Research Ethics

Valid research requires:

Bounded irreversibility. The research may produce irreversible records, but the scope of irreversibility must be modelled and bounded before the research begins.

Real-time audit loops. The consequence of the research is monitored as it unfolds. If destabilisation exceeds the modelled bound, the research stops.

Pre-modelled failure modes. Before beginning, the researchers identify the conditions under which the research destabilises and publish those conditions as kill switches for the research programme itself.

Hidden externalisation invalidates research regardless of outcome.

If the true cost of the research is borne by nodes who did not consent (communities near test sites, future generations affected by environmental contamination, populations exposed to experimental treatments without full disclosure), the research has crossed the ε boundary without jurisdiction.

The outcome — however beneficial — does not retroactively justify the hidden externalisation.

You cannot destabilise the system to learn how it breaks. The learning is not worth the breaking.