A recent seminar - Berggren's - as part of the Systems Engagements series at Linköping University on systems simulation raised a question I have been circling around in my own practice without having a good framework, or wider theoretical grounding for: at what point does the modelling and prototyping that designers do become simulation in any meaningful sense, and what - if anything - does calling it simulation add?
The seminar presented a typology of simulations - constructive (system performance assessment), virtual (individual and team decision-making), and live (real-life performance assessment) - each with different fidelity levels and different relationships to the referent system they model (Berggren, 2024; Laere et al., 2018). But the question that stayed with me was not taxonomic. It was functional: what does a simulation do for the people who use it, and what conditions determine whether it does that well or badly?
This post attempts an answer by reading Berggren's account alongside a concept from the psychoanalytic tradition that, as far as I can tell, the simulation literature has not engaged with: Winnicott's (1971) transitional object and the potential space it creates.
The intermediate area
Winnicott's contribution - developed through observation of infants but, as he was clear, applicable to cultural experience throughout life - was to identify a third area of experiencing that is neither purely internal (fantasy, wish, hallucination) nor purely external (objective reality, the material world). The transitional object - the child's blanket, the teddy bear - belongs to both domains simultaneously. It is real, but the child endows it with subjective meaning. It is not the mother, but it stands for the mother. The child knows, at some level, that it is "not real" - but the point of the transitional object is precisely that this question is not asked. As Winnicott (1971, p. 14) puts it, the transitional object exists in a space where the question of whether it was found or created is, by tacit agreement, left unresolved.
The potential space that the transitional object creates - between inner and outer, between fantasy and reality, between the subjective and the objective - is, for Winnicott, the space in which play, creativity, and cultural experience become possible. It is an intermediate area that depends on a particular relational condition: the infant needs to feel held - contained, in Bion's (1962) term - sufficiently that they can tolerate the ambiguity of a space that is neither wholly safe nor wholly dangerous. If the holding environment fails, the potential space collapses: either into pure fantasy (withdrawal from reality) or into pure compliance with external demands (the loss of creative agency).
Simulation as potential space
What struck me in reflecting on Berggren's material is that a simulation - any simulation, from a constructive system model to a live exercise - creates precisely this kind of intermediate area. A simulation is both real and not-real. It has consequences, but not full consequences. Participants act within it, make decisions, experience outcomes - but under conditions where the stakes are contained. The fire service's tabletop exercise, the military's war game, the healthcare team's scenario walkthrough: each creates a space where people can engage with aspects of their system that would be too risky, too complex, or too anxiety-provoking to experiment with in the live environment.
In my own work, I have been involved in building what might be called constructive simulations without using that language. The genetic algorithm for occupational rehabilitation that I described in Concept Modelling of Work Rehabilitation was, in effect, a model of a referent system - representations of clinical care pathways - that could be used to explore counterfactual possibilities: what if this patient had been assigned a different sequence of interventions? What if the resource constraints were different? Much of my design work around that algorithm involved mapping the referent system, identifying points of interface between model and reality, and working through - with clinicians and developers - what the model's outputs meant in practice.
What I did not have at the time was a vocabulary for the relational dimension of this work. The technical question - how to build a simulation that accurately models the referent system - is well addressed by the literature Berggren draws on. The question that is not well addressed is what happens between people when a simulation is introduced into a system, and why some simulations produce learning while others produce defensive retreat.
Service prototyping as simulation
Consider the common design research move of prototyping a service encounter - walking a group of staff and patients through a future-state care pathway using role-play, storyboards, or a staged environment. This is not a simulation in Berggren's technical sense; there is no computational model, no referent system being replicated at any formal fidelity level. But it is a simulation in a deeper sense: it creates a temporary, bounded space in which the dynamics of a service can be experienced and examined under conditions of containment. The interactions that unfold during the prototype session have real consequences for participants - the frustrations, the misunderstandings, the moments of recognition are not pretend - but the session is structured so that those consequences can be thought about rather than merely enacted. It is, in Winnicott's terms, a potential space: an intermediate area between the inner world of participants' assumptions about the service and the outer world of its operational reality, held in being by the session's boundaries, the facilitator's role, and the provisional status of the materials.
The key feature of service prototyping, from a psychodynamic perspective, is that when it works it does so not because it eliminates anxiety but because it contains it. A well-facilitated prototype session surfaces real tensions - between professional jurisdictions, between what the pathway promises and what the organisation can deliver, between how staff believe they act and how others experience them - but holds those tensions within a structure (the session design, the facilitator's authority, the time boundaries, the shared understanding that this is provisional) that makes it possible to learn from the experience rather than merely defending against it. What Lawlor and Sher (2021, p. 8) describe as providing "a temporary safe environment (Bion's notion of 'container') for the client to address and work through anxieties" applies as much to a two-hour prototype walkthrough as to a formal organisational consultation.
