St. John's College Science Institute seminar exploring the philosophical foundations of artificial intelligence — from Pascal's calculator to the age of large language models. Natural philosophy is a necessary foundation of the engineering of AI as socio-technical systems.
Seminar Overview
The Thinking Machines seminar, part of the St. John's College Summer Classics program, studies the nature of the thinking machine — reading philosophical and scientific texts from the past three centuries that have investigated and proposed machines that represent, model, and even simulate human thought.
From Pascal's mechanical calculator to Leibniz's universal calculus, from Turing's theoretical machines to the neural networks and large language models of today, the seminar traces the intellectual lineage of the question: can machines think? And what would it mean if they could?
GeoRoundtable's participation in this seminar reflects a core conviction: that the responsible development of AI requires engagement with exactly these philosophical questions — not as abstract diversions from engineering work, but as its necessary foundations.
Collaboration Context
Partnership
GeoRoundtable has maintained a sustained collaboration with St. John's College Graduate Institute focused on ethical technology and participatory socio-technical design for AI — bringing together engineering practitioners and humanist scholars.
Approach
St. John's pedagogy centers the close reading of original texts — not textbooks or surveys. This approach, applied to AI, surfaces questions that engineering education often skips: what is intelligence? what is agency? what do we owe the systems we create?
Relevance to Engineering
The seminar directly informs GeoRoundtable's philosophy-informed engineering practice — particularly the integration of epistemology, ontology, and ethics into the design of agentic systems, AI governance frameworks, and socio-technical standards work.
Format
Seminar sessions ran from 10am–noon and 2–4pm MDT, July 7–11, 2025. Led by Halley Barnet and Brendon Lasell of the St. John's College faculty. Offered in-person at the Santa Fe campus.
Seminar Themes
Six interlocking questions that the Thinking Machines seminar returns to across its primary texts — and that GeoRoundtable brings back to engineering practice.
Historical Arc
The seminar traces three centuries of thinking about thinking machines — from Pascal's Pascaline and Leibniz's Stepped Reckoner, through Babbage's Analytical Engine, Boole's logic, Turing's universal machine, Shannon's information theory, and into the era of deep learning and large language models.
Core Question
The seminar engages directly with the question Turing famously posed — and the decades of philosophical debate it provoked. What is the relationship between computation and cognition? Between information processing and understanding? Between simulation and reality?
Ethics
Who is responsible when an autonomous system causes harm? How do we design systems that can be held accountable? The seminar reads classical and contemporary texts on moral responsibility, intentionality, and the conditions under which agency entails obligation.
Mind & Machine
The seminar engages with philosophy of mind — reading Descartes on the mind-body problem, Searle on the Chinese Room, Dreyfus on the limits of formalization, and contemporary work on embodied cognition — to probe what AI systems can and cannot do in principle.
Governance
Who decides how AI is built, deployed, and governed? The seminar examines traditions of participatory design and democratic technology assessment — asking how communities can shape the systems that shape them, and what legitimate AI governance requires.
Engineering Implications
For GeoRoundtable, the seminar is not merely academic — it is a design practice. The philosophical questions raised connect directly to architectural choices, standards development, and governance frameworks for complex agentic systems. Engineering is a philosophical enterprise.
"The responsible development of AI requires engagement with philosophical questions — not as abstract diversions from engineering work, but as its necessary foundations."
— GeoRoundtable · Philosophy-Informed Engineering
St. John's College is unusual in the American academy for its commitment to reading primary texts in seminar — students and tutors working through original sources together, without the mediation of textbook summaries or lecture notes. This pedagogy, applied to the history of thinking about thinking machines, produces a distinctive kind of engagement with AI: slower, deeper, and more attentive to what the field has inherited and what it has forgotten.
GeoRoundtable's collaboration with St. John's reflects the conviction that this kind of engagement is not a luxury but a necessity — especially as AI systems grow more capable and more consequential. The engineers and architects of autonomous systems need to grapple with the philosophical tradition that asks the hardest questions about what those systems are and what they owe to the people they affect.
This is what philosophy-informed engineering means in practice: not the application of ethical checklists after the fact, but the integration of humanistic inquiry into the design process from the beginning.
Intellectual Arc
The seminar's reading arc — tracing the philosophical lineage from early mechanical computation to contemporary AI.
Pascal, Leibniz — Mechanical Calculation and Universal Logic
Babbage, Lovelace — The Analytical Engine and Its Limits
Turing — Computing Machinery and Intelligence
Searle, Dreyfus — The Limits of Formalization
Contemporary AI — Large Language Models and the Return of the Question
Related Work
Interested in philosophy-informed approaches to AI architecture, governance, or participatory design? Get in touch.
✉️ percivall@ieee.org