Coerced Commitment

A signature exists. The credential checks out. The timestamp is clean. The system records a principled commitment — and downstream systems treat the page as if a free agent genuinely resolved to what it says. In reality the mark was made under threat, deadline pressure so severe it functioned as coercion, or a power imbalance that left no meaningful choice. The act of signing happened; the commitment it is supposed to express did not.

In software systems

Workflows treat any authenticated submission as a valid L0 decision. A user clicks through a mandatory approval dialog to unblock their account. A contractor signs a digital waiver because the deployment pipeline will not proceed otherwise. A clinician accepts a default diagnosis code because the EHR will not close the encounter without it. Each event names an authority and a conclusion; none records whether the authority had meaningful discretion.

Audit trails show compliance: signed, timestamped, attributed. Post-incident review cannot distinguish willing commitment from compelled assent because the schema encodes only the expressive act, not the conditions under which it was made. When harm surfaces, accountability lands on the person who signed — though the record never captured that signing was the price of continuing work.

In human organisations

Organisations equate signature with consent. Employment contracts, safety attestations, and committee approvals accumulate marks made under implicit or explicit pressure: sign now or lose the deal, sign or you cannot return to the ward, sign or the project misses its gate. The document enters the file as decided truth. Later disputes discover that the named authority never freely committed to the content — only to the act of putting ink on paper.

Performance and legal processes then treat the signed artefact as proof of intent. The organisation has a decision on record. What it lacks is a decision that reflects what the agent actually resolved. The gap between expressive act and genuine commitment is invisible because both are stored as principled ground truth.

In socio-technical systems

Regulated domains compound the failure. A patient signs informed consent on a tablet handed over by a nurse who must clear the queue. A loan applicant accepts terms on a phone while a broker waits. A platform worker accepts updated terms because the app will not load otherwise. The socio-technical stack makes assent frictionless for the institution and costly for the agent — then stores the click as tier-one commitment with no field for duress, time pressure, or power asymmetry.

Downstream billing, liability, and compliance systems reason from the record as if free commitment occurred. Investigators reconstruct intent from a mark that may only certify survival under constraint. The failure is epistemic: coercion is real, the signature is real, and treating the second as proof of the first collapses a gap authentication was never designed to close.

Structural causes

Epistemic tier collapse

Coerced commitment is tier collapse at the commitment boundary: the system stores the expressive act (signature, click, credential use) with the epistemic status of the genuine inner commitment it approximates. No metadata distinguishes compelled assent from free resolution. The record is honest about what was captured and dishonest about what that capture certifies — unless separate mechanisms (witnessing, cooling-off periods, independent verification, appeal paths) narrow the coercion gap and are themselves recorded.