Observing the top level (was: most difficult ...)

[Martin Taylor 2008.01.26.10.26]

[From Bill Powers (2008.01.26.0049 MST)]

Kenny Kitzke (2008.01.24.1040CST)

I think that one of the problems we have in our discussion is that we're not really talking about the same things. When I try to talk about system concepts, I find it very difficult because we have to use words to communicate at all, and words don't exist at the system concept level that I'm trying to describe with them.

While lurking on the edge of this dialogue, I've been having a persistent thought or question, which your observation brings into focus: "Can the 'Observer' observe the top-level perceptions?"

Maybe this is off-base, but it seems to me that the only words on which we can develop an agreement are words for things that are accessible to the Observer in all the parties concerned. There's no need for words in the actual process of control (though I have no doubt they can often be a component of some control systems), so there's no theoretical reason why there should not be control systems for perceptions that are inaccessible to the Observer. It seems reasonable that such perceptions could not have words to describe them in ways that would mean anything to another person.

Since this is a subjective matter, the question has to be personal: Bill, does your Observer actually observe the varying current states of your controlled system-level perceptions, or is your belief in their existence derived from your logic-level understanding of HPCT, with their states unObervable?

My Observer does not see my top-level perceptions, so far as I know (though would I know?). But I can, I think, Observe what some of the reference values may be, perhaps by inference from the actions I do to maintain those perceptions at appropriate levels, or by Observing what seems to be "good" and what seems to be "bad" behaviour in myself and in others.

If your Observer does perceive the ongoing state of system-level perceptions, perhaps it might be possible by some training technique (using examples such as Zen training as a possible model) to help others to do the same. If that could be done, maybe it would be possible to develop words for the different possible states of those perceptions.

Martin

Maybe this is off-base, but it
seems to me that the only words on which we can develop an agreement are
words for things that are accessible to the Observer in all the parties
concerned. There’s no need for words in the actual process of control
(though I have no doubt they can often be a component of some control
systems), so there’s no theoretical reason why there should not be
control systems for perceptions that are inaccessible to the Observer. It
seems reasonable that such perceptions could not have words to describe
them in ways that would mean anything to another
person.
[From Bill Powers (2008.01.26.0930 MST)]

I am copying this to Elizabeth Huang at the University of Manchester, UK
– a friend who is professionally interested in these things.

Martin Taylor 2008.01.26.10.26 –

In my experience (a phrase that is exceedingly important in all that I
have to say), perceptions for which we have no words are quite accessible
to observation, or Observation. The way a fork-full of spaghetti looks,
for example, on its way to your mouth. All I can do it allude to it;
there is no word for it. If you had never seen it, you would have no idea
what I mean.

Since this is a
subjective matter, the question has to be personal: Bill, does your
Observer actually observe the varying current states of your controlled
system-level perceptions, or is your belief in their existence derived
from your logic-level understanding of HPCT, with their states
unObervable?

In every case, the Observer is aware of the current perceptions (and
sometimes the beliefs concerning them, since the beliefs are other
perceptions).

This is understandable if you consider how such things come to be
Observed. I start out thinking that this sentence I’m writing right now
doesn’t seem to have – ah, here it is – an object. Then I notice what I
just thought, and indeed still seem to be thinking. Before I noticed it,
it was as if I were the thought, or as if I were looking through the
thought at the sentence without being aware that the thought was
modifying the experience. Then, somehow, I stepped back, and there in my
field of awareness was the thought. I can recreate the thought now, this
time without identifying with it, seeing it as an object of awareness
separate from me.

If anything, it’s the Observer that is hard to observe; the Observed
thought is easy, once you notice it. The Observer does not perceive
itself in the usual meaning of perception. The best I can do is to select
words that seem to reflect my experience of Observing, of being in a
point of view at the center of everything, of being that into which all
information comes, of being Here, not There. The Observer is simply me,
stripped of all observable attributes including selfhood. The attributes
are what I Observe, once I notice them as having separate
existence.

My Observer does
not see my top-level perceptions, so far as I know (though would I
know?). But I can, I think,

That penultimate “I” is the one to which you refer when you say
“my” Observer. The next one is a report that a thought
(concerning the words, “I can”) is going on. “The
Observer,” in that case, is a lower-level symbol used to indicate
the previous “I”. It is very hard to refer to the Observer, to
the real “I”, because all the reference-terms are of lower
levels. At some point, all we can do is turn to the other person and ask
“Do you recognize what I’m talking about?” This is the top of
the layered protocols, isn’t it?

Observe what
some of the reference values may be, perhaps by inference from the
actions I do to maintain those perceptions at appropriate levels, or by
Observing what seems to be “good” and what seems to be
“bad” behaviour in myself and in
others.

Yes, you can Observe those perceptions, and you can also Observe the
thought, “Those are perceptions.” And so is that, and that, and
that … And then the principle comes into view, which you can find
words to indicate: “Aha,” you say, referring to the experience
of going up a level, “Infinite regress!” referring to the
principle also called recursion. Once you see that thoughts can be
recursive, you collapse the series and see that they are ONLY thoughts,
means of manipulating symbols any way you please. For example, “My
name is not Martin Taylor,” a thought that either of us can think,
but with different meanings. “The statement on the other side of
this card is true” — “The statement on the other side of this
card is false.” That’s a paradox at the program level, but not at
the principle level. If your Observer is (if You are) mobile enough, and
your program level contains knowledge about the levels and the Observer,
such paradoxes are no problem at all, Douglas Hofstadter notwithstanding.
Or maybe he knew that.

If your Observer
does perceive the ongoing state of system-level perceptions, perhaps it
might be possible by some training technique (using examples such as Zen
training as a possible model) to help others to do the same. If that
could be done, maybe it would be possible to develop words for the
different possible states of those perceptions.

I’m working on it. So were/are other people, no doubt, past and present.
Since the object of description is above the level of words, there is no
one right way to refer to system concepts, not even “system
concepts.” Communication about system concepts can be established
only between people who have Observed system concepts
(“government” or “religion” or “physics”).
Layered protocols again. Anyone can do so, I maintain, through the method
of levels. But it can’t be done using words alone. And it’s hard to do
without someone else helping.

Best,

Bill P.

P.S. Onlookers may Google the search term “Taylor layered
protocols” to see what that is about. Martin developed this idea
before hearing of PCT.

[Martin Taylor 2008.01.26.13.49]

[From Bill Powers (2008.01.26.0930 MST)]

I am copying this to Elizabeth Huang at the University of Manchester, UK -- a friend who is professionally interested in these things.

Fine with me.

Martin Taylor 2008.01.26.10.26 --

Maybe this is off-base, but it seems to me that the only words on which we can develop an agreement are words for things that are accessible to the Observer in all the parties concerned.

But your response addresses the complementary proposition...

In my experience (a phrase that is exceedingly important in all that I have to say), perceptions for which we have no words are quite accessible to observation, or Observation.

I think we all agree on that. Those perceptions could, however, be graced with words, if there were a communicative need for them. My question was prompted by the difficulty of communication between you and Kenny, and by long and unproductive thought about the nature of "the Observer".

The thought is that if there are controlled perceptions that cannot be Observed, it would be very hard to generate words for them outside of an analytic consideration of a model with which the communicating parties could agree. On the other hand, perceptions that can be Observed could be graced with verbal labels, even if their content is different across people.

The way a fork-full of spaghetti looks, for example, on its way to your mouth. All I can do it allude to it; there is no word for it. If you had never seen it, you would have no idea what I mean.

The clue there is "If you had never seen it". If you had seen it, and I told you that I called that look "svirtly", you might not know just what qualities were encompassed by "svirtly" -- whether it had to be red, for example -- but you would have a basis on which to build a more precise agreed meaning. If you had never seen it, I doubt you could do that so readily.

That was the reason for my question:

does your Observer actually observe the varying current states of your controlled system-level perceptions, or is your belief in their existence derived from your logic-level understanding of HPCT, with their states unObervable?

If neither you nor Kenny can Observe "the varying current states of your controlled system-level perceptions", then necessarily "words don't exist at the system concept level that I'm trying to describe with them." At least it would be necessarily true if my hypothesis is correct that you can't communicate words for the perceptions you can't Observe, but you should be able to create and use effectively words for the place of those perceptions in some cumminicable model.

Martin

Martin Taylor 2008.01.26.13.49

Maybe this is off-base, but it seems to me that the only words on which
we can develop an agreement are words for things that are accessible to
the Observer in all the parties concerned.

But your response addresses the complementary proposition…

In my experience (a phrase that
is exceedingly important in all that I have to say), perceptions for
which we have no words are quite accessible to observation, or
Observation.
[From Bill Powers (2008.01.27.0017 MST)]

I meant to add “… as well as perceptions for which we do have
words.” Of course we can always invent them.

I think we all
agree on that. Those perceptions could, however, be graced with words, if
there were a communicative need for them. My question was prompted by the
difficulty of communication between you and Kenny, and by long and
unproductive thought about the nature of “the
Observer”.

I’m not sure it’s unproductive – yet. The method of levels, if pursued
with persistence, leads to a state of contentless Observation, which is
an informative experience and can be talked about among those who have
been there (a lot of the terminology already exists).

The thought is that
if there are controlled perceptions that cannot be Observed, it would be
very hard to generate words for them outside of an analytic consideration
of a model with which the communicating parties could agree. On the other
hand, perceptions that can be Observed could be graced with verbal
labels, even if their content is different across
people.

Yes, labels can be invented – but picking out just what to label isn’t a
trivial problem. I think of perceptions AS the content of experience, but
there are unresolved issues there, because of all perceptual signals
being alike. Paul Churchland’s “Network theory of knowledge”
might offer a solution, but I don’t see what it is yet. The best picture
I have come up with is the analog computer, in which all variables are
represented in exactly the same form, as voltages, yet the computer as a
whole is a working model of something. The brain could work the same way,
but without labels on anything. Our computer simulations actually work
that way – under the symbols used in writing the programs there are just
numbers, all alike but for magnitude.

The way a
fork-full of spaghetti looks, for example, on its way to your mouth. All
I can do it allude to it; there is no word for it. If you had never seen
it, you would have no idea what I mean.

The clue there is “If you had never seen it”. If you had seen
it, and I told you that I called that look “svirtly”, you might
not know just what qualities were encompassed by “svirtly” –
whether it had to be red, for example – but you would have a basis on
which to build a more precise agreed meaning.

You’d at least be able to say, “Here comes svirtly
again!”

If you had
never seen it, I doubt you could do that so
readily.

I don’t think you could do it at all. Svirtly would have meaning only
when there is a perception for it to point to. “the way spaghetti
looks” has meaning only for people who have seen (or try, accurately
or not, to imagine) the way spaghetti looks.

If neither you nor
Kenny can Observe “the varying current states of your controlled
system-level perceptions”, then necessarily “words don’t exist
at the system concept level that I’m trying to describe with them.”
At least it would be necessarily true if my hypothesis is correct that
you can’t communicate words for the perceptions you can’t Observe, but
you should be able to create and use effectively words for the place of
those perceptions in some cumminicable model.

I meant to say (and thought I did say) that I can observe those current
states of controlled system-concept perceptions, at least the ones I’ve
identified. I don’t think they’re what Kenny is referring to as system
concepts, though. Hard to tell. You really need multiple ways to discuss
such things, so you can eliminate the irrelevant aspects and narrow down
to the essential ones.

Anyhow, using words is a matter of hierarchical functions, not awareness
(except for the reorganization aspects). A linguist develops a lot of
nice systematic ways of learning languages, which in principle could
operate without awareness once learned. I don’t know if they ever do. I
think lots of people use words without much awareness or editing, pretty
automatically. They use lots of words that evoke vague associations but
have no real meanings, like “intelligence.”

Best,

Bill P.