Working document — Brian Esty. Started 2026-07-01. Status: conceptual spine, stress-tested. Not yet descended into gait specifics or the teaching problem (both deferred by design).
Purpose
A coherent conceptual basis for continuously refining stance and gait — usable both for self-training and for coaching clients and other practitioners. The refinement is understood as continuous, never all-or-nothing. The framework must be mechanistic enough to instrument and drill, and philosophical enough to stay true to living systems rather than machines. Both faces are required; neither is decoration.
The design spec (Brian's words): the model should assist the user in self-training into greater resilience. The user is simultaneously the sensor, the actuator, and the interpreter. This is a first-person, closed-loop model, not a third-person description of the body.
The Block, Named
Three frameworks that each work in practice but don't reconcile as theory:
- Tensegrity / continuous-network model — holistic, explains emergence and self-organization, matches what the hands and one's own gait experience. But it is descriptive, not generative: all reference, no usable loop. It gives no perceivable signal to the user and no discrete adjustment.
- Reflex / neuromotor model (MNRI; the four functional bins — anterior/posterior, lateral, rotational, vertical/axial) — generative and testable, a library of adjustments with a developmental order and clear targets (mature reflex). But its feedback loop is slow and usually therapist-mediated, not self-run.
- Shock-load / heel-strike feedback (signal-processing, engineer's frame) — real signal, near-instant feedback. But it is a single scalar: an error detector without a reference. It can make you flinch, not refine.
The key move: these were being merged as competing theories. They are not competitors. They are the components of one control loop, and no single one is the loop. Each is strong exactly where the others are weak.
A self-training system needs four things:
- A signal the user can perceive (a felt or measured error)
- A reference — what "better" feels like, so the error means something
- An adjustment the user can make that moves the signal
- Feedback fast enough that the loop closes quickly enough to learn from
Mapping: shock-load owns (1) and (4); tensegrity supplies (2) as the "optimized web"; reflex/MNRI owns (3) and also holds references (mature reflex). The unifying object is the loop itself, not a better theory of the body.
Stop looking for the model of the body. Specify the loop.
The Spine (final form, after stress-testing)
The body is a self-organizing tensegrity that stays alive by riding its own error signals at every scale — productive disequilibrium. Self-training is consciously joining these loops.
There are two layers:
- an inner loop that refines organization within the currently-available state space (its edge is resolution; its error is felt as effort / inefficiency), and
- an outer loop that governs how large that state space is by rating regions as safe or dangerous (its edge is the protective response; its error is felt as threat).
Injury is a true constraint the inner loop routes around. Trauma is a defensive rating only the outer loop can revise. Resilience is the truthfulness of the map at the finest holdable resolution across the largest truthful territory — never mere size, never a fixed endpoint.
The craft, for practitioner and self alike, is to work each loop as close to its error edge as possible without tripping the protection that would contract the map.
The path is necessarily unique, because both the constraints and the ratings are written by an individual history.
Why It's Both Mechanistic and Philosophical (the same face, not a compromise)
- Mechanistic face: perceivable error → reference → adjustment → fast feedback, tightened over time. Clean control theory.
- Philosophical face: a living system is defined by disequilibrium — a state that "facilitates adaptation and the emergence of new forms" (Brian's own site language). A system at equilibrium has no error signal, nothing to perceive, nothing to refine — it is stuck or dead. The loop runs because the body is never settled.
These are the same claim at two scales. "Productive disequilibrium" is "living permanently at the error edge."
This is also why the "too mechanistic" objection dissolves rather than being argued away: a mechanistic loop feels reductive only when imposed from outside on a passive body (the Newtonian lever model, rejected). A loop the living system runs on itself, using its own tensegrity as reference, is not imposed — it is the system's own self-organizing behavior made legible to conscious attention. We are not adding control to the body. We are giving conscious attention a seat at a loop that was already running automatically. (This is the site thesis: the automaticities compensate constantly, below awareness; functional bodywork trains the automaticities.)
And it resolves "continuous refinement, not all-or-nothing": all-or-nothing is equilibrium thinking (fixed/broken). A loop has no final state. Resilience is the quality of the loop, never its endpoint. Refinement is the only mode it has.
Stress Tests It Survived
1. Injury — the fused femur/hip client
Naive prediction: removing degrees of freedom (a rigid strut where a tunable one was) should lower resilience. But the client shows strong resilience within the constraints.
Resolution — this sharpens the spine rather than breaking it: the reference is not a fixed ideal posture; it is optimal function given the current constraint set. The loop compares the body to the best available organization of this body now, not to an abstract ideal. The fusion shrank the state space; it did not degrade the loop. The client's loop found a high-quality attractor inside a smaller space.
