Working Paper 04 — What the WP01–WP03 program actually is and what it can legitimately claim to be

This document revises Working Papers 01–03 of the research program Discursive Signatures and Uses of Large Language Models.It clarifies the epistemic status of the series: the previous papers introduce analytical categories and testable hypotheses but do not yet constitute empirically validated results.The paper specifies the limits of the available corpus, distinguishes exploratory modelling from empirical validation, and defines the conditions under which the proposed protocol could become testable. This working paper is part of the research program "Discursive Signatures and Uses of Large Language Models". For a structured presentation of the program, see:https://yogatypic.github.io/batir-linvisible/andResearch Program Overview — https://doi.org/10.5281/zenodo.18794488