The Chladni Plate Alternative to Platonic Necessity: Douglas Brash, Bioelectric Morphogenesis, and Constraint-Based Explanation

When Rigorous Science Meets Comfortable Metaphysics

Douglas Brash’s presentation “Abstract Forms & Tangible Biology”, shared through the 2025 Platonic Space symposium, offers a concrete challenge to Michael Levin’s biological Platonism. It draws on vibration, boundary conditions, local construction rules, and a broader constraint-based research programme to show how elaborate forms can arise without an externally consulted template. Brash has not supplied a complete quantitative model of bioelectric morphogenesis. The sharper argument is that the success of pattern formation does not itself establish Levin’s additional, causally effective Platonic source.

My January 10 wording was compressed enough to invite a larger claim than I intended. I called Brash’s account “complete” and said it offered predictions Levin’s framework “cannot match.” Yet the very next sentences limited my target exclusively to the metaphysical interpretation added to Levin’s empirical results. “Complete” meant adequate to challenge the claim that those results require nonphysical Platonic ingress, not a final biological model. “Cannot match” was about the Platonic addition’s lack of a distinct prediction beyond the mathematics and biology, not an assertion that Levin’s actual computational models predict nothing. My shorthand deserved repair; the causal question it was reaching for still deserves an answer.

This distinction matters. Levin’s laboratory work on bioelectric patterning is genuine empirical research. His group reports striking effects of physiological interventions on planarian regeneration. Credit depends on the particular result: repetition within a collaborating team, independent reproduction of a phenotype, and independent validation of a predictive model carry different evidential weight. The inference from those results to minds or patterns that exist independently of physical realization and act through biological interfaces requires a separate causal and evidential bridge. Success at changing anatomy does not, merely by itself, identify the ultimate ontological source of anatomical possibilities.

What would distinguish an additional Platonic influence from physical constraint-based mechanisms that predict the same outcomes? Without a specified source-to-interface mapping, an intervention, and a result on which the rival accounts diverge, an extra ontology can remain compatible with every discovery while acquiring credit for none of the associated causal work. That is a serious evidential problem, though it does not make every ontological proposal meaningless or demand that one laboratory result deductively falsify an entire research programme. Brash’s examples expose the missing bridge; they do not finish the rival research programme.

The critique targets the additional metaphysical interpretation, not Levin’s laboratory achievements and not mathematical Platonism in general. Constraint-based explanations can establish local sufficiency or weaken a claim of Platonic necessity without establishing a final ontology. The relevant question is which account takes distinctive evidential risks and explains the controlled phenomena.

What Platonic Morphogenesis Actually Claims

Precision matters when criticizing a philosopher-scientist’s strongest position. Levin’s proposal is more than the ordinary practice of describing living systems mathematically: his public comments and subsequent published formulation discuss patterns or minds as additional to physical interfaces. But he also treats the interface and its physical organization as important, and allows a plurality of possible forms. The criticism must address those commitments as he states them, not an invented rule that every planarian must access exactly one canonical body.

“A complex pattern that is an additional ingredient to the physical interface, in the same way that the specific value of e… exist independently.”

— Michael Levin, documented exchange · primary source

“Minds are forms in that space, and they access each other, in that space (laterally) but also project into the ‘physical world’ through interfaces.”

— Michael Levin, documented exchange · primary source

“we facilitate them into the physical world … it’s substrate independent in the sense that the mind is not coming from the physical substrate.”

— Michael Levin, documented exchange · primary source

These are ontological proposals, not merely statements that equations describe organisms. Their distinctive content is the proposed additional source and its relationship to physical realization. The fact that a biological interface matters does not dissolve the source claim, but the interface must also do the work of making that claim empirically identifiable. Otherwise the same bioelectric result supports both the restrained physical account and the added Platonic interpretation without distinguishing them.

The charitable reading recognizes Levin’s genuine problem: regenerative systems preserve, revise, and sometimes recover large-scale anatomical organization despite changes to individual cells. These observations motivate research into physiological memory, control, and morphological possibility. They do not settle whether that possibility is an independently existing causal realm. The inference from demonstrated biological competence to a nonphysical source is precisely what remains to be earned.

The Thermodynamic Challenge: What It Establishes, and What It Does Not

Physical information has physical implementations and operation-specific thermodynamic constraints. Landauer’s principle identifies a lower bound on dissipation for specified logically irreversible operations, such as resetting an initially unbiased bit in an ideal thermal setting: k_BT ln 2. Experiments by Bérut et al. (2012), Hong et al. (2016), and Yan et al. (2018) test that principle in different implementations. Bennett (2003) shows why logical irreversibility, rather than the mere existence of information, is the relevant distinction. The broader point is that physical computation, storage, control, and erasure have describable physical conditions, not that every distinction carries an identical heat price.

