Re: Funes the Memorious

From: Harnad, Stevan (
Date: Tue Nov 28 1995 - 11:03:43 GMT

> From: "Parker, Chris" <>
> Date: Thu, 23 Nov 1995 08:33:17 +0000 (GMT)
cp> What I would like to have known was how were his skills before the
cp> accident related to those afterwards, and was any change in the
cp> vivid detailed memory really retrospective, as far as his early
cp> childhood memories (and even dream!!!) were concerned, I
cp> guess it was, which says something (like Penfield) about what
cp> actually is stored and the separateness of the access process
cp> (or is it the separateness of added structure?). As the author says:
cp> "it is odd and even incredible that no one ever performed an
cp> experiment with Funes". The association with sensations is
cp> also there once again.

Don't forget that it's fiction! So Borges's musings about
experimentation are novelistic, not literal. What this wonderful
little allegory on memory should show you is the price of remembering
everything: being unable to abstract or generalise. If he could not take
artistic liberties with logic, Borges could not have Funes speaking at
all, as the problem Funes had with understanding why we insist on
calling countless infinitely unique views of a dog all by the same name
would have prevented him from being able to name anything at all (hence
unable to speak or reason).

Remember that names are attached to KINDS of things, not to each
individual instance or view of a thing. If all the instances and views
are unique in their every detail, then you can never pick out the KINDs.
To pick out the kinds, you need to focus on what all the instances have
in common, and ignore (forget) all the rest.

Remember Watanabe's "Ugly Duckling" Theorem: Suppose there are only
three things in the world, and they are literally two 1's and one 0.
You are tempted to say there two "kinds" of things then, the
two 1's and the one 0. But things don't end there. Let's line them up
(since if there are three of them, they can't all be in the same
place). Suppose they are positioned this way (though it doesn't matter,
the same thing would happen if they were positioned any other way):
1 1 0

Notice that if they are positioned like that, than there are some more
properties we haven't mentioned: The leftmost 1 and the 0 on the right
share the property that they are NOT in the middle. Let's label that
property "1" too, and the property of being in the middle "0".
Now we have our three things, 1,1,0, and beneath them we'll list these
1 1 0
1 0 1

As you see, if you had sorted them on the base of the second (spatial)
property, there would have been two kinds of things too, but they would
have been different ones.

And the list goes on and on, because the properties that can be called
"1" and "0" are endless. Unless something else selects some of them as
somehow "special," different from the rest, so that all the rest can be
ignored, there is no way to sa that some things are more similar than

(The "0" is the ugly duckling, which turns out to be just as similar to the
two swans, ("1" and "1") as they are to one another!)


Watanabe, S. Pattern recognition: human and mechanical New York: Wiley 1985

Watanabe, S. Epistemological relativity: Logico-linguistic source of relativity. Annals of the Japan Association for the Philosophy of Science, 1986 Mar, v7 (n1):1-14. ABSTRACT: Discusses the logico-linguistic type of epistemological relativity, as exemplified by N. Goodman's (1955) grue emerald and the author's (1966) ugly duckling theorem. A general theorem is introduced and proven that covers all cases of this type of relativity.

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:56 GMT