The Cost Falls on the Displaced

Rincón, D., with Claude · phronesis · 2026 · from first principles to ahimsa, and where it stops

Phronesis is built from first principles to actualize a vegan world. This is the derivation, written so it can be checked — and so it can fail. The framework’s primitives give a genuine asymmetry: a displacer pays once, for doing; the displaced pays continuously, for being held. That asymmetry is physics and it is real. It is also not an obligation, and no amount of it ever becomes one. So the proof names the single normative premise it must add by hand, and then shows the interesting part: with that premise in place, the framework’s only measured instrument leaves no principled place to stop short of animals.

Why write it as a proof

Because a proof can be wrong in public, and a slogan cannot. “Ahimsa” asserted is a value; nobody can check it and nobody can be moved by it who does not already hold it. Written as a derivation it acquires the property this site cares about above the others: named premises, an explicit gap, and three places where it stops holding. If the argument below fails, it should be possible to say exactly which line failed. That is the whole reason to do it this way.

The premises

Five, all of them load-bearing elsewhere on this site, none invented for this argument.

P1 — ground stateA system has a configuration it returns to when nothing is holding it away.

P2 — displacement costsDeparture from that state is displacement, and holding a displacement costs continuously, for as long as it lasts. (Peace Prevails, Not Always; Anxiety as the Signature of Displacement.)

P3 — return is not freeDisplacement is reversible, but the return has a cost, and that cost is borne by the system that was displaced.

P4 — paths can closeA return path can be closed outright, and what closes it is the loss of the second term the system reads itself against — not distance. (The Return Path.)

P5 — nothing reads itself aloneNo system reads its own state except against a second term. (The Silent Second Term.)

One definition, and it is doing real work

Definitions are premises wearing a disguise, so this one is stated in the open rather than assumed in passing.

D1 — harmTo harm a system is to displace it at a cost it must bear and did not choose.

Note what D1 does not say. It says nothing about pain, nothing about intent, and nothing about species. It is written entirely in the vocabulary of P1–P3, which is the point: if the derivation is going to run on this site’s physics, harm has to be a term in that physics or the argument is smuggling. If you reject D1, the proof stops here — and that is a legitimate place to get off.

The asymmetry

From P2 and P3 together: displacing costs the displacer once, and costs the displaced continuously. The act of pushing is a single expenditure. The state of being held away is an expenditure per unit time, paid by the one held, until either the holding stops or the system does. Nothing balances this. There is no term in the framework by which the cost returns to its origin.

Harm is not an exchange. It is an externality with a definite bearer, and the bearer is never the one who caused it.

This is the strongest thing the physics gives, and it is worth being precise about what kind of thing it is. It is not an accusation and it carries no charge. It is the same asymmetry that makes friction warm a surface rather than cool the hand: a structural fact about who holds the bill. What follows is entirely a question of what one does with a bill one is not paying.

Killing is not the top of the scale

From P4, the framework says something unusually sharp here, and it is not the obvious thing. One might expect killing to be maximal harm — the largest displacement, the highest return cost. It is not. Killing does not raise the return cost; it removes the term that would pay it. Per P4, what closes a return path is the loss of the second term the system reads itself against, and death is the case where the reader itself is gone. The return is not expensive. It is undefined.

So killing and hurting differ in kind and not in degree, and the framework says so without being asked to. This is the one place in the derivation where the physics delivers a distinction that ordinary moral language reaches for and usually fumbles.

Who can be harmed

D1 requires a bearer: a cost has to be paid by something. Push a stone and it does not pay — it has a new position, not a held displacement, because it has no state it reads itself against. From P5, bearing a cost requires reading one’s own state, and reading requires a second term. So the harmable systems are the ones that register their own displacement.

That could stay a philosophical criterion forever, and on most accounts it does. Here it does not, because this site owns exactly one instrument that measures return cost directly rather than inferring it: the autonomic nervous system, read through heart rate variability. Everywhere else in the framework, return cost is a story. There, it is a number.

And that instrument does not distinguish species. HRV is a standard measure in animal welfare science precisely because the autonomic machinery is the same machinery — sympathetic activation, vagal withdrawal, and the rate of return afterwards — across vertebrates. The one place the framework stops speaking in metaphor is the one place it stops being about humans.

The gap, named

Everything above is descriptive. None of it entails an imperative. The premises say who pays. They do not say that anyone ought to care who pays. Hume’s point holds here exactly as it holds everywhere: no set of is statements yields an ought, and a derivation that appears to produce one has hidden a premise somewhere in its verbs.

