On participatory realism

Our views of the relation between science and the natural world tend to gravitate between two defining positions : either science offers us a true (although perhaps limited) representation of physical reality, or our representations of the world (although perhaps useful) simply lack the properties that would make them real. It is sensible to concede to the realist view that the power of scientific knowledge is likely not the outcome of a miracle, but of the intrinsic accuracy of the maps of the world it has developed. But any knowledge must be conceived by a human mind to begin with, and its hard to explain how our evolutionary and social history would have equipped us with just the kind of cognitive machinery that can conceive the ground nature of physical reality. Moreover, has a century of development in quantum physics not done away with the idea of an objective reality ?

The present article will articulate a third position that integrates the strengths of both the realist and anti-realist position, while giving us some grip on why we are here discussing the issue : participatory realism. Participatory realism is a brainchild of both recent cognitive science and quantum physics, and can be understood as the affirmation that while physical reality exist independently from us, we participate in constructing it through our attempts to understand and manipulate it. Think of the way humans have progressively constructed a niche that fits their bodily needs, while extending their cognitive abilities in the process through emotional care, economic specialization, urbanization, counting, writing, and ultimately the recent turn toward digital technologies. In the same manner, physical observers such as ourselves create a world where their constitutive assumptions remain true, while also expanding the space of accessible states – i.e., constructing physical reality.

We will first discuss the intricacies of realism, its variants and its negations, and establish how the dichotomy between the positions of “realism” and “anti-realism” is ultimately failed. Then, we will explain the meaning of participatory realism in cognitive science, and more particularly how it emerges from the necessity for cognitive agents to understand their world so as to maintain their structural identity. In particular, we will investigate how this prefigures the role of human sociality and material culture in constructing the ways we have of understanding the world. Finally, we will turn to why any fundamental theory of physics must account for active role of the observer in shaping physical reality, and why our account of the construction of human knowledge offers an appropriate model of this process. Participatory realism will then emerge as an onto-epistemological theory offering an appropriate grounding for both our understanding of the world and our reflexive understanding of our own knowledge.

Scientific realism and its problems

Realism, in and of itself, is a very broad statement that we represent things as they actually are. For example, we can call social realism an art movement that aims to show the day-to-day concerns of the working class, and moral realism the idea that moral values constitute an objective truth. The brand of realism we aim to discuss here can be called epistemic realism or scientific realism, and it claims very plainly that our knowledge of the world (as developed through the means of science) translate natural facts preexisting our observations. At first sight, it is hard to see why one would contest this view. If we deny that the external world is simply an hallucination produced by one’s own mind (which we should), then we must surely agree that objects have some kind of underlying reality that is accessible to investigation, and therefore that science can tell us about what they actually are.

While this argument is reasonable, it would be misleading to present it as a validation of realism. No one contests that there is indeed a physical reality and that science is studying it. What is at stake in the debates surrounding scientific realism is the relation between our models of reality and the natural world. We can agree that there is indeed a physical reality outside of ourselves, but deny that science has the capacity to decide on the “true nature” of the objects which compose it. A strong argument for this position is the underdetermination of theory by evidence (Lakatos 1976): any set of observations can in principle be explained by an array of different theories, and we can have no assurance that all involve the same kind of objects. Our best shot at truth is therefore to do away with theories that directly conflict with evidence, and be very prudent about any claims regarding the reality of the remaining ones.

The argument from underdetermination, as well as an array of philosophical and sociological concerns about the neutrality of researchers, have driven some researchers to articulate an explicitly anti-realist position about scientific knowledge. For example, Colombo, Hartmann, and van Iersel (2015) hold that explanation in biology simply does not need to commit to the reality of the objects it postulates. To clarify, a model that describes a biological structure (say, the electrochemistry of neuronal activation) as a mechanism works regardless of whether this structure is a mechanism. A mechanical model is explanatory because it tells us how a given biological system can reliably produce a phenomenon of interest, while being thoroughly coherent with everything we know about the system and its activity. But it does not commit us to actually believe that everything in the model is true, from the nature it lends to system component to the description it makes of their relations.

The idea that models can actually describe real structures while not being “real” in themselves has motivated the development of a bridging position known as structural realism (Ladyman 2020). According to this position, scientific theories and models do not bear onto natural objects in themselves, but onto the relations they have between them. For example, modeling a mechanical body as a point of mass m does not mean that we actually it is point-like, and equipped with an intrinsic mass property of value m. It does mean however that for the purpose of mechanical interaction, it will behave like it was a point equipped with an intrinsic mass m. This position relaxes the requirements for a scientific theory to be valid, as it does need to correspond to an underlying reality but simply to translate known pathways of causality. However, it also dissolves the distinction between realism and anti-realism, as it negates the idea that we can or should know what objects are in themselves.

