If one were to be asked which are the most demanding problems in today’s science, they would most likely be named as problems of multidisciplinarity. And it is indeed multidisciplinarity—or interdisciplinarity, or any other potential synonym—that appears in many research plans of new, trendy institutes. If this is so clear, why does it feel that multidisciplinary approaches are rather stuck? Why are current institutions not truly fostering these types of approaches?
Let’s start with the typical obstruction. Disciplines are definitively strong, and specialization prevails. It is probably common knowledge in the current scientific ecosystem that the more specialized, the better: “carve your own niche and hold it,” as one advisor—in less polite words—once told me. This is indeed good advice. Many follow it, and within their own niches they establish networks of codependence, their “discipline.” Disciplines are both good and bad, but in our context they tend toward the latter, as they exhaust options for multidisciplinarity, which they often interpret as a threat. Moreover, disciplines somehow rarify the multidisciplinary approach, reducing it to a mere sum of parts; something that, in practice, reinforces rather than solves organizational problems.
A related obstacle is funding. Many funding bodies claim to support multidisciplinary research. Is this effective? In most cases, it is not. The intrinsic problem is that the teams assembled under these schemes are still driven by specialists. The result is typically a collection of semi-independent research programs with little real integration. Yet these projects are often successful in terms of funding, which reinforces the very structures that limit multidisciplinarity.
By now, these points may sound familiar. Perhaps they are. But it is precisely here that one should attempt to define what “true” multidisciplinarity might look like, at least in the sense defended in this note.
Wasting time. Multidisciplinarity typically implies slowing down the pace of science. It requires reading, pausing, and lingering on ideas; something that is often perceived as inefficiency in a system optimized for output.
Connecting “disciplines.” The connection I refer to is not merely technical, for example, applying a tool outside its original domain. Rather, it is about identifying associations that were not previously visible. For instance, quantum computing redefines information processing by treating the bit as a quantum object, allowing superposition and fundamentally new interpretations.
Commonly unsuccessful. If multidisciplinarity is about opening new frontiers, then failure should be expected. Consider synthetic biology: designing genetic constructs that implement controlled structures such as toggle switches is notoriously difficult. Anyone who has worked with genetic manipulation knows how fragile such systems are, let alone when more complex logic is involved. Yet the dominant culture of science still privileges certainty over exploration.
Particularly adventurous. Multidisciplinarity is difficult, and not everyone is naturally inclined toward it. Many researchers are more comfortable within well-defined boundaries, and this is not inherently negative. In fact, multidisciplinarity depends on the progress of established disciplines. Still, those who move across them can be seen as particularly adventurous. The question is whether we can create environments where such profiles are not only tolerated, but developed.
Specialized influx. This may seem contradictory, but it is not. A multidisciplinary profile often risks losing depth—there are only so many hours in a day. For this reason, any multidisciplinary environment must remain attractive to specialists. Depth should not disappear; it should be accessible.
The “Johan Cruyff” scholar. Total Football is a tactical approach in which any outfield player can assume the role of another; no role is fixed, and players move fluidly across positions. This system became iconic largely through the influence of Johan Cruyff, who emphasized a holistic understanding of the game. A truly multidisciplinary scholar resembles such a player: not just versatile, but capable of shifting roles while maintaining the coherence of the whole.
All of the above suggests that multidisciplinarity is not simply a matter of intention, it requires a particular form of organization.
What I propose is what I call the toy room model. It is an organization where scientists pursue science for its own sake, deliberately stepping outside many of the constraints that define the current ecosystem—metrics, short-term impact, or strict disciplinary alignment. I call it the toy room model not to suggest triviality, but to emphasize a space for exploration, recombination, and trial.
The idea is to gather a limited number of researchers, provide them with sufficient resources and time, and allow them to explore without predefined objectives. They should be free to discuss what they want, move across domains, and access specialists when needed. At the same time, this environment should help form new generations of researchers with a more “Cruyff-like” profile: adaptable, structurally aware, and capable of navigating different roles without losing coherence.
This does not mean the absence of structure. A minimal organization is required:
Even under these conditions, failure is likely. A natural risk is diffusion: ideas that do not consolidate, discussions that remain open-ended, and a gradual loss of direction. This risk cannot be fully removed.
And yet, this is precisely the point. If multidisciplinarity is to produce genuinely new forms of knowledge, it must accept environments where outcomes are uncertain and direction is not fully predefined. The current system optimizes for reliability; something like the toy room would instead make space for possibility.
Whether such a model can exist within current institutions, or requires new ones, is an open question. But without experimenting with structures of this kind, multidisciplinarity risks remaining more a declared objective than a lived reality.