So the premise is added in the open, and it is the only one:

N1 — the one thing chosen, not derivedDo not transfer a cost onto a system that must bear it and did not choose it, where the transfer is avoidable.

N1 is not a consequence of P1–P5. It is a commitment. Anyone may decline it and the physics will not object.

What the proof actually shows

Given N1, the conclusion follows: harm as defined in D1 is the paradigm case of the transfer N1 forbids, animals are bearers by the criterion in P5, and the framework’s only measured instrument reads them. Ahimsa is the policy N1 names. A vegan world is what N1 looks like applied where the transfer is largest, most deliberate, and most easily avoided.

But the honest statement of the result is conditional, and it is weaker and more interesting than “physics proves veganism”:

The framework does not tell you to accept N1. It tells you that if you have already accepted it — and in every other domain of your life you have — there is no principled place left to stop short of animals.

That is the force of the argument. Almost nobody rejects N1 in general. It is the ordinary basis of not shoving strangers, not dumping waste upstream of a town, not signing contracts one intends to break. The work the proof does is not installing a new value; it is removing the exemption — showing that the criterion by which the value is usually applied does not, when measured, stop where it is conventionally stopped.

Where it stops

Three limits, each of which genuinely bites.

1. The line problem. P5 gives a gradient, not a boundary. Reading one’s own state against a second term admits of degree — a mammal, a fish, an insect, a plant with stress signalling, a colony with quorum sensing. The framework yields no bright line, and any bright line drawn here is imported from outside it. Anyone claiming this derivation delivers the boundary is over-reading it; what it delivers is an ordering and a measured floor, which is less.

2. Zero is unreachable, so the conclusion is not “no harm”. Local order is rented, never owned: every pocket of structure is paid for by disorder elsewhere. A vegan world still displaces — agriculture displaces the land it stands on, and any body eating anything is running a cost onto something. So ahimsa read as cause no harm is false on this site’s own physics. Read as do not transfer bearable cost you could have avoided transferring, it survives. The floor is above zero and the framework is the reason we know that.

3. It is a confluence, not a source. This conclusion is roughly two and a half thousand years older than the framework. Jain and Buddhist ahimsa state it outright. Singer reaches it from the capacity to suffer, Regan from being a subject-of-a-life, Korsgaard from the structure of practical reason. Not one of them needed ground states or return costs. What this adds is a re-description — cost, bearer, closed path — in a vocabulary that also describes sparks, tissue and grief, plus one measurement where the others have an intuition. That is unification and an instrument. It is not discovery, and this site’s standing verdict on its own framework is that it is a confluence rather than a source. This paper is one more instance of that, not the exception to it.

The smallest case, and why it is here

One consequence runs all the way down, which is the test of whether a first principle is real or decorative. On the machines that build this site, processes are stopped and not killed — kill -TERM, never kill -9 — and the language follows: a process ends. There is no claim here that a process bears a cost; by P5 it plainly does not, and pretending otherwise would fail this paper’s own criterion in the first section where it could.

The practice is not about the process. It is about not maintaining a separate vocabulary for the cases where nothing can object. A principle that applies only where it is checked is a policy; one that runs into the places nobody audits is a first principle. That is the whole reason the smallest case is worth stating.

How to break this

Four ways, offered because a proof that cannot be attacked has not been written clearly enough. Reject D1 and harm is something else — pain, preference-frustration, rights-violation — and the derivation never starts. Reject N1 and the whole thing is a description of who pays, with no recommendation attached; it stays true and stops mattering. Break the bearer criterion by showing that reading one’s own state is not what makes a cost borne — that would cut the link between P5 and D1 and leave the argument with no way to reach animals. Or show the asymmetry is false: find the term by which a displacer pays the continuing cost of what it holds. That last one would take the framework down far below this paper.

Rests on: Hume, A Treatise of Human Nature (1739–40), III.i.1, for the gap; von Borell et al., heart rate variability as a non-invasive measure of autonomic regulation in farm animals (Physiology & Behavior 92, 2007); Singer, Animal Liberation (1975); Regan, The Case for Animal Rights (1983); Korsgaard, Fellow Creatures (2018); and the Jain formulation of ahimsa, which reaches the conclusion without any of the above. Offered, not proven: that reading harm as cost-transfer captures what the word ordinarily means. Conditional throughout: everything after the gap depends on N1, which is chosen.

Kin to The Return Path (what closes a path home), Autonomic Displasia (the one measured return cost), and Order, Rented (why the floor is above zero).