Containment as a design variable
This is where the Winnicottian reading of simulation becomes potentially useful for design. If a simulation functions as a transitional object - creating a potential space where participants can engage with anxiety-provoking aspects of their system without the full consequences - then the critical design variable is not the fidelity of the simulation to the referent system. It is the quality of the containment.
A simulation without adequate containment collapses the potential space. It collapses in one of two directions, mirroring Winnicott's account of what happens when the holding environment fails. If the simulation is too disconnected from the real system - too abstract, too playful, too clearly "not real" - it becomes pure fantasy: participants engage with it as a game, enjoy the exercise, and return to the real system unchanged. The workshop goes well; nothing happens afterwards. If the simulation is too close to the real system - too high-stakes, too politically exposed, too personally threatening - it becomes indistinguishable from reality, and participants respond with the same defences they use in the live environment: avoidance, splitting, denial, fight-flight. The tabletop exercise that was supposed to reveal vulnerabilities instead becomes a performance of competence; the scenario walkthrough that was supposed to surface disagreements instead reproduces the existing power dynamics.
The design task, then, is to create simulations that sit in the intermediate zone - real enough to matter, safe enough to think in. This is not primarily a technical challenge. It is a relational and structural challenge: how to design the boundaries, roles, authority relations, and holding conditions of the simulation so that participants can tolerate the anxiety it surfaces.
What game-forming reveals
The seminar material on game-forming - "gaming-simulation is defined as a specific form of simulation" that "aims at designing a model of a system in a complex problem area in order to be able to experiment with the model" (Laere et al., 2018) - prompted a further reflection. Building a game simulation requires making explicit the rules, states, and dynamics of the system being modelled. In design terms, this is the work of concept modelling: identifying entities, relationships, states, and transitions. But what became clear to me in trying to think through what it would take to build a game simulation in the contexts I am working in is that the hardest part is not the technical modelling. It is deciding what to include and what to leave out - which is to say, deciding what aspects of the system can be safely surfaced and those that would overwhelm the participants.
This is a psychodynamic question, not a technical one. The referent system includes not only its formal structures and processes but also its defences, its anxieties, and its unconscious arrangements. A simulation that models only the formal system - the processes, the decision points, the resource flows - will be technically accurate but psychodynamically naive. A simulation that inadvertently models the defended territory - that surfaces the conflicts, the power relations, the unspoken agreements - without providing containment for what it exposes will produce defensive retreat rather than learning.
Prototypes and the question not asked
This connects to a broader question about design prototypes, which are - in functional terms - simulations of future states. A prototype is both real and not-real: it is an artefact that stands for something that does not yet exist. The user who interacts with a prototype is engaging in Winnicott's intermediate area: the prototype is not the final service, but both parties agree, tacitly, not to press the question too hard. The prototype works precisely because it occupies the space between what is and what might be, and because that space is held open by the design relationship.
When prototyping works well, it works as a transitional object: participants can project their hopes, fears, and needs onto the prototype, can experiment with what a different future might feel like, and can do so under conditions that are contained - the designer is present, the session has boundaries, the understanding is that this is provisional. When prototyping fails - when the prototype is received with indifference, or hostility, or is immediately absorbed into the existing organisational Image without producing change - one possibility, as I argued in the previous post, is that the prototype has entered a defended space. But another possibility is that the containment has failed: the prototype was too fantastical to take seriously, or too threatening to engage with honestly.
The Winnicottian addition to design's account of prototyping is the insistence that the intermediate space is not a given. It has to be created and maintained through relationships, boundaries, and what Lawlor and Sher (2021, p. 17) call "dependability" - the provision, both physical and psychological, of a holding environment in which the organisation-in-the-mind can be "exposed as fully and as richly as possible". The designer who prototypes without attending to these conditions is offering a transitional object to a system that may not have the relational capacity to use it.
References
Berggren, P. (2024). Systems Simulation. Lecture notes, Systems Engagements module, Linköping University.
Bion, W.R. (1962). Learning from Experience. London: Heinemann.
Laere, J.V., Berggren, P., Ibrahim, O., Larsson, A. & Kallin, S. (2018). A Simulation-Game to Explore Collective Critical Infrastructure Resilience. In Safety and Reliability - Safe Societies In A Changing World (pp. 1305–1312). CRC Press.
Lawlor, D. & Sher, M. (2021). An Introduction to Systems Psychodynamics. London: Routledge.
Winnicott, D.W. (1971). Playing and Reality. London: Tavistock.