Obituary: Professor H. C. Longuet-Higgins ; Chemist Who Switched Disciplines to Artificial Intelligence and Cognitive Science

Article excerpt

H. C. LONGUET-HIGGINS was an outstanding scientist who made lasting contributions in two separate fields - theoretical chemistry and artificial intelligence.

Transcending all disciplinary boundaries, he had an astonishing ability to recognise, where others could not, the susceptibility of well-known open problems to immediate exact solution by applied mathematics. His work - which was often spoken of as having been worthy of a Nobel Prize - opened up many new areas of research. Where he had shown the way, his students and others made their own research careers.

Hugh Christopher Longuet-Higgins was born in Lenham, Kent, in 1923, the second of three children of the Rev Henry Longuet- Higgins, and Albinia, nee Bazeley. His younger brother Michael became a distinguished geophysicist. Christopher was educated at Winchester (where he was a contemporary of Freeman Dyson, whose brilliance in physics he said led him to avoid going in for that subject) and at Balliol College, Oxford, where he studied both music and chemistry. (While his career lay in science, he was also an exceptionally fine pianist, capable of brilliant improvisation in the late Romantic style, with a deep love and understanding of music that sustained him all his life.)

While still an undergraduate, in 1943 he published an important paper with his tutor Ronald Bell in the Journal of the Chemical Society, "The Structure of the Boron Hydrides", on the structure of diborane, from which several subsequently confirmed predictions concerning the existence and structure of similarly electron- deficient molecules followed. Later, he provided an analysis for the structure in terms of a novel type of bond whose existence he proved on the basis of Mulliken's molecular orbital theory, and which led to a complete analysis of the structure of the boranes. This work formed part of his PhD at Oxford under Charles Coulson, a pioneer in applying statistical and quantum mechanics to the analysis of molecular structure.

He was briefly at the universities of Chicago and Manchester, before being appointed to the Chair of Physics at King's College London in 1952. Two years later, he moved to take up the John Humphrey Plummer Chair of Theoretical Chemistry at Cambridge, whose previous incumbent had been John Lennard-Jones, Coulson's adviser and another founding figure of British quantum chemistry. Longuet- Higgins stayed at Cambridge until 1967, applying mathematical techniques to problems in theoretical and practical chemistry, including a foundational group- theoretic analysis of the symmetry of non-rigid molecules. During this period he also began his work on the formal analysis of music. He was elected a Fellow of the Royal Society in 1958.

In 1967, Longuet-Higgins made a startling change of field, moving to Edinburgh University under a Royal Society Research Professorship, where he joined Richard Gregory and Donald Michie in founding what was then called the Department of Machine Intelligence and Perception (now merged with the School of Informatics). He began research on diverse topics in artificial intelligence and cognitive science (as he christened the field in his 1973 commentary on the Lighthill report on the state of AI research), including neural computation, automated musical analysis, and computational natural language processing. …