We know that computers compute. But can natural systems – like a cell, or even currents in a fluid – also “compute”? A study by the Complexity Science Hub (CSH) and the Santa Fe Institute (SFI) presents an approach for understanding natural systems as computers – and how this could potentially be used to solve real-world problems.
THE STUDY IN A NUTSHELL
- Nature computes too: It’s not just laptops and smartphones that compute – natural systems like cells, brains, and chemical reactions can also be viewed as computers.
- The problem: Until now, there was no formal framework to describe what exactly such systems “compute” and how that works.
- The approach: The study presents a new way to identify and examine computations in any dynamic system.
- The core idea: A system is considered “computing” if its dynamics can be mathematically described like a computer – regardless of whether it computes consciously.
- Why it matters: The approach could help harness natural processes to solve problems and deepen our understanding of the role that information and computation play in nature.
When we think of computers, we think of laptops or smartphones. Machines built by humans. We know that these computers perform calculations. We know that they are dynamic systems whose components interact to carry out processes, and that this requires computational power. We know that there is an underlying logic to all of this, that there is an input and an output, and that this comes with energy costs.
“Other computers are much harder to recognize,” says study author Jan Korbel from CSH. Can cells, brains, chemical reactions, or even currents in fluids also compute? And if so, what exactly do they compute? And how can we verify that? “Scientists argue that they can indeed be viewed as computers – that they can, in a sense, ‘compute,'” says the researcher. However, these “non-constructed” computers are still poorly understood. It is therefore not always clear what exactly they compute, how they do it, or how this can be used to solve problems.
CRITERIA FOR WHEN A SYSTEM COMPUTES
“The issue is how to define, formally, a set of criteria for identifying what computation(s) a given, arbitrary dynamical system does, in order to give us insights into these computational systems found in nature,” says study author David Wolpert from SFI. Such criteria would be a powerful tool for understanding what it fundamentally means to “compute.” They could reveal connections between so-called “constructed” computers – such as phones and laptops – and “non-constructed” computers, meaning natural systems capable of performing computations.
In the study, recently published in the Journal of Physics: Complexity, Wolpert and Korbel lay out one such possible path – a new framework for identifying and examining computations encoded in dynamic systems.
A NEW FRAMEWORK
“Our framework shows how computations performed by a non-constructed system can be mapped onto, or connected to, those of a constructed computer,” says Korbel.
The core idea of the study is this: a system can be considered “computing” if its behavior can be mathematically described like a computer. This does not mean the system is consciously computing – rather, that its dynamics can be represented in a way that corresponds to a computation.
AN EXAMPLE
A concrete example is chemical reactions, whose sequence can be viewed as a kind of non-constructed computer, the researchers say. The initial concentration of chemicals can be thought of as the input. The chemical reactions that take place are the processing, and the final concentrations of the chemicals make up the output. Looking at a reaction this way, says Wolpert, reveals that chemical reactions “encode” a broad set of computational tasks that align neatly with computing operations that are already well known.
WHY THIS MATTERS
A mapping that connects natural systems to their constructed computer counterparts will allow researchers, according to Korbel, to study the computational behavior of any dynamic system.
It also provides a way to concretely define computational processes. “We can say that some system can compute and what a particular instance of their dynamics does in fact compute, because now there’s an explicit mapping between an abstract computational device and a real dynamic system,” says Korbel.
In the long run, this approach could help us better understand how natural systems process information – from the brain to complex physical processes.
A formal description of computation could open the door to directly compare natural and constructed computers and identify their shared principles. This allows not only to more precisely determine what a system actually computes, but also how natural processes might be deliberately harnessed to solve problems. At the same time, this approach may help to sharpen the very concept of computing – and with it, our understanding of the role that information and computation play in nature.
FROM A WORKSHOP TO AN IDEA
Wolpert and Korbel have been collaborating for years on projects at the intersection of computing and complexity. In 2022, they co-organized a workshop on the thermodynamics of computing systems. In the summer of 2025, both were part of a working group on stochastic thermodynamics. And in October 2023, both participated in a meeting on the nature of computation hosted at CSH. Researchers from a range of fields came together to exchange findings on how various systems can be understood as computing. These fields included fluid dynamics, neuroscience, and cellular automata – mathematical models that evolve according to simple rules. “They were very different things,” says Korbel. During the meeting, researchers looked for definitions and rules that could describe common features of computation.
WHERE THE COMPLEXITY LIES
This work shows that systems evolving over time can be understood in terms of computation. This makes it possible to measure the computational complexity of a dynamical a system by asking what kind of computations it can perform. At the same time, the same computation can be implemented in many different physical ways, with very different energy or engineering costs—making complexity science essential to understand these differences.
This article is based on a news story by the SFI.
About the study
The study “What does it mean for a system to compute?” by D. H. Wolpert and J. Korbel was published in the Journal of Physics: Complexity (doi: 10.1088/2632-072X/ae3af8).
The article is part of a special, forthcoming issue of the Journal of Physics: Complexity featuring the work of Wolpert and Korbel, dedicated entirely to the question: “What does it mean for a system to compute?” The other articles in this special issue also draw on work presented at the researchers’ fall meeting at CSH, where participants focused on decoding computation in its many forms.
