Carl Hewitt (1945–2021) was an American computer scientist known for his pioneering work on the actor model of concurrent computation, the PLANNER programming language, and contributions to artificial intelligence and open information systems. He spent most of his career at the Massachusetts Institute of Technology (MIT), where he influenced distributed computing and formal methods.
1 Early life and education
1.1 Early life
Carl Hewitt was born in 1945. Details of his early childhood are not widely documented, but he grew up in the United States during a period of rapid technological change. He developed an early interest in mathematics and computation, which would later define his career.
1.2 Undergraduate studies
Hewitt attended the Massachusetts Institute of Technology (MIT) for his undergraduate education, where he earned a Bachelor of Science degree in mathematics. During his studies, he became involved with the MIT Artificial Intelligence Laboratory and was exposed to early work in symbolic computation and logic.
1.3 Graduate studies
He continued at MIT for graduate work, earning a Master's degree and a Ph.D. in mathematics. His doctoral dissertation, completed in the early 1970s, introduced the PLANNER programming language, a precursor to logic programming and an early system for automated planning. This work laid the foundation for his later research on concurrent systems.
2 Career
2.1 Massachusetts Institute of Technology (MIT)
2.1.1 Research in artificial intelligence
Hewitt was a key figure at the MIT Artificial Intelligence Laboratory from the late 1960s onward. He worked on knowledge representation, automated reasoning, and the development of programming languages for AI. His exploration of pattern-directed invocation and procedural reasoning influenced the design of early expert systems.
2.1.2 Teaching and mentorship
He served as a professor in the MIT Department of Electrical Engineering and Computer Science. Hewitt taught courses on programming languages, distributed computing, and artificial intelligence. He mentored many graduate students who later became notable researchers in their own right.
2.2 Other affiliations and collaborations
Beyond MIT, Hewitt collaborated with researchers at Stanford University, the University of Texas at Austin, and various international institutions. He participated in early ARPANET projects and was involved in the development of the Common Lisp standard. He also consulted for industry labs, including Xerox PARC and Hewlett-Packard.
2.3 Later career and retirement
In the later decades of his career, Hewitt focused on formalizing the actor model and exploring open information systems. He retired from active teaching but continued research at MIT as an emeritus professor. He remained active in academic discussions until the early 2010s.
3 Major contributions
3.1 PLANNER programming language
3.1.1 Design and goals
PLANNER was developed in the late 1960s and early 1970s as a language for automated theorem proving and planning. Its design emphasized pattern-directed invocation and backtracking. The language allowed programmers to specify goals and let the runtime system search for proofs or plans using declarative rules.
3.1.2 Influence on later languages
PLANNER directly inspired the logic programming language Prolog, developed by Alain Colmerauer and Robert Kowalski. The concept of pattern matching and backward chaining in PLANNER became fundamental to many AI programming languages. It also influenced the design of the assertion-based language Micro-Planner and other rule-based systems.
3.2 Actor model of concurrent computation
3.2.1 Theoretical foundations
The actor model, introduced by Hewitt and his collaborators in the 1970s, frames computation as the behavior of independent, communicating actors. Each actor can create new actors, send messages, and decide how to respond to incoming messages. This model provided a mathematical foundation for understanding concurrent and distributed systems.
3.2.2 Semantics and formalization
Hewitt developed a formal semantics for the actor model using denotational and operational approaches. He defined an operational semantics based on transition systems and later contributed to an algebraic theory of actors. The model's semantics addressed issues of non-determinism, fair execution, and message ordering.
3.2.3 Applications in distributed systems
The actor model became a cornerstone of modern distributed computing. It influenced the design of programming languages such as Erlang, Scala (with Akka), and the Swift concurrency model. The model is also used in cloud computing frameworks, simulation systems, and real-time processing architectures.
3.2.4 Evolution and later developments
Over time, Hewitt expanded the actor model to include notions of trust, accountability, and open systems. He introduced the concept of "clients," "servants," and "computational agents." Later work incorporated logical reasoning into actors, blurring the line between computation and knowledge representation.
3.3 Open information systems and reasoning
3.3.1 Direct logic
Hewitt proposed "Direct Logic" as a reasoning system for open, distributed environments. Unlike classical logic, Direct Logic allowed reasoning with incomplete and contradictory information. It supported non-monotonic inference and was designed to be used by artificial agents that must make decisions under uncertainty.
3.3.2 Inconsistency robustness
A key contribution was the idea of "inconsistency robustness"—the ability of a system to continue functioning despite the presence of contradictory information. Hewitt argued that classical logic's intolerance to inconsistency was unsuitable for large-scale open systems. He developed formal frameworks where inconsistencies could be managed rather than eliminated.
4 Recognition and legacy
4.1 Awards and honors
Hewitt received the ACM Programming Languages and Systems (PL) Paper Award for his work on the actor model. He was also a fellow of the American Association for Artificial Intelligence (AAAI). His contributions were recognized in several festschrifts and special journal issues dedicated to his work.
4.2 Influence on computer science
The actor model is one of Hewitt's most enduring legacies. It is foundational to the theory of concurrent computation and has been adopted widely in both academia and industry. PLANNER also had a lasting impact on logic programming. Hewitt's ideas on open information systems influenced research in semantic web, multi-agent systems, and robust computing.
4.3 Notable students and collaborators
Among Hewitt's doctoral students were Gerald Jay Sussman (co-inventor of Scheme), William Clinger (developer of the Scheme standard), and Gul Agha (who formalized the actor model further). He collaborated extensively with researchers like Henry Lieberman, Irene Greif, and Gordon Plotkin on concurrency and AI topics.
5 Personal life
5.1 Family and interests
Carl Hewitt valued intellectual exchange and often hosted discussions at his home. He was an avid reader of philosophy, logic, and science fiction. He maintained close ties with family, though details of his private life were kept out of the public eye.
5.2 Health and death
In his later years, Hewitt faced health challenges. He died on August 12, 2021, at the age of 75. His passing was marked by tributes from the computer science community, highlighting his unique vision and contributions.
6 Selected publications
6.1 Early papers on PLANNER
- Hewitt, C. (1969). "PLANNER: A Language for Proving Theorems in Robots." Proceedings of the 1st International Joint Conference on Artificial Intelligence.
- Hewitt, C. (1971). "Procedural Embedding of Knowledge in PLANNER." Proceedings of the 2nd International Joint Conference on Artificial Intelligence.
6.2 Foundational actor model papers
- Hewitt, C.; Bishop, P.; Steiger, R. (1973). "A Universal Modular Actor Formalism for Artificial Intelligence." Proceedings of the 3rd International Joint Conference on Artificial Intelligence.
- Hewitt, C.; Agha, G. (1988). "Actors: A Model of Concurrent Computation in Distributed Systems." MIT AI Memo 926.
6.3 Later works on open systems
- Hewitt, C. (2006). "Inconsistency Robustness." Proceedings of the 2006 Conference on Advances in Social Networks Analysis and Mining.
- Hewitt, C. (2012). "Actors: The Next Generation." Formal Methods in System Design, 40(1), 1–28.