Is the concept of [incorrect question] new?

[Peter Olcott]

An [incorrect question] is defined as any question such that a correct
answer does not exist from the set of all possible answers.

I have previously referred to this as a (semantically) ill-formed question.
An example [incorrect question]: What time is it (yes or no)?

Does anyone else have any knowledge or references to the concept of
a semantically incorrect question, or am I the originator of this concept?

[DKleinecke]

"What time is it (yes or no)?" would be called "an unfair question"
in normal, nontechnical parlance. Do we need a different term?

Do you consider "Have you stopped beating your wife?" a semantically
incorrect question?

[Peter Olcott]

On 2/1/2015 10:47 AM, DKleinecke wrote:
> On Saturday, January 31, 2015 at 5:58:34 PM UTC-8, Peter Olcott wrote:
>> An [incorrect question] is defined as any question such that a correct
>> answer does not exist from the set of all possible answers.
>>
>> I have previously referred to this as a (semantically) ill-formed question.
>> An example [incorrect question]: What time is it (yes or no)?
>>
>> Does anyone else have any knowledge or references to the concept of
>> a semantically incorrect question, or am I the originator of this concept?
> "What time is it (yes or no)?" would be called "an unfair question"
> in normal, nontechnical parlance. Do we need a different term?

If we want to formalize our semantic intuitions, then yes we need this
additional
term and its related concept. For example we know that there is something
wrong with the following sentence: Colorless green ideas sleep furiously.
Yet it seems that we lack a sufficient formalism to point out exactly
what is
incorrect about this sentence.

The last sentence of this reference most aptly describes what I think
the error is:
http://en.wikipedia.org/wiki/History_of_type_theory#G.C3.B6del_1944
By the theory of simple types I mean the doctrine which says that the
objects of thought (or, in another interpretation, the symbolic
expressions) are divided into types, namely: individuals, properties of
individuals, relations between individuals, properties of such
relations, etc. (with a similar hierarchy for extensions), and that
sentences of the form: " a has the property φ ", " b bears the relation
R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
fitting together.

> Do you consider "Have you stopped beating your wife?" a semantically
> incorrect question?

Only if you never beat your wife, (or never had a wife) then yes.

I am interested in the concept of semantically incorrect questions
(within the formal semantics of linguistics) because this concept
is useful to derive additional insights within the field of mathematics
and computer science.

Both Kurt Gödel's Incompleteness Theorem and Alan Turing's Halting
Problem proof are directly derived from the inability to correctly answer
semantically incorrect questions.

[Peter T. Daniels]

On Sunday, February 1, 2015 at 12:26:05 PM UTC-5, Peter Olcott wrote:
> On 2/1/2015 10:47 AM, DKleinecke wrote:

> > "What time is it (yes or no)?" would be called "an unfair question"
> > in normal, nontechnical parlance. Do we need a different term?
>
> If we want to formalize our semantic intuitions, then yes we need this
> additional
> term and its related concept. For example we know that there is something
> wrong with the following sentence: Colorless green ideas sleep furiously.
> Yet it seems that we lack a sufficient formalism to point out exactly
> what is
> incorrect about this sentence.

Chomsky's ENTIRE ENTERPRISE is based on the simple fact that THERE IS NOTHING
WRONG FORMALLY with that sentence, whereas "Furiously sleep ideas green colorless"
IS formally impossible.

If you don't understand that, then you are _certainly_ in no position to be
investigating formal semantics, whether Montague's version or anyone else's.

[Peter Olcott]

On 2/1/2015 12:26 PM, Peter T. Daniels wrote:
> On Sunday, February 1, 2015 at 12:26:05 PM UTC-5, Peter Olcott wrote:
>> On 2/1/2015 10:47 AM, DKleinecke wrote:
>>> "What time is it (yes or no)?" would be called "an unfair question"
>>> in normal, nontechnical parlance. Do we need a different term?
>> If we want to formalize our semantic intuitions, then yes we need this
>> additional
>> term and its related concept. For example we know that there is something
>> wrong with the following sentence: Colorless green ideas sleep furiously.
>> Yet it seems that we lack a sufficient formalism to point out exactly
>> what is
>> incorrect about this sentence.
> Chomsky's ENTIRE ENTERPRISE is based on the simple fact that THERE IS NOTHING
> WRONG FORMALLY with that sentence, whereas "Furiously sleep ideas green colorless"
> IS formally impossible.

