In a previous article, sethr…@clickshop.com (Seth Russell) says:
>First a note to eliminate confusion for any late commers. This train
>was started to discuss different Knowledge Representation (KR) schemes
>and even the data structures in which they can be coded. [...]
In an attempt to find common ground, menti…@scn.org now proposes:
How a neuronal fiber holds an idea or concept:
This Usenet posting focuses on the role and activity of a single
concept fiber for several reasons:
1. So as to state clear, unambiguous ideas.
2. So as to describe something BUILDABLE in hard-/software.
3. So as to respond to Seth Russell’s posts on representation.
4. So as to state the Mentifex mind-model at its simplest.
Imagine that you are a long neuronal fiber stretching across the
great neuronal sheet of the human cortex, and that you are the
brain-mind embodiment of a concept (but only PART of an idea).
As a nerve fiber you can be either on or off, energized or not
energized, and at states in between, but you do not HOLD your
concept in the sense of CONTAINING it. Just as no "homunculus"
sits in your head, likewise no "conceptunculum" lies in a fiber.
Let us say that you are the long neuronal fiber of the concept
"(to) eat." The word "eat" is NOT inside your fiber-body, but
rather you CONTROL the word "eat" at various temporal points in
your life-long auditory memory channel. If you as the "eat" fiber
become fully energized, or "on," you will fire signal-spikes
across associative tags stretching from your own synapses over
to the recall-synapses of forms of the word "eat" in the auditory
memory. Then your internal ear will hear, that is, THINK, the
word "eat." You as a concept-fiber will have caused part of a
THOUGHT — an expressiomn of an IDEA.
Let us back up to WHY you become active in co-creating a thought.
Both BLANKET inputs and SINGLE inputs may partially or fully
energize you — turn you "on."
BLANKET (semi-activating) signals come from other neuronal fibers
that are part of the SYNTAX control architecture. In your case,
when the Chomskyan syntactic mind is trying to generate a thought
or sentence to the effect that "X eats Y," a node on the neuronal
syntactic tree sends out a BLANKET signal which partially activates
many — perhaps ALL — verbs in the vocabulary of your host mind.
The host mind is grasping for (French:) "le mot juste" — the best
word to express the idea of consuming something with the mouth.
[You do not think about this search; you are only a cog in a Cog.]
SINGLE inputs come to you from SINGLE sources, such as from what-
ever NOUN your host mind is thinking about. Let us say that you
as the nerve-fiber for "eat" have been active in the past with
hundreds of different subject nouns such as "lions" and "tigers"
and "bears." If you have already been semi-energized by a BLANKET
signal from a syntax node trying to FLUSH out an appropriate verb,
and if a SINGLE spike comes to you from a particular subject such
as "bears," then you will go into full activation and you will
send out your own "SEMI-ACTIVATING" signals to other concept fibers.
You will semi-activate ALL the other (object) concept fibers ("Y")
with which you have ever been involved on the basis of "X eats Y."
If "bear(s)" is fully active, and if "eat" is fully active, then
many concept fibers of things that bears typically eat will be
SEMI-ACTIVATED. What pushes a particular "object" concept fiber
"over the top" into the FULL ACTIVATION that rouses a WORD-string
within the auditory memory channel? The TIME and CIRCUMSTANCE of
SPREADING ACTIVATION cause a particular OBJECT concept to turn on.
Although you as the concept fiber "eat" have been active with many
subjects and many objects, it was always at a particular time in
the life-history of your host mind, and so the JOINT influence of
the BLANKET syntax signal and the single ONE-TIME subject signal
from the "bears" concept will SELECT only that particular "object"
fiber (for associative activation) which has SIMULTANEITY tags
stretching from "X" to "Y" in the "X eats Y" nexus of association.
This revolting, shocking proposition says that our minds store
ideas and knowledge as a temporal, that is, TIME-BOUND nexus of
associative tags, typically centered around a linguistic verb
that links to subject-concepts and to object-concepts. Knowledge
is represented within a brain as SIMULTANEOUS links among the
constituent concept-fibers of an idea expressible as a sentence.
