Tuesday, October 25, 2011

John McCarthy's Legacy

I was glad to see that the Technology division of the BBC News Web site took the trouble to run an obituary for John McCarthy.  I never knew him beyond casual acquaintance, but he had a profound effect on my approach to conducting research from my earliest efforts as a undergraduate to the present day.  That effect came from his invention of the programming language LISP, which shifted the focus of computation from numbers to symbolic forms.  Because the LISP language emerged around the same time as the earliest efforts in time-sharing systems, through the simplicity of its foundation, it may well have been the first programming language whose forms could be interpreted in “real time,” rather than requiring preprocessing and translation by a compiler.  One might almost say that LISP was the first programming language in which one could have a genuine “dialog” with one’s computer.

This was the reason that LISP had such a powerful impact on so many of us.  While others would puzzle over designing complex software to achieve complex tasks, LISP provided us with the ability to tinker our way through a problem, beginning by writing component modules, observing their behavior when given different inputs, and eventually assembling them into larger constructs.  Back in the days when the first laptops were coming on the market, I held off until the appearance of the Macintosh PowerBook 170, because I knew it had enough computer power to support a LISP interpreter.  I wrote at least one of my papers using that laptop as my primary “laboratory equipment.”

My primary interactions with McCarthy took place when I was co-editor (with Mark Stefik) of the Book Review section of the journal Artificial Intelligence.  I had an active hand in shepherding through the review of Formalizing Common Sense, an anthology of McCarthy’s papers on his “common sense” approach to artificial intelligence edited by Vladimir Lifschitz.  Philosophically, I could not live in a world as exclusively objective as the one in which McCarthy pursued his investigations;  but I worked hard to make sure that this collection got a fair shake.  In the course of that effort, I was definitely aware of his contentious side;  but I always found the right path along which I could work with him.

The one time McCarthy was in a conversation in which I cited an example of music behavior, he tuned out very quickly.  I suspect that he thought as much of my decision to focus on the practice of music as a manifestation of intelligence as I did of his efforts to express all reasoning in terms of the sentential forms of a logical calculus.  However, in those days artificial intelligence was more a “republic of letters” than it is today, since we had the luxury to ponder the nature of the value of the questions themselves, rather than being forced by funding sources to generate answers as efficiently as possible.

No comments: