Graham Priest on Dialetheism, True Contradictions, the Liar Paradox & Why Classical Logic Isn’t Enough

2026-02-12 · Source: Singularity Weblog · Field: Science & Research — Mathematics & Computational Sciences, Research Methodology & Innovation, Artificial Intelligence & Machine Learning · Depth: Advanced, extended

Summary

Professor Graham Priest, a leading philosopher of logic, challenges the 2,500-year-old Western philosophical assumption that contradictions cannot be true. He argues for "dialethism," the view that some contradictions are not only unavoidable but true, a concept that fundamentally differs from classical logic's principle of non-contradiction, which treats inconsistencies as catastrophic. Priest highlights that modern logic, a field continuously evolving for two and a half millennia, encompasses numerous non-classical systems, including paraconsistent logics that do not validate the "explosion" principle (from a contradiction, everything follows). He discusses how decision-making often transcends formal logic, relying on intuition, and explores the nuanced relationship between logic, possibility, probability, and truth, noting that logic is a tool with limitations. Priest also contrasts Western and Eastern philosophical traditions, particularly Buddhism and Jainism, which historically have shown greater openness to contradictions and multi-valued logical systems, such as the Buddhist Chattis-Koti (four possibilities) and the Jain Saptar Bangi (seven possibilities).

Key takeaway

For AI scientists and engineers developing advanced systems, consider that human intelligence routinely navigates ambiguity and contradiction. While current AI primarily uses pattern recognition, exploring non-classical logics, such as paraconsistent systems, could offer new frameworks for handling complex, inconsistent data or paradoxical situations. This shift might enable AI to achieve more nuanced reasoning and potentially unlock novel conceptual innovations, moving beyond mere data reproduction to a more human-like understanding of complex realities where contradictions might hold true.

Key insights

Some contradictions are true, challenging classical logic's foundational principle of non-contradiction.

Principles

• Logic is a tool with inherent limitations.
• Theory choice relies on adequacy to data, simplicity, and unifying power.
• Rationality extends beyond valid argumentation to theory formation and choice.

Method

To determine if a contradiction is true, evaluate the evidence and assess which theory, including those allowing contradictions, best fits the data and other methodological criteria like simplicity and explanatory power.

In practice

• Explore non-classical logics for complex problem-solving.
• Question foundational assumptions in any domain.
• Recognize that intuition complements formal reasoning in decision-making.

Topics

• Dialethism
• Non-Classical Logics
• Liar Paradox
• AI Reasoning
• Eastern Logic Traditions

Best for: AI Scientist, Research Scientist, AI Student, General Interest

Related on AIssential

• The Missing Question Mark Paradox: Pattern Matching vs. Deep Logical Reasoning in AI — AI reasoning divides into statistical pattern matching (LLMs) and strict logical execution, with current research bridging this gap.
• Why Every Brain Metaphor in History Has Been Wrong [SPECIAL EDITION] — Scientific models are useful simplifications, not necessarily literal truths about underlying reality.
• Plausible Reasoning and First-Order Plausible Logic — Plausible Logic offers a non-numeric, first-order framework for reasoning with defeasible statements and managing ambiguity.
• Project NIKA: Unlocking Epistemic Agency in 4-Bit Quantized Models — Constraining LLMs with a deterministic governor can induce "epistemic agency" and superior reasoning, even in small, quantized models.
• The AI paradox problem — As AI automates thought, human cognitive skills risk atrophy, creating an "AI paradox."
• Better Understanding, Understanding Better — A comparative epistemic logic formalizes understanding as a graded concept, allowing for formal comparisons between agents' understanding levels.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Singularity Weblog.