language development-infant
Arthur T. Murray
In a previous article, ljmi...@iquest.net (Laura J Miller) says:
> I am hoping someone can help me, or at least direct me
> to some literature.........
> As a neurobiology graduate student, I have taken great
> interest in the neurodevelopment of my one year old
> baby (sigh, no longer a baby.....toddler). [...]
http://www.newciv.org/Mentifex/ is the archival Website of a lin-
quistic theory of mind based in part on ideas about infant speech
and language development. For instance, the Mentifex mind-model
attempts to explain a "spiral" of the infant acquisition of vari-
ous parts of speech, e.g., first nouns, then verbs, in a looping
spiral where the infant learns to build up standard (Chomskyan)
sentence structures. The following mind-diagram is an overview:
Theory of Child Language Learning
/^^^^^^^^^^^\ (spiral of additions & deletions) /^^^^^^^^^^^\
/visual memory\ semantic ________ / auditory \
| /--------|-------\ memory / syntax \ | speech stream |
| | recog-|nition | \________/<--|-------------\ |
| ___|___ | | flush-vector| | _______ | |
| /image \ | __|___ ___V___ | /stored \ | |
| / percept \ | /deep \<-----/lexical\<---|--/ phonemes\| |
| \ engrams /<--|-->/concepts\--->/concepts \---|->\ of words/ |
| \_______/ | \________/ \_________/ | \_______/ |
Spiros Triantafyllopoulos
Arthur T. Murray <ba...@lafn.org> wrote:
: Theory of Child Language Learning
: /^^^^^^^^^^^\ (spiral of additions & deletions) /^^^^^^^^^^^\
: /visual memory\ semantic ________ / auditory \
: | /--------|-------\ memory / syntax \ | speech stream |
: | | recog-|nition | \________/<--|-------------\ |
: | ___|___ | | flush-vector| | _______ | |
: | /image \ | __|___ ___V___ | /stored \ | |
: | / percept \ | /deep \<-----/lexical\<---|--/ phonemes\| |
: | \ engrams /<--|-->/concepts\--->/concepts \---|->\ of words/ |
: | \_______/ | \________/ \_________/ | \_______/ |
The flush vector is actually perfected after the third year when
they potty train... :-) :-) :-)
Spiros
--
at home in Central Indiana | software engineer (Tools/C/Unix/X11/Motif/GUIs)
www.primenet.com/~strianta | grad student,human factors engineering,Purdue U
stri...@primenet.com | "Reading, 'Rithmetic, and Readnews since 1983"
Arthur T. Murray
In follow-up, stri...@primenet.com (Spiros Triantafyllopoulos) says:
> Arthur T. Murray <ba...@lafn.org> wrote:
: Theory of Child Language Learning
: /^^^^^^^^^^^\ (spiral of additions & deletions) /^^^^^^^^^^^\
: /visual memory\ semantic ________ / auditory \
: | /--------|-------\ memory / syntax \ | speech stream |
: | | recog-|nition | \________/<--|-------------\ |
: | ___|___ | | flush-vector| | _______ | |
: | /image \ | __|___ ___V___ | /stored \ | |
: | / percept \ | /deep \<-----/lexical\<---|--/ phonemes\| |
: | \ engrams /<--|-->/concepts\--->/concepts \---|->\ of words/ |
: | \_______/ | \________/ \_________/ | \_______/ |
> The flush vector is actually perfected after the third year when
> they potty train... :-) :-) :-)
With my B.A. degree in ancient Greek and Latin I could have devised
a fancier term based on the Greek verb "skedannumi" ("to scatter"),
but "flush" meaning "to chase from a place of concealment" is fine.
In this mind-model of how a baby learns language, a syntax-node for
nouns forces out or "flushes" out the most highly activated lexical
concept among all the nouns which have taken residence in the mind.
Then the focus of linguistic control loops (spirals) back around to
a verb node and the infant says "Train go" or "Bird fly" or "I go."
The same spiral loop which generates sentences permits the addition
of new elements in the syntactic tree, or the deletion of erroneous
speech-patterns used by the child in a transition to proper speech.
Thanks to: any child psychologist who can help publish this model.
Spiros Triantafyllopoulos
: > they potty train... :-) :-) :-)
: With my B.A. degree in ancient Greek and Latin I could have devised
: a fancier term based on the Greek verb "skedannumi" ("to scatter"),
: but "flush" meaning "to chase from a place of concealment" is fine.
