Re: Symbol Grounding Problem (attn: Vend)

On Sat, 05 May 2007 00:57:24 GMT, Neil W Rickert
<rickert+nn@xxxxxxxxxx> wrote:

HMSBeagle <jsbach@xxxxxxxxxxxxx> writes:
<rickert+nn@xxxxxxxxxx> wrote:

I'm not even sure why you are arguing that. I have been opposed to
the dictionary definition idea of meaning, and have been criticizing
it hear for years. Yet you respond as if that is what I am assuming.

I said I don't beleive in ungrounded symbol systems anymore. You
challenged me about pure mathematics not existing. I showed that pure
mathematics is not contained in what I was talking about, and
therefore is not a counter-example to "prove me wrong".
ol and its defn, does not copy

No, you didn't show that. You claimed to show that, but your
argument was faulty.


Thus we are to conclude what? That pure math is an actual physical
example of an ungrounded symbol system? What are you going to do now
with that conclusion? Are you going to teach a computer how to read
English by handing it a large newspaper corpus and having it apply
bayesian stats to words that appear near each other? It has been
tried before.




I also demonstrated WHY a deductive system must be ungrounded.
To recapitulate: If a deductive system were grounded in the way
natural language is, it would become inconsistent rather quickly.
Exemplar: "This sentence is false."

That's the part that is laughable. Grounding isn't what allows
the liar paradox.

Grounded symbol systems (eg natural language) have a contextual
substrate of speaker and listener. Deductive proofs in math do not.

That seems to agree with my earlier point about ungrounded symbols
and pure mathematics.

What point was that?

You have an inability to differentiate between the "internal logic" of
a deductive proof and the THING that humans engage in called "pure
mathematics." Within the INTERNAL LOGIC of a deductive proof, yes,
all those symbols are ungrounded and ungrounded for specific reasons
that I pointed out again and again for you.

I do not beleive that ungrounded symbols systems exist. I no longer
beleive in them.


A hilbert-style proof does nothing more than connect the logical dots
between statements that are "already true" within the "platonic
system" of symbols from whatever-branch-of-math.

That doesn't seem to be a good description.

What is your point?


Natural language goes from the "wild panorama" of biological sense
organs up INTO the system of agreed-upon symbols. This is precisely
what groundedness means. Deductive math never leaves the symbolic
universe. If it did, you get these loopholes where one can make
self-referential statements, since the "wild panorama" would naturally
contain them.

However, there are plenty of self-referential statements within
mathematics. So something is wrong with your argument.

No. There are none. The ZF axiomatic system is widely known to be a
consistent deductive system.


Rather the mathematician integrates the defn into his RATHER
GROUNDED system of understanding things.

This is a typical non-mathematician's misunderstanding.

A veiled ad hom?

No. I'm just correcting the record. As far as I know, saying that
somebody has a typicl view isn't usually considered ad hominem.

I was about to say you know nothing about me or my level of education.

I can make a reasonable inference that you don't think like a
mathematician, based on your comments.

Oh that's nice. You are going to tell me how I think now.
Did you study Advanced Troll under Professor Zick?


Are you going to tell me you are a post-doc mathematician at Stanford?

I am certainly not at Stanford.

But what are we going to do now? Spiral into a litany of ad homs and
personal attacks?

I can only repeat that "typical" is hardly insulting. It certainly
was not intended as a personal attack.


Then tell us all what the INTENT was exactly, since you are so
innocent.


In mathematics the defns are stated unambiguously so that any
time a symbol appears a pure substitution can be made without any
loopholes for interpretation.

I guess you are not aware of the non-standard mathematics (based
on non-standard interpretations).

Non-standard analysis is still used in consistent axiomatic systems.

However, the existence of non-standard analysis contradicts your
assertion about "loopholes for interpretation."

I meant "mathematics" in terms of "what working mathematicians do".

Working mathematicians do many things. And some of them do
non-standard analysis.