Participatory realism and cognitive agency

For the next step of our argument, we will turn away from the question of what we can in principle know about nature, and discuss how we concretely get to know the world. Since we are natural beings, our knowledge of the world reduces to what we can infer from our perception, given the constraints of our cognitive architecture. And this architecture is not a passive receiver of information, like antennas or photographs. Cognitive agents learn about the world by “asking questions”, i.e. by actively looking for information of interest (Fields and Levin 2020). A very simple example is the development of chemical receptors in bacteria and eukaryotes, which enable them perceive and follow sources of nutriments. A more elaborate example is the way we humans flexibly allocate attention to precise their model of the world: what is the next word in this sentence? is this person looking at me? does the sky look like it’s going to rain?

Let us take the perspective of one of these cognitive agents, more precisely this of a hunter in a forest. The hunter knows a pack of deer is currently roaming this territory, and is looking for signs of their specific position. They notice fresh deer footprints, and immediately start following them. What they actually see is nothing but a simple geometrical pattern imprinted in the ground. It is not detailed instructions on how to go to a given place, or a line on a map. In other words, it only constitutes a path toward a deer for someone who knows what their footsteps look like and is currently looking for one. But the hunter would likely laugh to the suggestion that the deer track is not physical reality, but only a projection of their mind. For them, it is clear as day: they see a deer track just before their eyes, and they know for a fact it will lead them to their prey.

The deer tracks of the above example do not exist in either the hunter’s mind or the external world, they emerge as a percept at the interface of the two. To the hunter, everything appears as if the deer track was an actual physical object, and they just happened to notice it due to its observable properties (i.e. the presence of footprints). Therefore, they simply perceive the track as real, and use the knowledge of its presence to navigate the world. The same goes for any other physical object. We cannot glimpse at the intrinsic nature of the rain, trees, or dogs, but we can infer what kind of entities underlie our perceptions and then simply treat them as real (Froese 2022). And given the physical nature of cognition, there is no other way we can experience the world. However sophisticated, any cognitive agent experiences a reality of their own making, constructed on the basis of their anticipated percepts (whether explicit or not) (Ramstead et al. 2021; Clark 2019).

It is critical to understand is that this does not make the world we experience any less real. Whether the deer track preexists to the hunter’s perception does not matter to judging of its reality: for all practical purposes, everything appears as if there indeed was a deer track there, and the hunter just happened to find it. The role of subjective experience in shaping physical reality is most flagrant when agents modify their material environment to fit their expectations, as is prevalent in the human species. However, immaterial norms and knowledge (e.g. on tracking deer) also shape the way we act in the world, and therefore participate to create the kind of trace we imprint onto the world. Therefore, the perceptions we have of the world (up to scientific knowledge) are not either a neutral representation of what’s there, or a simple illusion projected onto physical reality. They fully participate to bring about physical reality by shaping the causal relations between the perceiving agent and the world they navigate.

Participatory realism and the construction of physical reality

Although it is most intuitive at the scale of biological agents, the notion that perception shapes physical reality goes much deeper. Let us say an agent A acquires information about a system of interest S. By definition, that means that the system’s entropy decreases. As a closed system’s entropy cannot decrease, this entropy has to be offloaded somewhere. As offloading it into either the agent’s or the system’s would instantly void the observation, they must somehow dissipate this entropy as heat in their environment E. But then, unless they can somehow cool down E by the same amount by forgetting about their observation (which no credible physical architecture would allow), it means that they have irreversibly lost information about E. In other words, what observations any given agent can make is constrained by the history of observations it has made in the past, as well as by the computational architecture underlying the observation process.

Our best efforts seem indicate that this path-dependency is not simply an emergent statistical law, but rather a fundamental property of physical observation. At a quantum level, physical interaction indeed leads to the phenomenon of entanglement where the interacting systems can only take certain value relative to each other. For example, we could constrain two photons to take opposite spin value without deciding what this spin value is. Both photons would remain in a superposition of both possible states, but any observation of one would instantly (i.e., faster than light speed) constrain any future observation of the other’s state. While famously denounced by Einstein as “spooky action at a distance”, this property was nonetheless demonstrated experimentally in the so-called Aspect experiment Aspect, Grangier, and Roger (1982). Even more disturbingly, inserting or removing a detector in similar experimental setting have been shown to retrospectively determine what path a photon has taken before the observation even occurred (Jacques et al. 2007).