: In this mind-model of how a baby learns language, a syntax-node for
: nouns forces out or "flushes" out the most highly activated lexical
: concept among all the nouns which have taken residence in the mind.
: Then the focus of linguistic control loops (spirals) back around to
: a verb node and the infant says "Train go" or "Bird fly" or "I go."
: The same spiral loop which generates sentences permits the addition
: of new elements in the syntactic tree, or the deletion of erroneous
: speech-patterns used by the child in a transition to proper speech.
Sorry for including the whole thing. This is fascinating. Back in '84 I
was a graduate student studying computer science, and of course, back then
being the golden era of AI, I decided to do a (restricted) natural
language processing program for my thesis.
In the process I had to study how language is acquired by infants and
found many of the same things you report above. Since I had a question
only sentence structure and only one verb to worry about, I used
the verb as the starting point of the structure parsing. Nouns and
modifiers could be added to the verb to form the question. A rather
simple recursive transition network was used as the syntax
representation, with nodes for verbs, nouns, modifiers (adjectives,
usually), and so on.
In addition, I had to follow the child's like approach of learning.
First, learn a whole bunch of words, then connect them to make very simple
sentences, then define new words in terms you already know. The program
could actually start with a fairly limited set of definitions (initial
knowledge of definitions and relationships) and ask, child-style, when a
word that was unknown was encountered. If the unknown word was defined in
terms of known words and relationships, it was added to the knowledge
base.
Check my presentation on "Knowledge Assisted Information Retrieval:
Techniques and Applications" in the proceedings of the 1985 ACM Computer
Science Conference in New Orleans, LA.
Joseph S. Wisniewski
[something that she only felt a need to post to sci.lang, about the
development of her child's linguistic ability]
Hello Mr. Murray. Why did you feel the need to chop part of
Ms. Millers post until all that remained was a paragraph that,
taken out of context, served only as an excuse to post one more
copy of your diagram? Why did you then crosspost this to several
other groups in which you are totally off topic?
Do you not realize that each repeated Mentifex post detracts from
whatever reputation you have as a serious scientist, and encourages
a preception (however false) among the internet community that your
theory is not valid?
Or, is it your purpose to actually discredit the Mentifex theory,
by appearing to be its strongest proponent, yet playing the fool?
> > I am hoping someone can help me, or at least direct me
> > to some literature.........
>
> > As a neurobiology graduate student, I have taken great
> > interest in the neurodevelopment of my one year old
> > baby (sigh, no longer a baby.....toddler). [...]
>
> http://www.newciv.org/Mentifex/ is the archival Website of a lin-
> quistic theory of mind based in part on ideas about infant speech
> and language development. For instance, the Mentifex mind-model
> attempts to explain a "spiral" of the infant acquisition of vari-
> ous parts of speech, e.g., first nouns, then verbs, in a looping
> spiral where the infant learns to build up standard (Chomskyan)
> sentence structures. The following mind-diagram is an overview:
The following diagram is more interesting:
---------
Drugs -----> / \
/ Brain \ ---> Mentifex
Sleep Deprivation --> \ /
\ /
---------
Sorry, folks, but I could not resist.
--
Joseph S. Wisniewski | Views expressed are my own, and don't reflect
Ford Motor Company | those of the Ford Motor Co. or affiliates.
Project Sapphire | LeMans, Daytona, Bonneville, and Sebring are
jwis...@ford.com | just races, won by people driving Ford cars!
Arthur T. Murray
David G. Mitchell deserves a response (below) to his objection:
> Arthur T. Murray <ba...@lafn.org> wrote:
>
>> Thanks. Your initial article was an invitation for my posting of
>> details of a theory of mind in newsgroups germane to linguistics.
> Is there anything you _don't_ regard as an invitation
> to post your picture yet one more nauseating time?
Please consider the rationales in descending order of importance:
1. The purpose of this never-before-seen-in-history worldwide fo-
rum is the exchange of ideas in an atmosphere of free speech.
2. Ideas suppressed in traditional forums will not here be silenced.
3. Antipathy to the unexamined message forces the issue of its truth.
4. It is easier to react against than to originate a theory of AI.
5. Since the Minsky-McCarthy Newell-Simon "founding fathers" AI
conference at Dartmouth in 1956, the long wait may now be over.
6. The Mentifex diagram of the public-domain theory of mind is
little larger than a signature but epitomizes 14 years of work.
