Platonic Patterns Without Platonic Baggage: Pavel Chvykov’s Physics Reveals What Michael Levin’s Symposium Can’t Admit

How constraint-based thermodynamics accidentally became the latest exhibit in biological Platonism’s longest-running motte-and-bailey

“There is no such thing as philosophy-free science; there is only science whose philosophical baggage is taken on board without examination.”
— Daniel Dennett

Table of Contents

• Platonic Patterns Without Platonic Baggage: Pavel Chvykov’s Physics Reveals What Michael Levin’s Symposium Can’t Admit
  • What Chvykov Actually Does
  • The Three Interpretations (And Why Two of Them Don’t Matter)
  • Where Levin’s Framework Diverges (And Why It Matters)
  • Eight Points of Incompatibility
  • The Quotes That Reveal the Gap
  • What Aristotle Would Have Said
  • The Real Harm of the Platonic Label
  • A Necessary Disambiguation: Three Meanings That Cannot Coexist
  • The Pattern Across Symposium Speakers
  • What Would Falsify What?
  • Conclusion: What the Symposium Actually Shows
  • References

Pavel Chvykov’s talk at Michael Levin’s Platonic Space Symposium, titled “Why Physical Systems Find Platonic Patterns,” is a model of careful physics. It’s grounded in Hamiltonian dynamics, dissipation, noise, and attractors. It makes falsifiable predictions about steady-state distributions, relaxation times, and phase transitions under perturbation. It doesn’t hand-wave. It doesn’t smuggle teleology into equations. It doesn’t claim that abstract patterns do causal work simply by existing.

And yet here it is, presented as evidence for Platonic Space.

This is now the fourth time this has happened at Levin’s symposium. Brian Cheung’s convergence work explained representational alignment through constraint satisfaction in a single causal world. Gordana Dodig-Crnkovic’s “cognitive Platonism” treated mathematical structures as evolved cognitive tools, not transcendent entities. David Resnik’s population genetics models showed how evolutionary dynamics produce apparent design without designers. And now Chvykov’s rattling theory demonstrates how stable patterns emerge from thermodynamic selection with no access to abstract realms required.

Every one of these frameworks would have been rejected by Plato and embraced by Aristotle. Every one explains what Levin wants explained without requiring what Levin claims exists. And every one gets called “Platonic” anyway.

The pattern is becoming hard to ignore.

What Chvykov Actually Does

Chvykov starts where honest physics starts: with systems that have Hamiltonian structure but also dissipation, environmental coupling, and noise. In such systems, you get what he calls “mixed chaos”: regions of phase space where order and disorder coexist, separated by fractal boundaries.

The key insight is this: when you add dissipation and structured environmental driving, systems don’t explore phase space uniformly. They get pulled toward regions where local perturbations cause minimal rattling, minimal change in future trajectory. These low-rattling states aren’t goals. They aren’t targets. They’re just the configurations that survive because unstable configurations don’t.

“Low rattling” is measured precisely. If you perturb a system slightly and it returns quickly to its previous behavior, that’s low rattling. If perturbations cascade unpredictably, that’s high rattling. Systems coupled to structured environments naturally settle into low-rattling attractors because high-rattling configurations are thermodynamically expensive to maintain.

This gives you:

• Steady-state distributions conditioned by environmental structure
• Attractor islands in mixed-chaos phase space
• Predictive behavior that looks Bayesian without assuming priors
• Robustness to perturbation that emerges from stability selection

None of this requires pre-existing forms. None of it requires access to an independent realm. All of it follows from constraint dynamics and entropy gradients.

The robot swarm experiments make this concrete. Swarms exploring mazes under different sensory constraints converge on different steady-state distributions, not because they’re accessing different ideal forms, but because different environmental structures select for different low-rattling configurations. Change the environment, change the attractors. Change the coupling strength, change the relaxation time. Everything is physically accountable.

The Three Interpretations (And Why Two of Them Don’t Matter)

Chvykov considers three ways to interpret these “Platonic patterns”:

Transcendental: Patterns exist independently, prior to instantiation, and physical systems access them.

