The core definitions
supporting Teamersive and Audrey Loos’ research

🔬📊

 We propose a communicational approach to teamwork. This requires rigorously defining five foundational concepts that structure our research: the team, the communication network, performance, resilience, and robustness.

The team

Although the term is widely used, rigorous academic definitions of the team remain scarce. Based on Weick (1993) and Mucchielli (2019), we define the team as a set of individuals in pursuit of a common objective, developing shared cooperation processes across time and space. Existing as a team above all implies that members collectively recognise it as a collective mind, which generates consented interdependence and co-responsibility toward the task. We therefore consider the team as an integrated system whose functionality emerges from the convergence of individual contributions toward a common horizon.

The communication network

To analyse a team, we model it as a communication network. Monge and Contractor (2001) define this network as the modes of contact constituted by the exchange of messages, characterised by their structures and flows. Inspired by the approach of Leenders et al. (2016), we substitute the static snapshot of the network with a continuous film of interactions: relational event networks. Team members are the nodes, and each verbal or non-verbal interaction, timestamped and directed, constitutes a relational event that feeds a dynamic adjacency matrix. We can thereby observe the communicational architecture of the team through structural signatures (such as density, centrality, or reciprocity) that operationalise latent dynamics such as cohesion, trust, and the stability of exchanges (Leonardi & Contractor, 2018; Leenders et al., 2016).

Performance

Performance is defined as a team’s capacity to achieve its objectives with effectiveness (achieving goals) and efficiency (optimal use of resources) (Drucker, 1999).

Beyond compositional characteristics, we can identify relational precursors of collective performance:

Social integration

Bell et al. (2015), in their study of long-duration space missions, show that socially well-integrated teams, whose members exhibit strong compatibility at the level of deep attributes (values, personality, emotional intelligence, cultural backgrounds), manage stress and conflict more effectively. Surface-level diversity (gender, age, nationality), by contrast, proves less determinant than deep diversity for cohesion over time. Furthermore, when a team grows beyond a critical threshold, the formation of subgroups increases, undermining coordination and collective performance (Mucchielli, 2019; Bell et al., 2015).

Shared mental models

Shared mental models (SMMs) constitute one of the most robust predictors of team performance. DeChurch & Mesmer-Magnus (2010), in their meta-analysis, define them as knowledge structures held collectively by team members, enabling them to form precise explanations and expectations regarding the task and, consequently, to coordinate tacitly, including under pressure, without the need for explicit communication. According to Gisick et al. (2018), these knowledge structures are always relative to an object: they concern the task (procedures, processes, planning), the technologies deployed (tools, systems), interactions within the team (roles, responsibilities, communication channels), or the team members themselves (competencies, personalities, expertise, hierarchical relationships). This plurality of objects is decisive: strong alignment on the task may coexist with misalignment on roles, and it is precisely the articulation between these different dimensions that conditions the team’s level of performance.

To measure them, DeChurch & Mesmer-Magnus (2010) distinguish three complementary levels of analysis. The elicitation method identifies the content of mental models — that is, the key concepts identified by each person. The structural representation assesses how these pieces of knowledge are organised and connected in the individual’s mind. Finally, the emergent representation captures how individual mental models combine to form a shared model at the collective level, characterised by a degree of similarity or consensus among members. Smith-Jentsch et al. (2001) operationalise SMMs along four dimensions: leadership, information exchange, supportive behaviour, and communication — the critical components of teamwork that members must hold a common understanding of to improve coordination and anticipate each other’s actions.

Leadership configuration

Leadership within a team is best understood as a relational configuration: a topology describing how influence relationships are distributed among members. It is at this structural level that leadership becomes a precursor of collective performance. Lungeanu et al. (2022) distinguish two major classes of configurations: connected networks, in which all members are engaged in direct or indirect influence relationships, and fragmented networks, in which subgroups or isolated members remain disconnected from the leadership dynamic.

It is shared and connected leadership that produces the strongest effects on shared mental models and, consequently, on performance. Teams in which multiple members mutually influence one another develop stronger cognitive convergence than teams with a single or fragmented leadership (Lungeanu et al., 2022; DeChurch & Mesmer-Magnus, 2010). This result stems in particular from the fact that a shared configuration engages more members in framing and negotiating collective representations, whereas hierarchical leadership imposes top-down convergence from a single reference framework.

