Perfinitive Research
WMM-01 — Human-First Water Mapping Method
A systematic exploratory method for identifying consequential but poorly visible structures in human-AI environments.
Making the surrounding condition visible
Many important AI-mediated effects are easier to experience than to see.
A person may notice that a new assistant no longer “works the same,” that deleting memory did not restore an earlier interaction, that AI-assisted performance did not survive without assistance, or that a manager acts on a summary the worker never saw.
Perfinitive Research studies these surrounding conditions: the persistent, distributed, mediated, and often poorly legible structures that increasingly shape human experience.
The central question
What consequential condition is a human experiencing because of the surrounding human-AI system that may remain difficult to see, name, control, or explain?
Discovery operators
- Asymmetry
- What can one side know, retain, infer, alter, reproduce, transfer, or control that the other cannot?
- Persistence
- What continues to matter after the immediate interaction ends?
- Carrier
- Where is the consequential effect carried?
- Boundary
- What crosses sessions, tasks, models, providers, people, institutions, or jurisdictions?
- Irreversibility
- What visible action can be reversed without restoring the prior condition?
- Adaptation
- How does each participant change because of repeated interaction?
- Intermediation
- What changes when AI stands between humans, institutions, audiences, or decisions?
- Authority
- Who can inspect, decide, correct, contest, revoke, transfer, verify, or exit?
- Human Consequence
- What happens to capacity, agency, judgment, attribution, correction, dependence, mobility, legibility, opportunity, or exposure?
- Assumption Inversion
- What becomes visible when the normalized description of the system is reversed?
The Legibility Gap
Where is the human experiencing a consequential effect without having a correspondingly clear representation of the system condition producing it?
The ten WMM operators are then used to diagnose why the gap exists.
People-first filter
A candidate structure advances only if understanding it could materially change a person’s ability to understand or exercise capacity, agency, authority, attribution, correction, mobility, independence, or control over consequential records and decisions.
Water test
A candidate is especially important when the consequence is easier to experience than the mechanism is to perceive.
Priority properties: ambient, cumulative, distributed, indirect, delayed, opaque, normalized, difficult to reverse, difficult to attribute.
Methodological boundary
WMM-01 may discover a candidate structure. WMM-01 cannot establish its causal mechanism. The method requires empirical validation; candidate novelty remains subject to literature review.