After resetting the universe, would life evolution tell a different story?

Fundamental constraints to the logic of living systems

If you find Complexity Thoughts interesting, follow us! Click on the Like button, leave a comment, repost on Substack or share this post. It is the only feedback I can have for this free service. The frequency and quality of this newsletter relies on social interactions. Thank you!

---

Dall-e 3 representation of this issue’s content

---

Introduction

Months ago, I was having with Ricard Solé one of our (usually) interesting discussions. Among other questions, he was wondering why we recurrently observe the existence of some functional sub-systems, such as an eye, across a variety of species.

It makes sense: the eye as we know it it’s essentially ubiquitous (with a few exceptions). Is it the only possible functional unit for a complex organism to visually capturing the external environment or is it some kind of frozen accident that happened at some point in the evolutionary history of life on Earth?

One of the key aspects of biological organization is the existence of distinct phases of complexity --and the transitions therein. While the possible microscopic configurations can define a continuum, or increase combinatorially with system size, the emerging meso- and macroscopic patterns take far fewer forms. The emergence of such a discrete and low dimensional universe of phases has a deep impact on what is possible within developing organisms and evolution — Bernat Corominas Murtra, coauthor of the paper

He pointed me to some articles by Pere Alberch about the “logic of monsters”. I warmly invite you to read those papers, they are fascinating, with visionary questions and insights. Here an extract from one abstract:

There are two philosophical approaches to interpret the orderliness of natural systems. […] Classical neo-Darwinism falls within the "externalist" tradition, with its emphasis in natural selection as the main ordering agent in evolution, this approach basically argues that the properties of the physical ant biotic environment determine the selective pressures and consequently dictate which form will be selected over others. Therefore, the discreteness and order of natural diversity is a direct reflection of the topography of the adaptative landscape. The internalist approach attributes some of the order observed in nature as the result of the emergent properties of generative rules.

In a nutshell: why you can instantaneously recognize that a creature like a chimera cannot be real?

Figure: a chimera (1590-1610) from Museo del Prado.

Well, because it’s somehow violating our intuition of the logic of living systems. But what does that mean at all?

The question is old and you might find it enlightening to read the ‘70s book from Francois Jacob on the “Logic of Life”, which started a different way of thinking about this problem.

Unraveling the logic of living systems

With the ambitious goal of trying to answer the above questions, Ricard was able to gather together an interdisciplinary dream team of complexity scientists with various expertise, from information theory to the origin of life, from thermodynamics to complex networks. The idea? Discuss some of the fundamental constraints that limit the space of evolutionary outcomes, focusing on areas that are most well-studied and most likely to be universal, involving several case studies that reveal deep constraints associated with the logic of the organization of living systems:

It has been argued that the historical nature of evolution makes it a highly path-dependent process. Under this view, the outcome of evolutionary dynamics could have resulted in organisms with different forms and functions. At the same time, there is ample evidence that convergence and constraints strongly limit the domain of the potential design principles that evolution can achieve. Are these limitations relevant in shaping the fabric of the possible? Here, we argue that fundamental constraints are associated with the logic of living matter. We illustrate this idea by considering the thermodynamic properties of living systems, the linear nature of molecular information, the cellular nature of the building blocks of life, multicellularity and development, the threshold nature of computations in cognitive systems, and the discrete nature of the architecture of ecosystems. In all these examples, we present available evidence and suggest potential avenues towards a well-defined theoretical formulation. — Solé et al

The resulting paper tackles a broad range of challenges to identify a rather few number of fundamental constraints that matter when dealing with life and its evolution. Can I summarize them in two lines?

Well, Ricard challenged me to achieve that, but I know that’s not possible. So let’s agree that the following points are representative enough of the content of our paper and then let’s try to explain in some more detail the important steps:

• Universal thermodynamic logic constraints

• Informational, algorithmic, and computational logic constraints

• Ecological logic constraints

• Functional evolutionary logic constraints

To some extent, these points are strictly related to major evolutionary transitions thought to describe life as we know it, but that discussion is definitely left for a future post.

We try to understand which aspects of biology -- if any -- are determined a priori by the laws of physics, logic, and mathematics. The ultimate goal is to find universal principles that govern life as a state of matter, both on this planet and beyond — Artemy Kolchinsky, coauthor of the paper

Universal thermodynamic logic constraints

Let’s start from thermodynamics, the only physical theory that is correct across any scale (you might want to check the post below for more details about this).

In fact, as argued already by Boltzmann in 1886 and later by Schroedinger in his book “What is life?”, the second law of thermodynamics states that the change in total entropy of a system and its environment must be positive:

\(\Delta S_{\text{total}} = \Delta S_{\text{env}} + \Delta S_{\text{system}} > 0\)

Since living systems can operate only if they reduce their entropy at the expense of the environment, we have:

\(\Delta S_{\text{system}} < 0, \quad \Delta S_{\text{env}}>0, \quad |\Delta S_{\text{env}}|>|\Delta S_{\text{system}}|\)

This is a very fundamental constraint, that we name universal thermodynamic logic of life. In fact, the second law is valid both globally and locally, and therefore even individual entropy-reducing reactions that characterize the metabolism of a living being must be locally coupled to a free-energy source: it’s unavoidable.

