Inconsistent theories (let peace return)

CSGnet Archive CSG2009
21

[From Bill Powers (2009.10.07.1005 MDT)]

Martin Taylor 2009.10.07.00.40 –

This is probably the basic
reason why Shannon divorced information from meaning.

Actually, Shannon said that because the communication channel had to
support the communication of meaning for all sorts of people and topics,
the communication engineer must not consider meaning. That’s a little
different.

Let’s back off this 'tis so 'taint so argument at too low a level. My
position is that meaning never gets transmitted at all, so we’re not
talking about the same things. You can see a character being encoded and
sent off down a wire and arriving at the destination and being decoded
and read. You can check and see whether the intended character was
received. You can define channel capacity that way.
We can’t see any of that happening with meaning. The way I see meaning,
it is what your own brain attaches to incoming perceptions, and it
consists of other perceptions of your own, and nobody else’s, that have
been recorded and are played back when the received signal appears. The
characters that are sent down a wire do not “carry” meaning.
They are intended to evoke meanings, and the meanings they evoke can be
vastly more (or less) complex than the signals intended to evoke them. I
can allude to “physics” with a single word; it’s meaning is
incomparably more detailed than the word – except in the mind of someone
who knows nothing about physics. The meaning to someone who is familiar
with physics couldn’t be transmitted in 7 characters, 56 bits.
Communication is more a matter of finding resonances or comparable
experiences than it is of sending anything from one brain to another. If
you are red-green color blind you don’t know what “yellow”
means to me. I can send that word to you, but the experience you attach
to it will not be my experience that I’m trying to remind you of. It will
be your experience it reminds you of. Dalton’s account of
discovering his own color-blindness (his first scientific paper) reveals
the many ways he managed to get meaning from other people’s allusions to
color phenomena – meanings which he discovered after many years, to his
shock and dismay, to be totally spurious.

This happens on CSGnet all the time. It happens between you and
me.

Here is a link to Galton’s paper, called “Extraodinary facts
relating to the vision of colour” (I hope its length doesn’t keep
the link from working):

[

](Readings in the History of Psychology - Wayne Dennis - Google Books)In case the link doesn’t work, I also attach a screen shot of the
first two pages of Dalton’s paper. I hope this doesn’t exceed copyright
limits; please use Google Books to get more of the paper (Google the
title: Extraordinary facts" etc… This paper is extremely
appropriate for the present discussion.

Best,

Bill P.

Dalton.jpg
Dalton.jpg1158×847 180 KB

···
22

[From Rick Marken (2009.10.07.1815)]

Bill Powers (2009.10.07.1005 MDT)--

My position is that meaning never gets transmitted at all, so we're
not talking about the same things.

This is a marvelous post, Bill. The paper by Dalton is wonderful.

The way I see meaning, it
is what your own brain attaches to incoming perceptions, and it consists of
other perceptions of your own, and nobody else's, that have been recorded
and are played back when the received signal appears.

I see it that way too (if I get your meaning;-))

The characters that
are sent down a wire do not "carry" meaning. They are intended to evoke
meanings, and the meanings they evoke can be vastly more (or less) complex
than the signals intended to evoke them.

I think there is another possible sense of the "meaning" for a
perceptual signal, and that is the perceptual function. If p = x+y
then variations in p "mean" variations in the sum of x and y; if p = x
* y then variations in p "mean" variations in the product of x and y.
Of course, p doesn't "carry" meaning given this meaning of "meaning"
either; the signal p is just variations in neural impulse rate
regardless of the function of the inputs to the perceptual function
that produces the variations in p.

Your theory of "meaning" is certainly the right one for the dictionary
definition of meaning: "the thing one intends to convey especially by
language". But mine is somewhat similar. It's the thing conveyed
about the environment especially by the perceptual;-)

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

23

[From Dag Forssell (2009.10.07. 20.00]

[Rick Marken (2009.10.07.1815)]

Bill Powers (2009.10.07.1005 MDT)--

<snip>

> The way I see meaning, it
> is what your own brain attaches to incoming perceptions, and it consists of
> other perceptions of your own, and nobody else's, that have been recorded
> and are played back when the received signal appears.

I see it that way too (if I get your meaning;-))

Seems you finally have come around to recognize that there has to be memory in perception. Glad to see that you have internalized this insight. You sure fought the idea long and hard, but hope for progress springs eternal [or some such].

