Voting and Expertise Weighting

docs/03-governance/00-overview.md names the Adaptive Governance Framework (AGF) — the network’s layered governance system — as combining “embedded representation and redundancy protocols” with the Dynamic Enterprise Agreement (DEA) — a versioned operating agreement replacing fixed bylaws — but does not describe a ballot mechanism. The only vote-weighting language already in the corpus is at the SEP (Strategic Enterprise Partnership — a joint venture between teams) level: docs/02-structure/03-strategic-enterprise-partnerships.md §2.3 lists “Proportional: Voting power or influence can be based on stake, effort, or domain expertise” among SEP governance structures. That describes voting between UMEs (United Micro Enterprises — small, self-managing venture teams, ≤ ~15 people) party to a joint venture. This section describes a different relationship — voting among the members of a single UME — and does not restate or extend the SEP-level material.

Source and provenance. As in §3.1, this section is drawn from ~/code/papr/misc/pio-001-p-context.md and ~/code/provide.io/content/english/governance.md, both read in full, plus two further documents checked for this topic specifically: ~/code/papr/misc/form.md and ~/code/papr/bits-form.md, near-duplicate drafts of the same pitch letter (confirmed by diff to differ in four minor wording points only). All four are unreviewed working notes. pio-001-p-context.md carries the only detailed structural description of the mechanism found across the four; form.md and bits-form.md each contribute a single confirming sentence and no further structural detail.

A. The Hybrid Ballot

pio-001-p-context.md states: “provide.io, and the ICN, are democratic entities. They will leverage a combination of one person, one vote, and a ranked/weighted, systems.” The same source continues: “This model will enable domain relevant systems experts to have more influence over a decision that is more relevant to their expertise. Yet allow for someone in a completely different specialization, such as education, to vote on decisions which need to be made which may have more of a systemic impact.”

As stated, the ballot combines two layers: a flat one-person-one-vote baseline, and a ranked or weighted layer that gives subject-matter experts more influence on decisions within their own domain. The flat layer is preserved for decisions the source describes as having “more of a systemic impact” — cross-cutting or network-wide questions, by contrast with domain-specific ones — where a member outside the relevant domain retains a full vote. The source’s own example contrasts a systems expert’s domain-relevant weight against an education specialist’s vote on a decision of broader, systemic reach.

governance.md corroborates that the mechanism applies at the UME level, in a single unelaborated sentence: “The members of the Cell will have weighted voting rights within the provide.io/ICN systems.” It adds no structural detail beyond confirming that weighted voting exists at the unit level. form.md and bits-form.md — identical at this point in their text — confirm the same pitch point in a sentence with an apparent dropped word: “The ability to vote throughout the organizational network while having a weighted system in place to ensure that subject matter experts are weight[ed more] than someone who has no clue what they’re voting for,” which this section reads as stating that subject-matter experts are weighted more than participants without relevant expertise. Neither document describes the one-person-one-vote half of the hybrid, the domain-versus-systemic distinction, or the learning purpose treated in §3.2.B below — those structural details appear only in pio-001-p-context.md.

None of the four sources specifies the mechanics of the weighted layer: no formula for computing a weighted score, no process for certifying who counts as a domain-relevant expert for a given decision, and no rule for combining the flat and weighted layers into a single outcome. This is a gap in the source material, not an omission of this section.

B. Anomalous Votes as a Learning Signal

pio-001-p-context.md states, in the sentence immediately following the domain-versus-systemic description quoted above: “This will also enable the SMEs to learn more about anomalous votes.” This is the only statement of this purpose found across the four sources, and it is the entirety of what the source material says about it.

The source frames this as a secondary effect of the weighting mechanism — “also enable” — rather than as its stated primary purpose, which is the influence-weighting described in §3.2.A. As stated, the mechanism’s direction of effect runs from members to experts: a vote that diverges from what subject-matter experts would expect is treated as something for those experts to learn from, not as a vote to be corrected, discounted, or excluded. Nothing in the source frames an anomalous vote as an error to be suppressed or as evidence against the voter; the sentence names only a learning benefit to the SMEs (subject-matter experts) who observe the divergence.

Beyond that single sentence, the mechanics are unstated. The source does not define what counts as “anomalous” — a numeric deviation from an expert-weighted result, a categorical disagreement, or something else. It does not describe what SMEs do with an anomalous vote once observed: whether it triggers a review, a conversation with the voter, a change to how the decision is explained, or nothing beyond passive observation. And it does not state whether the signal is captured or routed through any existing mechanism — for instance, the AGF or Value Alignment Monitoring (VAM) — ongoing checks that actions match stated principles, described in docs/11-deployment/02-provide-io-as-deployment-engine.md. No source names either as the channel for this signal; a connection to either would be an addition this section does not make.

This section, together with §3.1, is the first place in this corpus that states a concrete rule for how votes are weighted at the UME level. It supplies the flat and weighted ballot layers and the learning purpose attached to divergence between them; it does not supply the weighting formula, the expertise-certification process, or the anomalous-vote follow-through, because the source material does not supply them either.