A structured language for precision encoding

CIE expressions are built from two separable layers. The grammar defines how tokens are composed, how clauses are structured, and how meaning derives from their combination. The Registry — a managed, versioned vocabulary — defines what those tokens resolve to: the specific named codes for actors, institutions, positions, actions, domains, and locations that give CIE expressions their semantic content.

An expression is syntactically valid by reference to the grammar alone. It is semantically valid from its resolution by the Registry. This separation is foundational: it allows the grammar to be stable while the vocabulary can grow, and it ensures persistent consistency of exact meaning, independent of the time, place, or circumstances of the encoding.

The examples below are drawn from real events across different action types and information environments. They are not edge cases chosen to show the system's range — they are representative of what a working analyst encodes on any given day. The complexity varies; the underlying logic is the same throughout.

A bilateral meeting between heads of state with senior ministerial participation, covering multiple defined domains, at a specified location, over a bounded time period. This is among the most common event types in political encoding — and the one where prose reporting most reliably loses the structurally important detail.

A defense ministry spokesperson issues a public description of armed incidents at an international border, including approximate coordinates, unit-level force identifications, and casualty data. This encoding illustrates how CIE handles geographic precision, the separation of the statement act from the events being described within it, and quantified casualty reporting in a subsidiary clause.

A government meteorological agency issues a seismic event warning and a derived tsunami warning for named coastal prefectures. This example shows CIE at its most compressed — a high-information, time-critical event encoded in two subsidiary clauses. The brevity is not a limitation. When the event is discrete, the encoding needs nothing more.

Not all political information arrives publicly confirmed. A substantial share of what matters in intelligence and risk analysis comes through channels that require explicit sourcing and credibility assessment before the underlying claim can be used analytically. CIE encodes RUMINT as a first-class information type — with its own metadata header, its own credibility vocabulary, and its own rules for how source claims interact with the confirmed event record. The structure below is not a workaround. It is a designed capability.