Regardless of its configuration, leadership in a team fulfils four essential functions: support (maintenance of relational resources and cohesion), interaction facilitation (development of reliable group relations), work facilitation (regulation of roles and planning of tasks), and goal emphasis (maintenance of collective direction) (Bowers & Seashore, 1966, cited in Lungeanu et al., 2022).

Network configuration

Leonardi & Contractor (2018) add that network structure must adapt to the objective: high internal density favours efficiency, while access to diverse external ties supports creativity and complex problem-solving.

Temporal dynamics of interactions

Leenders et al. (2016) show that temporal patterns, such as inertia (maintenance of existing ties), reciprocity (tendency to return an interaction), and transitivity (propagation of ties through third parties), structure communicational dynamics and constitute underlying mechanisms of collective coordination.

Nevertheless, as we have established, the obsession with performance tends to produce hyper-efficient but rigid systems that are unable to absorb unforeseen shocks (Hamant, 2022).

Resilience

The literature on organisational teams under pressure has largely relied on the concept of resilience to conceptualise the capacity to navigate crises. Resilience can be defined as a system’s capacity to actively reorganise after a shock, preserving its essential functions and aiming for a return to a stable operating state, what Hollnagel et al (2006) call bounce-back. It is therefore a concept focused on recovery: the process of returning to stability is active and sequential, but the intended outcome remains the restoration of a prior state, not the transformation of the system.

Communicational sources of resilience

Weick (1993), in his analysis of the Mann Gulch disaster, shows that resilience is above all a communicational phenomenon. He identifies four resilience sources: (1) creativity and improvisation in the face of the unprecedented, as a driver of collective sensemaking (the members’ capacity to construct a shared meaning of the situation); (2) respectful interaction (listening, mutual trust) to maintain communication flows during a crisis; (3) clarity of roles, enabling each member to anticipate the actions of others and redistribute tasks in the event of failure; (4) an attitude of wisdom, as the capacity to question established norms and routines rather than blindly conform to them. Weick & Orton (1990) add loose coupling: the flexibility of links between system elements provides agility and limits the propagation of failures, while enabling local initiatives during a crisis.

Resilience structurally presupposes that a shock has already occurred and that a loss of function has taken place. It does not protect the system before the shock: it enables recovery after it.

Robustness

Where resilience responds to the shock, robustness aims not to collapse when it occurs. We define Robustness  as a system’s capacity to maintain its essential functions and resist disruptions without significant degradation of its structure and processes (Hollnagel et al., 2006; Grimm et al., 2023). In the team literature, this quality translates into the maintenance of a stable level of performance under pressure, made possible by robust communicational processes (DeChurch & Mesmer-Magnus, 2015; Edmondson, 2019).

Postulate

We start from the postulate proposed by Hamant (2022, 2024), a biologist inspired by living systems, according to which robustness rests on sub-optimality. Hamant defines sub-optimality as the ability to evolve over the long term by using internal weaknesses not as problems to be circumvented, but as levers enabling adaptability. Although this proposition is as yet seldom cited in organisational sciences, it constitutes a proposal to shift the paradigm according to which a team’s value is measured primarily by its ability to continuously improve its performance.

Sources

According to Hamant, apparent internal weaknesses such as redundancy, inefficiency, heterogeneity, or randomness are not waste: they are the very mechanisms of adaptation. Translated to the scale of a team, this means that redundancy in communication ties (receiving information through multiple channels) or member heterogeneity (diversity of perspectives, roles, and competencies) constitute robustness resources, provided they are preserved rather than eliminated in the name of efficiency (Hamant, 2022; Weick & Orton, 1990). A robust team is thus one that allows itself not to function permanently at its optimal level. By maintaining margins of manoeuvre, it preserves its capacity to respond to the unexpected. A system that eliminates all apparent inefficiency in the name of performance tends to become rigid and it is precisely this rigidity that makes it fragile when a shock occurs (Hamant, 2022). Optimality should therefore only ever be temporary.