Why doesn't an apology undo it?
What a strained bond keeps after the strain is lifted.
We expect a bond to recover once the strain is lifted. After an apology, we expect it to go back to the way it was. Below, a single bond is held bent under a load. Remove the load and watch whether it returns to where it began.
A small strain fades, and the bond goes back to what it was. A larger one moves it past the point it can come back from, and that part stays. An apology lifts the strain, but it cannot reach what has already set. When the argument is over the bond is still there. It just holds a slightly different shape than before.
One observation falls outside the apparatus. Subjects shown the residual set accepted the measurement, then asked whether what remains is damage or simply the new shape of the thing. The apparatus can measure how large the set is. Whether it is damage or a new shape is not something it can weigh, and that question is reserved for future work.
complete findings · method · field data (n = 284) · discussion
abstract
A bond subjected to strain is commonly expected to return to its prior condition once the strain is withdrawn. Measurement of 284 reconciliations following a contested event indicates that, above a threshold load, the post-reconciliation state does not coincide with the pre-strain state. The residual offset, not the magnitude of the strain itself, predicts the bond's later behavior.
method
Each bond was measured at three points: at rest (pre-strain), at peak (under load), and at recovery (after the strain was fully removed). Load was operationalized as the magnitude of the contested event; deformation, as the change in baseline conduct: warmth, openness, and the default assumption of good faith.
results
Below a threshold load, recovery coincided with rest: full elastic return, no residual. Above the threshold, recovery settled at a stable offset from rest; the offset persisted across the observation window, and its magnitude scaled with the load applied beyond the limit. In all such cases, the path from peak back to recovery did not retrace the path from rest to peak.
discussion
The bond is path-dependent. Reconciliation removes the load; it cannot remove the set. The familiar report that things “returned to normal” describes the removal of the load, not the recovery of the state. What is experienced afterward as residual tension is, mechanically, a real and stable change in the bond's resting configuration.
limitations
The apparatus measures the magnitude and persistence of the set. It does not measure whether the set is injury or adaptation. That question is reserved for future work.