You are beginning to cut whole paragraphs out of my repliein order to
take my sentences out of context. Earlier you were trying to
appeal to the "groundedness" as if it were a thing that is contained
within the sentential logic of mathematical deductive systems.
However, since I am a mathematician myself, I didn't fall for that
trick. This is the very reason I brought up Hilbert and his use of
"meta-mathematics". However, you deleted all of those paragaphs to
make your own reply seem relevant. I invite anyone who is reading
this to go back and read my ACTUAL COMMENTS that Mr. Rickert cleverly
deleted.


I meant within the context of a hilbert-style proof (aka a "rigorous
proof") there are no loopholes for interpretation. I did not mean
"mathematics" in terms of "foundational mathematics where we juggle
axioms however we feel like it." (Hilbert himself called this
meta-mathematics. )

During the substitution of defn for symbol, there IS NO loophole for
interpretation.

Mathematics is versatile, can be used to solve many practical
problems, precisely because we can interpret it in ways that are
useful to apply to practical problems. You don't have to be involved
with foundational issues to use interpretations.

Versatility does not equal room for interpretation.


---

That's all dogma. I expect the Chomskyans might agree. But I take

A general comment on Chomskyan linguistics. As is obvious from my
postings in this thread, I disagree with it. It is equally obvious
that you agree with it.

Chomskyan linguistics has many supporters. It is considered a
respectable position. There is no need for you to be defensive about
your support for it. We are bound to disagree over it. And it is
unlikely either of us will persuade the other. Let's not spend too
much time on a fruitless argument.

No. This is bullshit. You have already made wild attempts to wed me to
Chomsky like some sort of bride. I denied your attempts at every
turn. You want to address the specific content of an argument with me
go ahead. This is not a game where we pick Team Colors and ally
ourselves under banners and then fight it out.

I am not here as an acolyte of Chomsky. I am not a Chomsky-ite. I
have done nothing but post a particular argument that he made.
Nothing more than that has happened and nothing less than that has
happened.


Chomsky points out that turning a statement into a question would be
much easier by simply reversing the words in the sentence. Why does
no human language use this simple technique?

That's one of the sillier Chomsky arguments.

I find it rigorous and persuasive.

That's because you are beholden to Chomskyan dogma, despite your
protests to the contrary.

I like how you use this flowery language to avoid having to address
the actual content of the argument that I typed up. Is it wrong
because "Chomsky said it"? That's not an acceptable way to
demonstrate something.

There isn't an argument to address.

Chomsky claims that reversing the words would be a grammatically
simpler way of forming a negative. He is probably correct. Chomsky
takes that as evidence for an innate universal grammar. To me, it
supports the view that language is not a grammatical system at all.


Fine. But you are using "grammatical system" to refer to a specific
Chomsky architecture which I never subscribed to. And furthermore, you
don't have to subscribe to it to make certain arguments.

But there is grammar in language and demonstrably so. If I change the
order of the words in a sentence it changes its meaning.



It's a silly Chomsky argument, because most of the argument is
in the assumptions that Chomsky brings to the discussion. If you
make different assumptions, as I do, then the argument falls flat
on its face.

---

But this is precisely Chomsky's point. The brain of humans has
PECULIARITIES that are formed from the tug-and-pull of natural
selection.

If Chomsky actually said that, I would like a citation.

I may be confusing him with Pinker here.

That's more plausible. Chomsky avoids discussing evolution. Pinker
does try to argue that language evolved.

Perhaps I still wasn't clear. My position is that natural languages
are not grammatical systems, and that brains are not following
grammar rules. Grammars are invented by linguists, and imposed
on languages, to help those linguists in their systematic study of
languages. But the grammars never quite fit (much as Boyle's law
doesn't quite fit). You get ever more complex grammars when you
try to fix up your grammar to make it closer to fitting, instead
of simply accepting that languages don't fit our imposed grammars.

Right. This paragraph demonstrates what you meant by "complex".
"Complex" to you means the ENTIRE GRAMMAR SYSTEM is complex after you
collect all the rules up in a pile.

That is not what I meant when I used it. What I meant was a singular
rule of grammar, taken all by itself in one single sentence is
complex. In fact, it is more complex than it needs to be.

However, if natural language speakers are not actually following
any grammar (which is my view), then even the complexity of a
single grammar rule is not relevant to how we use language.


Yet if you move words around in a sentence it can change its meaning.
Hmm.. ya know what. Scratch that. Because that could be an anecdote
after all. I may end up agreeing with you on this point in the end.

Moving a word around in a sentence is almost gauranteed to make the
sentence meaningless. You have to carefully craft a sentence where
there is still meaning after moving the words around. Thus anecdotes.



Chomsky said "There is no way a child could have simply learned the
complexities of language by itself using imitation alone. There must
be a deep grammar in the genes of toddlers."

I not sure whether Chomsky used those words. But his view is some
approximation of that.

And I hope we don't disagree on THAT point since I'm going to
springboard off it now with this:

[1] Did Chomsky mean this complexity is in the GRAMMAR SYSTEM taken as
a whole?

Chomsky argued on the difficulty of the learning problem for
a child. According to Chomsky, a child could not learn to speak
grammatically, with the evidence available to the child. Chomsky
supports this with his "poverty of stimulus" argument. There is
actually a research paper by E.M. Gold, "Language identification
in the limit", Information and Control 1967. This paper (which I
have not studied in detail) claims to prove that a formal language
(i.e. the type of grammatical system that Chomsky studies) could not
be learned without negative evidence (presumably somebody pointing
out to the child that he got the grammar wrong). Yet, according to
Chomsky, children acquire language without such negative evidence.
Chomsky concludes that part of grammaticality must be innate.

You are correct that there are cases where the referent is
interpretational.
However, there has to exist a socially-agreed-upon proper referent,
somewhere in the mix.

Why does there have to exist such a "proper referent"? And wouldn't
the existence of a "proper referent" prevent meanings from changing
over time?

If there were no proper referent, communication between two human
beings would never exist.

I have heard that argument before. I don't find it at all
persuasive. In my opinion, the argument is based on mistaken ideas
about how language works.

Well I think at the very least there is a proper referent, even from
the standpoint of radical behaviorism. It's not like im arguing for
something even more suspicious, such as demanding that the meaning in
each person's head is identical in terms of chemistry.
.

Relevant Pages

  • Re: Symbol Grounding Problem (attn: Vend)
    ... It was not referring to "meaning" ... natural language is, it would become inconsistent rather quickly. ... They fit into a Chomsky hierarchy ... And you meant the "complexity" arrises because some of the grammar ...
    (comp.ai.philosophy)
  • Re: unnatural languages
    ... language by consciously "looking up" each word in a mental primer. ... Even reading is mastered by children before "grammar" instruction ... division of mathematics, but in the sense of adding, subtracting, ... The immediate level is probability, ...
    (sci.lang)
  • Re: The Minimalist Program
    ... neurolinguistics better than any other metaphor or analogy that might ... I don't think Chomsky himself has ever claimed that generative ... grammar has neurological under pinning. ... He has absolutely no interest in e.g. language ...
    (sci.lang)
  • Re: THE QUIZ...
    ... >>>You are of course free to use 'language' in this very broad sense. ... > Grammar is what it is, irrespective of the theoretical approach or model. ... by Chomsky, however, not a practical one; ... language in animals - I did not study linguistics and Chomsky but I ...
    (sci.lang)
  • Re: Symbol Grounding Problem (attn: Vend)
    ... challenged me about pure mathematics not existing. ... natural language is, it would become inconsistent rather quickly. ... Chomsky claims that reversing the words would be a grammatically ... takes that as evidence for an innate universal grammar. ...
    (comp.ai.philosophy)