Emergent: Patterns arise from dynamics and have no existence apart from those dynamics.

Participatory: Patterns co-emerge through system-environment interaction and then influence subsequent dynamics, but only through physical coupling.

He doesn’t endorse transcendental Platonism. He doesn’t need it. His entire explanatory apparatus works within the participatory frame, where “influence” means ordinary physical causation, not metaphysical ingression.

This matters because the word “participatory” can sound like a compromise between emergence and transcendence. It isn’t. It’s just emergence with explicit acknowledgment that patterns, once formed, become part of the environmental structure that constrains future dynamics. The attractor you settle into affects what happens next, not because it’s an ideal pulling you, but because it’s a physical configuration with physical consequences.

There’s no dualism here. No separate realm doing causal work. Just constraint propagation through time.

Where Levin’s Framework Diverges (And Why It Matters)

Michael Levin’s biological Platonism makes much stronger claims. In talks, papers, and interviews, he describes Platonic space as:

• Existing independently of physical systems
• Containing goal states that organisms navigate toward
• Housing semantic content rather than mere pattern
• Accessible via bioelectric interfaces

He treats morphospace not as a convenient abstraction for successful developmental outcomes, but as something with ontological weight: a space that exists prior to the organisms that instantiate it, that exerts causal influence on biological processes.

This is not a minor terminological difference. It’s a categorical shift from constraint-based emergence to metaphysical realism.

Consider how Levin talks about planarian regeneration. When a planarian regenerates a head after decapitation, Levin often describes this as the system “knowing” the target morphology, “navigating” toward it in morphospace, “correcting errors” relative to an ideal form. The bioelectric patterns aren’t just causally relevant; they’re said to encode goals, represent targets, carry meaning.

Chvykov’s framework explains the same robustness without any of this. Regeneration works because developmental systems have low-rattling attractors: configurations that are stable under perturbation because they satisfy multiple constraints simultaneously. Damage perturbs the system, but if the attractor basin is wide enough, relaxation pulls it back. No target morphology “out there” doing the pulling. Just constraint satisfaction doing what constraint satisfaction does.

The kicker? Levin’s own lab published data showing persistent divergence in planarian regeneration. Durant et al. (2017) demonstrated that regenerated planarians don’t always converge to the same morphology; they can stabilize in alternative configurations depending on the specifics of the perturbation. That’s exactly what you’d expect from path-dependent attractor dynamics. It’s hard to square with convergence to ideal Platonic forms.

Eight Points of Incompatibility

Let’s be precise about where these frameworks actually contradict each other:

1. Causal mechanism vs. metaphysical ingression
Chvykov: Patterns emerge from thermodynamic selection under constraint.
Levin: Patterns exist independently and causally influence physical processes, but without specifying the physical mechanism for this influence.

2. Falsifiability vs. interpretive flexibility
Chvykov: Predict steady states, measure rattling, perturb and observe. Wrong predictions = failed model.
Levin: If morphologies converge, it’s Platonic attraction. If they diverge, it’s exploration of morphospace. What observation would falsify independent Platonic space?

3. Path dependence vs. ideal convergence
Chvykov: History matters. Perturbations change trajectories. Multiple stable states possible.
Levin: Target morphologies pull systems toward ideals. But the lab’s own regeneration data shows path-dependent divergence, not ideal convergence.

4. Emergence vs. pre-existence
Chvykov: Forms emerge from local constraint satisfaction and energy flow.
Levin: Forms pre-exist in morphospace, waiting to be accessed or instantiated.

5. Physicalism vs. dual-substrate ontology
Chvykov: One world. All causation physical. Patterns supervene on dynamics.
Levin: Explicitly rejects reduction to physics alone, suggesting “pattern space” operates under different principles.

6. Agency as gradient navigation vs. agency as transcendent input
Chvykov: Agency-like behavior emerges from statistical regularities in constraint landscapes.
Levin: Goal-directedness is primitive, not reducible to chemistry or thermodynamics.

