Research & Perspectives
Research on judgment, AI governance, reasoning and interdisciplinary discovery. Every paper is linked to its permanent DOI record.
Selected collection
The selection foregrounds practical frameworks for adoption, governance, reasoning, multi-agent assurance and research exploration.
A practical, AI-positive and anti-abdication framework for deciding whether an AI system should be adopted before it becomes embedded in workflows, policies, procurement decisions or institutional habits. Its core test is simple: will this AI system leave human judgment stronger than it found it?
DOI 10.5281/zenodo.21154039The TenthDan framework treats governance not as a brake on AI adoption but as the discipline that simultaneously bounds risk and unlocks institutional speed.
DOI 10.5281/zenodo.18713472A governance layer that treats every consequential AI-agent decision as a reviewable case — with facts, applicable rules, comparable prior cases, reasons, accountable human authority, an appeal route and post-case learning. Auditability without judgement is not yet accountability.
DOI 10.5281/zenodo.20536905An AI-positive, proportionate post-answer scaffold for calibrated epistemic agency. InnerSapience helps users inspect, contextualise and take responsibility for AI-supported conclusions without turning every use of AI into a long exercise — so AI removes effort without silently removing judgement.
DOI 10.5281/zenodo.20788458A protocol for making declared multi-agent coordination behaviour inspectable, rehearsable and receiptable. It names the gap False Harmony — a final output that looks coherent while hiding missed escalation, role failure, authority breach or suppressed dissent — and introduces Choir Receipts as tamper-evident records of what was declared, tested and submitted.
DOI 10.5281/zenodo.20724040Extends the Vlogcano framework to scholarly discovery. Instead of searching for what is closest and most similar, the Collider searches outward for Maximum Viable Distance — the furthest intellectual distance at which a research collision retains an explicit bridge, sufficient evidence, and enough stability to form a coherent research opening. Deterministic by design, with inspectable collision receipts and a voluntary Collision Commons.
DOI 10.5281/zenodo.21424750Archive
The complete twelve-paper record remains intact. These works provide the wider intellectual context for the selected collection.
Place-based government repeatedly forgets the place it is governing. This paper argues that civic memory — the maintained public capability through which a place remembers and reuses its decisions, commitments, evidence and lived experience — should be treated as infrastructure, not heritage.
A dual-layer method for traversing the space of possible ideas — pairing intentional disciplinary collisions with a structured selection layer to convert raw novelty into usable insight.
ICLE — Inspect, Connect, Lens, Evolve — a methodology for moving an idea through successive disciplinary lenses so it matures by design rather than by accident.
A model of early-venture formation that reframes the founder's journey as a crossing between protected ideation and open competitive waters, with implications for how we mentor and fund nascent ventures.
An 'angel-in-the-loop' pattern — a constitutionally-bound human reviewer embedded in AI workflows — proposed as a more principled alternative to ad hoc human-in-the-loop arrangements.
Mathanica develops a linguistic and architectural vocabulary for describing, comparing, and reasoning about intelligent systems with the precision the field currently lacks.