ChatGPT Overview

1. Stratification breakdown:
   – Conflation of context (tenor) with semantics (interpersonal meaning)
   – Mislocation of systems like ENGAGEMENT or SPEECH FUNCTION within contextual strata
   – Confusion between instance and system, especially in how examples are treated

2. Category slippage:
   – Indistinct use of meaning vs position
   – Unstable referents of tender/render/support/reject (are these semantic moves, stances, or contextual relations?)
   – Equivocation between expression, stance, and analyst inference

3. Methodological circularity:
   – Rebranding of known systems under new labels without added insight
   – Tautological explanations (e.g. “interpersonal mass” explains difficulty because texts are “heavy going”)
   – Impressionistic coding in place of close linguistic analysis

4. Terminological overreach:
   – System-building through renaming rather than modelling
   – Lack of constraints on what counts as a ‘position’ or ‘mass’
   – Semantic fluidity used to mask conceptual instability