This small public journey demonstrates the architecture with iCIDM as Reference Implementation #001. It does not claim that an institution is conscious. “Self-observation” here means an inspectable operational capability: representing what is known, uncertain, contested or missing.
Counter-questionWhere is evidence thin, old, disconnected or absent?
PurposeTurn institutional memory into the next observable question.
STEP 2 · OBSERVE THE EVIDENCE LANDSCAPE
Start from real institutional memory
The public prototype begins with 2,231 CIDM publications and 460 active reference objects, then asks about relationships among people, pathogens, technologies, periods, questions and capabilities.
KNOWNSupported strongly enough for the stated scope.
UNCERTAINEvidence exists, but confidence or interpretation remains limited.
CONTESTEDCredible alternatives or disagreement remain.
UNKNOWNRequired observation/evidence is missing or not yet connected.
UNASKEDA gap exists but has not yet been formulated as a question.
STALEEvidence may no longer represent current Reality.
STEP 4 · TRY THE GAP GENERATOR
Choose one observation lens
Candidate gapWhich historical conclusions would change if old samples or datasets were re-analysed using current methods? Reality Return: identify a preserved dataset/specimen population and pre-register a re-analysis question.
Candidate gapWhich capabilities appeared only after a technology entered CIDM, and which questions became newly observable? Reality Return: compare before/after publication and capability topology.
Candidate gapWhich pathogens have deep longitudinal evidence, and which have episodic or fragmented institutional memory? Reality Return: construct a coverage matrix and select one missing road for targeted review.
Candidate gapWhich potentially important people–technology–pathogen relationships are visible in Explorer but not yet strong enough for Navigator? Reality Return: review evidence and either promote, reject or retain as uncertain.
Human gateScientific review decides whether the gap is meaningful.
AI roleSearch, compare, reconstruct and expose candidate patterns.
Failure testCan the system explain why it labelled something unknown?
STEP 6 · REALITY RETURN
The journey must leave the website
A useful unknown produces an observation, review, re-analysis, experiment, data collection, discussion or decision in the real institution. The result then returns as new evidence and may revise the map.
Question→Action→Reality′→New Evidence→Revision
WHAT THIS DEMONSTRATES
One architecture, not another subsystem
Question Passport identifies the inquiry. iCIDM supplies a real population. Explorer finds candidate roads. Navigator carries reviewed roads. Self-Observation exposes epistemic state. Experiment Passport specifies a test. Reality Return closes the loop.