Life is algorithmic. Is the Universe?

How fundamental is information and consciousness? Is structure information?

This post was prompted by John Carter’s essay on information and observation “The Eye at the End of Time”. That, and my irritation with Donald Hoffman’s interface theory of perception (to be discussed in a future post). Rather impertinently, I have a different read on observation in quantum mechanics than does John Carter. (If I am wrong, I hope I am at least usefully wrong.)

It is useful to start with some definitions:

Information: what is conveyed or represented by a particular arrangement or sequence of things. (Genes are packages of information that direct cells to grow and structure themselves in particular ways.)

Structure is not in itself information — it does not convey or represent — even if it can generate information and information is structured.

Living thing: something that uses information and resources to maintain its own functioning. (A cancer cell is a cell that defects from cooperating with other cells and spreads itself at their expense.)

Algorithm: (1) an unambiguous method of solving a specific problem; (2) a sequence of instructions; (3) an embedded if-then sequence.

Emergent: increased complexity in structure having causal effect. More technically, when quantitative changes in a system result in qualitative changes in behaviour (adapted from physicist Philip Anderson [1972].)

Entropy: system’s thermal energy per unit temperature that is unavailable for doing useful work. Because work is obtained from ordered molecular motion, the amount of entropy is also a measure of the molecular disorder, or randomness, of a system.

Assembly Theory says that life is existence/persistence and copying, which is another way of saying that a living organism maintains its own functioning and reproduction.

Quantum mechanics — especially since the advent of computers and even more since the advent of the internet — has led to serious consideration that information, or even consciousness, might be fundamental to the universe. This is because of the issue of measurement in quantum mechanics and the collapse of the wave function.

The Schrödinger equation (governing the wave function) at the heart of quantum mechanics treats measurement mathematically. There is no definition of what physical processes count as measurement.

At the quantum level, measurement affects what is observed. The question here is whether such measurement is structuring reality or is measurement increasing information. If it does the former, it will do the latter. But it can do the latter without doing the former. “Affects what is observed” is very much ambiguous between those two possibilities.

Unsplash in collaboration with Osarugue Igbinoba

There is also an ambiguity in something being indeterminate. It can be indeterminate in the sense of not having specific features prior to being measured or it can be indeterminate in the sense of information attainable being constrained, such that something cannot be determined with current information. Not yet having specific features constrains information available but something can have a specific feature and yet information about it is constrained.

As living beings, we use information. As conscious beings, we perceive (and respond) to information. As self-conscious beings, we interrogate and package information at a more complex level. As is discussed more below, it can be hard to keep clear differences between the information, what the information is about and what it is from — hence, the ambiguities in indeterminate.

There is a difference between perception—bacteria perceive—and the apprehension of perception. This is particularly true for self-conscious beings that interrogate and package information.

Descriptively versus actively algorithmic

Mathematics is the science of structure, the universe is stably structured, so features of the universe can be described mathematically. In that (loose) sense, the universe can be said to be algorithmic.

But that is being passively algorithmic. There are structures, that can be described mathematically. Predictions can be made from such mathematical descriptions. Measurements can be made to ascertain how well a prediction did.

But those processes of using mathematics, conducting experiments or observations, measuring the same, working out their implications, all involve active use of information, not mere passive generation of them.

Living things are actively algorithmic, are active users of information, in a way non-living things — apart from some constructed tools of living things — are not, because living things use information and resources to maintain themselves and propagate. Living cells operate to increase energy available to them and to retain, or even increase, their own order. The directedness of living things is actively anti-entropic in a way nonliving things are not.

We can think of living cells as being structured to be anti-entropic Maxwell’s demons: that is, to let in what maintains and expands their order and to block what does not. Death is highly entropic.

That living things are actively algorithmic in order to maintain themselves, and to propagate, is how purpose gets into the universe. It is such directedness — that purposiveness of living things — which sets biological phenomena (and phenomena emergent from biology) apart from the rest of physics and chemistry, from phenomena covered by (the sciences of) Physics and Chemistry.

Computers are in a sense actively algorithmic — they use information — but in a very limited way. Computers do not use information to maintain themselves, nor for purposes generated from within themselves. They are constructed tools. They use information in computations for purposes intended to be useful to their creators. All of which makes computers poor analogues for living processes.

That their processes are computations makes computers even weaker analogues for living process. To see that, it helps that physicist Sir Roger Penrose has provided by far the best statement of Godel’s Incompleteness Theorems I have come across.

