The current picture of matter that physics offers us is unable to explain how life evolved from inorganic materials. The problem seems to be with physics defining matter simply in terms of its building blocks. But a new theory suggests that the assembly history of material objects, how they came to have the complexity they do, is an essential property of all matter. Lee Cronin argues that this assembly theory of matter shows how evolution by modification and selection isn’t simply the preserve of life but describes how all matter develops.
For centuries, scientists have struggled to explain how life could have emerged and evolved from the inanimate matter described by physics and chemistry. On the surface, living organisms seem fundamentally different from non-living things. While physics perfectly describes the behaviour of fundamental particles like electrons and protons, it falls short when applied to complex phenomena like biological evolution. Curiously, many experts do not see this as a fundamental problem, as they expect that the solution to this problem will someday be found in fundamental physics. Thus far this has not been possible and there seems to be no prospect that a satisfactory bottom-up explanation, from elementary particles to life, will be found. Something is fundamentally wrong with our current view of reality.
Darwin’s theory of evolution elegantly explains why some organisms are better adapted than others. But it doesn’t address where the capacity for open-ended evolution came from in the first place. If we go far enough back in time, the first primitive living systems must have arisen from simpler prebiotic chemistry. But how could animate complexity appear without evolution already operating? To bridge this explanatory gap between physics and biology, scientists have explored how combinatorial space might accommodate the transition from inanimate chemistry to biology. Since biology is built on a subset of chemistry, the combinatorial space contains all possible molecules, which is virtually infinite. One way to represent these possibilities is to consider the "adjacent possible" which represents the space of possibilities that are contiguous with existing reality. But so far, no idea has fully integrated the emergence of novelty and selection into the framework of physics. This means we still lack a theory that can take particles following the known laws of physics and demonstrate how evolution can appear as an emergent phenomenon.
Instead of treating fundamental particles like atoms and molecules as the irreducible building blocks, Assembly Theory defines all objects by their capacity to be assembled or broken down using minimal paths.
To address these issues, Sara Walker and I developed a new framework called Assembly Theory (AT) that provides a unified approach for understanding how both physics and selection-driven evolution fit together. Assembly theory is not just a new theoretical framework, it’s rooted in experiment and was initially built to address the question ‘how can a complex molecule form spontaneously?’ It turns out, answering that question is the key to understanding how complexity of any kind emerges in our universe, both at the inorganic and organic level. As it turns out, life and technology share more in common than we had previously imagined.
Rethinking What Makes an Object
Assembly Theory is based on a deceptively simple shift in perspective. Instead of treating fundamental particles like atoms and molecules as the irreducible building blocks, Assembly Theory defines all objects by their capacity to be assembled or broken down using minimal paths. This means an object’s properties are determined not only by its components, but also its history of formation. For example, a protein molecule can be described in terms of its amino acid sequence – its component parts. But from the perspective of AT, the protein is more than just a list of its parts. It also carries evidence of the evolutionary history that selected that specific sequence over the astronomically large set of possible alternatives. By considering the shortest path to make the object from its building blocks, Assembly Theory can be used to see if the object was produced by a process of selection and evolution rather than random chance. This is because the larger the number of steps needed to make the object on the shortest path (i.e. the simplest way possible), the lower the probability that the object could have formed by chance. This means the shortest path helps us identify if a memory, like a gene, was required to construct the object.
By embracing objects as products of their history, not just their component parts, AT provides space for novelty generation and selection to operate within the bounds of physics. The theory quantifies an object’s complexity based on how many assembly steps were required to build it from simpler components. This “assembly index” measures meaningful functional information rather than just randomness; functional information is the complexity evolved by living systems to aid their survival. AT also tracks how many identical copies of a given object are present. Unique objects with high assembly indices are exponentially unlikely to appear by chance. So abundant complex objects must be the product of selection acting on variability, rather than mere randomness. Quantifying assembly index and copy number together distinguishes functional biotic phenomena from abiogenesis.