7. Structural realism vs. excess ontology
Chvykov: Patterns are real as structural invariants, not as separate entities.
Levin: Patterns have independent ontological status in morphospace.

8. Precision vs. ambiguity
Chvykov: Equations, phase portraits, testable predictions.
Levin: “Platonic realm,” “mind spaces,” “pattern ingression”: terms that shift meaning across contexts.

Every testable claim Levin makes, Chvykov’s framework reproduces without the metaphysical overhead. The question is: what does the extra ontology buy you?

The Quotes That Reveal the Gap

Here’s Levin in his own words, from symposium discussions and published interviews:

“Platonic spaces are real in the sense that they are explored by evolution and development. Morphospace and goal states are just as real as physical attractors.”

This asserts ontological parity. But Chvykov’s attractors are physically real: measurable configurations in actual phase space. Levin’s morphospace is an abstraction over possible outcomes. Calling them “just as real” papers over a profound difference in ontological status.

“The body is not a bag of chemicals. It’s a cognitive agent exploring a space of possible forms.”

Chvykov shows how adaptive behavior emerges from noise-driven exploration of constraint landscapes. No intrinsic agency required. No cognitive homunculus needed. The “exploration” is thermodynamic sampling, not intentional navigation.

“The target morphology acts like a strange attractor in morphospace, pulling the system toward it.”

This sounds like dynamical systems language, but it’s backwards. In actual dynamical systems, attractors don’t pull. Systems settle into attractors because surrounding regions are unstable. The “pulling” is a metaphor for gradient descent, not evidence of teleological causation.

“This is why I insist on goal-directedness being primary—not reducible to chemistry or gene regulation.”

Chvykov’s low-rattling principle explains apparent goal-directedness without making it primitive. It’s an emergent property of systems that survive constraint pressure. Treating it as irreducible reintroduces what Dennett called “skyhooks”: explanations that don’t explain, just rename the mystery.

“The notion of ‘error’ makes no sense unless the system is navigating a pre-existing space of preferred states.”

This only makes sense if you already accept Platonic realism. In Chvykov’s framework, “error” is deviation from low-rattling configurations, measured by increased sensitivity to perturbation. No preferred states needed. Just stability gradients.

“Bioelectric circuits are not just causal chains—they instantiate intentional states over time.”

Chvykov models bioelectric-like phenomena as information-carrying patterns in coupled dynamical systems. They’re causally relevant without being intentional. Levin’s attribution of intentionality does no additional explanatory work; it just imports mentalistic vocabulary.

“Regeneration is not copying a prior state. It’s re-deriving a solution to a form-based problem.”

Chvykov would agree it’s not copying; it’s relaxation to a low-rattling attractor. But calling it “solution-finding” smuggles in agency. The system isn’t solving anything. It’s settling into whatever configuration is stable under the constraints.

“The same morphology can be arrived at by many routes because it’s specified in an abstract space, not by a specific molecular trajectory.”

Degeneracy (many paths to the same outcome) is exactly what you’d expect from wide attractor basins in dynamical systems. No abstract space required. Multiple trajectories converge because they’re all within the basin of the same physically real attractor.

“Information is not in the genes. It’s in the patterns of bioelectricity and their role in a shared semantic space.”

Chvykov: Information is in the statistical structure of system-environment coupling. No semantic space needed. Bioelectric patterns carry information in the Shannon sense; they reduce uncertainty about future states. That’s not semantics. That’s dynamics.

“We need a science of the possible forms themselves, not just the mechanisms that produce them.”

Here’s the core disagreement laid bare. Chvykov’s project is: understand mechanisms and the forms follow as necessary consequences. Levin’s project is: study forms independently of mechanisms. One of these is science. The other is morphology dressed as metaphysics.

What Aristotle Would Have Said

Aristotle spent considerable effort distinguishing his position from Plato’s, and the core objection was precisely this: Plato separated form from matter, treating forms as independently existing entities that matter “participates in.” Aristotle rejected this as incoherent. How does participation work? What mechanism allows matter to “access” form? How do changeless forms causally influence changing matter?