So Chomsky only focused on syntax and ignored semantics...

There is nothing formally wrong with the syntax of that sentence,
yet there is something wrong with the semantics of that sentence
that is why it lacks a possible correct interpretation.

If a thing is colorless, then it is not green.
Ideas do not sleep.
I think that I may have slept furiously before, having an angry dream.

> If you don't understand that, then you are _certainly_ in no position to be
> investigating formal semantics, whether Montague's version or anyone else's.

I deg to biffer.

[Peter T. Daniels]

On Sunday, February 1, 2015 at 1:40:03 PM UTC-5, Peter Olcott wrote:
> On 2/1/2015 12:26 PM, Peter T. Daniels wrote:
> > On Sunday, February 1, 2015 at 12:26:05 PM UTC-5, Peter Olcott wrote:
> >> On 2/1/2015 10:47 AM, DKleinecke wrote:
> >>> "What time is it (yes or no)?" would be called "an unfair question"
> >>> in normal, nontechnical parlance. Do we need a different term?
> >> If we want to formalize our semantic intuitions, then yes we need this
> >> additional
> >> term and its related concept. For example we know that there is something
> >> wrong with the following sentence: Colorless green ideas sleep furiously.
> >> Yet it seems that we lack a sufficient formalism to point out exactly
> >> what is
> >> incorrect about this sentence.
> > Chomsky's ENTIRE ENTERPRISE is based on the simple fact that THERE IS NOTHING
> > WRONG FORMALLY with that sentence, whereas "Furiously sleep ideas green colorless"
> > IS formally impossible.
>
> So Chomsky only focused on syntax and ignored semantics...

Of course. He was firmly in the mainstream of American Descriptivist
linguistics, which regarded Meaning and semantics as something to be
dealt with in the distant future.

In the 1960s, his students began to deal with semantics within his framework,
and there soon emerged two distinct schools, the Generative Semanticists
and the Interpretive Semanticists. Each side called the other side "the bad
guys," though they remained friends and were united by e.g. their opposition
to the Vietnam War. This is the story told very well by Randy Allan Harris
in *The Linguistics Wars*, which represents his History of Science dissertation
at RPI. It's the best of the several accounts because he wrote as an outsider
and interviewed all the participants.

> There is nothing formally wrong with the syntax of that sentence,
> yet there is something wrong with the semantics of that sentence
> that is why it lacks a possible correct interpretation.

There have been quite a few interpretations, including several by poets.

> If a thing is colorless, then it is not green.
> Ideas do not sleep.
> I think that I may have slept furiously before, having an angry dream.

See? You're on your way to creating an interpretation.

> > If you don't understand that, then you are _certainly_ in no position to be
> > investigating formal semantics, whether Montague's version or anyone else's.
>
> I deg to biffer.

Dut you bon't know the territory.

[DKleinecke]

On Sunday, February 1, 2015 at 9:26:05 AM UTC-8, Peter Olcott wrote:
> On 2/1/2015 10:47 AM, DKleinecke wrote:
>
> > Do you consider "Have you stopped beating your wife?" a semantically
> > incorrect question?
>
> Only if you never beat your wife, (or never had a wife) then yes.

Now I am confused. You have used context to defines correctness. It
would seem that the sentence is neither correct nor incorrect of
itself (assuming your "incorrect" is a binary evaluation) but only
the pair sentence + context is incorrect. If that's what you mean
you should specify it.