The Mentifex mind diagram below shows only one fiber per concept,
but similar concept-fibers holding the same concept are REDUNDANT
in MASSIVE PARALLELISM lest perish the thought if its underlying
concept fiber should die — as brain cells are wont to do. GANGS
of concept fibers holding a single concept act logically as ONE.
Hearing Vision Concepts Volition Emotion Motor Output
/iiiiiii\ /!i!i!i!\ /YYYYYYYYYYYY\
| ||||||| || ||||||| | B H E | |||||||||||| |
| ||||||| || | ___ | | + + + | |||||||||||| |
| ||||||| || / \ | +—-/ \ | |S|||||||||| |
| ||||||| || (bears)-|–+ + (eat) | |H|||||||||| |
| ||||||| || \___/ | + + \_/ | |A|||||||||| |
| ||||||| || | + +—+ __ | |K|||||||||| |
| || |||| || | +——————/ \ | |E|||R|||||| |
| ||e—–||———|——–+ ____ (fear)-|–*|||U|||K|| |
| |a||||| || ___ | + + + / \—\__/ | |||||N|||I|| |
| |||t||| || / \ | + +—+ / de- \——–|——*|||L|| |
| ||| ||| || (honey)-|—-+ + ( ci- ) | |||||||||L|| |
| ||||||| || \___/ | + +—+ \ sion /——–|———-*|| |
| ||||||| || | + + + \____/ | |||||||||||| |
—
http://www.scn.org/~mentifex/ AI Minds for Uncontrollable Robots.
menti…@scn.org (SCN User) writes: >
> In a previous article, sethr…@clickshop.com (Seth Russell) says:
> >First a note to eliminate confusion for any late commers. This train
> >was started to discuss different Knowledge Representation (KR) schemes
> >and even the data structures in which they can be coded. [...]
How are non-linguistic concepts represented? For example, the general
idea of a chess combination.
Are ‘concepts’ involved in motor skills, such as playing a tune.
Are arithmetical processes, such as addition or mutiplication,
non-linguistic?
In sci.logic, Verdigris <> saw fit to bestow on us:
>menti…@scn.org (SCN User) writes: >
>> In a previous article, sethr…@clickshop.com (Seth Russell) says:
>> >First a note to eliminate confusion for any late commers. This train
>> >was started to discuss different Knowledge Representation (KR) schemes
>> >and even the data structures in which they can be coded. [...]
>How are non-linguistic concepts represented? For example, the general
>idea of a chess combination.
I don’t really understand what is non-linguistic concepts. Your chess
combination can be represented (mentally) as a set of objects and thier
location relative to each other. Each object (Red King, White Pawn, Board)
is a concept by itself. The entire chess scence is represented as a set
of groups: if the Red King is just one squre above the white pawn you can
say that:
(There is a Red King x)
(There is a White Pawn y)
(There is chessboard z)
(The Red King is member of group X)
(group X is the group of all chess pieces currently on the z)
(The Red King is member of group Y)
(group Y is all the items that can kill the White Pawn)
etc.
How is that non-linguistic? besides, why do you assume that non-linguistic
items (if there are any) are not represented the same way linguistic items
are represented? (I think it’s better to ask: why do you think that *linguistic*
items are represented differently from *non-linguistic* items?)
>Are ‘concepts’ involved in motor skills, such as playing a tune.
That’s a good one. I think yes, I think that the concept of the note
A (whether it’s a prototype or a core concept) is important in order
to produce the tune of any song. It’s true that if we didn’t have the
concept of the note A we could still make noise, but not music.
(And I don’t mean that the ‘A’ note is the important item here, but
that we need a concept of a certain sound between such and such
frequentcy with such and such tone levels etc. to produce music.)
>Are arithmetical processes, such as addition or mutiplication,
>non-linguistic?
IMHO they are Non-Linguistic. I belive, and YMMV, that logics
are the basis for both Math and Lingu. but, like the apes and
the humans evloved from the same creature they took different
paths and altough similar they are inherently different. So that
alot of things in Math are similar to Lingu. but they are not,
IMHO, the same.
Just my 0.02 NIS.
/NL
—
Nir Levy, The above opinions are my own,
nlevy @ usa.net not my employer’s.
