David Rumelhart, pioneer in cognitive neuroscience, dies at 68
The Stanford psychologist created computer models that simulated human perception, language understanding and memory.
BY ADAM GORLICK
David Rumelhart, a psychology professor who studied how people think and learn complex skills such as reading and the use of language, has died. He was 68.
Rumelhart, who died on March 13 in Michigan after suffering from a progressive debilitating neurological condition, was a pioneer in the field of cognitive neuroscience who explored the concept of connectionism – the idea that no single neuron in the human brain does its job alone in processing information.
Leading a team of researchers that included James McClelland – now chair of the Psychology Department – Rumelhart created computer models in the 1970s and 1980s that simulated human perception, language understanding, memory and a wide range of other cognitive tasks.
"Dave was interested in how we're able to bring thoughts together in our minds," McClelland said. "He wanted to know how we can achieve an insight or grasp what the right answer is to a subtle question."
Together, Rumelhart and McClelland wrote Parallel Distributed Processing: Explorations in the Microstructure of Cognition, a book that brought the concept of connectionism to a wider audience of psychologists, neuroscientists and computer scientists.
Born in Wessington Springs, S.D., Rumelhart did his undergraduate work in psychology and mathematics at the University of South Dakota. He earned his doctorate at Stanford in 1967 and immediately launched his teaching career at the University of California-San Diego.
While at UCSD, Rumelhart began honing his ideas on how the brain works. He became dissatisfied with a classic understanding of the human thought process that was based on the notion that cognition happens through the mind's manipulation of symbols.
"He became fascinated by the idea that our minds work at a sub-symbolic level," McClelland said. "The idea was that thoughts emerge from neural activity. They're the consequence of the interaction of neurons. He went behind the scenes to look for the actual basis for our thinking ability."
Rumelhart came to Stanford in 1987, where he continued developing the framework he and McClelland established in Parallel Distributed Processing. A shy professor unless one knew him well, Rumelhart often showed his competitive streak on the tennis court or makeshift volleyball court he would fill with students and fellow professors on the Oval.
Among his honors and awards, Rumelhart was a member of the National Academy of Sciences and a recipient of a MacArthur Fellowship. He and McClelland jointly received several awards as well, including the Warren Medal from the Society of Experimental Psychologists in 1993.
Struck with a neurological disease a few years after arriving at Stanford, Rumelhart stopped teaching in 1998. In 2000, the Glushko-Samuelson Foundation established the Rumelhart Prize, a $100,000 award given annually to scientists who make a significant contribution to the theoretical foundations of human cognition.
Rumelhart is survived by his former wife, Marilyn Austin; their sons, Peter and Karl Rumelhart; his brothers, Donald and Roger Rumelhart; and four grandsons. A memorial service will be scheduled this spring.
Adam Gorlick, Stanford News Service: (650) 725-0224, email@example.com