[Joe Fineman]

Peter Olcott <OCR4Screen> writes:

> On 2/1/2015 12:26 PM, Peter T. Daniels wrote:
>> On Sunday, February 1, 2015 at 12:26:05 PM UTC-5, Peter Olcott wrote:
>>> On 2/1/2015 10:47 AM, DKleinecke wrote:
>>>> "What time is it (yes or no)?" would be called "an unfair question"
>>>> in normal, nontechnical parlance. Do we need a different term?
>>> If we want to formalize our semantic intuitions, then yes we need
>>> this additional term and its related concept. For example we know
>>> that there is something wrong with the following sentence: Colorless
>>> green ideas sleep furiously. Yet it seems that we lack a sufficient
>>> formalism to point out exactly what is incorrect about this
>>> sentence.
>> Chomsky's ENTIRE ENTERPRISE is based on the simple fact that THERE IS
>> NOTHING WRONG FORMALLY with that sentence, whereas "Furiously sleep
>> ideas green colorless" IS formally impossible.
>
> So Chomsky only focused on syntax and ignored semantics...
>
> There is nothing formally wrong with the syntax of that sentence, yet
> there is something wrong with the semantics of that sentence that is
> why it lacks a possible correct interpretation.

If one admits the possibility of metaphor (which one had better do, lest
one abolish language altogether), then it is hard to imagine a
grammatical sentence that defies semantics. The one in question
obviously means "Uninteresting immature ideas undergo violent conflict
during their latency".
--
--- Joe Fineman jo...@verizon.net

||: If you're not paying for the service, you're not the :||
||: customer. :||

[Peter Olcott]

I never divide things up according to standard conventions, it seems to me
that these arbitrarily divisions just make things much more complicated than
necessary.

[Brian M. Scott]

On Sun, 01 Feb 2015 17:52:04 -0500, Joe Fineman
<jo...@verizon.net> wrote in
<news:84h9v58...@verizon.net> in sci.lang:

Oh, well done!

Brian
--
It was the neap tide, when the baga venture out of their
holes to root for sandtatties. The waves whispered
rhythmically over the packed sand: haggisss, haggisss,
haggisss.

[Peter Olcott]

Maybe your own colorless green ideas really do sleep furiously:

[Peter Olcott]

On 2/1/2015 3:45 PM, DKleinecke wrote:

Since an [incorrect question] is defined as any question such that a correct
answer does not exist from the set of all possible answers. Therefore the
answer to the question: Have you stopped beating your wife, lacks a correct
answer from the set of {yes, no} if either one has never beat one's
wife, or
does not have a wife, thus it is an incorrect question.

Since I started talking about "ill-formed questions" nine years ago on
these
and comp.theory forums it does seem that I am the originator of the
concept.

[Helmut Richter]

Am 01.02.2015 um 18:26 schrieb Peter Olcott:
> On 2/1/2015 10:47 AM, DKleinecke wrote:

>> Do you consider "Have you stopped beating your wife?" a semantically
>> incorrect question?
>
> Only if you never beat your wife, (or never had a wife) then yes.

The question is not semantically incorrect. Who has never beaten his
wife or has no wife at all has never had an opportunity to stop beating
her. So the answer is no.

What is indeed semantically incorrect is the inference from not-stopping
an action to continuing the action, forgetting the necessary condition
that the action has ever taken place.

I am writing from Germany although I have never returned from Mongolia.

--
Helmut Richter

[Peter Olcott]

On 2/2/2015 2:15 AM, Helmut Richter wrote:
> Am 01.02.2015 um 18:26 schrieb Peter Olcott:
>> On 2/1/2015 10:47 AM, DKleinecke wrote:
>
>>> Do you consider "Have you stopped beating your wife?" a semantically
>>> incorrect question?
>>
>> Only if you never beat your wife, (or never had a wife) then yes.
>
> The question is not semantically incorrect. Who has never beaten his
> wife or has no wife at all has never had an opportunity to stop
> beating her. So the answer is no.
>

In that case that means that he still beats her, and there is no her.

I have been working on the halting problem in computer science for
over two decades. That it the reason that I created the idea of
incorrect question.
One example of an incorrect question is asking if an English poem halts.
Since an English poem *never* begin executing, asking whether or not
it stops executing *is* an incorrect question.

> What is indeed semantically incorrect is the inference from
> not-stopping an action to continuing the action, forgetting the
> necessary condition that the action has ever taken place.
>

That incorrect inference is logically entailed, thus making the question
incorrect.
How many feet long is the justice up your nose?
How many square circles full of absurdity did you feed your pigs today?

> I am writing from Germany although I have never returned from Mongolia.
>

Then you are a liar.

