The Barrier of Meaning by Gian-Carlo Rota.
The author discusses the “AI-problem” with Stanislaw Ulam. Ulam makes reference to the history of the “AI-problem” and then continues:
Well, said Stan Ulam, let us play a game. Imagine that we write a dictionary of common words. We shall try to write definitions that are unmistakeably explicit, as if ready to be programmed. Let us take, for instance, nouns like key, book, passenger, and verbs like waiting, listening, arriving. Let us start with the word “key.” I now take this object out of my pocket and ask you to look at it. No amount of staring at this object will ever tell you that this is a key, unless you already have some previous familiarity with the way keys are used.
Now look at that man passing by in a car. How do you tell that it is not just a man you are seeing, but a passenger?
When you write down precise definitions for these words, you discover that what you are describing is not an object, but a function, a role that is tied inextricably tied to some context. Take away that context, and the meaning also disappears.
When you perceive intelligently, as you sometimes do, you always perceive a function, never an object in the set-theoretic or physical sense.
Your Cartesian idea of a device in the brain that does the registering is based upon a misleading analogy between vision and photography. Cameras always register objects, but human perception is always the perceptions of functional roles. The two porcesses could not be more different.
Your friends in AI are now beginning to trumpet the role of contexts, but they are not practicing their lesson. They still want to build machines that see by imitating cameras, perhaps with some feedback thrown in. Such an approach is bound to fail since it start out with a logical misunderstanding….
Should someone mention this to the EC Brain project?
BTW, you may be able to access this article at: Physica D: Nonlinear Phenomena, Volume 22, Issues 1–3, Pages 1-402 (October–November 1986), Proceedings of the Fifth Annual International Conference. For some unknown reason, the editorial board pages are $37.95, as are all the other articles, save for this one by Gian-Carlo Rota. Which as of today, is freely accessible.
The webpages say Physica D supports “open access.” I find that rather doubtful when only three (3) pages out of four hundred and two (402) requires no payment. For material published in 1986.
You?