A Platonic Space background for questions in consciousness

I was recently asked to write a short preamble outlining my perspective on four specific topics, for participation in a conference on consciousness. Here it is:

I’m not a consciousness researcher, and my lab’s work at the intersection of developmental biology, computer science, and cognitive science mostly studies 3rd-person observable behavior (writ large, in all kinds of spaces). But I do think there are aspects relevant to the study of consciousness, even if I don’t make any strong claims about it (because it is not yet tightly impacting our experimental work). Much more is here, here, and here.

Background: diverse intelligence across embodiments

We explore the idea that the interesting properties studied in cognitive and behavioral science form a continuum, not a set of crisp categories. In other words, the right question about intelligence, consciousness, etc. is “what kind and how much”, not “is it or isn’t it” [1-3]. I think we seek models of transformation and scaling of the cognitive light cone, not sharp boundaries [4], as would be relevant for our journeys through embryonic development and evolution. I think neuroscience is not about neurons, and intelligence and consciousness are not really about brains [5]. We view bodies as multiscale collective intelligences, with layers from molecular networks to ecosystems that all have agendas, goals, and different degrees of ingenuity in meeting those goals. We seek models of scaling and alignment of competent parts towards enabling larger-scale minds operating on larger goals.

At the bench, we apply tools and concepts of behavioral and neuroscience to all sorts of things that aren’t brains: slime molds, cells, tissues, organs, inorganic materials, molecular networks, minimal computational processes, patterns in excitable media, hybrid beings composed of both evolved and engineered components, etc. etc. We also study basal cognition, tracing the origin of learning, problem-solving, representation, etc. to its roots on embryological and evolutionary scales, and develop models of how competencies scale and project into different problem spaces, including weird ones (like anatomical morphospace) that are hard for humans to visualize.

This research program has allowed us to discover a number of new approaches in regenerative medicine and engineering, etc. as we communicate with cells about anatomical goals, read and re-write anatomical and physiological memories, etc. Thus, this is not a linguistic or philosophical project to paint human-style hopes and dreams onto rocks; it’s an empirical effort to see what new discoveries and capabilities are unlocked when we drop commitments to ancient philosophical categories (and departmental boundaries) and do experiments to see how widely the tools of behavioral science can usefully apply [6, 7].

Implications for Consciousness

1) For exactly the same ~4 reasons we attribute consciousness to each other as humans (overcome the problem of other minds), we should take very seriously the presence of (some kind/degree of) consciousness in other body organs, for example. The fact that you don’t feel your liver being conscious is irrelevant – you don’t feel me being conscious either. It may not have the linguistic chops of your left hemisphere and can’t speak up for itself, but we’re working on that (interfaces to communicate with diverse intelligences).

2) epiphenomenalism, causal efficacy: I think consciousness goes all the way down, and has causal efficacy. Briefly, my model (described in more detail in [8] and here) includes the following:

• a) where-ever you start (questions in physics, biology, or cognitive science), if you keep asking “why” long enough, you end up in the math department: with truths that do not belong to the physical world in the sense that they cannot be changed by tweaking the fundamental constants of physics (e.g., Feigenbaum’s constant, the difference in behavior between quaternions and octonions, etc.). Moreover, we don’t create these patterns, we discover them: start with “empty set + successor of empty set” and eventually you are given, with no choice about it, things like the specific value of e. All this is long-recognized by Platonist mathematicians. I realize there are alternative metaphysical views of math. This is the one I think makes the most sense, for whatever that’s worth.

• b) if we drop the universal assumption that the only patterns in that latent space are low-agency static (unchanging) patterns, we can hypothesize a spectrum in which some patterns are dynamic/active and have different degrees of competency along the Wiener-Rosenblueth hierarchy [9]. Our work on novel organisms, which don’t have an evolutionary history of selection for their large-scale features that can be used to predict their form and behavior, suggests that some patterns in this space are behavioral policies and homeostatic setpoints (goal states). In other words, this space contains levels of patterns amenable to the formal models of mathematics, and also others that are more complex and active policies - or, “kinds of minds” studied by behavioral scientists. I think all our departments are just different kinds of behavior science, working in different media.

• c) thus, the relationship between mind and brain/body is the same as between math and physics: fundamentally, a non-physical set of patterns ingresses functionally into the physical world, affecting it at multiple sales. This is my answer to the interactionism problem: we already had interactionism in the time of Pythagoras, in the relationship between mathematical truths and physics.

