ABSTRACT
I remember a conversation with Terry Winograd when we were both graduate students at M.I.T. in the Artificial Intelligence Laboratory. Terry was explaining to me how the transformational linguists refused to consider all of the various phenomena that illustrate the need to take world knowledge into account when discussing language. I was amazed by this (having by then assimilated the MIT Artificial Intelligence Lab’s view of the world) and I asked Terry several times, “but what do they say when you point out X,” where X was some case where real world knowledge came into play. His reply was always something like:
They will say, “That’s interesting. Someday we will have think about that.” Of course, someday may be quite a ways off.
