Arthur T. Murray
uj797 at victoria.tc.ca
Wed Sep 22 15:55:19 EST 1999
http://www.geocities.com/Athens/Agora/7256/mind-fpc.html Mind.Forth PD AI
is a free public-domain artificial intelligence program based on concepts:
Hearing Vision The Evolution of Concepts Motor Output
/iiiiiii\ /!i!i!i!\ Primitive Verbal Abstract /YYYYYYYYYYYY\
| ||||||| || ||||||| | + + + + + | |||||!|||!|| |
| ||||||| || | ___ | | + + / \ + + | |||||!|||!|| |
| ||||||| || / \ | + + / \ + + | |||||!|||!|| |
| ||||||| || (image)-|--+ + (blood) +---+ | |||||R|||!|| |
| ||||||| || \___/ | / \ + \ / __+__ + | |||||U|||!|| |
| ||||||| || |(red) + \ / (dan- )+ | |||||N|||F|| |
| ||||||| || | \ / +----+ \ger/-|----|------*|||I|| |
| ||||||| || | +---|----+ \ / _+_ | |||||||||G|| |
| ||||||| || | + + +------+ / \ | |||||||||H|| |
| ||||||| || | + / \ + +(cour-) | |||||||||T|| |
| ||||||| || | + (wet) + + \age/--|----------*|| |
| ||||||| || | + \ / + + \_/ | |||||||||||| |
| ||||||| || | + + + + + | |||||||||||| |
Chaumont Devin, devil at lava.net, wrote on 22 Sep 1999:
>There seems to be a lot of confusion and controversy about this matter of
>concepts and what I said about it, so I will try to shed some light on the
>problem. Before I do so, however, let me say that I am 57, and cannot
>always remember everything I know about something instantaneously as
>before. Anyhow, here are what I see as some serious problems with the
>idea of primitives or concepts:
>1. In the very first place, I am pretty sure they don't even exist.
>2. After spending years working out my theory of Panlingua, it is clear
>that all of the internal structures of language can be modeled without
>using anything else except links and nodes, where a node is defined as
>nothing but a connecting point. Because I look for elegance and
>simplicity at the fundamental levels of nature, I am forced to perceive
>the idea of concepts/primitives as problematic because it doesn't fit
>naturally into the larger scheme of things.
But if we E-L-O-N-G-A-T-E the nodes and map them onto nerve fibers
(actually, redundantly reliable gangs or fascicles of neurons), then
we still have nodes but we have solved the problem of massively
parallel *connection* to the nodes, because any concept anywhere on
the mindgrid can connect *at some point* to any other concept fiber.
> Of course I have no int>So just take a hint and read carefully what I have written on this
>subject, and then decide!
> >> And, of course, these concepts are something that /is/ *somewhere*
> >> in the brain or mind.
> CD> And so you come full circle right back to the myth of concepts in
> CD> the brain, and this, I believe, is a major fallacy underlying
> CD> many linguistic theorizings that look good at first but will not
> CD> work as complete models.
>>Sorry, but you did not get me - when I need a "concept" I can make it
>>up and talk about it. And, of course, the concept is somewhere in my
>>brain - but not necessarily as a permanent structure or something
>>locatable as brain cells xy300-af798. But somewhere, somehow it is
>It is precisely this where and how that I am after, because it is by
>understanding this that we will be able to model it correctly on computers
>in systems that really work as expected.
>>Why do I insist on the existence of concepts ? Because I can
>>clearly separate of what "belongs" to the concept and what not - I can
>>tell why "green ideas" can't "sleep furiosly".
>But what I am saying is that these "concepts" may only be a figment of
>your imagination, and that in fact what is really happening is that you
>have this or that link that links in this or that way to this or that
>English word, and links internally to this or that semantic node which
>links in this or that way to this or that other semantic node, which in
>turn links in this or that way to some other English word, etc., and that
>by creating this structure in your brain in fact you have done what you
>think of as setting up this "concept", when in fact there is no "concept"
>at all, but only these various links and nodes. And then to ME what is
>important is to know precisely how each of these links works, so that I
>can then transplant them to my computer model, which will then work the
>same way. This is the only foolproof way I know of for really dealing
>with CL. It doesn't appeal at all to high-minded philosopher types, but
>it works, and that is all I am after. But as it turns out, it also solves
>a whole slew of linguistic problems that have never been solved before,
>and so it will probably get me the nobel prize for something or other when
>people finally stop arguing with me and figure out what it is that I have
>>One related point: please have a look at people suffering from what I
>>think is called aphasia, some kind of brain damage in which the
>>ability to speak is damaged (I'm on slippery rocks here as I don't
>>have any references at hands, however a colleague of mine works on the
>>issue so I can ask him). There are different, well, versions of this:
>>there are people which have problems with making grammatically
>>well-formed sentences ...
