Chapter 6 of 12 · 2 min read
Attribution and Contribution Graphs
A model for tracing participation without collapsing every contribution into money.
Word Wallet Web · v0.2 · Moses Sam Paul
Attribution and Contribution Graphs
From leaderboards to evidence
Contribution is relational. A researcher may define a construct, a designer may make it understandable, a participant may reveal a failure, and a maintainer may preserve the service. A single score erases these differences.
A contribution graph represents claims as edges between participants, artifacts, expressions, versions, and institutions. Each edge can carry a role, time, evidence reference, consent boundary, and review status. The graph records *what relationship is claimed* rather than announcing a final ranking of people.
Classes of participation
Useful classes may include authorship, facilitation, implementation, testing, stewardship, translation, critique, governance, and informed participation. They are not interchangeable. Communities should define what evidence is sufficient for each class and how a person can contest or remove a claim.
Receipts are not biographies
An invocation receipt can prove that an authorized event occurred without retaining the participant's raw text. A contribution receipt can identify an artifact and role without publishing private workshop material. Data minimization is therefore part of attribution design, not a later compliance layer.
The Phase 1 service uses HMAC-based receipts and sanitized saved observations. These mechanisms reduce unnecessary exposure, but they do not make every downstream inference safe. Access controls, retention limits, and deletion remain necessary.
Value remains plural
Some contributions may later participate in governed value flows. Others should never be monetized. Care, disclosure, dissent, and identity workshop material are particularly vulnerable to coercion when payment becomes the default frame.
The proposed graph therefore separates recognition, access, governance voice, and financial distribution. Any future conversion rule should be visible, versioned, appealable, and tested for gaming and exclusion.
Research questions
- Can contribution claims remain useful without becoming a surveillance graph?
- Which roles can be verified without exposing sensitive evidence?
- How should collective, inherited, and anonymous contributions be represented?
- What dispute process is credible across unequal institutions?