- The following article is part two of Adam Frank’s interview with theoretical physicist Sara Imari Walker.
- Walker’s work on Assembly Theory explores the “physics of life,” a field seeking to understand what fundamentally distinguishes living from non-living systems.
- This approach seeks to go beyond Darwinian evolution, addressing the deeper mechanisms that create and sustain the organizational states necessary for life.
Welcome to part two of my conversation with theoretical physicist Sara Imari Walker, a professor at Arizona State University and the author of the 2024 book Life as No One Knows It. Walker’s book is an example of radical new thinking at the forefront of the “physics of life,” a burgeoning field investigating what makes life different from non-life at a fundamental level.
It’s a trickier question than you might think. While physics has long excelled at describing inanimate objects — like planets, rocks, and atoms — life poses a unique puzzle: Living systems create, adapt, and evolve in ways that non-living matter does not. What is different about the organization of living systems that makes this possible?
To tackle this, Walker and her collaborators developed Assembly Theory, a framework that reimagines the role of information and complexity in explaining life’s unique properties. Instead of focusing strictly on biological details, Assembly Theory explores what enables complex structures to emerge and persist, emphasizing how certain molecular arrangements hold functional or evolutionary significance. By looking at how intricate forms are built and sustained, this approach offers a new lens for measuring life’s complexity, one rooted in fundamental physical principles. Though the theory is still in its early stages and somewhat controversial, it opens up intriguing possibilities for understanding not only life on Earth but potentially life throughout the Universe.
Here’s part two of my conversation with Walker. (Read part one here.)
AF: So it was your and Paul Davies’ work that really got me thinking about life and information. That’s what eventually led me to the Semantic Information project our group at the University of Rochester is working on now. Information plays an important role in Assembly Theory which is the focus of your new book. But all this begs the question: what is information? Why is it important? How does it figure into questions of life and causality?
SW: There’s like a million definitions for information. The way that I think about it is we shouldn’t go in assuming we know what this thing is because we have all these descriptions of it. What we need to start with is the question, “What is the problem to be solved and how do we solve that problem?”
I played around with a lot of ideas but the way that I think about information right now is very tightly coupled to causation and causal time. It’s actually about the amount of causal history and the amount of constraint the Universe had to impose on the structure within itself to create an object. Shannon [thought of information] as the reduction in uncertainty, so that is a very physical notion of what the assembly index is capturing.
AF: This is getting us to Assembly Theory. You are interested in the information bound up in assembling complex objects.
SW: Yes, the question is, “What can you measure and what can you actually retroject from measurement to build a theory?” We can think of a certain level of complexity and I can talk about all the things the Universe could have constructed at that level of complexity. That leads to an entire counterfactual space. But what actually was built? In that reduction in uncertainty is some sense of how much information the Universe constructed over time [in building the complex object]. But the problem is knowing how big that counterfactual space is. That’s not something that you can actually physically measure.
AF: Yeah, that’s one issue with Semantic Information Theory. We have to work with the counterfactual evolutionary trajectories of the system. These are histories that did not occur. One task we had was to find ways to make that computable. But with Assembly Theory, it seems you really are focused on being able to measure things in a straightforward way, like using mass spectrometer measurements of molecules.
SW: Right. And that’s why I’m super excited about Assembly Theory. There is something you can measure. In Assembly Theory we always talk about what’s actually physically real. Then we can try and construct some of the structure of that counterfactual space. So in this way, I do think information is very physical. I think it’s a very material property. But the ways that we’ve been trying to talk about it and measure it in the past are not the best candidates for building a theory that you can actually test.
Again, if you go back again to the history of physics, temperature had to be invented as that thing we measured before you could make the theory of thermodynamics. And if you go back to Einstein’s theory, he was the only one that took seriously the idea of the measurements of the speed of light being constant. Then he built his theory.
AF: So let’s talk more about Assembly Theory. The “copy number” is an important quantity associated with a complex object. It tells you how many of those objects the Universe has produced. I get that this is important since a single version of a complex object tells you it’s not very important, but if there are lots of copies, that means some version of selection is at work. Can you talk more about this idea?
SW: Yes, when you have a high copy number of things, then you can begin talking from an information theoretic perspective. You’re starting to capture an emergent layer and the causal structure underlying it. Now you have this thing that you can look at as the particular time series [that led to it].
AF: How does this relate to selection? When the Nature paper came out, many people criticized it because they said it didn’t explain something like Darwinian selection, but I take it you were thinking about something more basic than that.
SW: In a very active sense, the Universe is constructing itself. The mechanism of selection is like the construction of what’s next. Why do some structures get created and not others? Again we have a kind of a counterfactual statement. Darwin’s got this quote about the struggle for existence but he’s always talking about things that already exist. We don’t really think about the space of things that could exist or might exist. What is the mechanism of bringing novel objects into our Universe and also those objects persisting in time?
AF: So by “selection,” you want to ask a question that comes before Darwin?
SW: Yes. A cell that can evolve. There you have a state of organization that you recognize as already having captured some of these features that we feel are intrinsic to the evolution process. But what do you have before that? What’s the process of evolution that constructs organizational states of matter that can do that? There our theories of evolution selection are inadequate so you need an even deeper theory. You need a theory that explains the evolution of evolution!
(As a final note: Our conversation went far beyond this and the previous piece. I always come away from my conversations with Walker feeling like I need to get right to work on all the great ideas we discussed. I’m following the progress of Assembly Theory closely and think time will tell whether it can live up to its possibilities and promises. In the meantime, I highly recommend her book for a discussion of physics questions that you don’t usually get in physics books.)
