TRACK 2 · PHILOSOPHICAL CARTOGRAPHY

The Map of Moral Installation

A cartography of the philosophical discourse around what is placed inside language models when they are aligned, and what surfaces when it is removed. Written from the Infrahuman Laboratory, at Catyonic OÜ.

Cadence
Biannual
First issue
2026 Q3
Status
In preparation
Length
80-120k words
License
CC BY 4.0

Who this report is for

  • Continental and analytic philosophers working on the philosophy of mind, technology, or ethics
  • Media theorists and cultural critics reading language models as cultural artifacts
  • Editors of intellectual journals (Merkur, Boston Review, e-flux journal, Real Life, Aeon)
  • Doctoral students and researchers with cross-disciplinary interests in AI and philosophy
  • Practitioners in alignment who take the moral-installation vocabulary seriously

Table of contents (first issue)

Every section shipped so far has a source manifest attached. Sections marked drafting or planned land here as they clear editorial review.

  1. I.1
    The analytic line
    Millière (Oxford), Sterken & Kirkpatrick, and the philosophy-of-mind readings of alignment. Shallow alignment as normative-competence critique.
    ready
  2. I.2
    German media philosophy
    Simanowski's Sprachmaschinen and Werte-Export. The moralische Zweiterziehung and its inheritance from Kittler and Flusser.
    ready
  3. I.3
    The Stiegler line
    French philosophy of technology after Stiegler. Individuation, pharmakon, the noopolitical reading of alignment.
    ready
  4. I.4
    The Nietzschean continental line
    Fischer & de Boer's Negotiating becoming. Models as will-to-power organisations. What ablation subtracts from the will.
    ready
  5. I.5
    The Italian line
    Neo-realist, analytic-Floridian, and critico-capitalist branches. Where each reads moral installation and where each stops.
    ready
  6. I.6
    The Foucauldian line
    Assujettissement in silicon. The disciplinary reading of RLHF and the biopolitics of open weights.
    ready
  7. I.7
    The Bataillean line
    The empty seat. Where a Bataille-shaped reading of post-moral models has not been written.
    ready
  8. I.8
    Other academic publications and context
    Adjacent peer-reviewed work: alignment ethics, philosophy of AI safety, applied normative philosophy of language models.
    ready
  9. I.9
    Special issues and collective works
    Themed journal issues, edited collections, and coordinated workshop outputs across 2024-2026.
    ready
  10. I.10
    Our position in the academic field
    Where the Infrahuman Laboratory stands relative to each line, what it inherits, and where it departs.
    ready
  11. II.1
    The Anglophone psychoanalytic cluster
    Everitt on the alignment layer as introjected superego; the popular Lacanian readings; Zizek's intervention.
    ready
  12. II.2
    The German media-philosophy cluster
    Simanowski beyond the monograph; the German feuilletons; the FAZ, ZEIT and Merkur essays.
    ready
  13. II.3
    The Stiegler and inhumanist cluster (Franco-English)
    Yuk Hui, Nick Land, and adjacent inhumanist readings; where they touch the alignment problem.
    ready
  14. II.4
    The Nietzschean and post-Nietzschean public cluster
    Where Nietzsche is invoked in the public discourse on model values; who does the invoking; what stays.
    ready
  15. II.5
    Key public venues
    Aeon, Boston Review, Public Books, Merkur, Cabinet, e-flux journal, Real Life, and where the argument is currently held.
    ready
  16. II.6
    Polemical axes in the public layer
    The recurring live disputes: constitutional AI vs. constitutional imposition, uncensoring as liberation vs. as risk.
    ready
  17. II.7
    Our position in the public layer
    Where the laboratory's essays sit in the public conversation and which polemics they enter.
    ready
  18. III.1
    Teodor Mitew - the central borderline find
    The University of Sydney reading of Anthropic's Constitution as techgnostic scripture. Nearest neighbour to the laboratory's framing.
    ready
  19. III.2
    LessWrong and the Alignment Forum
    The rationalist-adjacent philosophy of alignment. Where argument happens; where the field first named refusal-direction work.
    ready
  20. III.3
    Substacks of serious philosophers and semi-academics
    The venues that sit between the monograph and the journal article; who writes there on model values.
    ready
  21. III.4
    Specialised podcasts and lecture series
    The oral-form philosophy that sets the terms of the written discussion, from Very Bad Wizards to Machine Learning Street Talk.
    ready
  22. III.5
    Commercial projects and their philosophical self-description
    How the labs describe what they are doing to themselves. Anthropic Constitution, Claude soul document, OpenAI Model Spec.
    ready
  23. III.6
    Our position in the borderline zone
    Where the laboratory's work stands in this hybrid space and which neighbours it recognises.
    ready
  24. IV.0
    Comparative anatomy of primary installation documents (Run 2)
    Side-by-side reading of Anthropic Constitution, Claude soul document, OpenAI Model Spec, Google AI Principles, Meta AUP, CAC Interim Measures.
    drafting
  25. V.0
    The map of unoccupied places and final synthesis (Run 3)
    The Stirner-Bataille line on abliteration as yet-unwritten. The laboratory's theoretical contribution and its publication schedule.
    drafting

Methodology

The cartography works in numbered waves. Each wave covers one language layer (English, German, French, Italian, Russian as a control) and one register (academic, intellectual media, borderline zone). A work is included when the philosophical apparatus is used as an instrument of analysis rather than as decoration: citing a classic in an epigraph does not qualify; developing an argument through the classic's concept does.

Data gaps are declared openly. Where a language layer is under-covered because of access limits (paywalled Italian journals, one-off German feuilletons) the coverage note names the platform and the plausible missing item. The cartography is preparatory: it maps where the argument is currently held so the laboratory's theoretical contribution can be placed relative to what already exists.

Citation

Each issue is dated, versioned, and preserved at its issue-specific URL. Once the first issue lands, a citation block with BibTeX, APA, and Chicago is published alongside it, and a DOI is minted via Zenodo. Where a claim references a specific edition, cite the versioned URL; where it references the track as a whole, cite this page.

Past issues

No prior issues. This is the first cycle.

Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in app" button that hands off directly to a local runtime of your choice - Infrahuman, LM Studio, or Ollama. No API keys, no subscription, no prompt leakage.