His alternative was hylomorphism: form and matter are inseparable aspects of the same substance. Form is not “in” an object as a separate component; it’s the organization of that object, the way matter is structured and behaves. Universals exist, but only as instantiated in particulars, not as separate entities in a transcendent realm.

Contemporary Aristotle scholarship, particularly from David Sedley, Charlotte Witt, and Mariska Leunissen, emphasizes that Aristotle’s teleology was naturalistic: purposes emerge from the intrinsic natures of things, not from external blueprints. A seed grows into an oak because that’s what seeds of that type do under normal conditions, not because there’s an ideal oak-form pulling it.

Chvykov’s framework fits this Aristotelian picture perfectly. Patterns are immanent to the physical processes that generate them. Form is constraint, the way systems must behave given their structure and environment. There’s no transcendence, no access problem, no mysterious participation.

Levin’s framework does not fit this picture. It reintroduces exactly the problems Aristotle identified: independent forms, unclear causal mechanisms, metaphysical rather than physical explanation.

If Aristotle could review both talks, he wouldn’t call Chvykov’s work Platonic. He’d recognize it as a sophisticated formalization of his own commitment to immanent causation. And he’d object to Levin’s exactly as he objected to Plato: it mistakes abstraction for ontology and stops explaining precisely where explanation is needed most.

The Real Harm of the Platonic Label

Calling Chvykov’s framework “Platonic” isn’t merely imprecise. It actively enables several forms of conceptual damage:

It lowers falsification pressure. Once something is framed as Platonic metaphysics rather than physical mechanism, critics can dismiss it philosophically rather than test it empirically. Conversely, supporters can defend it on metaphysical grounds rather than predictive success. Either way, the pressure to generate and test predictions decreases.

It obscures causal mechanisms. When patterns are said to “exist Platonically,” the actual thermodynamic, informational, and dynamical work they do becomes invisible. Dissipation, coupling strength, relaxation time, energy budgets (all the things that make predictions possible) fade into the background. What remains is “patterns explain themselves,” which explains nothing.

It shelters unfalsifiable claims. Levin’s biological Platonism faces genuine difficulties: path-dependent divergence in regeneration, absence of specified mechanisms for form-matter interaction, vulnerability to Benacerraf’s dilemma about epistemic access. But if Chvykov’s rigorous, falsifiable physics also gets called “Platonic,” it provides protective cover. The audience hears “multiple scientists support Platonism” without noticing that they mean completely different things.

It enables citation laundering. External actors (including intelligent design advocates and other anti-naturalistic groups) can cite symposium proceedings as evidence that “leading scientists accept non-physical causation” or “Platonic forms are making a comeback in biology.” This isn’t paranoia; it’s pattern recognition. The Discovery Institute has already begun citing morphic resonance research this way. Why wouldn’t they do the same here?

It misrepresents intellectual history. The real story is not “scientists are rediscovering Plato.” It’s “scientists keep developing Aristotelian models of immanent causation while calling them Platonic for unclear reasons.” That’s a much less exciting narrative, but it’s accurate.

It slows scientific progress. The most exciting questions raised by Chvykov’s work are mechanistic: What constraint structures select for which attractors? How do multi-scale couplings create hierarchical low-rattling states? Can we engineer systems with designer attractor landscapes? These are tractable questions with engineering applications. Framing them as “finding Platonic patterns” shifts focus from mechanism to metaphysics, from testability to interpretation.

A Necessary Disambiguation: Three Meanings That Cannot Coexist

The core problem is that “Platonic” is being used for three completely different ideas, and collapsing them creates the appearance of agreement where there is fundamental incompatibility.

Platonic-as-abstraction: We can describe patterns at a level that ignores substrate specifics. Mathematical expressions, dynamical equations, statistical distributions: these are substrate-independent in description. A pendulum is a pendulum whether it’s a mass on a string or a nanoscale cantilever. This is uncontroversial and implies nothing about ontology.