Note: Intriguingly, the emerging dissipative cyclic structures work out of equilibrium due to external driving: the organization into cycles is required, according to Morowitz, for any nonequilibrium chemical system in steady state.

Informational, algorithmic, and computational logic constraints

Of course there is something else beyond thermodynamical constraints. It is the case of information, since many ubiquitous (potentially universal?) characteristics of life as we know it are related to its informational, algorithmic, and computational properties. In fact, information carriers can reliably code for a large number of phenotypic states their information can be replicated.

Intriguingly, the behavior of linear polymers such as RNA and DNA can be understood in terms of classical models of computation, such as Turing machines. The ubiquity of such structures for biological information processing — together with the fact that more complex (higher-dimensional) architectures do not provide any form of practical advantage — suggests that there is something special about one-dimensional linear polymers that leads them to be a universal solution to the information-carrying problem. Indeed, it is tantalizing that such structures build the backbone of a cell and that those cells duplicate the carried information by means of processes that have been theorized as a general solution to replication: the Von Neumann universal constructor for self-replicating machines, which can be (partially?) understood even in terms of self-organization of soft-matter.

From so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved. — Charles Darwin

From a single cell, let’s move to the emergence of multicellular forms: is there a truly endless universe of multicellular form, or is the universe of what is actually observed limited due to fundamental constraints affecting the evolution of complex life?Physics can help to provide a more solid ground to early intuitions about the mechanics of biological form and its changes (see D’Arcy Thompson). Cells attach to neighbors more if it lowers the energy per unit area, and the dynamics can be mapped into the typical ones routinely used in statistical mechanics (e.g., resembling algorithms such as simulated annealing).

At a higher scale, storing and accessing information is necessary for the emergence of cognitive agents who learn from their environment and make decisions. A general feature of many cognitive systems is the presence of mechanisms that transform analog signals into digital responses. The need for signal discrimination and the fact that analog computation is more prone to noise might be two crucial constraints on the evolution of cognition. It is, once again, tantalizing that the general architecture for processing such information follows a logic of nonlinear response based on integrate-and-threshold dynamics (the McCulloch-Pitts model’s conceptual framework) such as

\(\frac{\partial x_{i}(t)}{\partial t} = f\left( \sum_{j=1}^{N} w_{ij}x_{j}(t) - \theta_{i} \right)\)

that can be used also to characterize other information transmission systems, like genetic networks. Remarkably, the underlying idea is to integrate information propagation like in a network cascade process and, if the integration leads to a value above some threshold (here, indicated by theta) then there is a response. Why this mechanism is so ubiquitous (or, at least it is able to explain so many processes)? Because, in practice, it allows to construct any logic Boolean circuit to operate similarly to our computers. [ndr: Also the computer metaphor deserves a dedicated post, be patient.]

Ecological logic and functional evolutionary constraints

Arrived at this point, we are already out of space and likely I lost my bet with Ricard. Nevertheless, it’s worth mentioning that the laws of evolution and ecological systems impose some additional layers of constraints to the logic of living systems. Models have provided great insights to reproduce observations in natural and digital ecosystems, from coevolution to cooperation and punctuated equilibrium. Nevertheless, we are still struggling with understanding why biodiversity is so extensive and how this can be reconciled with the apparent stability of large ecosystems. In fact, the dynamics of ecosystems consists of neutral, positive and negative interactions that contributes to shape the evolution of single organisms and their abundances. Once again, it’s not completely surprising that the most powerful recent approaches to ecology connects allometric scaling laws with fundamental physical constraints.

Ecosystem architectures are deeply constrained within a finite set of possible classes of ecological interactions. Current and past ecosystems reveal such a discrete repertoire of possibilities, and in silico models of evolving ecologies support this constrained repertoire. Among other regularities, the widespread presence of parasites suggests that they are an inevitable outcome of complex adaptive systems.

Towards a unifying picture: phase transitions?

It is remarkable that major evolutionary transitions, such as the emergence of multicellularity and language, can be understood through the lens of phase transitions in statistical physics. These transitions represent significant shifts in biological organization and complexity, similar to phase transitions observed in physical systems: accordingly, we can hope that the language of statistical physics can help to build an elegant mathematical and computational framework to describe those shifts. Intuitively, a special role is expected to be played by criticality, describing where systems are poised between order and disorder. In fact, it seems that criticality is crucial for the evolution of living systems, optimizing information transfer and sensitivity to external signals. We argue, therefore, that understanding these transitions can shed light on macroevolutionary processes and the fundamental laws driving biological complexity, such as self-organization and emergent phenomena, beyond just natural selection.

---

What’s next?

Well, despite the attempt to make some order, the research roadmap is still rich and long. Throughout the paper, the concept of network is ubiquitous: one possible limitation is that existing results could be put in context with the recent advances of network theory and how they are helping us to make sense of the observed phenomenology. We have gained a lot of insights from 2+ decades of network science, and our paper only partially succeeds in building on that knowledge.

A potential follow up for the next future might be to show how the tools of statistical physics can be used together with network theory for the ambitious goal of building a first formal framework to characterize the constraints to the logic of living systems.