Interpretation and Sequencing

Executives preparing for a contested decision spend a surprising amount of thought on order: which argument to make first, or whether to see finance before the meeting or after it. Some of that thought is well spent. Much of it goes on choices that leave the end of the maneuver where it was, and it comes out of the attention owed to a few choices that move the end and cannot be taken back.

The Lam System for Political Interaction tells the two apart. It leaves two kinds of judgment to the actor. Interpretation reads the world: whether the actor holds enough for an ask, whether the target read a lever as projected. Sequencing chooses among the moves the record allows. The Boundary of Computation argues that the two are one boundary seen from two sides. They also meet at particular places, and those places are where an actor's mistakes stop being correctable. Every choice of order that matters closes something off, and every one is made on a reading.

A reading leaves nothing to choose

Interpretive judgment answers questions that have answers. Does the actor's standing with the target bear this ask? Did the target's reading of a lever come out as the lever projected? Has a lever's purpose already been met by something else? Only the actor is placed to answer, and the answer is true or false about the situation. Once it is given, the system has nothing left to decide. A reading that validated is written as a perception and one that did not as a failure, and the actor's account of what they hold sets the gate's verdict. More than one entry could have come next at each of these points, and the world picks which, through the actor's reading of it.

Sequencing judgment answers questions with no answer in the situation. Two actors who read the target identically can order their moves differently for good reasons, as Stances Toward Uncertainty describes, and at those points nothing in the world picks among the continuations the record admits.

Keeping the two apart is what lets a post-mortem find the fault. An actor reviewing a maneuver that went wrong wants to know whether a reading was wrong, or whether a sound reading was followed by a poor choice. The remedies differ: the first calls for reading the next target better, the second for choosing differently next time. The record can tell them apart only because each judgment is written as the actor's, at its own moment.

Most order is free

Much of what feels like sequencing is scheduling. Two levers whose gates have both passed, each spending a different resource, can be executed in either order, and the actor ends holding the same things with the same exits open. Their gates can be tested in either order, and their readings written in either order. A pending lever can be resourced before or after work on another lever. Two levers planned at the outset can be declared in either order, provided the actor's budget, the number of levers they have decided a cycle will take, has room for both. In each pair, neither move reads anything the other writes, so the maneuver arrives at the same place whichever goes first.

Once the actor's judgments are fixed, then, the order of independent moves never changes the outcome. The record is exact for everything sequenced and exact up to order for moves that ran side by side: it writes one order, because a record is a sequence, and makes no claim that the order happened.

This settles a question post-mortems often circle. "Would it have gone differently if I had shown the second case first?" has a definite answer when the two levers were independent: given the same readings, no. An actor who believes the order would have changed how the target read the cases is making a different claim, about the target, and it belongs with the readings. An actor who thinks the second case lands better once the first has prepared the ground acts on that belief by reading the first lever before playing the second, and then something is learned between the two moves, and their order is no longer free. Order matters to the system only where something is learned or something is closed between two moves.

The forks that close something off

Choices of order that change the outcome sit at six points, and each removes something the maneuver could otherwise still do. Giving up a lever closes it for good, whether it is given up at a failed gate, before its resources have been fetched, or after another lever has covered its purpose. Mooting or abandoning the cycle closes the cycle, and another attempt is a new cycle with a record of its own. Revising the aim runs detection and the objective again, so a lever declared just before the revision was chosen against the old aim and one declared just after against the new. When the budget has room for one more lever and two are on the table, declaring either takes the last place. When two gates have failed, whichever is escalated first runs while the other waits on what it brings back. And when the last live lever has been read, adding another lever and letting the target decide are a fork of their own, because the target decides once.

These are the points the system's determinism is conditioned on. Fix the actor's readings and their elections here, and the rest of the maneuver follows: the entries the record gains, what the actor holds at the end, how each cycle ends. The machine keeping the record has freedom only in scheduling, and scheduling never moves the end.

Every fork is made on a reading

Each of those elections rests on something the actor has judged about the world. A lever is given up because its purpose seems already met, or its resource out of reach, or its cost above what it can return. A cycle is abandoned when the target looks immovable or the interest no longer worth having. An aim is revised when the actor learns something that changes what the target can be brought to do. The last place in the budget goes to the lever aimed at the concern the actor reads as decisive, and the first escalation to the holder who looks likelier or quicker to release what they hold. The target is left to decide when the actor reads the comparison as already won.

So the two kinds of judgment arrive together, a strategic election immediately after the interpretive judgment it stands on, and a misreading behaves differently depending on what follows it. With no fork behind it, a misreading can still be caught as the maneuver goes on. An actor who overrates their standing with the target passes a gate on standing they do not have, plays the lever and finds out at the reading, and a new lever is still open to them. A misreading that licenses an irreversible election is locked in by it: a lever given up on a hopeful reading cannot be revived when the reading turns out wrong.

Illustrative. The people and the firm are invented. Odette Marais runs litigation support at a law firm and wants the managing partner, Bastien Kowalczyk, to fund an in-house e-discovery team under her, in place of the outside vendors the firm now pays. His adverse alternatives are to renew the vendors or to give the work to the IT department. She declares two levers at the outset: show Bastien what the firm paid vendors on its three largest matters, from billing records she holds; and have Hester Ng, who leads the largest practice group, back the plan at the partners' meeting. The first gate passes. The second fails, since Odette does not hold Hester's endorsement. She plays the first lever before deciding what to do about the second, and Bastien reads the vendor spend as a cost he can cut and the team as the way to cut it, as the lever projected.

That puts her at a fork with a reading under it. If Bastien is already persuaded, escalating to Hester spends Odette's standing with her and shows Hester the plan before the meeting, to buy nothing. If he is not, giving the lever up throws away the one most likely to carry the partners. She reads him as persuaded, gives up the second lever and, at the next fork, lets him decide. He funds the team.

Had she misread him, the record would hold a validated perception, a mooted lever and a selection of the vendors, in that order, and the error would be easy to place: the reading at the first lever, carried into an election that took the second lever away. What Hester would have said is the one thing it could not hold.

The room the system leaves

The rules are permissive at each of these forks because the choice there is the actor's, and they hold themselves to a strict test in being so. A fork counts as the actor's only if every legitimate way of proceeding at that point can be written through it. If some legitimate behavior cannot be reached whatever the actor chooses, the fork is a defect and the system has to change. Points where a reading of the world decides which entry follows are kept out of the test, as interpretation, so that a guard the world settles is never passed off as a choice the actor makes, or the reverse.

The actor is left knowing where deliberation pays. Time spent choosing the order of two independent cases buys nothing the record will ever show. Time spent testing a reading in the moment before it licenses an election goes where an error would otherwise become permanent, since that moment is the last at which the reading can be changed at no cost. Afterward the record will show exactly which reading the election rested on, beside the election itself, and the lever, the cycle or the place in the budget that it closed will stay closed.