The Control Grammar

From what control is to how it is made

Two coaches watch the same exchange. One says the top player had great pressure; the other says the bottom player lost their frames. Both are pointing at something real, and neither description lets the other check it, teach from it reliably, or say what would have had to change for the outcome to change. The ontology says what kinds of thing exist in that exchange and Adversarial Viability Control Theory says how the contest evolves under intelligent opposition. The grammar answers the third question: what is the smallest usable language that keeps the mechanically and adversarially important structure of that contest?

The answer has to compress without erasing nine things: causal mechanism, forceability, viability, interruption, timing, differences in what each athlete can execute, what each athlete can see, the rules, and the ways a contest can end. So the grammar is built in typed layers, and the layers are related without being interchangeable.

Four relations between two bodies

Four primary mechanical relations describe the physical situation. They are about the relationship rather than about either person: not that you are strong, but that a particular connection exists between you and them, carries a particular load, and reaches a particular target. Each is stated as a question you can ask of a position.

• Access Can the relevant control reach and act upon its target? Your choking arm can reach the neck from the seatbelt, or it cannot because the chin and the far hand are in the way. Reach is relative to the task: the same hand that reaches the wrist may not reach the far hip.
• Coupling What is connected to what, and what can that connection transmit or constrain? Grips, hooks, clamps, frames and wraps. A seatbelt that the opponent can shrug is a grip name without a coupling; a body triangle that moves their hips when yours move is coupling that has not been named.
• Support and load What is supporting the system, where is the load, and what can currently move? A pin is a load question as much as a positional one. Chest weight through the sternum with the opponent’s elbows pinned is one load picture; the same chest contact with their elbow-knee frame intact is another, and the second one can move.
• Connectivity What is the target still mechanically connected to? The arm you are attacking is still connected to the torso through their own grip, and to the floor through a post. Isolation is relational: a limb that looks separated may have borrowed support from somewhere else.

Two of these carry a warning that decides a lot of diagnoses. A grip name never guarantees coupling function, so “I had the underhook” says which grip existed and not what it did. And isolation is relational, so an arm you have separated from the torso may have found support from a post on the mat, in which case the connectivity you removed has been replaced and the armbar is further away than it looks.

There is deliberately no fifth relation for turning. Which task-relevant turns remain, for the bottom player under a pin or the defender in back control, is a real coaching question, and the grammar answers it compositionally, through access, connectivity, dynamics, the mechanism and viability, rather than with a separate atom. Keeping the inventory at four is a discipline: an element stays only where removing it would lose a distinction, break a causal explanation or hide a branch.

Two relations about relations

Above the four physical relations sit two that describe how controls fit together. Dependency asks what has to become true before the next relation can work: prerequisites, gates, substitutions, bypasses and foreclosures. Within the usual triangle policy the finish waits on the shoulder line; the knee cut needs the far hip stapled or it needs a substitute for it. Feasibility asks whether two controls can coexist in the same body and state. A hand cannot post on the mat and hold a grip at the same time; a crossface and a far-side underhook can coexist, which is why they are taught together. Pairwise compatibility does not guarantee that three controls fit, which is a large part of why complicated positions collapse.

What has to become true before the next relation can work?

Can these controls coexist in this body and state?

How the effect is produced

A relation says that a connection exists. The mechanism layer says how it produces its effect, in three parts: the contact topology, the mode of force transfer, and the process it targets. A crossface produces its effect through head-to-shoulder contact, applied by wedging and load transfer, targeting the opponent’s ability to turn towards you. A rear naked choke works through neck access and a force-transmitting coupling, applied by compression, targeting the vascular neck process. The modes are a short list the grammar keeps deliberately sparse:

• compression
• tension
• torque
• shear
• bending
• wedging / blocking
• frictional fixation
• impulse / collision
• load transfer

Mechanism matters diagnostically because two actions with the same relation can fail for different physical reasons. A pin that relies mainly on wedging fails when the wedge is displaced; one that relies mainly on frictional fixation fails when the surfaces slide. Telling a student to “put more pressure on” addresses neither.

Motion is part of the state

Two rolling scrambles can match exactly in contact: the same torso contact, the same hand coupling, the same leg position. In one the bottom player is rotating into exposure and in the other they are rotating out of it, and the next half second is different in every respect that matters: what is reachable, what can be interrupted, which reorientation is still viable. Double-leg finishes, sacrifice throws and transitional collisions behave the same way. So the grammar carries a sparse dynamics annotation, in ordinary words where numbers are not available: moving forward, retreating, rotating into, rotating away, loaded into the collision, falling through, reversing direction.