[Peter T. Daniels]

You are astonishingly confused. In "Have you stopped ... ?" there is a
_presupposition_. It has nothing to do with "correctness." There are also
"inferences" and "implicatures" and several other pbenomena, all of them
covered in linguistic semantics classes 40+ years ago. ISTR John Lyons'
two-volume *Semantics* was the standard catalog of such things.

You are mixing together all sorts of different semantic problems.

[DKleinecke]

The original example was, and I copy and paste,

> I have previously referred to this as a (semantically) ill-formed question.
> An example [incorrect question]: What time is it (yes or no)?

As we all know, that "question" has a easy answer
yes or no
We did that kind of thing in grade school. It's really a command and
not a question.

But the original question was

> An [incorrect question] is defined as any question such that a correct
> answer does not exist from the set of all possible answers.

> Does anyone else have any knowledge or references to the concept of
> a semantically incorrect question, or am I the originator of this concept?

The question is well-posed (not incorrect) even if the set in question is
empty. However the set in question doesn't seem to be interesting - hence
no one has given it a name. You are free to attempt to use it.

The definition you give for "incorrect question"is far too vague for me to
guess what it means. You have not defined "question" "answer" "correct" or
"possible" - all of which you are using in a technical sense.

[Peter Olcott]

A presupposition is equivalent to a premise in logic and if the premise
is false an
unsound conclusion would logically follow, regardless of the reasoning
following
this false premise. Unsound means incorrect.

[Peter T. Daniels]

As David has already intimated, you are not in a position to redefine
technical terms that have been in use for half a century at least, and
you are not entitled to invent new labels for familiar concepts.

"Have you stopped beating your wife?" is perfectly acceptable and utterly
unexceptionable when asked of someone who is known to have beaten his wife.

Someone who has never been to Mongolia can quite truthfully say "I have
never returned from Mongolia," but it is conversationally odd to do so.

There is far, far more to human language than truth-values.

[Peter Olcott]

Meaning that the answer must only come from elements of the set
{yes, no} *and* these elements must specify the current time.

The solution set is restricted to elements of the specified set, and none
of the elements of the specified set are correct answers to the question.

It the empty set is not an element of the specified set then the empty
set is an incorrect answer, likewise with {none} and {neither} et cetera...

What is an integer N such that N > 5 and N < 3
the answer is restricted to elements of the set of integers.

So it looks like the concept of [incorrect question] or its
earlier name [ill-formed question] (meaning semantically ill-formed)
is a brand new concept that I created (according to Google) in 2006.

[Peter Olcott]

I do not think that I am creating new labels for existing concepts or
redefining
technical terms. I have created a brand new concept called [incorrect
question]
previously called an **[ill-formed question]. (Since 2006)
** semantically rather then syntactically.

In some cases examples of incorrect questions involve false presuppositions.
It seems to me that anyone that considers any sentences based on false
presuppositions as semantically correct, are simply not bothering to look
deep enough into what semantically correct must mean.

>
> "Have you stopped beating your wife?" is perfectly acceptable and utterly
> unexceptionable when asked of someone who is known to have beaten his wife.

Yet the semantic meaning that is logically entailed from either element
of the set of possible
answers to this polar question is incorrect for both answers.

Yes) Logically entails that you have beaten you wife and have a wife.

No) Logically entails that you continue to beat your wife and have a wife.

Neither) Does not logically entail that you ever have or continue to
beat your
wife or that you have or ever had a wife.

When the answer is restricted to the elements of the set {yes, no} and
neither
is explicitly prohibited, then the question itself becomes incorrect in
some cases.

>
> Someone who has never been to Mongolia can quite truthfully say "I have
> never returned from Mongolia," but it is conversationally odd to do so.

Yes, you got me there, my mistake.

[Peter T. Daniels]

On Monday, February 2, 2015 at 3:28:40 PM UTC-5, Peter Olcott wrote:
> On 2/2/2015 1:35 PM, Peter T. Daniels wrote:

> > "Have you stopped beating your wife?" is perfectly acceptable and utterly
> > unexceptionable when asked of someone who is known to have beaten his wife.
>
> Yet the semantic meaning that is logically entailed from either element
> of the set of possible
> answers to this polar question is incorrect for both answers.
>
> Yes) Logically entails that you have beaten you wife and have a wife.
>
> No) Logically entails that you continue to beat your wife and have a wife.
>
> Neither) Does not logically entail that you ever have or continue to
> beat your
> wife or that you have or ever had a wife.
>
> When the answer is restricted to the elements of the set {yes, no} and
> neither
> is explicitly prohibited, then the question itself becomes incorrect in
> some cases.