What those Incompleteness Theorems tells us is that our understanding exceeds computation: things are true, and can be known to be true, that cannot be computationally demonstrated. Our brains are not just biological computers. Of course, consciousness is a big hint that they are not.

Equations are models

Mathematics is the science of structure; Physics, the science of physical processes, of matter, energy and force; Chemistry, the science of material processes, of matter; Biology the science of living processes, of life and living organisms. We can distinguish between Physics (the science) and physics (the world); between Chemistry and chemistry; between Biology and biology. But we cannot distinguish between Mathematics and mathematics, because there is no mathematics out in the world, there is only structure.

Information is the stuff of science, but it is not the stuff of reality, it is not structure itself. Information is what living organisms take from the rest of reality to do, to act: including to do science. Information is what makes a map, a map, but not the territory.

Mathematics is how we map structure with a certain rigour. An equation is a (potential) model of reality, it is not reality. A key element in using applying mathematics is working out what mathematical map is appropriate to what structure. When economist Paul Romer critiqued mathiness in Economics, he was pointing out that mathematics could be used to give a pretense of rigour and accuracy. This is a particular manifestation of the mathematical map not being the territory, of Mathematics being the science of structure, not structure itself.

It was a shock when Relativity overturned Newtonian Mechanics. Except, it didn’t really. The equations of Newtonian mechanics work fine as maps of reality across most of reality, just not for very small, and very large, phenomena. Relativity did not make those Newtonian equations false, it just demonstrated that there were limits to the range over which they were accurate.

The wave function of quantum mechanics is a map of reality. A map of reality based on a certain level of information. Measurement adds information. What is unclear at the level of information before measurement adds information is no longer unclear after measurement has done so.

Schrodinger came up with his famous cat thought experiment—the cat that is neither alive nor dead until measurement collapses the wave function—to indicate that there was something wrong with his own equation and how folk were conceiving quantum mechanics. If we remember the wave function is an equation, and an equation is an (hopefully accurate) map of our existing information about structure, there is no metaphysical or ontological indeterminacy involved in how the cat is, there is only informational (epistemic) indeterminacy about how the cat is that measurement resolves.

On quantum mechanics

Measurement is something an actively algorithmic thing does. Quantum mechanics is based on a linear wave function. The updating of information via measurement = collapse of wave function into a specified outcome. Unfortunately, quantum mechanics does not define what such a detector is.

If measurement is not defined, then contradictory interpretations are possible, generating contradictory results. Hence it would good if it could be replaced by a theory that does define measurement as a physical process.

Many physicists claim that measurement is not a physical process but an updating of information. Clearly it includes a physical process: information needs a physical conveyor. To have Mathematics requires things to write equations on and mechanisms for equations to be perceived.

Claiming that the collapse of the wave function is not a physical process but an informational one runs into the difficulty that quantum mechanics does not agree with various observations, such as the particularly chaotic oscillation of one of Saturn’s moon. What is the physical process needed to make it fit? This is unknown.

We can see all this is a very weak reed to rest deep claims about the universe being information “all the way down”, let alone consciousness being central to the universe.

Sir Roger Penrose does not believe consciousness imposes itself on the wave function. The Schrodinger’s cat example was for arguing that there was something incomplete about Quantum Mechanics. It was not for postulating some radical indeterminacy or to embed consciousness into the universe.

Nevertheless, the measurement problem in Quantum Mechanics — the collapse of the wave function — is one bit of Physics that is not computational. Godel’s Theorem tells us that understanding — which involves consciousness — is not merely computational. Intriguingly, as I have discussed elsewhere, consciousness is unable to fully observe itself.

Emergence: structure and consilience

In considering these issues, we are dealing with emergent properties in two senses. The first is in the structuring of the universe, the second in being beings who can study the universe.

Chemical phenomena are emergent from basic physical structures — such properties and behaviour are constrained (and enabled) by basic physical structures of the universe. So, Chemistry has to be consilient with Physics.

Biological phenomena are emergent from chemical phenomena — such properties and behaviour are constrained (and enabled) by basic chemistry of the universe. So, Biology has to be consilient with Chemistry.

Social and human sciences are the study of the properties and behaviour of a specific species: Homo sapiens. Human behaviour, including and social phenomena, are emergent from our biology — such behaviour and phenomena are constrained (and enabled) by our biology. So, social and human sciences have to be consilient with Biology.

Each further level of study is dealing with more complex phenomena that have their own causal processes, but whose structure and causal processes are constrained (and enabled) by more basic structures of reality.