It is included only where leaving it out would change the branches, and that rule is the point. Dynamics is not a fifth relation. It is the difference between a photograph of a scramble and the scramble.

Six questions about what a state permits

The relations describe the situation. The predicates say what the situation permits, and this is where the grammar earns its keep, because these six are routinely run together in ordinary coaching language.

• Reachable Can the target occur under at least one admissible joint interaction? The target can happen at all, under at least one way the exchange could go. From the front headlock, the back is reachable: there is a sequence in which the opponent’s turn gives it to you.
• Enforceable You can force the target across the class of responses you have declared or can model. Strictly stronger than reachable. The back is enforceable from that headlock only if the go-behind survives the re-square, the roll and the standing escape.
• Viable You can keep the safe or non-losing relation. Viability does not require the position to be improving. Under side control with the inside elbow-knee connection kept, the bottom player is viable: nothing is improving, and nothing is being lost.
• Restricted Which meaningful opponent prevention or counter-control policies remain? Which meaningful prevention or counter-control options the opponent still has. Inside a well-locked back control, many of the things the defender can still do change nothing that matters; the meaningful answers narrow to a few, such as the hand fight for the choke.
• Interruptible Can the opponent prevent completion of this process before it becomes effective? Whether what you are doing can be broken into before it takes effect. An escape that needs three movements can be broken into after the first; a single adjustment often cannot be.
• Terminal No mechanically viable timely prevention policy remains against the declared terminal process. No timely prevention remains against the process. A property of the state, kept separate from the contest ending. A rear naked choke at the point where no timely prevention remains is terminal whether or not the tap has arrived.

A route existing on a map does not mean you can force someone down it. The grammar makes that a definition rather than an argument, and it gives the opponent’s remaining options a name of their own, which is what lets a coach say “you took away their frame and left them the underhook” as a precise statement.

Races and robustness

When both athletes have something live, the question is who becomes effective first. The passer’s knee cut and the guard player’s underhook are both mechanically available; whether the cut lands before the underhook turns the passer can decide the exchange, and the grammar calls that a race between your time to effect and their time to interrupt. Robustness is a different question again: does this relation or this policy survive a declared disturbance or a declared class of opponent? A body-lock pass that works on everyone in the room may not survive a strong wrestler, and saying so is a statement about the pass against that class, never a score.

Who becomes effective first?

Does the declared relation or policy survive the declared disturbance or opponent class?

The athlete inside the state

The objective state does not decide what one athlete can actually do, see or choose. A choke can be mechanically live for a grappler who cannot execute it, or who has not seen it, or who has seen it and prefers the mount. So the grammar keeps a separate layer for the athlete: what they can reliably execute from here, what their picture of the situation contains, and what they selected. Being able to force something against the modelled responses is a property of the game; being able to do it is a property of the person.

This layer is where deception lives. A feint changes almost nothing mechanical. It changes the opponent’s inferred state: they see an attack that is not coming, select a response to it, and their response changes the physical state. The grammar lists the mismatches that matter: a false sense that access exists, a missed dependency, a hidden coupling, an unrecognised terminal threat, a branch believed viable that is not, and a restriction believed to be there that is not. Each is a real event in a match and each has a place in the parse.

Rules and clocks

A policy can be mechanically available, executable by this athlete, illegal, and strategically poor, all at the same time. A heel hook entry that is accessible, executable and enforceable from the current position is the same physical object under every ruleset; under one it is legal and under another it is a foul, and the score and the clock change whether it is the right choice either way. The grammar keeps contest context as its own layer so that a change of rules never rewrites the mechanics. What changes is admissibility, value, the stopping condition and the reset.

Terminality and the tap

Mechanical terminality is a property of the state: no timely prevention remains. A contest can end before that point, at it, or after it, through different routes, and the grammar types the endings separately: a recognised concession, a loss of consciousness, a referee’s intervention, or the clock and the score. The straight ankle lock is the case that forced this. Pain and the sense of injury threat can change what the defender perceives and selects, and produce a tap, long before the lock is mechanically finished. Painful does not mean terminal.

For a coach this is two lessons at once. A tap is evidence about the defender’s decision, not proof that the mechanism was complete; and the defender who learns to tell threat from terminality can recognise a lock that is threatening before it is finished, and concede early rather than late. Pain is a reason to concede, never evidence to hold out against.