The causal debt is at the interface. If an independently existing Platonic pattern changes what a developing organism does, where does it enter the physical causal chain? Which variables, couplings, boundary conditions, or interventions distinguish its influence from bioelectric states, cell histories, and constraints already in the model? “Where is the energetic signature?” is a demand for an accountable physical causal pathway and its observable consequences, not a presupposition of a separate, fixed heat pulse for every imported pattern. Levin has proposed that “a better science of Platonic forms + their interfaces will force a re-do of Landauer’s Principle (or at least, its limitations)” in a public reply. That proposal remains a hypothesis about the interface until it specifies which physical accounting assumptions fail, how they fail, and what measurement would show it.

It is crucial to keep three propositions separate: mathematical structures can describe a system; physically realized structures can constrain its dynamics; and a mind-independent Platonic domain can supply additional causal organization. The first two claims do not entail the third. If the third is advanced, a retreat to the first cannot by itself vindicate it. This is an evidential-credit asymmetry, not a verdict about Levin’s motives or a claim that every use of a heuristic is a deliberate motte-and-bailey.

Brash’s Framework: Patterns from Dynamics Under Constraint

Brash asks how stable organization can emerge through physical dynamics operating under constraints. His Chladni-plate demonstration is an instructive example, while his 2020 BioEssays article sketches broader tools for studying biological causation, feedback, and attractor-like behavior. Neither amounts to a completed replacement for developmental biology. Their force here is as counterexamples to the assumption that patterned outcomes must be retrieved from an outside template. The same comparison matters when evaluating Levin and Resnik’s Mind Everywhere papers: experimental utility alone does not establish an independently supplying Platonic realm.

“Where did this pattern come from? There’s clearly no map here… Newton’s laws will control the salt grains at the micro level but they don’t cause the pattern at the macro level. The pattern comes from constraints at the macro level.”

— Douglas Brash, “Abstract Forms & Tangible Biology” (36:50-37:56) · primary source

“You’ve got nondirectional motion going on at the micro level. The directionality comes at the macro level and it comes from these constraints.”

— Douglas Brash (38:21-38:33) · primary source

The insight is about levels of description and causal organization. Micro-level laws, material properties, boundaries, resonance, and driving frequency jointly produce the pattern. A description at the pattern level can still capture genuine constraints and intervention targets; it need not be causally empty. What the demonstration undercuts is the inference from successful mathematical description to an additional external cause. Its explanatory burden remains with the mechanics and constraints we can measure and manipulate.

Stigmergy as a Counterexample to the Necessity of Platonic Access

Complex structures can emerge through local agents responding to what earlier agents have built. Brash uses wasp nests and stigmergy to illustrate this: the partially constructed environment helps regulate each subsequent act. Formal models of paper-wasp nest construction show how an algorithm interacting with existing geometry can generate different nest shapes (Karsai & Pénzes, 1998).

“Wasps do not have a map of a wasp nest in their head. The brains aren’t big enough. What they do is they have an algorithm. They land on a nest, a partly built nest, and if they see pattern P, they spit here. If they see pattern Q, they spit there.”

— Douglas Brash (44:43-45:00) · primary source

“This is a third kind of law that’s neither the pattern nor a criterion of constancy. It’s sort of an algorithm of micromotion plus external constraints, a lot like the Chladni plate.”

— Douglas Brash (45:35-45:48) · primary source

The important conclusion is limited but potent: intricate construction does not logically require each participant to represent a final blueprint, much less access a nonphysical form. Local feedback and inherited behavioral capacities can jointly explain relevant features. That does not prove that no wasp has internal representations; nor does it make termite construction, slime-mold networks, bone remodeling, and vascular growth one identical mechanism. Those systems require their own causal studies.

The Palanquin-Prince Inversion

Brash’s prince-and-palanquin analogy illustrates a possible inversion in explanation. From above, the prince appears to move through space as a unified actor. From below, the motion is realized by the coordinated bearers, the frame, and their constraints. The analogy cautions against treating a higher-level pattern as an independently pushing force. It does not imply that higher-level organization is fictional or that a real prince could never direct the bearers; those would be separate causal questions.

