Patrick J. Hayes

Patrick J. Hayes (born 1944) is a British-born computer scientist and philosopher whose foundational contributions span artificial intelligence (AI), cognitive science, and logic. He is best known for his work on knowledge representation, naive physics, the frame problem, and non‑monotonic reasoning. Hayes held academic positions at the University of Essex, the University of Illinois at Urbana–Champaign, and the Florida Institute for Human & Machine Cognition (IHMC). He is a Fellow of the Association for Computing Machinery (ACM) and the American Association for Artificial Intelligence (AAAI).

1 Early life and education

1.1 Childhood and early influences

Patrick J. Hayes was born in England in 1944. Little is publicly recorded about his early family life, but his intellectual development was shaped by a strong interest in mathematics and logic from adolescence. He was influenced by the burgeoning field of cybernetics and the writings of philosophers such as Ludwig Wittgenstein, which later informed his interdisciplinary approach.

1.2 Undergraduate studies at the University of Cambridge

Hayes studied at the University of Cambridge, where he earned a Bachelor’s degree in mathematics. During his undergraduate years, he was exposed to the work of Alan Turing and the early development of computing, which solidified his intention to pursue a career at the intersection of logic, philosophy, and machine intelligence.

1.3 Doctoral work at the University of Edinburgh

Hayes completed his Ph.D. at the University of Edinburgh in 1973. His doctoral dissertation, titled *The Frame Problem and Related Problems in Artificial Intelligence*, examined the challenge of representing the effects of actions without explicitly listing all unchanged conditions. This work laid the foundation for his later contributions to common‑sense reasoning and non‑monotonic logic. His supervisors included John McCarthy and Donald Michie, both prominent figures in AI.

2 Academic career

2.1 University of Essex (1972–1984)

After receiving his doctorate, Hayes joined the University of Essex as a lecturer in computer science. During his twelve years there, he developed courses in AI and logic, supervised several doctoral students, and began collaborating with other researchers on knowledge representation. His 1979 paper “The Naive Physics Manifesto” was written during this period, establishing him as a leading voice in common‑sense reasoning.

2.2 University of Rochester (visiting)

In the early 1980s, Hayes spent a visiting period at the University of Rochester. There he worked with researchers in cognitive science, exchanging ideas about mental models and the representation of spatial knowledge. This visit broadened his perspective on the intersection of AI and psychology.

2.3 University of Illinois at Urbana–Champaign (1984–1996)

Hayes moved to the University of Illinois at Urbana–Champaign in 1984 as a professor of computer science. He became a core member of the Beckman Institute for Advanced Science and Technology, contributing to the laboratory’s interdisciplinary AI research. During this time, he deepened his work on non‑monotonic reasoning, co‑authored key papers on the frame problem, and mentored a generation of AI researchers.

2.4 Florida Institute for Human & Machine Cognition (1996–retirement)

In 1996, Hayes joined the Florida Institute for Human & Machine Cognition (IHMC) in Pensacola, Florida, as a senior research scientist. He remained at IHMC until his retirement, focusing on conceptual spaces theory, metaphor in AI, and the development of cognitively inspired knowledge representation systems. His later work at IHMC solidified his reputation as a bridge between AI and cognitive science.

3 Research contributions

3.1 Artificial intelligence and knowledge representation

Hayes’s research in AI centered on how machines can represent and reason about the world. He advocated for formal, logical frameworks that capture common‑sense knowledge, challenging early AI approaches that relied solely on ad‑hoc programming.

3.1.1 The frame problem and common‑sense reasoning

3.1.1.1 Formalization of the frame problem

The frame problem asks how a reasoning system can efficiently determine which facts remain unchanged after an action. Hayes formalized this problem by showing that a naive logical representation of actions leads to an explosion of “frame axioms.” His work demonstrated that special non‑monotonic reasoning mechanisms—such as circumscription or default logic—are necessary to handle persistence of facts without explicit enumeration.

3.1.2 Naive physics and the “naive physics manifesto”

3.1.2.1 Everyday physical knowledge representation

In his 1979 paper “The Naive Physics Manifesto,” Hayes argued that AI systems must be equipped with a comprehensive theory of everyday physical knowledge, such as how objects fall, liquids flow, or containers hold. He proposed building large, formalized ontologies of “naive physics” that capture the qualitative reasoning people use effortlessly. This manifesto influenced subsequent work in qualitative physics and common‑sense knowledge bases.

