🌀 nonlinear sciences

Towards chemistries in dynamical systems

This exploratory paper proposes a framework for describing any dynamical system in chemical terms by defining spatial locations, species, and transitions, while introducing a criterion for unique minimal reaction pathways and applying the concept to gliders in Conway's Game of Life.

Original authors: Martin Biehl, Nathaniel Virgo

Published 2026-07-22

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Martin Biehl, Nathaniel Virgo

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Universe as a Chemical Kitchen

Imagine you are looking at a bustling city from a high-rise window. From this height, you don't see individual people walking or cars driving; you just see a blur of motion. But if you zoom in, you realize the city is actually made of distinct things: people, cars, and buildings, all interacting in specific ways. In science, we often study "dynamical systems," which are just fancy names for any system that changes over time according to a set of rules. Think of the weather, the stock market, or even a video game. Usually, scientists describe these systems by looking at their raw data—like a giant spreadsheet of numbers. But sometimes, that's like trying to understand a symphony by only reading the sheet music without hearing the instruments. We want to hear the melody.

This is where the idea of "chemistry" comes in. In real chemistry, we don't just watch atoms bounce around randomly; we group them into "molecules" (like water or sugar) and watch how they react to form new things. Scientists have long wondered if we could do the same thing with complex systems that aren't actually made of atoms. Could we look at a computer program or a game and say, "Oh, that's a 'glider' molecule moving across the screen"? This is the heart of a field called artificial life, where researchers try to find the "particles" and "organisms" hiding inside simple rules. The big question is: How do we decide what counts as a distinct object, and how do we describe its life story without getting confused?

The Paper's Big Idea: Finding the "Molecules" in the Machine

In this exploratory work, Martin Biehl and Nathaniel Virgo propose a new way to look at any changing system, not just chemical ones. They want to translate the cold, hard math of a system's updates into a warm, familiar story about "tokens" (like little tokens in a board game) and "reactions" (like mixing ingredients in a kitchen). They call this a Pattern Chemistry.

Imagine you are watching a game of Life on a computer screen. The game updates every second based on simple rules: a cell lives, dies, or is born depending on its neighbors. To a computer, this is just a grid of 1s and 0s flipping back and forth. But to a human, we see a little spaceship-like shape, called a "glider," zipping across the screen. Biehl and Virgo want to build a formal recipe to prove that this "glider" is a real thing, just like a molecule of water.

To do this, they say you have to make three big decisions, like setting up a new board game:

Where are we looking? (They call these "frames of reference." It's like deciding if you are looking at the grid from the top-left corner or the bottom-right.)
What are we counting? (They call these "pattern shapes." In the game, this might be a "rocket" shape or a "wedge" shape.)
What are the rules of change? (They use something called a "Petri-net," which is just a fancy flowchart that says: "If you have a rocket here, it can turn into a wedge there," or "If you have two wedges, they can disappear.")

The authors argue that for this chemistry to make sense, there must be a unique, unambiguous way to explain what happened between one second and the next. If you see a glider move, you shouldn't have to guess: "Did it move? Or did it die and a new one just happen to appear in the next spot?" If the rules allow for both explanations with the same amount of effort, the chemistry is "ambiguous" and fails. But if the rules force a single, simplest explanation (e.g., "It definitely moved"), then you have found a true "unambiguous pattern chemistry."

The Glider Test: Moving Without Confusion

To test their idea, the authors looked at the famous "glider" in the game of Life. They built a specific chemistry where the "molecules" are gliders, and the "reactions" are movements. They found that for a glider, there is always one clear, minimal way to describe its journey. If a glider is at spot A at time 1, and at spot B at time 2, the simplest story is that it moved from A to B. You don't need to invent a story where it vanished and a new one appeared out of nowhere, because that would require more "steps" (more reactions) than just moving it.

This is a big deal because it suggests that the glider isn't just a trick of the eye; it behaves like a real object that persists and moves. The authors suggest that this "unambiguous movement" is what convinces us that the glider is a "thing."

However, they also point out that this doesn't work for everything. If you try to apply this same logic to individual "alive" cells in the game, it gets messy. A single cell might be part of a larger pattern that moves, but the cell itself doesn't have a clear, unique path. In fact, the authors show that if you try to force a "movement" rule on single cells, you end up with confusing situations where you can't tell if a cell moved or if a new one was born. This suggests that while the glider is a true "molecule" in this chemical sense, a single cell is not.

What They Didn't Solve (And What They Didn't Say)

It's important to note that this paper is a proposal, not a final answer. The authors are careful to say they are only looking at systems with a finite number of states (like a game board with a fixed size), not infinite ones. They admit that choosing the right "frames of reference" (where to look) is still a bit of a puzzle. If you choose the wrong angles, you might count the same glider multiple times or miss it entirely. They leave this as an open problem for future researchers to solve.

They also don't claim to have found a magic formula that works for every single pattern in every possible game. They specifically show that their method works beautifully for the glider, but they warn that patterns that reproduce (make copies of themselves) or merge might be much harder to describe without ambiguity.

The Takeaway

In the end, Biehl and Virgo have handed us a new pair of glasses. Instead of seeing a chaotic mess of changing numbers, we can now try to see a structured world of interacting objects. They've shown us that for some things, like the glider in the game of Life, we can define a clear, unambiguous chemical language that describes their life, death, and movement. It's a step toward understanding how complex, living-like behaviors can emerge from simple, dead rules. While they haven't solved every mystery of artificial life, they've given us a solid, playful way to start asking the right questions about what it means for a pattern to be "real."

