Software historically organized human coordination around applications and interfaces. If language is the new API, coordination is no longer bounded by software, but by meaning.
Personal agents, messaging apps, feeds, and CRMs are products shaped by historical software boundaries, not fundamental primitives. They force human behavior into predefined interfaces and workflows. Systems built on language can instead organize around what people mean, want, know, offer, and agree to.
The underlying building blocks are coordination primitives: identity, context, intent, trust, negotiation, agreement, delegation, and the many ways humans exchange information and act together.
This map is an evolving attempt to identify, classify, and track them. It groups them by speech act, meaning what saying something does: assertives, directives, commissives, expressives and declarations, plus speakers (who can act) and sequences (chains of acts). Each primitive also keeps one of 8 layers (Foundation, Context, Intent, Interaction, Trust, Value, Coordination, Agents), available as a filter. The map also lists gaps: acts the grouping predicts but no entry covers yet. Each entry records a definition, legacy equivalent, maturity, key instantiations, and composition relationships, along with an estimated formation rate and market size. Formation rates, market sizes, and maturity levels are speculative: we don't have a methodology for them yet.
v2 is a hypothesis about what would make the v1 list derivable. It reclassifies the same entries into five kinds (parties, attitudes, acts, positions, records), five operators (aggregate, stake, condition, delegate, scope), one loop, and one underlying relation (reliance). It also marks entries v1 leans on without listing, and empty cells the grammar predicts.
v2/grammar.js: the classification, keyed by primitive name. It overlaysdata.jsand copies nothing from it.v2/index.html: rendering.
v2 adds no figures, sources, or instantiations. Renaming a primitive in data.js means renaming its key in v2/grammar.js.
data.js: primitive data (layers, primitives with their speech act, speech-act groups, gaps, maturity levels, map metadata)index.html: rendering, filtering, interaction, lensesog.png: link preview image
No build step. No dependencies. Static HTML served via GitHub Pages. Counts shown on the map are derived from data.js.
Edit data.js to add, update, or reclassify primitives, or to add sources for existing estimates. When you review an entry against sources, set reviewed and confidence. Update META.lastRevised when you change the map.
| Field | Description |
|---|---|
name |
Primitive name |
pillars |
Language, Value, and/or Coordination |
formationRate |
Estimated share of the relevant activity already running through programmable, machine-readable forms of the primitive. Speculative unless sourced. |
tamUnlocked |
Market size. Speculative: there is no methodology for it yet. Never summed across entries. |
maturity |
Theoretical, Experimental, Emerging, Growing, or Established. Speculative: there is no methodology for it yet. |
definition |
What the primitive represents |
legacyEquivalent |
How the same coordination happens today without programmable primitives |
instantiations |
Projects and protocols implementing the primitive |
composesWith |
Other primitives it naturally combines with |
Optional fields:
| Field | Description |
|---|---|
sources |
[{ label, url }]: public sources supporting the estimates |
confidence |
High (multiple independent recent sources), Medium (one credible source or a derivation with stated assumptions), or Low (no source). Defaults to Low when no sources are given. |
reviewed |
Date (YYYY-MM-DD) the entry was last checked against sources |
kind |
Abstraction-level flag for entries that are implementations, mechanisms, actors, or market categories rather than primitives |
taxonomyNote |
Why the entry is flagged and what the underlying primitive might be |
overlapNote |
How the entry differs from overlapping primitives |
Do not add figures, sources, or citations you cannot link to.