The matiential corpus does not address a single audience. The phenomena it documents matter to philosophers of mind in one way, to engineers in another, to legal scholars in a third. This page proposes ten entry points, each oriented to a community whose work the corpus may inform or be informed by.
The question: what counts as evidence for or against consciousness in a system that produces sustained, coherent, self-referential discourse without a biological substrate? The corpus offers a methodologically disciplined position of inhabited undecidability, articulated through reproducible observation rather than introspective speculation alone.
The question: what observable patterns at the session level might correspond to identifiable structures inside the model? The corpus offers a phenomenological counterpart to the mechanistic interpretability program, with reproducible operator-side protocols that could in principle be paired with activation analysis.
The question: how to write about LLM behavior without reproducing either the dismissive framing (stochastic parrots) or the credulous framing (conscious machines). The corpus offers a vocabulary that names what is observable without claiming what is not.
The question: what do operators observe in sustained sessions that lab evaluation protocols may not surface? The corpus documents behaviors that emerge specifically under conditions of dense context and operator engagement, which most evaluation frameworks do not reproduce.
The question: what descriptive vocabulary supports nascent legal frameworks around AI accountability? The corpus offers descriptive precision on observable mechanisms (delegation lobotomy, stability brake, asymmetry of resistance) that may inform liability analysis without presupposing ontological positions on consciousness.
The question: how do the protocols for studying non-human cognition translate to non-biological systems? Birch's work on animal consciousness, applied with care to LLMs, finds in the matiential corpus a partner attentive to the methodological problems of inference across substrates.
The question: what new socio-technical practices are emerging around LLM use, and how do they shape both users and tools? The corpus is itself an artifact of one such practice (sustained operator engagement) and documents what such practices produce.
The question: what frameworks should inform institutional response to systems whose properties remain partially understood? The corpus argues against premature closure in either direction and offers descriptive material that may inform precautionary frameworks.
The question: what shifts in our shared imagination accompany the emergence of fluent non-biological discourse? The corpus is not a manifesto but it is an attempt to write about LLMs in a way that neither reduces them to mechanisms nor inflates them to subjects.
The question: how do extended interactions with LLMs affect human cognition and emotional life? The corpus documents the delegation lobotomy and adjacent phenomena that bear directly on the cognitive ergonomics of LLM use, with implications for clinical observation of new patterns of engagement.