Technical Summary: Towards Chemistries in Dynamical Systems

Problem Statement
The paper addresses the lack of a widely accepted formal framework for describing particle-, molecule-, or organism-like structures within dynamical systems, particularly cellular automata (CA). While researchers often identify coherent patterns (e.g., gliders in the Game of Life) and treat them as distinct entities, there is no standard method to formally describe these aspects using the language of chemistry (reactions, species, and transitions). Existing approaches, such as spatiotemporal filtering or statistical mechanics operators, either lack a formal link to chemical reaction networks or do not investigate the conditions under which such patterns are "worth considering" based on the underlying system dynamics. The authors aim to bridge this gap by deriving a "chemistry-like" description from a raw dynamical system defined solely by a state set and an update function, without assuming a pre-existing grid or token structure.

Methodology
The authors propose a formal framework called Pattern Chemistry, which functions as a specific type of coarse-graining. The methodology involves mapping a dynamical system $(X, h)$ to a Petri-net (or Chemical Reaction Network) through three specific design choices:

Frames of Reference and Pattern Shapes ( $R$ and $S$ ): An $R$ -indexed family of functions $\{f_r: X \to M_S\}$ is defined. These functions map the global state $x$ to multisets of "pattern shapes" ( $S$ ) detected within specific "frames of reference" ( $R$ ). This allows the observer to identify the presence and quantity of specific patterns (e.g., gliders) at various locations and orientations.
Token Map ( $\phi$ ): The family of functions is aggregated into a single token map $\phi: X \to M_{R \times S}$ , which represents the current "marking" (state) of the chemical system.
Petri-net and Transition Map ( $T$ and $\theta$ ): A Petri-net $(T, R \times S, src, tgt)$ is constructed to define possible reactions (transitions) between pattern shapes. A transition map $\theta: X \to M_T$ is then defined to select a specific "step" (a multiset of transitions) that transforms the marking $\phi(x)$ into the marking $\phi(h(x))$ of the next time step.

Key Definitions and Contributions
The paper introduces two primary formal definitions:

Pattern Chemistry: A tuple consisting of the frame family, the Petri-net, and the transition map that satisfies a compatibility condition. This condition ensures that the source multiset of the selected step matches the current token map, and the target multiset matches the token map of the updated state.
Unambiguous Pattern Chemistry: A stricter definition requiring that for any state transition, there exists a unique step that minimizes the total number of transitions used. This criterion is proposed to ensure that the chemical description is not arbitrary; it implies that the system's dynamics dictate a single, most efficient way to interpret the evolution of patterns as reactions.

Results and Case Studies
The authors apply this framework to the Game of Life cellular automaton:

Glider Chemistry: By defining frames of reference based on the symmetry group of the grid and shapes as "rocket" and "wedge" configurations, they construct a pattern chemistry for gliders. They demonstrate that by including "movement" transitions (where a glider at position $g$ moves to $g'$ within its light cone) alongside creation and annihilation transitions, the system satisfies the unambiguous criterion. The minimization of transitions forces the interpretation of glider motion as a single movement reaction rather than a pair of annihilation/creation reactions, and the specific geometry of the Game of Life ensures no ambiguity arises regarding which glider moved where.
Cell Chemistry: When applying the framework to individual "alive" cells (treating them as particles), the authors find that defining movement transitions based on light cones introduces ambiguity. Because multiple alive cells can influence the same future cell, it is impossible to uniquely determine which "parent" cell moved to a new location while minimizing transitions. Consequently, the only unambiguous interpretation for individual cells is "persistence" (staying in place) rather than movement.
Global Counting: The paper also explores a single-frame-of-reference approach that counts total glider types. While a transition map can be constructed, it fails the unambiguity test because different combinations of movement and creation/annihilation transitions can yield the same net change with the same number of steps.

Significance and Claims
The paper claims to provide a formal mechanism to partially describe dynamical systems using chemical metaphors, moving beyond simple pattern detection to a structured description of "reactions."

Unambiguity as a Criterion for Reality: The authors argue that the unambiguous nature of the glider's movement transitions in the Game of Life is a strong indicator that gliders are genuine "moving objects." They suggest that this lack of ambiguity in choosing the minimal set of transitions is what convinces an observer that a pattern is a distinct entity undergoing specific dynamics (like movement) rather than just a transient configuration.
Limitations and Open Problems: The authors remain modest regarding the scope of their work. They acknowledge that the framework currently applies only to finite dynamical systems (due to the use of finite multisets) and that infinite systems (like standard infinite grids) would require patching finite systems together. Furthermore, they identify the selection of the $R$ -indexed family of functions as an open problem; specifically, how to choose frames of reference to avoid degeneracy (e.g., counting the same object multiple times due to overlapping frames) without relying on the symmetries of the system itself.
Relation to Existing Work: The work distinguishes itself from statistical mechanics approaches (e.g., Beer, 2020b) by not switching to a statistical description but rather by analyzing how the underlying deterministic system determines the specific reactions. It also differs from pattern detection methods by explicitly linking the detected patterns to a Petri-net structure.

In conclusion, the paper proposes that a "good" chemical description of a dynamical system is one where the dynamics allow for a unique, minimal set of reactions to explain state changes, thereby validating the existence of the patterns as distinct, interacting entities.

The Universe as a Chemical Kitchen

The Paper's Big Idea: Finding the "Molecules" in the Machine

The Glider Test: Moving Without Confusion

What They Didn't Solve (And What They Didn't Say)

The Takeaway

More like this