“The business end is going on at the micro level and you get these patterns and abstractions up at the level of the princes being carried around.”

— Douglas Brash (41:50-42:04) · primary source

“It looks to me like there’s a possibility that causality has this structure rather than the pattern one arrow pattern two causality, particularly in hierarchical systems like you see in biology.”

— Douglas Brash (43:32-43:43) · primary source

In a constraint-based account, large-scale organization can emerge from interacting physical processes and then make further differences through stable boundaries, feedback, and control. Its real causal roles are realized within those processes. The proposal challenges a second causal level supplied by a separate Platonic domain, not the usefulness or causal relevance of higher-level biological explanations.

Brash’s “Platonish” Distinction and Its Limits

Brash distinguishes mathematical relations, biological models, and traditional Platonic talk, and sometimes calls useful representations “Platonish.” His language is exploratory. It would overstate his position to treat that term as a definitive rejection of every Platonic ontology: parts of the same presentation entertain speculative connections with abstract forms.

“Plato’s forms, as I understand, were not really mathematical objects. So, these timeless forms or abstract patterns are just there… There are no laws connecting them. Whereas mathematical objects are connected by mathematical operations.”

— Douglas Brash (9:43) · primary source

“This isn’t really Plato… I’m going to call them Platonish representations. And that sounds a little bit pejorative, but in fact, I think it’s probably ultimately going to turn out to be more useful than the pure Plato’s representation.”

— Douglas Brash (11:47-12:15) · primary source

“Platonish” is not a secret synonym for “not Platonic at all.” The philosophically decisive distinction is instead between mathematical structures as tools or descriptions and a claimed external source of causal influence. Brash’s account helps us work with the former; whether an additional source exists is an independent question requiring an independent argument.

His observation about anthropocentric categories gains support from the classic study “What the Frog’s Eye Tells the Frog’s Brain” (Lettvin et al., 1959): a frog’s visual system responds selectively to ecologically relevant features. This cautions against assuming that human category systems are privileged descriptions of nature. It does not prove that no observer-independent mathematical structure exists.

“These human forms of Plato’s are not necessarily privileged basis vectors of truth and they’re actually anthropocentric.”

— Douglas Brash (15:19-15:26) · primary source

The Empirical Crux: What Durant et al. 2017 Actually Shows

An important, replication-sensitive set of results comes from Levin’s own laboratory. In Durant et al. (2017), transient disruption of gap-junctional signaling with 1-octanol changed the regenerative outcome of planarian fragments. The reported protocols include treatment during the first three days after amputation; it would be misleading to present all relevant interventions as one 48-hour exposure. The authors examined later recuts after the treatment had ended.

Independent evidence has to be sorted by which result it corroborates. Marsh and Beams (1952) showed that applied electric fields could change planarian regenerative polarity, decades before Levin’s programme. Petersen and Reddien (2008) independently demonstrated posterior-head regeneration after altering β-catenin signaling. These findings support the broader plasticity of regenerative polarity, but neither reproduces Durant’s particular transient octanol protocol, cryptic-state distributions, or voltage-based reset account.

An earlier Oviedo et al. (2010) experiment reported multiheaded regeneration and recutting after octanol, but Levin and collaborators were among its authors. It supplies within-programme continuity, not independent replication.

There is also a more direct, partial outside check. In an OSF preprint (2025), unaffiliated researcher Patrick Seebold reports producing two-headed Dugesia japonica after gap-junction inhibition. This is independent in authorship and experimental setting, but the accessible record is a preprint, not peer-reviewed, multi-laboratory confirmation. It does not establish external replication of Durant and colleagues’ specific cryptic 75:25 transitions, bioelectric-state measurements, or reset tests. Those further claims remain original-team findings pending comparable unaffiliated tests.

Consider the stronger auxiliary hypothesis that the pre-existing normal one-headed form is the privileged target to which an unperturbed worm must return. Repeated regeneration of two heads would then be a serious challenge, requiring that hypothesis to be revised. But a general claim that many forms are available in a structured possibility space does not entail compulsory convergence on one canonical anatomy. Levin explicitly allows two-headed planaria as a possible form. The observed persistence weakens a rigid ideal-template interpretation; it does not deductively refute all Platonism.

The deeper problem begins where that flexibility is treated as confirmation. Levin asks, “how do you know the 2-headed form is not a manifestation of a pre-existing pattern?” Fair question: the observed outcome is compatible with that possibility. But compatibility alone cannot establish that the extra pattern caused the outcome. If both one-headed and two-headed morphologies can be retrospectively assigned to an unspecified Platonic repertoire, which independently specified property of that repertoire selects the result before the intervention? What does Platonic access predict that a physiological-memory model does not?