7. Many germane posts cry out for a repetition of this message:
/^^^^^^^^^^^\ Concepts Interrelating With Words /^^^^^^^^^^^\
/visual memory\ interassociative ________ / auditory \
| /--------|-------\ memory / syntax \ |episodic memory|
| | recog-|nition | \________/<--|-------------\ |
| ___|___ | _V__flush-vector| spiral| _______ | |
| /image \ | /core\ ___V___ loop| /engrams\ | |
| / percept \ | /"self"\<-----/lexical\<---|--/ "you" \| |
| \ engrams /<--|-->/ "self" \--->/nonverbal\---|->\ "I" / |
| \_______/ | \ "other" )<--\concept- /<--|---\"I" / |
| | \"other"/---->\router /--->|--->\"you"/ |
| | \_____/ \_____/ | \___/ |
http://www.newciv.org/Mentifex/ run periodic Mentifex Web search.
J.P. Openshaw
In article <1997Feb10....@lafn.org>,
Arthur T. Murray <ba...@lafn.org> wrote:
>
>Please consider the rationales in descending order of importance:
< Stuff deleted... >
>6. The Mentifex diagram of the public-domain theory of mind is
>little larger than a signature but epitomizes 14 years of work.
In that case you must have some results and publications to show us?
Just repeatedly posting some convoluted ascii diagram is not a
convincing argument.
Please reveal what your diagram actually does, how we can use it
and why it will improve things. Otherwise dont expect people to take
any serious notice of you.
What is anyone supposed to make of this?
> /^^^^^^^^^^^\ Concepts Interrelating With Words /^^^^^^^^^^^\
> /visual memory\ interassociative ________ / auditory \
>| /--------|-------\ memory / syntax \ |episodic memory|
>| | recog-|nition | \________/<--|-------------\ |
>| ___|___ | _V__flush-vector| spiral| _______ | |
>| /image \ | /core\ ___V___ loop| /engrams\ | |
>| / percept \ | /"self"\<-----/lexical\<---|--/ "you" \| |
>| \ engrams /<--|-->/ "self" \--->/nonverbal\---|->\ "I" / |
>| \_______/ | \ "other" )<--\concept- /<--|---\"I" / |
>| | \"other"/---->\router /--->|--->\"you"/ |
>| | \_____/ \_____/ | \___/ |
Please could you just post when you have something to say, as Mentifex
is becoming a bit tiresome. As is your cross-posting.
John Openshaw
Joseph S. Wisniewski
Arthur T. Murray wrote:
>
> David G. Mitchell deserves a response (below) to his objection:
No, he deserves an answer to his question. You did not answer his
question.
He only posted his question to scil.ang, sci.cognitive and
comp.ai.nat-lang. He did not post it to comp.speech,
soc.history.science, or alt.memetics. The readers of those groups
deserve a response to this question: why do you keep crossposting
to inappropriate groups? Why do you use multiple freenet accounts
to do it?
David asked:
> > Is there anything you _don't_ regard as an invitation
> > to post your picture yet one more nauseating time?
"Arthur" responded:
> Please consider the rationales in descending order of importance:
>
> 1. The purpose of this never-before-seen-in-history worldwide fo-
> rum is the exchange of ideas in an atmosphere of free speech.
Invalid point. After several hundred posts of your diagram, it
it can hardly be said to have never been seen. And, considering
it's similarity to other oversimplified diagrams that have been
kicked around for decades, "never-before-seen" is totally
inappropriate.
> 2. Ideas suppressed in traditional forums will not here be silenced.
Also invalid. By generating sufficient negative public opinion of
yourself, you silence your own ideas in this forum. Now, provide
some evidence that your idea has been supressed in more traditional
forums.
> 3. Antipathy to the unexamined message forces the issue of its truth.
Again invalid. We've examined it.
> 4. It is easier to react against than to originate a theory of AI.
Since most consider your theory to not be original, that would put
you in the catagory of those who "react against."
> 5. Since the Minsky-McCarthy Newell-Simon "founding fathers" AI
> conference at Dartmouth in 1956, the long wait may now be over.
This is not a point in itself, meerly a lead in.
> 6. The Mentifex diagram of the public-domain theory of mind is
> little larger than a signature but epitomizes 14 years of work.
This is unclear; is it an apology for extremely low productivity?
> 7. Many germane posts cry out for a repetition of this message:
---------
Drugs -----> / \
/ Brain \ ---> Mentifex
Sleep Deprivation --> \ /
\ /
---------
--