Turing; factors; spontefaction

[From Bill Powers (960131.0800 MST)]

RE: Turing Test

Hans Blom, 960131 --

     An X behaves like a human if its visible outputs "looks like" what
     would be generated by a human. So is X a human? This has always
     been my problem with Asimov's Three Laws of Robotics, of which the
     essential element is how a robot can recognize whether an X is a
     human.

Excellent point. The nature of the system producing the behavior can't
be determined just by observing the output of the system.

Suppose the behavior in question is ringing somebody's doorbell. This
behavior consists of applying a force to the button beside the front
door. How could you distinguish (from inside the house) the behavior of
a human being from the behavior of a trained chimpanzee from the
behavior of an icicle blown by the wind so it hits the button on the way
to the ground? You can't even tell if the system that pushed the button
is alive, much less whether it is human.

I would like to propose a relaxed Turing Test. The objective is only to
tell whether the system is a spontefactive system or some other kind of
system (note change of name at Bruce Abbott's request, to avoid
trendiness).

     Anyway, all I see of others is appearances, behavior.

This is not quite all. You see the appearances created by their actions,
_in relation to other processes in the world_. Some of those processes
you can cause yourself, so you can perform experimental operations and
see how the apparent behavior of others changes. So you can look for
regularities in the observed relationships, and form and test hypotheses
about the organization responsible for these relationships.

···

-----------------------------------------------------------------------
Remi Cote 310196.0716 --

     Can you describe the frame of the protocol that can allow one to
     affirm beyound doubt that gain factor ara ALWAYS modulated by
     factor.

Are you referring to the equation that says the output is some factor
times the error signal? If the output is a continuous measure, like a
force measured in kilograms, and the error signal is also a continuous
measure expressed in impulses per second, then there will always be SOME
factor relating these quantities. For a signal of 10 impulses per
second, for example, you might get an output force of 2.5 Kg. Then the
factor relating them is 4: output = 4*error. How could there NOT be a
factor?

     Do anybody please remember me the name of that bacteria that react
     to certain chamical that stop forwarding when me thing the right
     chemical as I do when meeting the right girl? Please help me
     remember!

The bacterium is Escherichia coli, or E. coli. E. coli always swims at a
constant speed in approximately a straight line. At intervals, it
reverses some of its flagellae and tumbles briefly at random, then
starts swimming in a straight line again. When it is swimming in a
gradient of an attractant, the delay between tumbles depends on whether
it is experiencing a rising concentration or a falling concentration. If
the concentration is rising, the next tumble is delayed; if the
concentration is falling, the next tumble occurs sooner. So E. coli
spends less time swimming the wrong way than the right way, and it
progresses up the gradient quite efficiently.
-----------------------------------------------------------------------
Bruce Abbott (960131.0835 EST) --

     Yes, exactly -- an insufficient retrofraction! No doubt due to the
     icy condition of the road that prevailed at the time. Not to be
     confused, of course, with the retrofaction fraction or, as it used
     to be called, loop gain.

We can still use most of the other words, like loop gain. The only word
that has caused trouble is "control," so that is the word I have changed
to spontefaction (I thought you were rejecting "retro" as being too
trendy). Spontefaction comes from the Latin root "sponte" meaning "of
one's free will; voluntarily" and giving rise to "spontaneous;" and
"facere," to make or do, giving rise to such terms as factory and
facile.

"Spontaneous" is given these main meanings in my Webster's:

1. Arising without external constraint or stimulus
2. Arising from momentary impulse
3. Controlled and directed internally.

Except for the second, these meanings fit quite well the way a negative
feedback system works. A spontefactive system, as soon as it is supplied
with energy and a reference signal, will generate action that brings its
input to a specific state without needing any external direction or
stimulus (external to the loop). The dictionary adds a synonym:
"natural." A spontefactive system acts naturally, out of its own
organization, rather than being forced to act by external forces.

No one element of a spontefactive loop is capable of spontefaction.
Spontefaction is a property of the whole loop. So the slipperiness of
the road did not spontefact the car into the ditch. That is, if the car
had not headed for the ditch, the slipperiness of the road would not
have changed in such a way as to bring the car's path back toward the
ditch.
-----------------------------------------------------------------------
Best to all,

Bill P.