In the authors’ experiments, manipulating bioelectric connectivity changed regenerative outcomes, including reported persistent and cryptic anatomical-memory effects. A physical constraint-and-memory model can represent such results using altered state variables and attractors, but it must still predict the observed probabilities, persistence, and reset conditions before claiming a complete account. A metaphysical model must meet the same standard and then show where its additional source improves the result.

Scholarly Convergence: Brash Is Not Alone

Brash’s argument is part of a broader tradition of studying how organized behavior follows from interactions, feedback, and constraints. This matters because the alternative is a positive scientific research direction, not just a refusal of Platonism. But convergent vocabulary across fields is not proof that their models are mathematically identical, or that one model already explains every biological target.

One useful comparison is Montévil and Mossio’s (2015) account of biological organization as closure of mutually maintained constraints. It offers a way to ask which processes generate, depend on, and maintain a living system’s constraints. Friston’s free-energy principle concerns variational models of perception and action under particular assumptions; it cannot simply be equated with literal thermodynamic free energy, organizational closure, or Brash’s proposal. These are potentially complementary but distinct research programmes.

Daniel Dennett’s crane-versus-skyhook distinction remains a useful warning. A crane supplies an intelligible route from processes to a pattern; a skyhook invokes a source of apparent design whose explanatory work remains unspecified. The distinction is not a proof that Levin’s additional source is false. It demands that a source said to influence regeneration do more than rename the outcome. Dennett’s “Real Patterns” also offers a reason to take higher-level descriptions seriously when they improve prediction.

Brash’s peer-reviewed discussion of causal constraints and concrete wasp-nest construction models supply better support for this point than a borrowed aphorism. One need not establish that all possible forms are physically reducible to show that particular patterns can be explained by measurable local interactions, collective organization, and structural feedback.

Lakatos’s lesson is subtler than attaching a “degenerating” label to a research programme after one case study. Programmes can adjust auxiliary hypotheses legitimately; their historical standing depends on whether the revisions subsequently yield independently corroborated, novel results rather than merely accommodating known outcomes. Levin’s biological experiments are productive. The unsettled comparison concerns what additional, risky predictions the Platonic-source claim contributes beyond those experiments. Brash’s conceptual alternative must face the same prospective test.

Prospective Prediction: Where the Credit Actually Goes

Two comparisons need separate scoreboards. Brash’s 2020 account offers candidate mechanisms involving feedback, distributed constraints and organism-level states. His symposium examples illustrate pattern formation without an external map, but they do not constitute a prospective, head-to-head biological test against Levin’s models. That kind of numerical predictive superiority remains to be demonstrated.

The comparison I was making concerns the incremental claim. Let D be the supported physiological, mathematical and computational explanation at a stated target, and P the further claim that independently existing Platonic forms causally ingress. If D and D + P make the same predictions on that target, the experiments give P no additional empirical credit. Where D works, it is the more economical evidential explanation; where it remains incomplete, the unanswered mechanism is still an open research question. Was a second causal realm actually needed to generate the result, or did it merely arrive in time to take the bow? That is a comparison of explanatory warrant, not a victory lap over Levin’s experimental science.

A model developed by Levin and collaborators has produced a reported prediction followed by experimental testing within the same research project. In Lobo, Morokuma, and Levin (2016), a reverse-engineered regulatory-network model predicted the effects of previously unperformed hnf4 knockdown combinations, including rescue of an abnormal phenotype following hh knockdown. The authors then tested those predictions in living planarians and reported the predicted outcomes. That sequence merits provisional credit as a within-project prospective test, assuming the prediction preceded the intervention as reported; it is not equivalent to an unaffiliated replication or an externally preregistered test. Its explanatory variables belong to a computational and molecular model, however; the experiment does not independently identify a nonphysical source of the predicted form.

For the 2016 hnf4 prediction specifically, model construction and experimental testing were reported within a coauthored collaboration by Lobo, Morokuma, and Levin. An author working at another institution does not turn that shared project into an independent replication. I have not located an unaffiliated laboratory test of the same predicted hnf4/hh rescue. The result therefore has prospective within-project value, with external reproducibility, model transfer, and specificity still open.