Best, Dag

24

[From Bill Powers (2009.10.08.0917 MDT)]

Rick Marken (2009.10.07.1815) --

I think there is another possible sense of the "meaning" for a
perceptual signal, and that is the perceptual function. If p = x+y
then variations in p "mean" variations in the sum of x and y; if p = x
* y then variations in p "mean" variations in the product of x and y.
Of course, p doesn't "carry" meaning given this meaning of "meaning"
either; the signal p is just variations in neural impulse rate
regardless of the function of the inputs to the perceptual function
that produces the variations in p.

I was thinking more specifically about communication through the transmission of symbolic perceptions, where what is to be communicated is not the perception transmitted, but something else that it "means." Consider the message "The only letter between A and C". The meaning of that message is a perception of the letter B, which is not transmitted. You have to interpret the message and solve the little puzzle by figuring out what letter is between A and C, and imagine B. That meaning of the message comes totally from inside you -- unless you're Russian and only do Cyrillic in which C is a long way from A. Then the message has no one meaning.

Even the meaning of the word "red" has to come from your own memories of sensations.

Your theory of "meaning" is certainly the right one for the dictionary
definition of meaning: "the thing one intends to convey especially by
language". But mine is somewhat similar. It's the thing conveyed
about the environment especially by the perceptual;-)

Yes, I can see that as covered by the word "meaning." It's not specifically part of communication, however, where one perception evokes another at the same level or other levels. Perhaps this works if you communicate a set of lower-order perceptions as a way of demonstrating a higher-level perception such as "before". If you see A and then a little later you see B, how do you fill in the blank in "A happened _____ B"?

Best,

Bill P.

25

[From Rick Marken (2009.10.08.1030)]

Dag Forssell (2009.10.07. 20.00)

Rick Marken (2009.10.07.1815)--

Bill Powers (2009.10.07.1005 MDT)--

> The way I see meaning, it
> is what your own brain attaches to incoming perceptions, and it consists
> of other perceptions of your own, and nobody else's, that have been
> recorded and are played back when the received signal appears.

I see it that way too (if I get your meaning;-))

Seems you finally have come around to recognize that there has to be memory
in perception.

I think I understand what Bill said somewhat differently than you do.
I don't think Bill was saying anything about whether or not there _has
to be_ memory in perception. I think he was talking about just one
particular phenomenon: meaning. I think Bill was pointing to the fact
that meaning seems to involve the attachment (or association) of
remembered perceptions to other perceptions, such as the perception of
a word. So the meaning of a word perception like "apple" consists of
remembered perceptions, such as the remembered image, crunch and taste
of an apple, that are evoked by the perception of the word "apple".

It's true that the "meaning" perception (p) that results when a word
evokes a remembered perception can be conceived of as a higher level
perception that is a function of the lower level perceptions -- the
word apple (w) and the remembered image of an apple (i) say. It could
be modeled this way:

i -----> | |
           > f() |----> p
w -----> | |

where f() is the perceptual function that combines the word (w) and
remembered image (i) perceptions into a perception of the meaning (p)
of the word "apple". So this is a case where a high level perception,
of meaning (possibly a category type perception?), is a function of
both a "normal" lower level perception (of the word "apple") and of a
remembered perception (the remembered image, i).

What I questioned (and still question) is whether memory is
significantly involved in all perception. I'm certainly willing to
believe it is. I was (and still am) opposed, however, to just taking
your (and Bill's) word for it that memory is always involved in all
perception because: 1) so far I have found it unnecessary to include
memory in the perceptual component of my control models; the models
typically fit the data almost perfectly so it's not clear that the
addition of memory would improve things significantly, if at all. 2)
there is no research I know of that demonstrates the involvement of
memory in the perceptual side of control; it would be great if someone
did such research but I'm not going to be the one to do it since there
seems to be plenty to do in terms of research on the regular, plain
vanilla control model described in B:CP. 3) my own experience tells me
that memory is sometimes involved in "filling in" perceptions, and I
think it may sometimes affect my controlling; for example, perceiving
a room to look like it usually does (as remembered) so it's hard to
control for finding the glasses that are sitting on the mantel, where
they never usually are; I'm just not sure how prevalent this is and,
even if it is going on all the time, how much it affects our ability
to control.

So I'm not denying that memory is involved in perception, possibly
significantly. All I'm saying is that we don't know much about it.
There is no research that I know of on this; if you know of some then
I'd like to see it. But I don't believe things just because people say
they are true, even if they are people as brilliant as you and Bill.

