[quoting Padgett, 21 Feb 96.b] An example: In the Gorgias, Socrates leads a young student through the process of discovery, so that the student ends up with knowledge about the nature of a rectangle. Socrates doesn't ever offer a fact about the rectangle, only asks questions of the learner, that leads the learner to knowledge. Whoopee, discovery, yes? And yet, at the same time, it could be said that all the knowledge about the shape was being taught, because of its representation in a picture. (I believe Plato used the example to show that all knowledge is already up there somewhere in our heads, only needing to be brought to the surface; I don't buy that, though.)
Wasn't that Socrates with a slave boy and a triangle in the Meno? Or Mr. Mustard in the library...?
Anyway, why don't you buy Plato's idea that knowledge is in the head to begin with? It seems intuitively obvious that an empty computer couldn't learn anything. Computers that learn (like OCRs that learn fonts, or in fact any computer that extracts information from data) have to be programmed with all sorts of knowledge about the world. Dogs can't do geometric proofs because they lack the innate knowledge of geometry that humans have. (Obviously, what is innate is not constructed.)