Were the principles of life invented or discovered?

The concept of emergence just seems to rub some people the wrong way. The idea that there may be more in the whole than there is in the parts can seem like it’s cheating somehow, like we’re getting a free lunch, going beyond the basic laws of physics or even somehow violating them. This is especially true for varieties of emergence that claim that assemblies of parts can have causal power that the parts themselves do not have.

(Ernst Haeckel)

The basic premise of reductionism – the prevailing principle in much of modern science – is that all the apparent causality that we see in the arrangements of parts is really reducible to and fully explained by the basic forces operating on all the bits – the laws of physics playing out in ways that are certainly complex, but that require no other level of explanation. Any apparent purposiveness is an illusion, an epiphenomenon that plays no causal part in the dynamics of the system.

The alternative to reductionism seems to require an abandonment of the more basic premise of physicalism and an embrace of an almost mystical holism, a vague and slippery concept, too hard to pin down to satisfy hard-nosed scientists, especially many physicists. It is a woo-merchant’s dream, often accompanied by vague allusions to Eastern philosophies, implications of a cosmic teleology in evolution, or some appeal to dualism to explain the human mind.

This isn’t true, of course. It’s perfectly possible to formulate a non-reductive kind of physicalism, that involves emergent functions that depend on the arrangements of things, not only on the things themselves. However, this idea of emergence seems to over-rule the widely held, even if often tacit, axiom that “there cannot be more in the effect than there is in the cause”. This idea dates back to the ancient Greeks and has been described by A.O. Lovejoy as the “preformationist assumption about causality”.

The basic assertion (or assumption) expressed in this axiom seems to be that there is no causal free lunch. I like to think of this as analogous to the principle of the conservation of energy, but for causal power – you can’t get more causality out than you put in.

Now, when it comes to evolution, this poses a bit of a problem. Where does “life”, and life’s sub-functions, come from, if there is no life in the biochemical bits that living things are made of? Where does purposive, goal-directed behaviour come from if the bits of behaving organisms have no purpose? How can biochemical pathways, genetic circuits, and neural systems do things that proteins, genes, and neurons can’t do?

Is there a true creativity to the process of evolution? Can it really invent things capable of acting as causal systems or even as agents? Isn’t this getting some causality for free, if that causal power is not already in the bits that evolution puts together?

Even if you are not a reductionist and believe, as I do, that there is genuine causal power and functionality in the arrangements of things that is not contained in the things themselves or fully explainable by their low-level interactions, you are still left with the question of where that functionality comes from. It does feel like it somehow violates the conservation of causal power for evolution to have simply created systems that can do things – truly amazing, incredibly powerful things – from the mundane, purposeless matter that we are all made of.

But there is another way to think about it. We can think of the architectures of functional systems – arrangements that are capable of doing things – as having a free-floating “existence”, in an abstract sense. There are, in the first instance, certain ways of organising things that are self-sustaining, or self-organising, or even self-replicating. (And there are billions more ways of organising things that do not have those properties).

On a smaller scale, there are functional motifs – feedforward or feedback loops, with various delays, convergent or divergent connectivity – that can do things, that can act as filters, integrators, regulators, oscillators. And there are ways of combining these primitive computational or regulatory or mechanical elements into distributed networks or hierarchical systems that can mediate homeostasis, instantiate memory, model the environment, make predictive inferences, encode goals, drive reinforcement learning, make decisions. (And there are billions more ways that cannot).

The functionality is not invented or created by evolution – it is discovered. Those design principles just are. Those functional motifs exist like Platonic forms, whether or not they are ever discovered or instantiated. It doesn’t matter what the constituent bits are – they could be genes, proteins, neurons, brain regions (or cogs, pulleys, valves, semiconductors) – if you put them together in certain ways, with certain types of interactions, you’ll get some particular functionality “emerging” out of the system. It’s not quite that the functionality comes for free – it has no cost at all, as it already exists.

Finding it doesn’t come for free, however. It just looks that way in retrospect, when we consider only the successful pathways and ignore the billions of moves that did not land in a functional, robust, self-sustaining part of systems space.

We can think of evolution as blindly exploring that space of workable arrangements, but in a progressive fashion enabled by the storage of information in the genome and the ratchet-like action of natural selection. If a mutational event allows a move to a different part of systems configuration space, with some additional functionality, and if that functionality is beneficial, then the change will be kept. It will be passed down to the next generation, and may eventually become fixed in the population. That will generate a new starting point for additional exploration – to points in systems space that were previously unreachable. 

Of course, every generation will have lots of mutations that lead to points in systems configuration space that are not beneficial, that do not instantiate new, useful functions, or that do not lead to stable and robust system organisation. We don't see those when we look at the grand sweep of evolution, which is why its progressiveness can seem to be driven by some cosmic teleology – some divine design that propels life towards ever-greater complexity. This purposiveness of evolution is an illusion, though the purposiveness of organisms and the functionality of their subsystems is not.

You could therefore argue that, in addition to random mutation and natural selection, we need something else to explain the evolution of life – abstract, independently pre-existing principles of self-sustaining, self-organising, and eventually self-replicating complex systems. Those properties don’t just emerge by default when you start throwing more and more components together. The principles don’t emerge at all from the lower level dynamics – they exist in and of themselves and constrain the lower level organisation and dynamics.

The apparently emergent functions that characterise many complex systems and that many people find so metaphysically troubling are thus not mystical at all. This functionality isn’t invented or created de novo, but discovered, and the constraints by which its causal influence is mediated do not violate any lower-level physical laws. 

That shift in perspective can, in my opinion, underpin a more principled approach to both research and teaching in Biology. Rather than merely describing what is happening in a system, as if it just happens to work that way, we can recognise functional motifs and architectures that do different types of things. Approaching things in terms of the abstract engineering principles underpinning various functional architectures can thus offer a unifying foundation for the study of both evolved and designed systems.

With thanks to Jag Bhalla (@hangingnoodles) for very helpful comments and edits.

Related reading:

Climbing Mount Improbable. Richard Dawkins, 2006 https://www.amazon.co.uk/Climbing-Mount-Improbable-Richard-Dawkins/dp/0141026170

Freedom Evolves. Daniel Dennett, 2004 https://www.amazon.co.uk/Freedom-Evolves-Daniel-C-Dennett/dp/0140283897

Order Out of Chaos. Ilya Prigogine and Isabelle Stengers, 1984 https://www.amazon.co.uk/Order-out-Chaos-PRIGOGINE/dp/0553343637

Dynamics in Action – Intentional Behavior as a Complex System. Alicia Juarrero, 2002 https://mitpress.mit.edu/books/dynamics-action

Emergence, Complexity, and Self-Organisation: Precursors and Prototypes. Edited by Alicia Juarrero and Carl A. Rubino, 2008 https://www.amazon.co.uk/Emergence-Complexity-Self-Organization-Precursors-Prototypes/dp/0981703216

General System Theory: Foundations, Development, Applications. Ludwig von Bertalanffy, 1969 https://www.amazon.com/General-System-Theory-Foundations-Applications/dp/0807604534