Emerging complexity

We are living beings who are created by, and live in, that reality. We also perceive and study that reality.

We are information-saturated beings. We are living things who use information and resources to maintain ourselves, down to the cellular level. We receive, perceive and respond to information. We interrogate and package information: attempting to do so accurately and rigorously when engaged in scholarship and science.

We perceive patterns and structures via information. We respond to, and about, patterns and structures via information. As living beings, we have active response to information pervasively built into us: in our cells, within our cells, with our genes.

But, can non-living things be said to “use” information? Physical causation operates as consequences of structure. But it is not an active (so informational) response to structure in the sense that, for example, viruses respond to structure because they have sufficient directedness to do so.

We, as living organisms — and not merely living organisms, but self-conscious living organisms — use information to identify patterns and structure in the physical world. It does not mean that those patterns and structures are information in a causal sense within those (non-living) patterns and structures. Living beings have mechanisms of perception precisely so they can assemble and use information.

Non-living things do not have embedded algorithmic processes to maintain and replicate themselves. They just have structure and the dynamics that structure generates. Structures than can be described algorithmically, but are not actively algorithmic.

Living organisms have problems to solve if they are going to survive and reproduce. They have directedness. They need embedded algorithms—embedded if-then chains—to do that.

Non-living things have no problems to solve, because they lack directedness. They have no “point of view”, no “point of action”, they just are. The striking thing about life is precisely that the universe is not “information all the way down”, it is not (actively) algorithmic all the way down, as living things are.

Living things are very much algorithmic beings — in our genes, our cells, our perception, our cognition — and use embedded algorithms to respond to the world around us. The extent of algorithmic structures increases with increased complexity of organisms.

Information, and embedded algorithms, come from having directedness — using information and resources to maintain and reproduce — thereby facing problems to solve. In our responses to our physical environment, we use information to perceive and respond. We need information to interrogate and communicate. Our directedness requires information so we can act, even at a cellular level.

But it is a mistake to then project that back on to physical reality. To project our information-saturation onto physical reality. We use information to perceive reality but that is a long way from reality being fundamentally information.

We need to be clear about differences between the information, what the information is about and what it is from. We need to be clear that the wave function—like all equations in Physics—is a map of the territory, a model of reality. The equations are not the territory, they are not the reality.

Reference horizon

Our self-consciousness enormously increases our ability to create and use information by allowing us to interrogate and package information. But it also limits our ability to do so in a key way. It limits our ability to interrogate the basic structure of the thoughts and words by which we create, use and convey information.

As I tried to convey with limited success here and here, and rather more success here and here, to be consciously thought, a certain level of coherence has to be achieved, otherwise we cannot be conscious of it.

We therefore cannot consciously think about the sub-coherent constituents of thought. There is an information “black hole” — a reference horizon — at the centre of our thought and language.

That also makes it harder to separate our use of information — including observing patterns and structures — from those patterns and structures themselves.

If information having causal power is limited to organisms that actively respond and replicate — and their more sophisticated tools — then the universe is not information “all the way down”. Nor does information eliminating the information indeterminacy of the wave function provide any sort of evidence that consciousness is central to reality.

That a ball is consciously thrown does not change or affect the physics of a thrown ball. That the wave function model of quantum reality is informationally limited—a limitation that measurement resolves—and that measurement is consciously done does not change the physics—as distinct from the Physics—of the quantum realm.

References

P. W. Anderson, ‘More is Different,’ Science, New Series, Vol. 177, No. 4047. (Aug. 4, 1972), 393-396. https://www.tkm.kit.edu/downloads/TKM1_2011_more_is_different_PWA.pdf

Ned Block, ‘Consciousness, Big Science and Conceptual Clarity,’ in The Future of the Brain: Essays by the World’s Leading Neuroscientists, Jeremy Freeman and Gary Marcus (eds.), Princeton University Press. https://www.nedblock.us/papers/2014.consc.brain.pdf

Acer Y.C. Chang, Martin Biehl, Yen Yu, and Ryota Kanai, ‘Information Closure Theory of Consciousness,’ arXiv:1909.13045v2 [q-bio.NC] 11 Jun 2020. https://pmc.ncbi.nlm.nih.gov/articles/PMC7374725/

Arman Rezayati Charan, Shahriar Gharibzadeh, S. Mahdi Firouzabadi, ‘Realism is almost true: A critique of the interface theory of perception,’ arXiv:2111.03864v1, Nov. 2021. https://arxiv.org/abs/2111.03864

Marc O. Ernst & Martin S. Banks, ‘Humans integrate visual and haptic information in a statistically optimal fashion,’ Nature, Vol. 415, 24 JANUARY 2002, 429-433. https://www.researchgate.net/publication/11550808_Humans_integrate_visual_and_haptic_information_in_a_statistically_optimal_fashion

James Franklin, What Science Knows: And How It Knows It, Encounter Books, 2009.