However, these paradoxical violations of the most basic physical constraints imaginable only hold if we abstract the observer from the domain of physical interactions. We can instead attribute the loss of quantum coherence in these experiment to entanglement between the observer and the target system during the process of measurement. Then, there is no need for action at a distance or retrospective causality: as an observer, we are simply constrained to observe states of the world that are consistent with what we observed already (Griffiths 2019; Bacciagaluppi 2020). Entropy then reemerges as the information a given observer can’t access due to his history of past interactions (Tegmark 2012). Quantum states, in themselves, reduce to the knowledge an observer has about the world - although a kind of knowledge that physically constrains the reality we can experience. Because it recognizes an active role of the observer in shaping physical reality, this position was also characterized as “participatory realism” by quantum information theorists (Fuchs 2017).

The critical aspect of the present discussion is that the cognitive science and quantum information brands of participatory realism share more than their names. In the same way that cognitive agents need to act on their environment to “ask questions”, quantum observers first need to prepare the system of interest to measure them along specific reference frames (Fields et al. 2021). Most importantly, both processes seem to reduce to the same mathematical view of inference as synchronization-driven construction of the “statespace” which constitutes the natural representation of physical determinants of the process at hand (Guénin–Carlut 2022). In this view, the space of causally effective physical states is constrained by the process of observation itself, based on the imprinting of the agent / observer’s internal dynamics onto the external world. In other words, observation constitutes an actual driver of the construction of physical reality in the strongest possible sense, rather than a passive process of information gathering.


We have hereby introduced the position of “participatory realism” as a synthesis of arguments coming from both cognitive science and quantum information theory. In cognitive science, the argument concerns the way our knowledge of the world is constrained by the processes through which we come to understand it. In quantum information theory, the argument concerns the way physical reality itself is constrained by the physical process of observation. At first sight, these are entirely different statements: one is about what is actually in the world, and the other is one what we can know about it. However, the mathematical unity between those two statements suggest a deeper relation between the processes that construct the world and those by which we understand it. As per the present position, the process by agent actively constrain the world to be able to gather information interest is precisely what drives the unfolding of physical reality.

This does not mean, of course, that we live in an imagined world where anything we wish can become reality. Humans do not have an exclusive, nor even a privileged, claim to the status of physical observers. It is not likely that whatever processes drive the emergence of spacetime or sexual reproduction will simply go away if we ask them firmly, or that we can even imagine a reality where it would be the case. It is still the case however that our minds and bodies have the capability to project their structure onto observable physical reality, as they do in classical quantum experiment or in our attempts to build a niche that fits our biological needs and extend our cognitive abilities. It is still the case, if this view is in essence correct, that the question of what there is in the world cannot be separated entirely from the question of how we come to know about it.

This admittedly implicates a huge shift in our understanding of both physical reality and the status of scientific knowledge. However, there is no way in which it contradicts the reality of the world or the validity of scientific knowledge. Participatory realism recognizes there is a world out there, it simply insists that we participate in building it. Because the process of inference it describes is still constrained by the structure and states of the target system, it is still able to account for scientific knowledge as valid representations of causal structure in the natural world. The only reason why it seems so groundbreaking is because it clashes with the intuition that our knowledge is an objective, true representation of the world as it is, detached from the interactions we have with it. But as we have seen here, we have no reason to believe such a representation could even exist, let alone be cognitively accessible to a specie of ultrasocial apes.

Acknowledgement and references

The present work is a companion article to Guénin–Carlut, Avel. 2022. “Physics of Creation - Symmetry Breaking, (En)Active Inference, and Unfolding Statespaces.” OSF Preprints. https://doi.org/10.31219/osf.io/68947. In contrast to the highly formal and radical cosmological position I articulated, the present work is meant as an accessible explainer on the implication of “participatory realism” for our understanding of physical reality and our own knowledge, which is grounded in the assessment of uncontroversial positions in quantum physics, cognitive science, and epistemology.

Thanks to Iona Brenac, Alexis Rozanski and Maxwell Ramstead, whose editorial work helped me make it more accessible to a broader audience than my formal treatment.

I warmly recommend readers who’d like to understand more about the physics at hand to take a look at the PBS Spacetime series, especially their episode on participatory realism (PBS Space Time 2022).


This article was updated on October 29, 2022