La diversité de l'offre disponible sous le mot « coaching » est source de confusion pour les clients, les commanditaires et les coachs eux-mêmes. Dans une visée de didactique professionnelle, à partir d'une démarche de recherche socio clinique compréhensive auprès de six cas de coachs, il est montré que le référentiel est davantage un indicateur qu'un organisateur du dit « coaching ». L'absence de conceptualisation du coaching a pour effet l'impossibilité d'un consensus autour d'une définition, d'un référentiel, d'un mode de transmission et de reconnaissance. A partir de quatre organisateurs des métiers de l'humain que sont le rapport au savoir, le rapport aux normes, le rapport aux valeurs et le rapport à l'image du Soi professionnel, les résultats relatifs à l'activité des cas de coachs et à leurs dynamiques identitaires montrent que l'usage du référentiel est un indicateur de leur posture : guidage ou accompagnement. Puisque le mot « coaching » ne dit rien de la posture de celui qui s'en prévaut, la dénomination désignant l'activité d'accompagnement professionnel dans le champ des ressources humaines et ceux qui la pratiquent devra évoluer. Former ces accompagnateurs aux usages contraires du référentiel permettra que s'élaborent le genre du métier, et le style de ces accompagnateurs. / Market supply over the word « coaching » is confusing for customers and partners, and for coaches themselves. With occupational didactics as a line of sight, and a socio-clinical comprehensive approach to study the case of six coaches, this research comes to the conclusion that the reference table appears to be more an indicator than an organizer of what is called « coaching ». Because coaching has not yet been conceptualized, there is no consensus on its definition, nor on a reference table, or on ways of teaching it and it lacks recognition. Using four organizers of human businesses, which are the relationships to knowledge, to norms and standards, to values, and to one's professional self, it is possible to analyse the coaches' activities and identity dynamics. The obtained results show that the use of a reference table is an indicator of the position they adopt: either in a guidance or accompanying position. Since the word « coaching » says nothing about the position adopted by the ones who refer to themselves as coaches, the denomination referring to the professional -accompanying activity in the field of human resources, and to those who undertake it, will have to evolve. Training these accompanying professionals to the contrary usages of the reference table will allow the elaboration of the profession type and the style of these accompanying professionals.

Episodes of aggression and violence continue to beset our public spaces. This thesis explores how well we understand the transition to violence—and how aggression and violence in public spaces can be managed or controlled. We begin by arguing that established social psychological approaches to aggression and violence are inadequate for the task. Existing models explain violence through the failure of individuals to inhibit their own impulses or control their own emotions sufficiently. At best the models allow for the importance of dyadic interactions as individuals provoke each other as part of an escalation cycle. We argue that public space aggression and violence involves multiple parties and more complex sets of social dynamics. We suggest that, at the very least, the roles of third-parties and social categories need to be at the heart of theorising about violence in public spaces. To support our arguments, we examined violence directly through detailed behavioural microanalyses of real-life aggressive incidents captured on CCTV footage. We also built agent-based models (ABM) to explore different theoretical approaches to the impact of groups and third-parties on aggression and violence. The thesis contains seven studies. We begin with a CCTV behavioural microanalysis (Study 1) that showed collective group self-regulation of aggressive and violent behaviour in both within- and between-group conflicts. This study demonstrated an ‘intergroup hostility bias’, showing a greater likelihood of aggressive, escalatory acts towards outgroup members in intergroup conflicts than towards ingroup members in intragroup conflicts. Furthermore, this study demonstrated an ‘intragroup de-escalatory bias’, showing a greater likelihood of peace-making, de-escalatory behaviours towards ingroup members in intragroup conflicts than towards outgroup members in intergroup conflicts. Overall, we found that the majority of coded actions were acts of de-escalation performed by third-parties. With evidence stressing the importance of social dynamics, we compared dyadic models of aggression against an alternative social model (which allowed normative influence of others) in a dynamic agent-based modelling environment. We modelled the dynamics of metacontrast group formation (Studies 2 and 3), and found that group processes can produce both escalation of violence and inhibition of violence (Study 4). We found greater polarisation of violent positions in intergroup interactions than in intragroup interactions (Studies 5a and 5b). However, an emergent intergroup hostility bias did not emerge from this polarisation process. In Study 6, we re-examined the intergroup hostility bias present in our CCTV footage. We found an intergroup hostility bias for non-physical escalatory acts but not for physical escalatory acts. We examined the standardised number of actions contributed by third-parties and assessed the relationship between specific third-party conflict management strategies (policers and pacifiers) and conflict violence severity (Study 7). Overall, our results showed that third-parties and groups are integral features of the dynamics of violence. Third-parties largely attempt to de-escalate conflict, and the conflict management strategy they employ has a direct relationship to the violent outcome. Groups have a tendency to de-escalate their own members, and self-policing and collective inhibition take place. These findings have importance for current models of aggression and violence and also for evidence-based violence reduction initiatives.