Re: AI-Mind in Forth

On Jun 21, 11:03 am, Jonah Thomas <j2tho...@xxxxxxxxxx> wrote:
"Ed" <nos...@xxxxxxxxxxx> wrote:
"Jerry Avins" <j...@xxxxxxxx> wrote
Ed wrote:
"Jerry Avins" <j...@xxxxxxxx> wrote in message
news:xJWdnfO7J8lYe-rbnZ2dnUVZ_h6vnZ2d@xxxxxxxxxx
pablo reda wrote:
thinking IS a languaje based

The languaje is another skin of a physical body,
Humberto Maturana study this, he has a teory of knoledge
I believe that language is only the skin of thought. I know that
I have> conceived of some things "whole"; that is, without breaking
them down to> language. ...

Until someone pinpoints what a thought is (good luck on that)
saying it has a skin is as useful as as saying there is a "you"
which has a body.

I am sorry that my metaphor is lost on you. I would need a treatise
to express the idea without it, and the subject doesn't warrant
that.

Because one had no "words" to describe something doesn't mean
there wasn't, or it wasn't, language. If "language" is defined as
any form of communication then distinctions such as verbal,
non-verbal, internal, external are merely descriptive and largely
abitrary.

You may define language any way you like, just so long as you are
explicit about your definition and don't expect everyone to share it
as a matter of course.

With your sound example, there is at least an identifiable source.
That's not so with thought which science has yet to measure/isolate.
Consequently phrases such as 'skin of thought' are open to
interpretation. A nice turn of phrase to be sure, but not really
useful in discussions on AI etc.

I say that language is mostly metaphor. As long as we're talking about
the same exact thing we talked about and mutually understood before,
then it isn't. Then we have known symbols that refer precisely to known
things, and there's no ambiguity at all.

But when we apply what we said about one thing to something else, then
-- metaphor. Apply what you know about one fish to another fish, one
basket to another basket, one god to another god.... It gets even
murkier when you think of a fish's mouth as a basket to scoop up food,
or a god's cornucopia as a basket, etc. The metaphors get strained. But
what else can we do? Make up new words for everything, and teach each
other precise meanings for those words by demonstration?

I don't know whether the "skin of thought" metaphor will be useful for
AI. If you find a way to use it, then it's useful to you. If somebody
else doesn't get anything from it then it isn't useful to him. If it
turns out you can use the metaphor to get a new AI technique, then
somebody else might figure out the metaphor by working out how your code
works. That's a way that computer language can be clearer than other
languages -- in terms of *what* it says to do, it is utterly
unambiguous. But when it gets to what it means, and what the result is
good for, then we're no better than anybody else. We communicate with
computers much better than with end users....

Human memory is something like a database. A whole bunch of keys
connected to values. You do a search and each key is connected to other
keys, and you look for, among other things, a sort of *structure* to the
values, and you zero in on what you want while discarding lots of things
you don't want. The more of the keys you remember for the memory, the
easier it is to find it. When it was stored, other things you were
thinking about at the time, how you felt, where you were, thoughts it
led to, thoughts and circumstances that led you to think it -- the more
keys you have already the easier it is to retrieve it and the easier it
is to call up the other keys that confirm it's the one you want.

Language happens when people can share database keys. That's
communication. But do the keys refer to the same values in different
databases? No, it at least seems real implausible that different people
would have the exact same memories. (Even if they looked at the sme
Disney film they saw it in different circumstances and had different
thoughts about it.) For a key to be language, you need first a way to
show the key to somebody else. And then you need a way for them to get
close enough to the same responses for you to not notice they don't have
the same values.

Clearly, a big part of thought involves keys that we can't communicate
effectively, and those probably shouldn't be considered language.

You're metaphors is close enough to my understanding of language and
thoughts, but I'm afraid that comparing throughts to a database is and
keys to language referencing items in the database is far from being
accurate (the parallels are interresting though).

To comment on "communication" between a human and a computer vs a
human vs a human, programming a computer is straightforward and simple
as you mention because no ambiguities can exist (otherwise, the
computer simply can't process the request). Very strict rules of
predicate logic need to be followed for whoever wants to convey a
meaning. That's the only thing that a computer can handle because it
is fundamentally thought out to process information in a logical
manner.

The natural mode of thinking of human beings is NOT predicate logic in
a strict sense. We can follow rules and abide by them, but our thought
processes are much looser and seem to lean more towards hierarchical
organizations (because of the natural discrimination property of
neural networks).

Communication with a human being is a messy process, but it is
possible to converge towards an understanding through discussions and
a real intention of understanding the ideas that are being explored
(on both sides).

More often than not, the discussion is one sided on both side and it
is a real treat to find somebody who understand this messy process
enough to be aware that mutual understanding (to a certain extent) can
only be achieved through a serious discussion and a willingness to
understand.

You mention:

"Then we have known symbols that refer precisely to known things, and
there's no ambiguity at all."

From what I understand, this is incorrect. No symbol has the same
exact meaning for any individuals (even in mathematics; although the
explored concepts are much sharper in nature and thus easier to
explore and converge towards). There are as many different
understandings (internal representations) about a single symbol than
there are humans on the planet! We refer to words that are commonly
accepted. The accepted definitions set the norm, but we are all
slightly deviant from the norm, some more than others, in our
perception, and in any case, and to a certain extent, the norm is a
somewhat arbitrary line drawn in the sand used as a buoy that anybody,
hopefully, can refer to.

Regards
Jean-Francois Michaud

.

Relevant Pages

  • Re: AI-Mind in Forth
    ... I believe that language is only the skin of thought. ... I am sorry that my metaphor is lost on you. ... A whole bunch of keys ...
    (comp.lang.forth)
  • Re: In Office 2007 the single quotation mark changes the following
    ... However, for more information on language formatting, see ... "Graham Mayor" wrote: ... All other keys seem to work as designated. ... u i o a and c letters after the ' & " keys. ...
    (microsoft.public.word.docmanagement)
  • Re: I want to get rid of blue underlining in a table cell.
    ... You can use an international keyboard even when the language is set to U.S. ... This turns the quote keys and some others into "dead keys" that produce ... >> Any chance Word is interpreting your spine labels as "Internet or ...
    (microsoft.public.word.tables)
  • Re: Skandinavisk gruppe?
    ... If you set your language system to a foreign ... different keys. ... further modifies the US keyboard/language output. ... I'm familiar with the concept of the "country" settings ...
    (alt.smokers.cigars)
  • Re: How do i add or remove accents from text?
    ... and are getting what's call "dead keys" behavior. ... Language bar, ... Right click the EN and choose Settings. ... Windows not to use any keyboard shortcuts for switching languages. ...
    (microsoft.public.word.docmanagement)
Loading