Levin has also proposed a distinctive programme to quantify apparent “free compute” and map latent pattern spaces. Those are testable ambitions worth following, not established demonstrations that a Platonic source performs computation beyond the physical account. At present, the comparison therefore has three verdicts: Brash’s constraint framework provides important explanatory strategies but has not won the claimed prospective contest; Levin’s collaborating team has reported a prospective test whose independent predictive reproduction remains unestablished here; and the additional Platonic ontology has not yet earned an independently discriminating predictive success. If two completed models make equally accurate predictions under matched tests, their differing ontological commitments become relevant to explanatory economy. Until then, parsimony is a reason to withhold extra credit, not a substitute for a head-to-head experiment.

What the Constraint Framework Actually Explains

Constraint-based emergence already gives us some real explanatory resources; it does not yet explain all of morphogenesis. The contrast below states the support earned by each example and the added burden a Platonic causal account would have to meet.

Path dependence. The planarian results demonstrate that a transient intervention can alter subsequent regenerative outcomes, including hidden persistent states. Constraint and memory models are natural candidates for explaining the transition and reset dynamics. They have not yet been shown here to predict every outcome. An external-form hypothesis gains support only if it forecasts more than the physically modeled history and interface.

Multiple stable morphologies. Biological systems can reach different viable forms depending on initial conditions, material organization, and intervention. This is compatible with landscape or attractor descriptions. It does not by itself rule out a Platonic repertoire of many forms; equally, listing every observed form in that repertoire does not identify it as a cause. The unresolved test concerns selection among possibilities.

Thermodynamic accounting. Vibrations, biochemical activity, and material construction are physically realized and constrained. Landauer’s bound applies to specified irreversible information operations. The broader causal question concerns the independently identifiable effects of the proposed Platonic coupling: which measured variable or intervention outcome differs from the physical-interface account?

Emergence without a global map. Stigmergic models demonstrate how local feedback and construction history can yield an organized nest without a complete blueprint located in every builder. They neither establish the absence of all internal representation nor show that every developmental process is the same as nest building. They undercut a necessity inference from structured output to externally accessed form.

Across scales, with translation work. Constraints and feedback matter at molecular, cellular, organismal, and collective scales. Their variables, timescales, equations, and interventions differ. Reusing the word “constraint” is not proof of an identical scale-free mechanism; each cross-scale explanatory step needs a construct mapping and independent tests. The same demand applies when Platonic access is asserted at multiple scales.

My Discriminating Tests and Conditions for Revision

Intellectual honesty requires stating what evidence would change one’s confidence. The following are proposed future comparisons, not four predictions already uniquely derived and verified from Brash’s talk. Each needs operational models, boundary conditions, matched controls, and preregistered outcomes before it can adjudicate rival causal claims.

The question is no longer the slogan “What single observation falsifies Platonism?” It is: What could establish that this particular Platonic causal source is specifically right, instead of merely compatible with discoveries whose measured causal explanation does not require it?

Cranes All the Way Down

Brash’s demonstration clarifies a genuine explanatory possibility: patterned organization can follow from local dynamics, forcing, and constraints, without consulting a transcendent template. His wider research programme offers tools for biological study. This is a meaningful challenge to claims of Platonic necessity, even though it is not a completed replacement for Levin’s bioelectric models and cannot by itself falsify abstract objects generally.

This is a process-relational naturalistic working position, not a proof of exhaustive materialism. Higher-level patterns can be real and causally useful at the scales where they organize and constrain processes. What remains unearned is the stronger claim that such patterns must originate in an additional causally effective realm. Mathematical description, physical realization, and independent source-ontology are different claims, and evidence should not be transferred among them without a bridge.

Levin’s programme has produced substantive reported experiments, with stronger external support for some broad phenomena than for its specific cryptic-state and model-derived predictions. Credit should follow that uneven evidence. The interpretive question remains open: do the data discriminate a nonphysical Platonic source, or are that source’s proposed effects accommodated by revising an unspecified mapping to the very physical interface doing the experimental work? Creationist and Intelligent Design publications have appropriated Levin’s metaphysical vocabulary; that is a real public-communication concern, not evidence that he endorses their conclusions or that his hypothesis is therefore false.

Brash’s Chladni plate makes the point memorably: the mathematics describes what material constraints make possible; a pattern is not thereby a command issued from Platonic heaven. Organisms also have histories, feedback, and evolving constraints. The details must be discovered, not inferred from the elegance of a diagram.

The physical explanation has earned the right to stand on its own where it succeeds. Any additional causal ontology must earn its additional explanatory credit.

That is the standard I want both approaches, including my preferred alternatives, to meet.