Chris D. Frith, ‘The role of metacognition in human social interactions,’ Philosophical Transactions of the Royal Society B, 2012, 367, 2213–2223. https://pmc.ncbi.nlm.nih.gov/articles/PMC3385688/

L. Gabora, M. Steel, ‘A model of the transition to behavioural and cognitive modernity using reflexively autocatalytic networks,’ Journal Royal Society Interface 2020, 17: 20200545. https://royalsocietypublishing.org/doi/10.1098/rsif.2020.0545

Griffiths TL, Lieder F, Goodman ND. ‘Rational use of cognitive resources: levels of analysis between the computational and the algorithmic,’ Topics in Cognitive Science, 2015 Apr;7(2):217-29. https://onlinelibrary.wiley.com/doi/10.1111/tops.12142

Kevin Healy, Luke McNally, Graeme D. Ruxton, Natalie Cooper, Andrew L. Jackson, ‘Metabolic rate and body size are linked with perception of temporal information,’ Animal Behaviour, 86 (2013) 685-696. https://www.sciencedirect.com/science/article/pii/S0003347213003060

Cecilia Heyes, ‘New thinking: the evolution of human cognition,’ Phil. Trans. R. Soc. B, 2012, 367, 2091–2096. https://pmc.ncbi.nlm.nih.gov/articles/PMC3385676/

Heather Heying & Bret Weinstein, A Hunter-Gatherer’s Guide to the Twenty-First Century: Evolution and the Challenges of Modern Life, Swift, 2021.

Erik P. Hoel, Larissa Albantakis, and Giulio Tononi, ‘Quantifying causal emergence shows that macro can beat micro,’ PNAS, December 3, 2013, vol. 110, no. 49, 19790–19795. https://www.pnas.org/doi/10.1073/pnas.1314922110

Philip Low, The Cambridge Declaration on Consciousness, Jaak Panksepp, Diana Reiss, David Edelman, Bruno Van Swinderen, Philip Low and Christof Koch (eds), July 7 2012. https://fcmconference.org/img/CambridgeDeclarationOnConsciousness.pdf

Rainer Mausfeld, ‘The Physicalistic Trap in Perception Theory,’ in Perception and the Physical World, D. Heyer & R. Mausfeld (eds.), Wiley, 2002. https://philarchive.org/archive/MAUTPT#:~:text=The%20physiologistic%20trap%20thus%20slices,transformations%20of%20the%20sensory%20input.

Paul M. Romer, ‘Mathiness in the Theory of Economic Growth,’ American Economic Review 105 (5): 2015, 89–93. https://www.aeaweb.org/articles?id=10.1257/aer.p20151066

‘Ecological Theory of Perception’, Sage Edge Publications, no date. https://study.sagepub.com/sites/default/files/Ecological Theory of Perception.pdf.

Erwin Schrödinger, What Is Life? The Physical Aspect of the Living Cell with Mind And Matter & Autobiographical Sketches, Cambridge University Press, [1944, 1958, 1992] 2013.

Arthur P. Shimamura, ‘Toward a Cognitive Neuroscience of Metacognition,’ Consciousness and Cognition 9, 313–323 (2000). https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=cfd323548dc3b1739e164594552ba52a20794a73

Naotsugu Tsuchiya and Christof Koch, ‘Attention and consciousness,’ Scholarpedia, 2008, 3(5):4173, revision #129964. http://www.scholarpedia.org/article/Attention_and_consciousness

Jason Wei, Rishi Bommasani, Colin Raffel,  Barret Zoph, Sebastian Borgeaud, Dani Yogatama, Maarten Bosmam  Denny Zhou, Donald Metzler, Tatsunori Hashimoto,  Oriol Vinyals, Percy Liang, Jeff Dean, William Fedus, ‘Emergent Abilities of Large Language Models,’ arXiv:2206.0762v1 [cs.CL] 15 Jun, 2022. https://arxiv.org/abs/2206.07682

John Archibald Wheeler, ‘Information, physics, quantum: the search for links,’ in Proceedings III International Symposium on Foundations of Quantum Mechanics, 1989, 354-358. https://philpapers.org/archive/WHEIPQ.pdf