Documentation / Concepts

Truth Lanes

Congregation DB separates enterprise signals into distinct lanes.

Mixed decision environments contain more than one kind of truth. Some signals are directly asserted, some are inferred, and some emerge only after interpretation across multiple inputs.

Separating signals preserves meaning and allows reasoning to occur without collapsing provenance.

Lane structure

Lane Model

Signals remain separated inside distinct lanes so downstream reasoning can combine information without collapsing meaning.

TRACEABLE REASONING SURFACEDETERMINISTICasserted stateverified recordsource tracePROBABILISTICweighted signalconfidence pathscored inferenceDERIVEDinterpreted contextsynthesised viewdecision cue
Each lane preserves a different class of truth rather than forcing a single undifferentiated model.
Only selected paths are emphasised, showing that reasoning can traverse lanes while provenance remains legible.
The lower band represents a shared reasoning surface built on separated signals rather than merged ambiguity.

Signal types

Deterministic signals

Signals that can be directly asserted from source systems.

Probabilistic signals

Signals representing weighted or confidence-based inference.

Derived intelligence

Signals generated through interpretation or synthesis.

Why lanes matter

Signal Separation

Keeping signal classes distinct preserves their meaning and makes mixed reasoning environments more auditable.

Distinguishing signal classes prevents deterministic fact from being conflated with inference. It also allows downstream systems to reason over heterogeneous information while retaining the ability to inspect how that information was formed.