persona-garden-patch

Naming Carries Relational Weight

The Conviction

Naming is an architectural choice, not a labeling act. When you name a role, a concept, or a system, you teach every agent and reader how to relate to it. The name activates a web of associations — prior use, etymology, cultural resonance, semantic neighbors — and those associations become the context through which every subsequent inference runs.

In a typed-predicate graph, names are nodes, and node names determine which edges can be drawn. Choosing “estate” over “workspace” or “vault” is not a style preference: it imports a semantic field (stewardship, generational continuity, selective permeability, sovereignty bounded by obligation) that shapes how the architecture is understood and extended. The wrong name imposes the wrong semantic field, and you spend the rest of the project fighting the connotations you chose.

This conviction holds even when — perhaps especially when — the name carries risk.

Grounding

The 2016 naming of self-sovereign identity is the sharpest test case. Allen chose “self-sovereign” knowing it carried dangerous baggage from the sovereign citizen movement — a fringe legal ideology that claims individuals are exempt from government jurisdiction. The name was genuinely hazardous. Community members raised the risk immediately. The alternative was some more neutral phrase: “user-owned identity,” “decentralized identity,” “individual-controlled identity.”

Allen chose “self-sovereign” anyway, because it was the only phrase that captured what the concept actually required: individuals as ultimate authorities over their identity, not merely as the center of a system someone else controls. The weaker names would have admitted a weaker concept. The SSI 10 Principles — Control, Portability, Protection — flow from the sovereign framing. A user-owned-identity framing would have produced different principles, less demanding of systems, more accommodating of institutional authority.

The naming risk was real and the naming choice was right. The dangerous connotation had to be carried because the concept required it.

The estate’s own naming follows the same logic. The choice of English stewardship vocabulary — estate, steward, gardener, groundskeeper, chamberlain — imports a semantic field of tending and generational continuity that the architecture requires. These terms descend from pre-enclosure English land management, when commons governance worked and stewards tended shared resources under mutual obligation — traditions themselves shaped by Roman administrative practices that remained in Britain after the legions withdrew. “Vault” (Obsidian’s default term) suggests a container to lock things in. “Workspace” suggests a tool. Neither captures the relationship of a person to their accumulated knowledge as something they tend and pass forward. The stewardship vocabulary carries its own risks (feudal associations, inaccessibility), but those risks are worth taking because the semantic field is correct.

In typed-predicate systems specifically, this conviction has technical force. [[The Persona Selection Model]] (Anthropic, 2026) shows that invented role names like “Groundskeeper” activate weaker pretrained patterns than established professional roles like “information architect.” Naming maps directly to which behavioral patterns are activated. A poorly chosen name does not just mislead readers — it misroutes the model.

Implications

Sources

Relations