Consequences:
- Uniqueness-to-the-individual becomes mechanically required, not a soft caveat: every constraint set defines a different state space → a different reference → a necessarily individual loop. The model predicts uniqueness.
- The last trace of the Newtonian ideal-posture ghost dies. There is no universal target — only "best organization available to this net today," which the body already knows how to seek.
- Injury = a permanent boundary condition the loop routes around.
2. Trauma — a different animal from injury
Injury changes the plant (the physical system controlled). Trauma changes the loop itself. A held trauma pattern is not a smaller state space — the degrees of freedom are physically present — but the loop has locked a variable it is physically free to move, because entering that region once meant threat. The error signal from that region is read not as "refine here" but as "danger, do not perceive here."
This is genuinely dangerous to the model, because it means the loop can be pathological: efficient at defending exactly the wrong attractor. Efficiency is not goodness. A high-quality loop can be excellent at keeping you stuck.
Resolution — the two-layer distinction:
- Inner loop refines within what the system currently treats as available. Currency: resolution.
- Outer loop governs what counts as available at all — marks regions safe/dangerous. Currency: safety.
Injury lives in the inner loop's plant; trauma lives in the outer loop's map. This is why you cannot refine your way out of a trauma pattern — more inner-loop reps just make you better at avoiding the barrier. The barrier moves only when the outer loop re-rates the region as safe. That is precisely the job of craniosacral, reflex work, energetic/emotional modalities, and NLP: they operate on the rating, not the movement. The site's "de-stressing releases resources tied up in adaptation as a burst of energy" = the outer loop un-defending a region, freed degrees of freedom flooding back, the state space growing. That felt burst is the state space enlarging.
3. Can the outer loop over-open? — the failure mode in both directions
Yes. If the outer loop rates as safe a region that is not actually survivable (a range the fusion can't support; a protective barrier that was doing real structural work), reality corrects it — hard.
Brian's decisive observation: when the outer loop over-opens, the client experiences a protective response, which often seems like a setback. The work is to take the client as close to this edge as possible without overstepping.
This is the keystone. The protective response is not a setback — it is the outer loop's error signal. The outer loop is itself a loop, continuously testing its ratings against reality; the protective response fires when a rating is tested past what is currently true. It reports the true location of the edge. So there is no crack: every layer is the same creature — a loop that learns by riding its own error signal — operating on a different variable.
Therefore resilience = truthfulness of the map, not size of the territory. Over-opening is not "more resilient"; it is a false rating reality will violently correct. The fused hip is resilient because its map is true: it rates the real limits correctly and organizes optimally inside them.
The Craft, In One Line
Work each loop as close to its error edge as possible without tripping the protection that would contract the map.
- Too timid → no error is presented → the loop doesn't learn.
- Too far → protection fires → the state space contracts → you've taught the outer loop that the region is more dangerous, not less.
- The skill is living where the error is maximal and the protection is not yet triggered.
This is one skill wearing different clothes — the same in tissue depth, reflex challenge, emotional exposure, and gait load.
Definitions (for teaching and self-use)
- Inner loop — refines organization within the available state space. Edge: resolution. Error felt as effort/inefficiency.
- Outer loop — sets how much state space is available by rating regions safe/dangerous. Edge: the protective response. Error felt as threat.
- State space — the set of organizations the body can currently enter. Shrunk truly by injury, defensively by trauma.
- Reference — best available organization of this body under its current constraints and ratings. Not an ideal posture.
- Resilience — truthfulness of the map, at the finest holdable resolution, across the largest truthful territory. Never mere size; never a fixed endpoint.
- Productive disequilibrium — living permanently at the error edge; the condition that makes learning/refinement possible at all.
Deferred (not yet ready — Brian's call)
- Descend one level — the actual gait loop. What is the reference for gait? What are the perceivable errors beyond heel-strike/shock-load? What is the adjustment vocabulary? Make it concrete enough to self-train.
- The teaching problem. How to transmit "join the loop" to a client who has never felt their own tensegrity as a reference — because the reference is a felt sense you cannot simply hand someone.
Open sub-thread under (3)/(1): the practitioner's live skill of finding the edge without overstepping — how it's sensed, and whether/how it can be taught rather than only accrued over decades.
Provenance
Developed in conversation, 2026-07-01. Path: three-frames problem → the loop reframe (the four control-loop components) → mechanistic/philosophical unity via productive disequilibrium → stress test on injury (fused hip → reference-is-constraint-relative) → stress test on trauma (two-layer inner/outer distinction) → over-open failure mode → keystone (protective response = outer loop's error signal) → resilience as truthfulness of the map.
To save as PDF: open in a browser and use Print → Save as PDF.