We were asking
Can a fixed early action accumulate commitment if it is repeated at later ages?
The action that separated high- and low-commitment siblings at q36 stopped pointing along that coordinate when replayed later, even though the repeated pulses continued to move state and form.
Can a fixed early action accumulate commitment if it is repeated at later ages?
It cannot reliably do so; the original information ordering closed within four steps, and later replays drove large changes along other directions.
Can feedback re-evaluate the current state at every pulse and recover the future-constraining effect?