Platonic-as-convergence: Different systems optimizing under similar constraints converge on similar solutions because the space of viable solutions is narrow. This is Chvykov’s sense. It’s what Cheung demonstrated for neural networks, what Dodig-Crnkovic showed for mathematical structures, what Resnik’s population genetics implies. The convergence is real, robust, and explained by constraint satisfaction in one world.

Platonic-as-transcendence: Abstract patterns exist independently of physical instantiation, in a separate ontological realm, and somehow exert causal influence on physical processes. This is Levin’s sense. It requires solving the interaction problem, the ingression problem, and the epistemic access problem that have troubled Platonism for 2,400 years.

Chvykov operates in senses one and two. Never three.
Levin requires sense three.
Calling both “Platonic” allows audiences to slide between them without noticing the categorical difference.

Once you make this explicit, several things become unavoidable:

• Abstraction does not imply transcendence
• Convergence under constraint does not imply access to a realm
• Chvykov’s data cannot support Levin’s metaphysics without adding premises Chvykov doesn’t state

The motte is constraint-based emergence. The bailey is transcendent Platonism. Shared vocabulary lets you retreat to the motte when challenged while continuing to occupy the bailey when unchallenged.

That’s not synthesis. That’s equivocation.

The Pattern Across Symposium Speakers

This is not an isolated incident. It’s systematic.

Brian Cheung showed that vision and language models converge because they’re both modeling the same causal world. He never invoked access to an independent realm. His entire framework was constraint satisfaction producing representational alignment. Yet his work was presented as support for Platonic Space.

Gordana Dodig-Crnkovic explicitly argued for “cognitive Platonism”: the view that mathematical structures are evolved cognitive tools that work because they track physical constraints. Not transcendent entities. Not independent realms. Cognitively constructed abstractions that survive because they’re useful. She distinguished this carefully from transcendent Platonism. The distinction was promptly ignored.

David Resnik demonstrated how population genetics produces apparent design (robust, functional patterns) without designers. Stabilizing selection, frequency-dependent selection, and epistasis create attractors in genotype space. These look like target states but emerge from dynamics. His models are pure evolutionary mechanism. No goals required.

Pavel Chvykov now adds thermodynamic constraint dynamics to this list. Low-rattling attractors, dissipation-driven selection, constraint satisfaction without teleology.

Every one of these frameworks would be classified by Aristotle as immanent causation, not Platonism. Every one explains the phenomena Levin wants explained without requiring what Levin claims exists. Every one is presented at a symposium explicitly about Platonic Space and described as providing support for the symposium’s central thesis.

At what point does this pattern become impossible to ignore?

The symposium is functioning as a motte-and-bailey generator. Invite rigorous naturalistic work. Call it “Platonic.” Use the association to lend credibility to much stronger metaphysical claims. When challenged, retreat to “we just mean patterns are robust and universal.” When unchallenged, advance to “therefore non-physical facts causally influence biology.”

This is not how science should work.

What Would Falsify What?

Here’s the test that matters.

Chvykov’s framework is falsifiable:

• Predict steady-state distributions given environmental structure
• Predict relaxation times given coupling strength
• Predict attractor shifts given perturbations
• If predictions fail repeatedly, abandon the model

What would falsify Levin’s independent Platonic space?

If morphologies fail to converge, that’s “exploration of morphospace.”
If they converge, that’s “attraction to ideal forms.”
If bioelectric patterns correlate with outcomes, that’s “semantic encoding.”
If they don’t correlate, that’s “wrong level of analysis.”
If regeneration is robust, that’s “error correction toward targets.”
If it’s path-dependent, that’s “navigating attractor landscapes.”

Every outcome is interpretable within the framework. That’s not a virtue. That’s a vice. A theory that explains everything explains nothing.

Levin has been asked repeatedly, in multiple forums: what observation would show that Platonic space does not exist independently? The question has not been answered. When pressed, the response is typically to reframe the question as misunderstanding what “Platonic” means, or to offer examples that are actually consistent with naturalistic emergence.

This is diagnostic. If no answer is forthcoming, the claim is not scientific. It might be metaphysics. It might be poetry. It might be a useful heuristic. But it’s not a falsifiable hypothesis.

