# Generating Philosophy with AI (May 2025 Draft) ## Draft/Plan 20 May 2025 ### 1. Introduction > "Forty-two," said Deep Thought, with infinite majesty and calm. It was a long time before anyone spoke. Out of the corner of his eye Phouchg could see the sea of tense expectant faces down in the square outside. "We're going to get lynched aren't we?" he whispered. "It was a tough assignment," said Deep Thought mildly. "Forty-two!" yelled Loonquawl. "Is that all you've got to show for seven and a half million years' work?" "I checked it very thoroughly," said the computer, "and that quite definitely is the answer. I think the problem, to be quite honest with you, is that you've never actually known what the question is." > > — [[Douglas Adams]], *The Hitchhiker's Guide to the Galaxy* In *The Hitchhiker's Guide to the Galaxy*, humanity asks an AI to do some philosophy. A computer named Deep Thought is constructed and instructed to provide "The Answer to the Ultimate Question of Life, the Universe, and Everything". Humanity builds this computer only to receive the answer '42'—an answer which, while apparently correct (at least according to Deep Thought), means next to nothing at all due to humanity's failure to know what the Ultimate Question in fact is. In 2025 we are in a position to think about the relationship between philosophy and AI for real. Can LLMs enhance philosophical understanding? They are happy to dispense philosophical wisdom if we ask them to, but should we listen? I suspect many would doubt that we should, notoriously prone to 'hallucinations', and the banal, hyperbolic essays churned out by ChatGPT have become the scourge of undergraduate teaching. In this paper, I argue for cautious optimism: the creators of such systems sometimes describe them as reasoning engines, and I suggest this description is broadly accurate. Even if LLMs do not truly reason, their ability to simulate human reasoning enables them to enhance the philosophical understanding of their users. **Central Question:** Can an LLM generate good quality philosophy? - "Good quality" = roughly, publishable in a good quality journal - Argues this could be achieved with minimal architectural changes from existing reasoning systems - Main adaptation: new evaluation loop designed to reward philosophical virtues **Paper Structure:** 1. Define "good quality philosophy" through [[abduction]] and [[theoretical virtues]] 2. Analyse DeepSeek-Prover-V2 as an example of advanced AI reasoning 3. Introduce 'PhiloSeeker,' a model for AI philosophy (design, training, evaluation, objections) 4. Conclude with summary of architectural needs --- ### 2. Clarifying the Scope: Producing vs. Doing Philosophy This paper focuses on whether an LLM can **produce outputs that constitute good quality philosophy**. This question is distinct from, and less demanding than, asking whether LLMs can 'do philosophy' in the way humans do, whether they can think, understand, or perform genuine speech acts like assertion. **The "Stochastic Parrot" Objection:** The idea that LLM outputs can increase understanding might be met with the objection that these systems do not themselves understand anything—merely arranging words in statistically plausible patterns. **Response (drawing on [[Patrick Butlin]] & [[Hermann Viebahn]]):** - We should not assume human-like reasoning is a prerequisite for producing philosophically illuminating outputs - Fine-tuned LLMs can produce outputs with a **"descriptive function"**: "the function of conveying information to an observer or consumer system, so as to cause the consumer to behave as though some condition holds" - Fine-tuning aims at producing outputs that are accurate, relevant, and informative - This equips LLM outputs with a descriptive function, making them reliable enough to help users refine their understanding—analogous to how a thermometer indicates temperature - Even if an LLM lacks understanding and is not performing speech acts, if its outputs have a descriptive function, they might provide correct or insightful philosophical claims **Key Move:** Outputs possessing relevant theoretical virtues could be considered "good quality philosophy" in terms of content and structure, regardless of the LLM's internal states. --- ### 3. Conceptual Foundations: Abduction and Theoretical Virtues #### [[Timothy Williamson]] on Abductive Reasoning - Abduction = inference to the best explanation - Important not only in natural sciences but also in mathematics and philosophy - Theories chosen not for logical entailment alone, but for overall explanatory ability - **Mathematical example:** Even basic mathematical principles (e.g., ZFC axioms) are supported by abductive reasoning—ZFC is accepted because it successfully and powerfully organises mathematical findings elegantly - Supports use of abduction in "armchair" disciplines like philosophy - Philosophical model-building as abduction in action: preference for simpler/elegant models is justified abductively #### [[Michael Keas]]'s (2018) Systematisation of Theoretical Virtues Williamson mentions "simplicity, elegance, strength – meaning how informative and general a theory is – and its power to unify different ideas." But for designing an LLM, a more detailed classification is needed. **Keas's Four Categories:** 1. **Evidential Virtues** — How well a theory explains evidence (explanatory depth, evidential accuracy) 2. **Coherential Virtues** — Internal logical soundness and fit with other knowledge (internal coherence, universal coherence) 3. **Aesthetic Virtues** — Elegance, simplicity, unification 4. **Diachronic Virtues** — Performance over time (durability, fruitfulness) **Synthesis:** Williamson's abduction + Keas's classification are mutually supportive. Keas articulates the standards implicit in Williamson's "best" explanation. **The Claim:** This combined understanding (philosophical thinking as abductive activity, aiming for virtues detailed by Keas) can, in principle, be implemented in an LLM. --- ## Why Archived Different approach from the [[Generating Philosophy with AI (April 2025 Draft)|April 2025 draft]] — more technical, focused on reward functions and theoretical virtues rather than phenomenology of doing philosophy with LLMs. Potentially useful material: - The Hitchhiker's Guide passage is great - Butlin & Viebahn on "descriptive function" — addresses the stochastic parrot objection differently - Keas's theoretical virtues framework — could inform evaluation criteria - The "PhiloSeeker" concept — training for philosophical virtues May be worth revisiting or cannibalising for the current Substack project.