2001 : WHAT QUESTIONS HAVE DISAPPEARED?

[ print ]

author (with George Lakoff) of Where Mathematics Comes From; Philosophy of the Flesh
Do computers think?

This question was at the heart of heated debates for decades during the recently past century, and it was at the ambitious origins of the Artificial Intelligence adventure. It had profound implications not only for science, but also for philosophy, technology, business, and even theology. In the 50's and 60's, for instance, it made a lot of sense to ask the question whether one day a computer could defeat an international chess master, and if it did, it was assumed that we would learn a great deal about how human thought works. Today we know that building such a machine is possible, but the reach of the issue has dramatically changed. Nowadays not many would claim that building such a computer actually informs us in an interesting way about what human thought is and how it works. Beyond the (indeed impressive) engineering achievements involved in building such machines, we got from them little (if any) insight into the mysteries, variability, depth, plasticity, and richness of human thought. Today, the question "do computers think?" has become completely uninteresting and it has disappeared from the cutting edge academic circus, remaining mainly in the realm of pop science, Hollywood films, and video games.

And why it disappeared?

It disappeared because it was finally answered with categorical responses that stopped generating fruitful work. The question became useless and uninspiring, ... boring. What is interesting, however, is that the question disappeared with no single definitive answer! It disappeared with categorical "of-course-yes" and "of-course-not" responses. Of-course-yes people, in general motivated by a technological goal (i.e., "to design and to build something") and implicitly based on functionalist views, built their arguments on the amazing ongoing improvement in the design and development of hardware and software technologies. For them the question became uninteresting because it didn't help to design or to build anything anymore. What became relevant for of-course-yes people was mainly the engineering challenge, that is, to actually design and to build computers capable of processing algorithms in a faster, cheaper, and more flexible manner. (And also, for many, what became relevant was to build computers for human activities and purposes). Now when of-course-yes people are presented with serious problems that challenge their view, they provide the usual response: "just wait until we get better computers" (once known as the wait-until the-year-2000 argument). On the other hand there were the of-course not people, who were mainly motivated by a scientific task (i.e., "to describe, explain, and predict a phenomenon"), which was not necessarily technology-driven. They mainly dealt with real-time and real-world biological, psychological, and cultural realities. These people understood that most of the arrogant predictions made by Artificial Intelligence researchers in the 60's and 70's hadn't been realized because of fundamental theoretical problems, not because of the lack of powerful enough machines. They observed that even the simplest everyday aspects of human thought, such as common sense, sense of humor, spontaneous metaphorical thought, use of counterfactuals in natural language, to mention only a few, were in fact intractable for the most sophisticated machines. They also observed that the nature of the brain and other bodily mechanisms that make thinking and the mind possible, were by several orders of magnitude, way more complex than what it was thought during the hey-days of Artificial Intelligence. Thus for of course-not people the question whether computers think became uninteresting, since it didn't provide insights into a genuine understanding of the intricacies of human thinking. Today the question is dead. The answer had become a matter of faith.

RAFAEL E. NÚÑEZ, currently at the Department of Psychology of the University of Freiburg, is a research associate of the University of California, Berkeley. He has worked for more than a decade on the foundations of embodied cognition, with special research into the nature and origin of mathematical concepts. He has published in several languages in a variety of areas, and has taught in leading academic institutions in Europe, the United States, and South America. He is the author (with George Lakoff) of Where Mathematics Comes From: How the Embodied Mind Brings Mathematics into Being; and co-editor (with Walter Freeman) of Reclaiming Cognition: The Primacy of Action, Intention, and Emotion.