Endnotes
[NOTE: After reading an "endnote", there are two
ways to get back where you were reading:
- Press the "Back" button of your browser. This is
usually satsifactory, but in some circumstances it may take you
to the top of the page, not to the sentence where the reference
occurred.
- Click the link which follows the note. This always repositions
you at the sentence where the reference occurred. The drawback
is that it deposits another entry on the trail which is laid
down for the browser's "Back" button. That is, if you
subsequently use the "Back" button, you'll come through
the note again enroute to the page from which it digressed.
It's a matter of preference, do as you find convenient.]
- Semantic Memory, by Ross Quillian,
Bolt Beranek and Newman, Inc., Cambridge, Mass., 1966. Hearsay
has it that the situation cited is less dramatic in 1970 -- but
that just confirms the point I am making. [Back]
- See, for example, Real English
-- A Description of its Operation, by David Klappholz, March
1970, University of Pennsylvania, Moore School of Electrical
Engineering. [Back]
- Recognition of a limited vocabulary
is considered practical already. Cornell University has been
experimenting with a system in which the statements of a simple
programming language are spoken into a phone and recognized by
the computer. A "moment of silence" is required between
each pair of words, and the computer must "know" the
speaker. See "On the Feasibility of Voice Input to an On-Line
Computer Processing System," Communications of the ACM,
Vol. 13, No. 6, June, 1970. [Back]
- Accomplishments Summary 1968/1969
of the Biological Computer Laboratory, ed. Heinz von Foerster,
Illinois University, Urbana, Illinois, June, 1969. [Back]
- Brain, Mind, and Computers, by
Stanley L. Jaki, Herder and Herder, NYC, 1969. The book may be
recommended for a lively, if somewhat paranoid style, and for
bringing in a sense of philosophical history that is rare in
technical discussions. As for the rest, it is consistently wrongheaded.
I've been trying to secure a copy in a spirit of investment:
I figure that in twenty years it will be a valuable addition
to my "famous losers" collection. (I've regretted all
my life that I didn't buy a copy of the one I saw in a library
back in 1946, "proving" that space travel was impossible.)
[Back]
- "Robot Control Strategy,"
by Leonard Friedman, in Proceedings of the International Joint
Conference on Artificial Intelligence, May, 1969. [Back]
- Both the terminology and the point
of view are borrowed from ethology, the study of instinctive
behavior. Cf. Studies in Animal and Human Behavior, by
Konrad Lorenz, Harvard University Press, Cambridge, Mass., 1970.
This is another whole field whose findings suggest programming
as a basic part of behavior. Lorenz distinguishes among three
kinds of behavior: instinctive, conditioned, and insightful.
The first two can be construed as programming with no difficulty
at all; the last is more controversial. [Back]
- See A Mobile Automaton: An Application
of Artificial Intelligence Techniques, by Nils J. Nilsson,
in Proceedings of the International Joint Conference on Artificial
Intelligence, Washington, D.C., May, 1969. [Back]
- I could be pretentious and let on that
I saw this in the AFIPS Conference Proceedings, Vol. 34, 1969
Spring Joint Computer Conference ("Assembly of Computers
to Command and Control a Robot," by Louis L. Sutro and William
L. Kilmer). To tell the truth, though, I saw it first in Analog
Science Fiction / Science Fact, where it was reprinted. [Back]
- It's safe to assume that the U.S.S.R.
has a similar project. Lunaxhod, the explorer robot which
they have already put on the moon, is probably not much smarter
than a remotely-controlled garage door. When they get to Mars,
however, they'll have the same reasons as M.I.T. to want a machine
that can supplement its instructions with a little independent
"judgment." [Back]
- Computer-Controlled Robots, by
H. A. Ernst, IBM Research, Yorktown Heights, N.Y., 1970, report
#RC-2781. Perhaps it's a mite unfair to ask that engineers give
robots what Socrates couldn't define for Charmides. [Back]
- The scene is from Mad magazine.
I'm just speculating what was said. [Back]
- op. cit. (Brain, Mind, and
Computers, by Stanley L. Jaki). [Back]
- At least, she offered to. The local
juice men decided that the royal word would do. [Back]
- And all those Indian souls saved, into
the bargain! [Back]
- One often has to eat one's words in
these matters almost before one can utter them. I had no sooner
written the above, than my wife overheard a news item and called
it to my attention: "Dr. José Delgado of Yale University,"
it said, "has achieved direct communication between brain
and computer."
I wrote for information and hit a mother lode: a reprint of an
article from Excerpta Medica International Congress Series
No. 180 ("The Present Status of Psychotropic Drugs --
Proceedings of the VI International Congress of the C.I.N.P.,
Tarragona, April, 1968.")
I think my statement can still stand: in the ambitious sense
which was intended, nobody has read out and decoded any
message from the nervous system. What Delgado reports, however,
is that we can read out certain signals which are regularly
associated with particular types of (chimpanzee) behavior, or
with particular moods. By electrically stimulating the areas
from which these signals arise, we can trigger the same sort
of behavior or the same sort of mood. That, of course, is not
the same thing as cracking the code; the signals that are fed
in don't have to match the ones that are normally emitted.
I don't count the following item: "The first man to communicate
with a computer via brain waves appears to be Dr. Edmond Dewan
of the Air Force Cambridge Research Labs, who, controlling the
alpha rhythms of his brain, sent Morse code to the computer (aided
by an EEG) to spell out 'cybernetics.'" (Quoted in the ACM
SIGART Newsletter, No. 19, Dec., 1969.) That's an amusing
stunt, but it doesn't use the signals as the nervous system uses
them; the meaning is imposed by artifice. Presumably, all Dr.
Dewan had to do was roll up his eyeballs, which is known to produce
bursts of alpha waves, for the duration of each "dot"
or "dash." [Back]
- You can read it for yourself in an
Everyman's Library edition, 1963, E. P. Dutton & Co.,
New York. [Back]
- Reprinted in Science-Fiction Thinking
Machines: Robots, Androids, Computers, edited by Groff Conklin,
the VANGUARD Press, Inc., NYC, 1954. [Back]
- War With the Newts, by Karel Capek.
There's a paperback edition by Berkeley Publications, 200 Madison
Ave., New York, N.Y. [Back]
- Asimov's robot stories are collected
in two anthologies: I, Robot, 1963, and The Rest of
the Robots, 1964, both by Doubleday & Co., New York.
In the introduction to the latter of these two books, the author
states explicitly what he was up to. [Back]
- "Turing's test" may well have
been suggested to Turing by some of Asimov's stories, for in
them, robots that try to impersonate humans are forever getting
tripped up by questions that they can't answer the way a human
would. [Back]
- The Living Machine, National
Film Board of Canada. Don't miss it. [Back]