Autonomy -Reply

[Hans Blom, 960422]

(Rick Marken (960419.1000))

I agree that it is possible, in principle (assuming no limits on the
loop variables), to get r11 and r12 to reach all possible
combinations of values by suitably manipulating d1 and d2; the
environment certainly _influences_ what we want.

In the example, the equations indicate that _nothing but_ the
disturbances vary the goals; all else is constants. So what else
"influences"?

What I have been saying is simply that the inanimate environmental
does not _control_ what we want, because the inanimate environment
is not organized as a control system.

This is circular reasoning. If the environment is taken as not being
a control system, by definition it cannot control. If, however, we
assume (just for a minute, because we're going to drop that heretical
thought immediately after this thought experiment :wink: that the
environment COULD POSSIBLY control, we might try to find out how the
"disturbances" of the world change an organism's goal. That's the
thought experiment that I did. I'm not telling you that any of this
is "true"; it's just modelling...

Other control systems _can_ control what we want, to some extent,
but they cannot control what we want _arbitrarily_. That is, an
external controller cannot control what we want without taking into
account the fact that we have other wants that might conflict with
the want that the external controller wants us to have; ...

In the analyzed example, there were no other goals. For argument's
sake, however, it is easy to show that N goals can be set with N
disturbances. So let a manipulator manipulate a large number of
disturbances, equal to the number of the organism's intermediate
goals, and ALL the organism's (intermediate) goals are controlled by
the environment.

I think you are unwilling to call people "autonomous" if an external
controller can control people's wants _at all_.

No. In the example hierarchical controller that I analyzed, ALL goals
could be simultaneously manipulated. A sufficiently powerful
manipulator (Gaia? God?) can control ALL goals. Of course, we mere
mortals can't ;-).

Greetings,

Hans

···

================================================================
Eindhoven University of Technology Eindhoven, the Netherlands
Dept. of Electrical Engineering Medical Engineering Group
email: j.a.blom@ele.tue.nl

Great man achieves harmony by maintaining differences; small man
achieves harmony by maintaining the commonality. Confucius

[From Bruce Gregory (960422.1130 EDT)]

(Hans Blom, 960422)

I am confused by many things, but in particular by your statement

  If, however, we assume (just for a minute, because we're going
  to drop that heretical thought immediately after this thought
  experiment :wink: that the environment COULD POSSIBLY control, we
  might try to find out how the "disturbances" of the world change
  an organism's goal.

Granted we assume that the environment could be a control system.
The next step, it seems to me, would be to test for the controlled
variable. Do you have in mind any environmental variable that
would pass this test?

Regards,

Bruce G.

[Hans Blom, 960423e]

(Rick Marken (960422.1300))

From your post only this:

Hans' claim (about the ability to control wants with disturbances)
is _only_ true if we ignore reality.

Yes, you can take that as a fact ;-). Unless I temporarily and mis-
takenly forsake my own principles, I only talk about models, not
about reality.

So there is not much of a range over which wants can be controlled
when those wants are part of a high gain control hierarchy. The
higher the gain of the control hierarchy, the more autonomous (less
controllable) it is.

That's welcome progress after the black-and-white views of autonomy!

Greetings,

Hans

···

================================================================
Eindhoven University of Technology Eindhoven, the Netherlands
Dept. of Electrical Engineering Medical Engineering Group
email: j.a.blom@ele.tue.nl

Great man achieves harmony by maintaining differences; small man
achieves harmony by maintaining the commonality. Confucius

[From Rick Marken (960423.0730)]

Me:

The higher the gain of the control hierarchy, the more autonomous (less
controllable) it is.

Hans Blom (960423e) --

That's welcome progress after the black-and-white views of autonomy!

Aren't you the fellow who said (Hans Blom, 960416):

Thus there are no goals that an organism can call "its own" in any meaningful
way.

Sounds pretty black (or white) to me.

I can't tell you how nice it is to hear that I have made "welcome progress
after the black-and-white views of autonomy" from the fellow who started
this whole thing off by saying that control systems are _not_ autonomous.

Well, actually, I can tell you how nice it is: it's as nice as getting a
lecture on values from a corportate executive who just got a million dollar
raise after laying off 20% of his employees.

Best

Rick