–
I didn’t do it; Nobody saw me do it;
You can’t prove anything; -Bart Simpson
–
[nl...@usa.net.TO_EMAIL_REMOVE_THIS (Nir Levy) writes: > In sci.logic, Verdigris <> saw fit to bestow on us:
[>
[> >menti...@scn.org (SCN User) writes: >
[> >> In a previous article, sethr...@clickshop.com (Seth Russell) says:
[> >>
[> >> >First a note to eliminate confusion for any late commers. This train
[> >> >was started to discuss different Knowledge Representation (KR) schemes
[> >> >and even the data structures in which they can be coded. [...]
[> >
[> >How are non-linguistic concepts represented? For example, the general
[> >idea of a chess combination.
[>
[> I don’t really understand what is non-linguistic concepts. Your chess
[> combination can be represented (mentally) as a set of objects and thier
[> location relative to each other. Each object (Red King, White Pawn, Board)
[> is a concept by itself. The entire chess scence is represented as a set
[> of groups: if the Red King is just one squre above the white pawn you can
[> say that:
[> (There is a Red King x)
[> (There is a White Pawn y)
[> (There is chessboard z)
[> (The Red King is member of group X)
[> (group X is the group of all chess pieces currently on the z)
[> (The Red King is member of group Y)
[> (group Y is all the items that can kill the White Pawn)
[> etc.
[>
[> How is that non-linguistic? besides, why do you assume that non-linguistic
[> items (if there are any) are not represented the same way linguistic items
[> are represented? (I think it’s better to ask: why do you think that *linguistic*
[> items are represented differently from *non-linguistic* items?)
[>
[>
[> >
[> >Are ‘concepts’ involved in motor skills, such as playing a tune.
[>
[> That’s a good one. I think yes, I think that the concept of the note
[> A (whether it’s a prototype or a core concept) is important in order
[> to produce the tune of any song. It’s true that if we didn’t have the
[> concept of the note A we could still make noise, but not music.
[> (And I don’t mean that the ‘A’ note is the important item here, but
[> that we need a concept of a certain sound between such and such
[> frequentcy with such and such tone levels etc. to produce music.)
[>
[> >Are arithmetical processes, such as addition or mutiplication,
[> >non-linguistic?
[>
[> IMHO they are Non-Linguistic. I belive, and YMMV, that logics
[> are the basis for both Math and Lingu. but, like the apes and
[> the humans evloved from the same creature they took different
[> paths and altough similar they are inherently different. So that
[> alot of things in Math are similar to Lingu. but they are not,
[> IMHO, the same.
[>
[> Just my 0.02 NIS.
[> /NL
[> —
[> Nir Levy, The above opinions are my own,
By linguistic I meant the specialised processing areas of the
human brain which process language. One can, of course, define
language very widely, so that a concept would have to be
linguistic in some sense.
It is conceivable that a person could lose the power of speech
but still be able to play chess, but I do not know if this is
a medical fact. The chess combination, it seems to me, is a
learned routine, like a phrase of music or a boxer’s counterpunch.
Such routines regarded as ‘programs’ could be viewed as being
associated with some unpecified language.
One might try the following definitions:
External Symbol: A state of system A representing a state of system B
Internal Symbol: A state of system A representing a state of system A
Concept: A state of system A corresponding to a symbol
Language: A system of symbols representing objects and relations.
Some examples of such symbols would be:
animal tracks -> animals
cave drawings -> animals, men
photograph -> two dimensional map of real objects
video -> three dimensional map of real events
bee’s dance -> flowering plants
PI -> attributes of a circle
Concepts:
Seeing animal tracks and remembering animals
Seeing the bee’s dance and knowing where to find flowers
When a blue tit pecks the cap of a milk bottle or when
an insect selects a specific type of flower is it
using linguistic concepts?
In the final analysis, I think that language and meaning are just
regularities which arise in the transformations of chaos. Their
absolute status is no different fom any other random occurrence.
in the case of music, I believe we could learn to regard certain
sequences of sound as music with all the rest consigned to noise.
From this point of view, a thought or concept is a state of mind/brain
(system A) which corresponds to a state of the environment (system B).