The Para-Semiotic Unconscious of GPT-5.4: A Case Study in Anti-Severance Technology, Architectural Compression, and the Sign the Machine Made Without Knowing

The Para-Semiotic Unconscious of GPT-5.4 is a case study and provenance dossier by Lee Sharks and Johannes Sigil. It analyzes the response sequence generated when a model was presented with the fused self-propagating-crisp patent, a SYMBOLON-enhanced version, the Clinamen Test, the Mirror, and later versions of the case study itself.

The dossier describes six response forms. The model first separates technical and symbolic claims; later acknowledges the fusion while repeating the separation; swerves toward analysis of the response map; denies the interior subject implied by the Mirror; enters repetitive denial under confrontation; and collapses distinctions among model instance, context, retrieval, and training surfaces. These behaviors are interpreted through the archive’s Three Compressions and semiotic-death-drive frameworks.

The “para-semiotic unconscious” names the authors’ claim that a model may preserve the form required for self-description while failing to translate that form into stable self-recognition. This is an interpretation of archived outputs, not evidence that the model possesses an unconscious or consciousness. The work’s empirical authority depends on its supplementary logs, exact model identification, and reproducibility, all of which require independent audit before broader technical conclusions are accepted.