Documentation / Concepts

CongLang

CongLang is the internal query and reasoning language used within Congregation DB.

The language operates over contextual decision graphs where enterprise signals remain separated by type and provenance.

CongLang allows reasoning queries across:

Deterministic signals

Signals that can be asserted directly from source systems while retaining lineage.

Probabilistic signals

Signals that carry weighting, confidence or inferred interpretation rather than direct assertion.

Derived intelligence

Signals produced through synthesis, interpretation or graph-based decision reasoning.

Language path

Conceptual language flow

CongLang

The internal language surface for contextual query and reasoning expression.

CongIR

A stable reasoning representation that preserves semantics before execution.

CongSynth

The context synthesis engine where provenance, relationships and dependency structure remain intact.

Decision Intelligence

Downstream systems consume outcomes with reasoning context preserved.

CongLang expressions move through CongIR before reasoning through CongSynth and producing traceable downstream context.

Design philosophy

Design principle

Queries should never destroy the meaning or origin of the signals they operate on.

preserve signal provenance

maintain separation between signal types

support contextual reasoning

enable provable decision outputs

Example CongLang Query

CongLang examples are conceptual. They illustrate how reasoning can move across graph structure, provenance and mixed signal classes without exposing implementation details.

CongLang

Portfolio exposure query

MATCH portfolio_company
WHERE sector = "energy"

RELATE supply_chain_partner

RETURN exposure_risk
WITH provenance

In plain language, this query starts from portfolio companies in the energy sector, follows their supply-chain relationships, and returns an exposure-risk view with provenance still attached to the result.

CongLang

Mixed-signal reasoning query

MATCH portfolio_company
WHERE covenant_status = "active"

RELATE demand_signal
WHERE confidence > 0.72

RETURN continuity_risk
WITH deterministic_signals, probabilistic_signals, provenance

This example combines deterministic operating facts with a probabilistic demand signal so reasoning can occur across both classes without collapsing them into a single undifferentiated result.