Techniques and positions, redefined

In this language a technique is an entry signature, a policy family, a response-family structure and an exit signature. It is not one exact trajectory. The knee cut is the state you enter from, the branches you take on each of the guard player’s answers, and the states you exit into. A position is a surface label with an optional state signature attached, and the grammar keeps the labels while limiting what they may imply:

surface label + optional semantic state signature

“Mount, with the defender’s left elbow-knee reconnection still viable, right-arm isolation present, attacker chest coupling stable, and terminal arm access asymmetric” contains more usable information than “mount”, and it needed no new position name to say it. The same discipline keeps the vocabulary small at the bottom. None of the following is a new primitive; each is a composition of the relations above:

• posture
• pressure
• inside position
• leverage
• base
• chest connection
• back exposure
• kuzushi
• structure
• rotational reserve

Pressure, for instance, can be described through support and load, coupling and a mechanism. Saying so is what lets two coaches who use the word differently discover that they mean different things.

Reading an exchange in the grammar

Chest-to-chest, parsed. The objective relations are anterior torso contact, trunk and hip coupling, and the top player’s external support. The mechanism is anterior torso-to-torso contact applying compression, load transfer, frictional fixation and wedging, targeting trunk reorientation and frame reconstruction. The game consequence is fewer viable reorientation pathways, less frame reconstruction, and a restricted escape family. Nothing in that parse needed the word pressure, and every part of it can be checked against what the bottom player can still do.

Back control, parsed. Rear orientation, torso and hip coupling, privileged access for the attacker, and a defender who lacks reciprocal facing. The mechanism is frictional fixation, wedging and redirected load, targeting the defender’s re-facing. The consequence is that the defender’s viable re-facing branches are reduced, their meaningful prevention responses are restricted, and the attacker’s terminal branches to the neck and arm are preserved. The seatbelt, the hooks, the body triangle and chest-to-back contact are implementations of that structure, and the grammar represents what sits beneath them without making any one of them fundamental.

Diagnosing why something failed

The grammar’s most practical output is a failure signature: which objective relation failed, what game consequence followed, where in the athlete the failure sat, and what the context contributed. “The armbar failed” is not a diagnosis. In the grammar it becomes one of several different events, and each calls for a different week of training.

The objective failure classes are access, coupling, support and load, connectivity, dependency, feasibility and mechanism. The game failures are enforceability, viability, response restriction, interruption, the race, and a terminality misclassification. The athlete-side failures are execution, information or perception, and selection. A context failure is a mechanically valid choice that was illegal, poor under the score or the clock, or needlessly risky. And a termination failure is an analyst’s error: treating pain, a tap, unconsciousness, a stoppage and mechanical terminality as one thing.

Translating what coaches already say

The grammar does not replace a gym’s vocabulary. It sits under it. Base, posture, alignment, connection, inside position, frames, wedges, anchoring, attachment, dominant angle, leverage, structure and kuzushi all translate into the relations above, and the translation is many-to-many and depends on context: “base” may mean support and load in one sentence, connectivity in the next, and dynamics in a third. Different lineages name one policy differently, split one policy family into several techniques, or fold several stages under one name, and the grammar can line them up only where the deeper relations match. Two techniques are equivalent when they preserve the same control proposition, target process, game consequence and response structure, never because practitioners say they do the same thing.

For the room, the whole language compresses to twelve questions, and they are the grammar’s own:

1. Can you reach what matters?
2. What are you connected to, and what does that connection do?
3. Where is the load and what is supporting it?
4. What is the target still connected to?
5. Which meaningful turns remain?
6. What has to become true first?
7. Can these controls coexist?
8. Can you actually force this?
9. What can they still preserve?
10. What can they still do that matters?
11. Can they stop this before it works?
12. If this is a finish, is any timely prevention left?

A formal parse is not the right thing to say to an athlete. The grammar is explicit that the parse, the coach’s compression of it and the cue the athlete hears may use different words, provided the cue keeps the meaning that matters for the task. Depth of analysis and depth of communication are separate dials.

What the grammar hands on

The grammar is the diagnostic and translation interface the rest of the theory speaks through. The capability architecture derives the things a grappler must become able to do from these relations and predicates, so that “restore connectivity” and “deny reorientation” are capabilities rather than a list of moves. The coaching method uses the failure signature to diagnose before it prescribes. What the grammar does not do is choose: it selects no priority, no pedagogy, no curriculum and no training load. It says what is happening, precisely enough to argue about.