Terry Winograd

• 1. Introduction

  ›

  Terry Allen Winograd was born on the 24th of February 1946, and he would go on to become one of computing's most quietly influential figures. He is a professor at Stanford University, co-director of the Stanford Human-Computer Interaction Group, and the man who once taught a computer to understand toy blocks. That last detail sounds modest. It wasn't.

  The question Winograd spent decades wrestling with is deceptively simple: can a machine truly understand language, or is it only pretending? What happens when a program runs out of the narrow world you built for it? And what do a Chilean philosopher, a search-engine pioneer, and a simulated robot arm all have in common? The answers thread through one career that kept changing direction, always following the same hard question.

• 2. The Blocks World Experiment

  ›

  Winograd grew up in Colorado and graduated from Colorado College in 1966 before heading east to MIT, where between 1968 and 1970 he wrote his PhD thesis in the form of a program called SHRDLU. The program was his attempt to give a computer enough understanding to actually use natural language, not just shuffle symbols.

  To make the problem tractable, Winograd confined the program to a simulated world of toy blocks. Within that miniature universe, SHRDLU could accept commands of genuine grammatical complexity. A user might type, "Find a block which is taller than the one you are holding and put it into the box," and the program would carry out the action using a simulated block-moving arm. When it lacked enough information, it could reply in plain English: "I do not know which block you mean."

  Winograd later concluded that SHRDLU illustrated a hard truth about hand-built semantic memory. The program worked beautifully inside its toy world and collapsed the moment anything outside that world appeared. Computer scientists came to call this brittleness, and SHRDLU became a classic demonstration of just how limited that approach was.

• 3. A Philosophy Intervenes

  ›

  In 1973, Winograd moved to Stanford University, intending to build a full AI-based framework for understanding natural language. He envisioned a series of books. Only the first volume, covering syntax, was ever published in 1982.

  Something had intervened. Winograd encountered the critique of cognitivism developed by philosopher Hubert Dreyfus, which challenged the foundational assumptions behind classical AI. Around the same time, he met Fernando Flores, a Chilean philosopher whose thinking was rooted in phenomenology. The meeting changed the direction of his work.

  Winograd came to articulate his shift in terms of communication itself. As he put it, the success of communication depends on real intelligence on the part of the listener, and there are many other ways of communicating with a computer that can be more effective, given that it lacks that intelligence. Together, Winograd and Flores published their critique in 1986 as "Understanding Computers and Cognition: A New Foundation for Design." In the latter part of the 1980s, the two also collaborated on an early form of groupware, grounding their approach in what they called conversation-for-action analysis.

• 4. Computing With a Conscience

  ›

  In the early 1980s, Winograd stepped outside the laboratory entirely. He became a founding member and national president of Computer Professionals for Social Responsibility, a group of computer scientists who were troubled by the nuclear weapons debate, the Strategic Defense Initiative, and the growing involvement of the U.S. Department of Defense in academic computer science.

  The group put Winograd at the intersection of technical expertise and public advocacy, a position that reflected something consistent in his career: a belief that how computers are built matters as much as whether they can be built. That concern would soon take a more formal shape inside Stanford itself.

• 5. Design as Its Own Discipline

  ›

  In 1991, Winograd founded the "Project on People, Computers and Design" at Stanford to promote teaching and research into software design as a field distinct from software engineering. His argument was that design occupies its own space between analysis and programming, and that it should draw on traditions from textile design, industrial design, and other professions.

  The book "Bringing Design to Software" came out of that project and describes its core ideas. Winograd also helped found the Hasso Plattner Institute of Design at Stanford, known informally as the "d.school," which would go on to become one of the most visible design-education programs in the world. His association with the d.school placed him at a junction between engineering culture and design culture that Stanford was deliberately trying to create.

• 6. Advising the Search Engine

  ›

  Starting in 1995, Winograd served as adviser to a Stanford PhD student named Larry Page, who was then working on a research project involving web search. In 1998, Page took a leave of absence from Stanford to co-found Google.

  In 2002, Winograd took his own leave, spending a sabbatical at Google as a visiting researcher. There he studied the intersection of theory and practice in human-computer interaction, examining how the ideas he had spent decades developing inside universities played out inside one of the largest-scale computing environments ever built.

  Winograd received the SIGCHI Lifetime Research Award in 2011, and was named a Fellow of the Association for Computing Machinery in 2009. He has continued researching collaborative and ubiquitous computing at Stanford, asking questions about how people and computers work together that trace directly back to the toy blocks he animated in a Cambridge laboratory more than fifty years earlier.