Theory Audit — Field Note 02: Legibility Drift

Phase 2 — Theory Audit

Audit question

Is Legibility Drift anything more than familiar jargon, expert compression, the curse of knowledge, grammatical abstraction and status signalling? If a residual remains, is it the narrower condition in which readers can follow the concepts yet cannot reconstruct the actors, actions and event being described?

37 papers across 6 folders

Folder A: Specialist compression, jargon and discourse communities (6 papers). Folder B: Expertise, common ground and the curse of knowledge (6 papers). Folder C: Situation models, context, abstraction and concreteness (6 papers). Folder D: Nominalisation, grammatical metaphor and compressed academic prose (6 papers). Folder E: Processing fluency, expertise, power and status signalling (6 papers). Folder F: Longitudinal drift and AI-assisted scholarly language (7 papers).

Verdict: D — Observational diagnostic retained, narrowed

Legibility Drift is not a newly discovered causal mechanism. Its components are well established across jargon, expertise, text comprehension, grammar and status research. The useful residual is a field-level diagnostic: internally coherent explanation can become less referentially recoverable even when intended readers understand its vocabulary. The note survives because it notices a gap that readability and jargon measures usually miss — not whether the reader knows the words, but whether the reality described remains reconstructable.

Provisional definition

A longitudinal reduction in the referential recoverability of professional explanation: as shared language becomes more compressed, abstract or recursively self-referential, intended or boundary audiences become less able to reconstruct the actors, actions, conditions and outcomes being described, even when the vocabulary remains intelligible and the claims are not factually false.

Minimal mechanism chain (6 stages)

Shared work → Compression → Audience narrowing → Referential thinning → Recovery loss → Social reinforcement. The sequence is conditional; shared language can remain highly referential and without time-series evidence the case is opacity or abstraction, not drift.