  • if you stick with classic billiard-ball models of causation, then almost by definition anything that isn’t physical is going to be seen as epiphenomenal. But I think philosophers (not to mention quantum physicists) have moved on from that anyway. Thus, my preferred definition of causation:

  • A causes B iff A is the most insightful explanation of why we got B and not C. I realize this is observer-relative and cognition-first (emphasizing insight and understanding); I think that’s a feature, not a bug. Plus,

  • A causes B iff tweaking A would have caused B to be different (Judea Pearl’s counterfactual notion of causation). I think the patterns from the Platonic space meet both criteria. There is more (how significant patterns can enter even deterministic algorithms, providing an alternative to the usual false dichotomy about causation – chance vs. necessity) but you can find it elsewhere.

• d) taken together, the above reasoning can be used to suppose: consciousness is what we call the view from the Platonic space into the physical world. We are not physical bodies impacted by patterns, we are the patterns, ingressing into physical bodies, whether biological, robotic, or hybrid (interfaces). And “life” isn’t the only medium that is amenable to it – we’ve found it in even very minimal computational systems [10].

• e) thus: we don’t make consciousness (either during reproduction or in robotics/AI) - we facilitate it into the physical world by building embodiments or interfaces for it. I’m well aware that dualist models like this are very out of favor in physics, biology, and computer science. Although, neuroscience has its share of relevant phenomena (e.g., [11]) and computer engineering is based on the idea of a non-physical thing called an “algorithm” which functionally controls how electrons flow [12] (if you think it’s epiphenomenal, you’ll never be hired at a software company).

3) binding problem: there are deep questions here I can’t answer yet, but I think the binding problem is not what we think it is. I think bodies are like an ecosystem, which contains huge numbers of low-tier patterns, some mid-tier patterns, and perhaps ~1 human-scale pattern (for example). Maybe there is something about physical interfaces that require a whole ecosystem of layers in order to host a high-level intelligence. But I don’t think the actual minds are bound from subunits – I think bodies contain many, highly diverse, minds at the same time.

4) research agenda: we are actively pursuing, among many other things:

• determine the kind and amount of “free lunch” provided to evolution (and to engineers) by this space. By quantifying the amount of effort put in on the physical end and the resulting form and function, we can determine (using minimal computational models and biological ones like biobots) what the space offers that our current models of computation, learning, and evolution leave out. Here is the first of a set of papers coming shortly on this.

• characterize the basic taxonomy of the kinds of patterns inhabiting that space

• study the mapping between properties of the physical interface and the patterns that ingress through it, to understand/predict/control what kind of mind can be facilitated when we build things, as engineers, bioengineers, parents, etc.

References cited

1. Levin, M. and D.B. Resnik, Mind Everywhere: A Framework for Conceptualizing Goal-Directedness in Biology and Other Domains—Part One. Biological Theory, 2026.

2. Levin, M. and D.B. Resnik, Mind Everywhere: A Framework for Conceptualizing Goal-Directedness in Biology and Other Domains—Part Two. Biological Theory, 2026.

3. Levin, M., Technological Approach to Mind Everywhere: An Experimentally-Grounded Framework for Understanding Diverse Bodies and Minds. Frontiers in Systems Neuroscience, 2022. 16: p. 768201.

4. Levin, M., The Computational Boundary of a “Self”: Developmental Bioelectricity Drives Multicellularity and Scale-Free Cognition. Frontiers in Psychology, 2019. 10(2688): p. 2688.

5. Levin, M., Bioelectric networks: the cognitive glue enabling evolutionary scaling from physiology to mind. Anim Cogn, 2023.

6. Davies, J. and M. Levin, Synthetic morphology with agential materials. Nature Reviews Bioengineering, 2023. 1: p. 46-59.

7. Levin, M., The Multiscale Wisdom of the Body: Collective Intelligence as a Tractable Interface for Next-Generation Biomedicine. Bioessays, 2024: p. e202400196.

8. Levin, M., Ingressing Minds: Causal Patterns Beyond Genetics and Environment in Natural, Synthetic, and Hybrid Embodiments. preprint, 2025.

9. Rosenblueth, A., N. Wiener, and J. Bigelow, Behavior, purpose, and teleology. Philosophy of Science, 1943. 10: p. 18-24.

10. Zhang, T., A. Goldstein, and M. Levin, Classical sorting algorithms as a model of morphogenesis: Self-sorting arrays reveal unexpected competencies in a minimal model of basal intelligence. Adaptive Behavior, 2024. 33(1): p. 25–54.

11. Kofman, K. and M. Levin, Robustness of the Mind-Body Interface: case studies of unconventional information flow in the multiscale living architecture. Mind and Brain, 2025. in press.

12. Ellis, G. and B. Drossel, How Downwards Causation Occurs in Digital Computers. Foundations of Physics, 2019. 49(11): p. 1253-1277.