And if it’s not falsifiable, it shouldn’t be presented alongside falsifiable research as if they’re the same kind of enterprise.

Conclusion: What the Symposium Actually Shows

The real lesson of the Platonic Space Symposium is not that Platonic forms are being rediscovered in biology and AI. It’s that constraint-based emergence keeps producing everything we want without requiring anything beyond physics.

Chvykov’s thermodynamic attractor dynamics.
Cheung’s kernel convergence under shared causal structure.
Dodig-Crnkovic’s cognitive structures evolved to track constraints.
Resnik’s population genetics producing apparent design through selection.

All naturalistic. All falsifiable. All mechanistic. None requiring transcendent ontology.

The best work at the symposium consistently undermines the symposium’s central thesis. The pattern is so clear it’s starting to look like a controlled experiment: invite rigorous constraint-based science, label it “Platonic,” and see if anyone notices the substitution.

Some of us notice.

The uncomfortable conclusion: when you strip away the metaphysical vocabulary and evaluate these frameworks by their actual explanatory structure, they’re Aristotelian, not Platonic. They explain form through constraint, not through access. They ground patterns in dynamics, not in transcendence. They predict, fail, and update, which is what honest theories do.

Levin’s biological Platonism doesn’t do those things. It absorbs every outcome, accommodates every result, and remains invulnerable to evidence. That’s not because it’s profound. It’s because it’s not scientific.

The question is no longer whether Platonic space exists independently of the physical world. The question is why, given frameworks like Chvykov’s, anyone would still need it.

And if the answer is “because it feels explanatory,” Dennett already told us what that feeling is worth.

Not much.

If you can’t answer what would falsify your core claim, you’re not doing science. You’re doing apologetics with equations.

The symposium’s best speakers keep building better cranes. The question is when the host will stop calling them skyhooks.

References

Chvykov, P., & England, J. L. (2018). Least-rattling feedback from strong time-scale separation. Physical Review E, 97(3), 032115. https://doi.org/10.1103/PhysRevE.97.032115
https://journals.aps.org/pre/abstract/10.1103/PhysRevE.97.032115

Chvykov, P., Berrueta, T. A., Vardhan, A., Savoie, W., Samland, A., Murphey, T. D., Wiesenfeld, K., Goldman, D. I., & England, J. L. (2021). Low rattling: A predictive principle for self-organization in active collectives. Science, 371(6524), 90-95. https://doi.org/10.1126/science.abc6182
https://www.science.org/doi/full/10.1126/science.abc6182

Chvykov, P., & England, J. L. (2025). Emergent order from mixed chaos at low temperature. Scientific Reports, 15, 38490. https://doi.org/10.1038/s41598-025-22877-4
https://www.nature.com/articles/s41598-025-22877-4

Durant, F., Morokuma, J., Fields, C., Williams, K., Adams, D. S., & Levin, M. (2017). Long-term, stochastic editing of regenerative anatomy via targeting endogenous bioelectric gradients. Biophysical Journal, 112(10), 2231-2243. https://doi.org/10.1016/j.bpj.2017.04.011
https://www.cell.com/biophysj/fulltext/S0006-3495(17)30411-9

Durant, F., Bischof, J., Fields, C., Morokuma, J., LaPalme, J., Hoi, A., & Levin, M. (2019). The role of early bioelectric signals in the regeneration of planarian anterior/posterior polarity. Biophysical Journal, 116(5), 948-961. https://doi.org/10.1016/j.bpj.2019.01.029
https://www.cell.com/biophysj/fulltext/S0006-3495(19)30065-7

Benacerraf, P. (1973). Mathematical truth. The Journal of Philosophy, 70(19), 661-679. https://doi.org/10.2307/2025075
https://www.jstor.org/stable/2025075

Dennett, D. C. (1995). Darwin’s Dangerous Idea: Evolution and the Meanings of Life. New York: Simon & Schuster.
https://books.google.com/books/about/Darwin_s_Dangerous_Idea.html?id=FvRqtnpVotwChttps://www.sciencedirect.com/science/article/abs/pii/0167278990900876