3.2 Logic and philosophy

Hayes contributed to the philosophical foundations of AI, particularly in understanding the limitations of classical logic for representing uncertain, changing, or context‑dependent knowledge.

3.2.1 Non‑monotonic logic and default reasoning

Classical logic is monotonic—adding new premises never invalidates old conclusions. Hayes helped develop non‑monotonic logics that allow conclusions to be retracted when new information arrives. His work on default reasoning provided formal systems that mimic human common‑sense inference, where typical assumptions (e.g., “birds fly”) can be overridden by exceptions (e.g., “penguins do not fly”).

3.2.2 Situations, contexts, and mental models

Hayes explored how knowledge is situated in particular contexts. He contributed to the theory of mental models, arguing that humans reason by constructing internal, dynamic representations of situations. His formal treatments of context influenced the development of context‑aware computing and situation calculus.

3.3 Cognitive science and conceptual spaces

3.3.1 Conceptual spaces theory (collaboration with Peter Gärdenfors)

Working with cognitive scientist Peter Gärdenfors, Hayes helped develop conceptual spaces theory—a geometric framework for representing concepts and their relationships. In this model, concepts correspond to regions in multi‑dimensional spaces defined by quality dimensions such as color, size, or shape. The theory provides a bridge between symbolic AI and connectionist approaches.

3.3.2 Metaphor and analogy in AI

Hayes studied how metaphors and analogies enable human reasoning and how they can be implemented in AI systems. He argued that analogical mapping, grounded in conceptual spaces, allows machines to transfer knowledge from familiar domains to novel situations. This work informed later research in analogical reasoning and creative AI.

4 Selected publications

4.1 Books

• *Principles of Knowledge Representation* (co‑editor, with Ronald J. Brachman, 1985)
• *Formal Theories of the Commonsense World* (co‑editor, with Jerry Hobbs, 1985)
• *Conceptual Spaces: The Geometry of Thought* (co‑author with Peter Gärdenfors, 2004)

4.2 Journal articles and book chapters

4.2.1 Key papers on the frame problem

• Hayes, P. J. (1973). “The Frame Problem and Related Problems in Artificial Intelligence.” *Artificial Intelligence*, 4(3–4): 141–183.
• Hayes, P. J. (1977). “In Defence of Logic.” *Proceedings of the 5th International Joint Conference on Artificial Intelligence*, 559–565.

4.2.2 Papers on naive physics

• Hayes, P. J. (1979). “The Naive Physics Manifesto.” In D. Michie (ed.), *Expert Systems in the Micro‑Electronic Age*, Edinburgh University Press, 242–270.
• Hayes, P. J. (1985). “The Second Naive Physics Manifesto.” In J. R. Hobbs & R. C. Moore (eds.), *Formal Theories of the Commonsense World*, Ablex, 1–36.

5 Awards and honors

5.1 AAAI Fellow

Hayes was elected a Fellow of the American Association for Artificial Intelligence (AAAI) in 1993 in recognition of his pioneering work in knowledge representation and common‑sense reasoning.

5.2 ACM Fellow

He was named a Fellow of the Association for Computing Machinery (ACM) in 1994 for his contributions to AI, logic, and cognitive science.

5.3 Other recognitions

Hayes received the Lifetime Achievement Award from the International Conference on Knowledge Representation (KR) in 2012. He also served as president of the Society for the Study of Artificial Intelligence and Simulation of Behaviour (AISB) from 1987 to 1989.

6 Personal life

Patrick J. Hayes has been married to his wife, Mary, for several decades. The couple has two children. Hayes is known among colleagues for his dry wit and his enthusiasm for sailing and classical music. He has remained active in retirement, occasionally writing and advising on AI policy.

7 Legacy and influence

Hayes’s work shaped the development of AI by insisting on rigorous logical foundations for common‑sense reasoning. His “Naive Physics Manifesto” inspired projects such as Cyc and the OpenCYC knowledge base. The frame problem, as formalized by Hayes, remains a central challenge in AI planning. His later collaboration on conceptual spaces has influenced cognitive science, robotics, and the philosophy of mind. Through teaching and mentorship, he helped train many leading AI researchers, and his writings continue to be cited across computer science, philosophy, and psychology.