I'm not sure where you got
the idea that anyone was proposing the sort of refusal you describe.
John's point, and mine, was only that it would be nutty to use Nand and Nor in
place of the usual boolean connectives in a language for knowledge
engineering. But if you *want* to use Nand and Nor, given the usual
booleans it is trivial to introduce them by definition (and conversely, of
Then we are in violent agreement about the ends, but I have not reached
you with the notion that we perceive what we EXPECT to perceive, for example,
what we perceive if we choose to limit our expectations a priori. Actively looking
for one _expression_ of a phenomenon ignores the contribution of other expressions
to a complete knowledge of that phenomenon, and can result in peculiar misconceptions.
That is the only point I was trying to make - let's agree that this is enough
frittering of time on the topic.