Best

Rick

PS. There is a very interesting article on "Scientific Management" in
the most recent _New Yorker_. Apparently it's a review of a book that
is quite an indictment of "Scientific Management" and it's current
incarnation as "Management Consulting". The article is available on
the net at Not So Fast | The New Yorker.
I haven't finished it yet but I'd be interested in hearing what you
(and Kenny and any other Management Consultants on this list) have to
say about it, if you get a chance to read it.

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

26

[From Fred Nickols (2009.10.08.1400 EDT)]

[From Rick Marken (2009.10.08.1030)]

PS. There is a very interesting article on "Scientific Management" in
the most recent _New Yorker_. Apparently it's a review of a book that
is quite an indictment of "Scientific Management" and it's current
incarnation as "Management Consulting". The article is available on
the net at
http://www.newyorker.com/arts/critics/atlarge/2009/10/12/091012crat_atlarge_lepo
re.
I haven't finished it yet but I'd be interested in hearing what you
(and Kenny and any other Management Consultants on this list) have to
say about it, if you get a chance to read it.

The New Yorker piece is very interesting. Thanks for pointing me to it. I finished Stewart's book - The Management Myth - a couple of weeks ago so the New Yorker piece is a welcome addition.

I don't think Stewart's book is an indictment of "scientific management." It is instead an indictment of the view that management is or can become a science. Taylor saw work and "the system" at the center of things; Stewart posits people at the center and claims there can be no science there. Moreover, there's more to management consulting than just tools or techniques traceable to Taylor's work. The Human Relations thread traces to Elton Mayo who Stewart argues was a fraud (just as he argues Taylor was a huckster). Stewart pretty well demolishes strategy and strategic planning but Henry Mintzberg did that long before Stewart took out his ax. And, of course, there is the financial piece of management consulting and the accountants long ago taught us all we needed to know about cooking the books.

Both pieces are interesting - even fun reading - but neither is likely to change much of anything.

Management is sorely missing the PCT view of things but I have a nagging suspicion that fully comprehended PCT is far from appealing to most managers and downright frightening to at least some management theorists.

Anyway, thanks for the pointer.

Regards,

Fred Nickols
nickols@att.net

27

[From Bill Powers (2009.10.10.0931 MDT)]

Rick Marken (2009.10.08.1030) --

>Dag Forssell to Rick: Seems you finally have come around to recognize that there has to be memory in perception.

RM: I think I understand what Bill said somewhat differently than you do.
I don't think Bill was saying anything about whether or not there _has
to be_ memory in perception. I think he was talking about just one
particular phenomenon: meaning.

BP: Yes, that's true. It may be true that perceptions always have some components of memory in them, but I see that as happening at the input to the perceptual input function. Any perception is a function of multiple lower-order perceptions. When some of the lower perceptions are missing, the higher system might go partly into the imagination mode and get the missing perceptual inputs that way, rerouting the reference signal outputs to the missing lower systems to serve as the missing perceptual signals to its own perceptual input function. This would involve machinery that isn't shown in the standard model diagram.

Of course if all the inputs to a perceptual input function are present as signals from lower systems, no memory has to be involved. I haven't figured out a way to model this. What I would like here is a simple way to assure that if a given input signal is missing it will automatically be replaced by the corresponding reference signal -- without needing a little demon standing by and constantly checking to see if all the usual inputs are there (how would it perceive them, and on what switching system would it act?). We can't have both the real and the imaginary inputs present all the time, because then we'd have twice as much signal as desired, and if there was a disturbance it could be canceled just by the imaginary input without actually doing anything at lower levels to counter the real disturbance.

RM: It's true that the "meaning" perception (p) that results when a word
evokes a remembered perception can be conceived of as a higher level
perception that is a function of the lower level perceptions -- the
word apple (w) and the remembered image of an apple (i) say. It could
be modeled this way:

i -----> | |
          > f() |----> p
w ------> | |

That's how I conceived the "naming" process at the category level. Your symbol i stands for some set of lower-order nonverbal perceptual signals reaching the category level input function, while w is a word-configuration signal that is part of the same category. For example, the signal i indicating the word "Red" is part of the same category as the signal w indicating the sensation of redness. Either the sensation input or the word input can generate the category perception p, which is the meaning of the word or equally the perception of a color category recognized without words.

To put that differently, the perceptual signal p, indicating that one color category is present, can be turned on by either the word "red" (w) or the sensation of redness (i). So p is the meaning of either "red" or redness. This accomplishes the equivalent of memory association but without memory. However, the association is not directly between a word and a sensation, but occurs in the form of a meaning signal that can be evoked either by the word or by the color.

Lots of missing research here, so we can only guess.

Best,

Bill P.