>I would call this part "text generation".
>> ... and which similarily have problems judging
>>whether a sentence is grammatically well-formed or not.
>And I would call this one "analysis".
>>version is just the other way round: they cannot tell a grammatically
>>well formed but semantically nonsense sentence from a grammatically
>>and semantically well formed sentence. Well, as I am no expert, I will
>>make no further claims about what should be inferred from this ...
>I don't think I can make any claim, but if you subject this to Panlingua
>theory interpretation, then here are the results:
>Inability to parse or form grammatically correct sentences: damage to
>those parts of the brain that convert Panlingua representations to strings
>of words and vice versa.
>Inability to detect incongruousness: probably some kind of damage to the
>I do wish I had more time to look at these things, because this theory
>covers all of them, and I suspect that I might be able to make
>improvements to my theory as well as explain something about what is goind
>on in all cases by referring to the theory as it is. It is my hope that
>others will take this thing further. I am only one person,a nd certianly
>cannot possibly do everything int his field. Maybe sometime people will
>start listening to what I am saying and pick up the challenge.
>> ... - but I
>>think there is something in the brain of those people damaged that has
>>a lot to do with mapping of words to "concepts".
> >> Why else could it be that we are able to talk about concepts at
> >> all and know about their possible relationsships ?
> CD> You simply imagine them, just like you can readily imagine
> CD> anything else in or not in the universe.
>>I don't see how this contradicts what I have been saying. Perhaps I
>>wasn't as clear as I intended.
>No, you have been perfectly clear. What I am getting at is simply a
>refinement in our understanding of the internal processes involved.
> >> ./How/ .they are made up in the mind or brain is something that is
> >> a very important issue to some scientists, to me the /symbols/
> >> /representing/ the concepts will do.
> CD> And what I have been saying is that these may be fine for
> CD> linguistic imaginings, but that they will NOT do if you have any
> CD> serious intention to model languages on computers
>>Why ? What will people using this model of concepts etc. miss ? Or, in
>>what respect is it wrong and why ? (Please, when answering to the
>>"why" part, make some examples where common theories don't give
>>intuitive results or empirical results showing where a theory is wrong
>>- having intuitions is a good starting place, but only the proof of
>>such intuitions is scientific research. This comment is not intended
>>to shove any miscredit on your work - it is just that I have read to
>>many claims without proofs in these discussions here in c.a.n-l.)
>This is a good and reasonable challenge, and I am not sure that I can meet
>it head on. I have approached it obliquely above, falling back upon such
>things as elegance, universality, etc., and yet I feel these have been
>inadequate to meet this challenge you have posed. I think that in fact I
>AM relying somewhat upon gut feeling, but it is educated gut feeling, and
>not just thin air. How about the "Ockham's razor" principle? Would it be
>sufficient in this case? In other words, the idea of internalized
>concepts must not be right because it is possible to explain everything
>fully without them.
> CD> Something similar is now required in order to progress further in
> CD> our understanding of language.
>>I think "progress" is a broad term. You take your ideas as something
>>that you see as progress, I take mine.
>The reason is that I cannot see language in these terms, as I have
>explained above. This would be okay if in fact language were like other
>things in which many approaches were possible, but I an convinced that it
>isn't, and this is why I dare say that I have made concrete progress which
>people will ultimately be forced to recognize.
>>I am more interested in
>>understanding some aspects of language, most notably discourse
>This is of great importance, but studying it off by itself will get you
>nowhere because that is precisely the way it has been stodied many
>thousand times before. It is only by fitting it into the structure of
>Panlingua that it will make sense.
>>Computing this is something that is by no means obvious,
>>and the same goes for other problems as well.
>In fact it IS obvious when you understand Panlingua, because it is the
>traversing of previous Panlingua thought structures in order to interpret
>the present structure during parsing. I have not dared touch this one yet
>because of the limitations of time, but I need information about this very
>>I would like to see some
>>more discussions about such things in this group, BTW. There is too
>>much meta-discussion and flame-throwing going on here for my taste.
>Definitely, and I think I know why. As a friend told me some days ago,
>this is because some of us are here to discuss science, while others are
>here in order to cope, where coping means dealing with personal problems
>by posting on the Internet. I think this answer makes sense. Meantime I
>am getting the lion's share of publicity, so I cannot complain. In this
>way the knowledge that I have contributed will slowly percolate outwards
>despite the censorship imposed upon my work by academia (whether
>deliberately or not matters nothing), and the inevitable may yet happen
>while I am alive, so that unlike Mendel, I will see my contribution
>accepted and used before I die.
More information about the Neur-sci