By quantifying information in terms of physical assembly operations, Assembly Theory represents selection as a physical phenomenon rather than an extraneous evolutionary ingredient.
Imagining “Assembly Spaces”
To explain how AT incorporates selection into physics, we propose the concept of “assembly spaces”. These conceptual spaces contain all the possible pathways where objects could be assembled step-by-step from their constituents. The vastness of these assembly spaces captures the combinatorial explosion of options generated by each additional assembly operation which increase exponentially as a function of the assembly index (this is the number of steps required to make the object on the shortest path). While in theory every possible structure exists within the space, only a tiny fraction can be produced with finite resources and time. Physical and historical constraints determine which trajectories are followed through these assembly spaces. Along certain paths, the emergence of an innovation or replication mechanism can profoundly reshape subsequent exploration. When viewed in reverse, the objects encountered along any path reflect evidence of the selection pressures that shaped their evolution. By quantifying information in terms of physical assembly operations, Assembly Theory represents selection as a physical phenomenon rather than an extraneous evolutionary ingredient. Assembly indices distinguish functional complexity from randomness without presupposing units of selection or reproduction.
Reconciling Physics With Open-Ended Evolution
To demonstrate the power of AT, we modelled simple polymer systems (these are chains of molecules linked together) where chains assemble step-wise from individual molecules. When assembly proceeds randomly, each new step multiplies the number of possible outcomes exponentially. This quickly exhausts all resources without producing many high-index copies. But when the models incorporate a basic self-replicating polymer that preferentially selects certain reactions, the systems generate both high diversity and highly replicated polymers. By quantifying the resulting “assembly” of these ensembles, AT confirms that selection is required to produce this end state.
Unlike traditional simulations of evolution, these abstract polymer models implement selection as a generalized physical phenomenon, rather than something that only applies to organic matter. Their behaviour mirrors real chemistry where self-replicating polymers induce specificity in reaction networks which controls which molecules react together like a many keys trying to find a specific lock. AT offers a language to detect and measure selection even in unfamiliar systems by simply observing what structures are produced. The models also illustrate how AT integrates open-ended growth of complexity within physical limits. When new innovations expand the diversity of possible components and assembly operations, the encompassing space swells exponentially but by including knowledge of the past limits the expansion being so vast it is impossible to explore. In other words, viable trajectories through assembly space are strongly shaped by historical contingencies. So, expansions proceed unevenly along selected paths, rather than uniformly filling all possibilities.
Together, these features of AT provide the components necessary to translate a system governed strictly by physics into an evolvable open-ended process. By reformulating objects as assemblies, the emergence of life-like behaviour in physical systems becomes not only possible, but quantifiable and predictable. We think this approach may finally provide the sought-after bridge between physics and biology.
With its radically expanded notion of objects, Assembly Theory offers an opportunity to generalize Darwin’s “descent with modification” as a principle that pervades all physical reality.
Implications Beyond the Origin of Life
Assembly Theory promises to reshape our understanding of how complex phenomena like life can emerge from simple particle interactions. But its implications may reach even farther. For example, the assembly perspective could offer new insights into our own advanced technologies, which are also assembled from component parts. Applying AT to human innovations like computers or aircraft could uncover parallels to biological evolution. Common principles may govern the open-ended generation of novelty in both nature and human engineering. Shared assembly dynamics may manifest across any substrate where combinatorial complexity arises through historical contingency.
The assembly lens also forces us to rethink fundamental assumptions about the nature of physical matter. Living organisms are not the only “objects” bearing evidence of historical constraints. All stable matter is formed through assembly processes. Thus, a select subset of the possible exists in our universe due to selection effects reaching back to the origins of particles, atoms, and chemistry.
From this perspective, the emergence of open-ended evolution was not something imposed upon physics when life appeared. Rather, rudimentary evolutionary processes may operate all along, at every scale where complex objects come into being. With its radically expanded notion of objects, Assembly Theory offers an opportunity to generalize Darwin’s “descent with modification” as a principle that pervades all physical reality.