to hurt the parents and have them experience the pain the adolescent feels in this period of desired and feared separation. This desire to “communicate to the parents the pain they experience” can become a fatal “suicidal collusion” when a member acts out the fantasy of the group.

Collusions between a group and a leader or a scapegoat or other individuals, such as group therapists, are often but not exclusively mobilized by narcissistic issues.

Collusions between groups and their leaders can be symmetrical, when a group and its leader resort to mutual idealization and crossed projective identification to defend against an unresolved issue related to self-worth. Such collusions can occur in families with very hierarchical organizations (Petriglieri & Stein, 2012) or in politics. Messianic hope in times of loss of confidence, collusively produced between narcissistic group members and their narcissistic leader (Bateman, 1998), was exemplified by the Trump presidency in the USA. Mentzos classified such collusions as “narcissistic-reparative”, “narcissistic boosting” and “narcissistic-destructive collusions”. In the last subtype, unwanted parts of the self are projected onto others outside of the group who are devaluated, desingularized and described with stereotypes (Mentzos, 1988). Such collusions are especially dangerous when the leader incorporates not only the ego ideal but also the superego, “assumes” responsibility, and frees group members of guilt (Main, 1975). Trump’s famous words that he could shoot someone on Fifth Avenue and get away with it and his cheered devaluations of individuals and subgroups within and outside the USA are the manifestation of a narcissistic-destructive collusion. An example of a complementary collusion is a leader idealized by a group he despises (as has been reported regarding Hitler, who stated when war was lost that Germany did not deserve him). In both cases, the idealized leader incorporates the ego ideal of group members, who vicariously experience grandiosity and ward off low self-esteem and shame. In both symmetrical and complementary collusions, the leader needs the group’s admiration to prevent the erosion of their own self-esteem. The addictive call of some politicians for rallies to take in the experience of a crowd can be a manifestation of such a need. A recent psychoanalytic perspective on leaders convincingly described how different types of leaders (e.g., “controller”, “therapist”, or “messiah”) endowed with different qualities (e.g., pragmatic, romantic, or rebellious) can coexist in the same person or be distributed among different leaders (Western & Wilkinson, 2010).

Collusions also occur between a group and a scapegoat. Scapegoating may or may not be collusive, with collusive scapegoating being much less frequent and requiring a different therapeutic approach. The victim of ordinary scapegoating needs to be recognized, protected, and compensated; the victim of collusive scapegoating needs therapeutic intervention. In collusive scapegoating, the targeted individual shares with the group an unresolved issue and introjects the unwanted parts projected from the group, providing opportunities, through projective identification or behaviors provoking reactions from the group, to confirm the group’s view and to continue scapegoating. Collusive scapegoating circulates under different terms, such as “identified or designated patient”, “scapegoat” or “symptom carrier”, with or without reference to collusion (Wangh, 1962). Children do not have inborn unresolved issues, but when they—out of feared or real danger of separation—internalize the projected self- or object-representation of their parents, they may become participants of a collusion.

In the therapeutic setting, polyadic collusions are observed in group therapy. A therapist officially called to solve a problem may serve to ease inner-group tensions, increase group cohesion and externalize conflicts: the group avoids the unresolved issues regarding aggression (and separation anxiety). Collusion is at work if a therapist integrates the projected aggressive feelings and—paralyzed by own unresolved issues regarding aggression—reacts by distancing him or herself from the group, herby fueling the collusive spiral (Loeser & Bry, 1953). Such phenomena also occur in group supervision when supervisors collude aggressively with the supervisees who consider the supervisor insufficient (collusion over feelings of impotence).

Collusions within groups (Gemmill & Elmes, 1993) may manifest as splitting, as described in the introductory example of supervision: crossed projective identification with increased tensions provoked aggressive attitudes between the two subgroups, confirming their view that the other’s position is aggressive and potentially harmful to the patient.

In collusions between groups, the same unresolved issue is avoided by crossed-projective identification (Young, 1992). For example, the unresolved issue concerning separation (anxiety) prevents intragroup aggression and conflicts, which are collusively externalized with the help of another group that operates in the same way for the same reasons. Mutual accusations of therapeutic inefficacy between psychoanalysts and cognitive therapists may serve as an example (of repressed anxiety related to being inefficient). In addition to collusive intergroup aggression, unresolved narcissistic issues might come into play through collusive rivalry. Collusive rivalry may be observed in the hospital setting, with members of medical disciplines viewing members of other disciplines as inferior and vice versa. From a general point of view, collusions occur more easily in settings where the unconscious is not part of the work or neglected. In the psychoanalytic setting, collusion is seen as a way of entering into communication to gain a deeper understanding. However, collusions with negative effects on the therapeutic relationship have been described even in the psychoanalytic setting.

Mentzos stretched the concept of polyadic collusion to suggest that externalizations by means of cross-projective identifications between nations may explain certain wars (Mentzos, 1988). He conceived wars as narcissistic crises in which shame (low self-esteem) is avoided through collusive intergroup aggression: revenge takes the place of shame, and blood the place of tears. We find it questionable to understand wars solely from a psychological perspective, thereby anthropomorphizing nations (Petriglieri & Petriglieri, 2020).