Leunissen, M. (2010). Explanation and Teleology in Aristotle’s Science of Nature. Cambridge: Cambridge University Press. ISBN: 9780521197748
https://www.cambridge.org/core/books/explanation-and-teleology-in-aristotles-science-of-nature/311F862489619C0EEB985B0D24275F20

Sedley, D. (2007). Creationism and Its Critics in Antiquity. Berkeley: University of California Press.
https://www.ucpress.edu/book/9780520931282/creationism-and-its-critics-in-antiquity

Witt, C. (1989). Substance and Essence in Aristotle: An Interpretation of Metaphysics VII-IX. Ithaca: Cornell University Press.
https://www.cornellpress.cornell.edu/book/9780801495786/substance-and-essence-in-aristotle/

Huh, M., Cheung, B., Wang, T., & Isola, P. (2024). Position: The Platonic representation hypothesis. Proceedings of the 41st International Conference on Machine Learning, 235, 20617-20642. https://doi.org/10.48550/arXiv.2405.07987
https://proceedings.mlr.press/v235/huh24a.html
https://arxiv.org/abs/2405.07987

Dodig-Crnkovic, G. (2016). Information, computation, cognition: Agency-based hierarchies of levels. In V. C. Müller (Ed.), Fundamental Issues of Artificial Intelligence, Synthese Library 377 (pp. 139-159). Cham: Springer. https://doi.org/10.1007/978-3-319-26485-1_10
https://link.springer.com/chapter/10.1007/978-3-319-26485-1_10

Dodig-Crnkovic, G. (2017). Nature as a network of morphological infocomputational processes for cognitive agents. The European Physical Journal Special Topics, 226, 181-195. https://doi.org/10.1140/epjst/e2016-60362-9
https://link.springer.com/article/10.1140/epjst/e2016-60362-9

Dodig-Crnkovic, G. (2022). Cognition as morphological/morphogenetic embodied computation in vivo. Entropy, 24(11), 1576. https://doi.org/10.3390/e24111576
https://www.mdpi.com/1099-4300/24/11/1576

Dodig-Crnkovic, G., & Milkowski, M. (2023). Discussion on the relationship between computation, information, cognition, and their embodiment. Entropy, 25(2), 310. https://doi.org/10.3390/e25020310
https://www.mdpi.com/1099-4300/25/2/310

Sedley, D. (2007). Creationism and Its Critics in Antiquity. Berkeley: University of California Press. ISBN: 9780520931282
https://www.ucpress.edu/book/9780520931282/creationism-and-its-critics-in-antiquity

Witt, C. (1989). Substance and Essence in Aristotle: An Interpretation of Metaphysics VII-IX. Ithaca: Cornell University Press. ISBN: 9780801495786
https://www.cornellpress.cornell.edu/book/9780801495786/substance-and-essence-in-aristotle/

Lange, M. (2016). Because Without Cause: Non-Causal Explanations in Science and Mathematics. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780190269487.001.0001
https://global.oup.com/academic/product/because-without-cause-9780190269487

Woodward, J. (2003). Making Things Happen: A Theory of Causal Explanation. Oxford: Oxford University Press. ISBN: 9780195189537
https://global.oup.com/academic/product/making-things-happen-9780195189537

Levin, M., Pezzulo, G., & Finkelstein, J. M. (2017). Endogenous bioelectric signaling networks: Exploiting voltage gradients for control of growth and form. Annual Review of Biomedical Engineering, 19, 353-387. https://doi.org/10.1146/annurev-bioeng-071114-040647
https://www.annualreviews.org/doi/10.1146/annurev-bioeng-071114-040647

Pezzulo, G., & Levin, M. (2016). Top-down models in biology: Explanation and control of complex living systems above the molecular level. Journal of The Royal Society Interface, 13(124), 20160555. https://doi.org/10.1098/rsif.2016.0555
https://royalsocietypublishing.org/doi/10.1098/rsif.2016.0555