And that's why such questions are not permitted in legal proceedings. However,
ordinary human communication has no such restriction on the set of possible
answers to questions.

An appropriate response to such a question, after the expression of indignation,
might be, "How dare you assume that I ever beat my wife! I can't have stopped doing something I never did."

[Peter Olcott]

On 2/2/2015 4:11 PM, Peter T. Daniels wrote:
> On Monday, February 2, 2015 at 3:28:40 PM UTC-5, Peter Olcott wrote:
>> On 2/2/2015 1:35 PM, Peter T. Daniels wrote:
>>> "Have you stopped beating your wife?" is perfectly acceptable and utterly
>>> unexceptionable when asked of someone who is known to have beaten his wife.
>> Yet the semantic meaning that is logically entailed from either element
>> of the set of possible
>> answers to this polar question is incorrect for both answers.
>>
>> Yes) Logically entails that you have beaten you wife and have a wife.
>>
>> No) Logically entails that you continue to beat your wife and have a wife.
>>
>> Neither) Does not logically entail that you ever have or continue to
>> beat your
>> wife or that you have or ever had a wife.
>>
>> When the answer is restricted to the elements of the set {yes, no} and
>> neither
>> is explicitly prohibited, then the question itself becomes incorrect in
>> some cases.
> And that's why such questions are not permitted in legal proceedings. However,
> ordinary human communication has no such restriction on the set of possible
> answers to questions.
>
> An appropriate response to such a question, after the expression of indignation,
> might be, "How dare you assume that I ever beat my wife! I can't have stopped doing something I never did."

From a human relations point of view you would be entirely correct from a
pure logical point of view you would be incorrect.

It would be an appropriate response, yet would be incorrect for exceeding
the bounds of the constraints placed on the solution set.

I developed this concept for the purposes of applying it to two very famous
problems in logic: (1) The Halting Problem (2) The Incompleteness Theorem.

[Peter T. Daniels]

Problems in logic are to be solved in logic. It's rather arrogant to decree
that human language must be changed in order to make ir fit logic.

Since you have nothing to say about human language, this is not a suitable
forum for you.

[Peter Olcott]

http://plato.stanford.edu/entries/montague-semantics/
There is in my opinion no important theoretical difference between
natural languages and the artificial languages of logicians; indeed I
consider it possible to comprehend the syntax and semantics of both
kinds of languages with a single natural and mathematically precise
theory. (Montague 1970c, 222)

[Helmut Richter]

Am 03.02.2015 um 16:13 schrieb Peter Olcott:

> http://plato.stanford.edu/entries/montague-semantics/
> There is in my opinion no important theoretical difference between
> natural languages and the artificial languages of logicians; indeed I
> consider it possible to comprehend the syntax and semantics of both
> kinds of languages with a single natural and mathematically precise
> theory. (Montague 1970c, 222)

To a large extent, this is true.

But one should not expect that the rules underlying a natural language
and the rules underlying mathematical logic are the same.

Example: When a Bavarian (bar) says something, here with (de) and (en)
literal translation, it may mean the exact opposite of what a logician
would understand:

bar: Bei uns hod no nia koaner koan Hunger ned glittn.
de: Bei uns hat noch nie keiner keinen Hunger nicht gelitten.
en: With us has still never no-one no hunger not suffered.

The meaning is "at our place no-one ever had to go hungry at any time"
and not the opposite, as the four negations cancelling each other would
suggest.

In many languages, the phrases "before he comes" and before he does not
come" mean exactly the same.

Some people would call that illogical, I'd call it logical according to
a logic with rules specific to these languages -- rules that happen to
disagree with common logical rules.

According to the latter, we have the principle "ex falso quodlibet"
which allows to utter sentences like "after each return from Mongolia, I
died withing two weeks" and "after each return from Mongolia, I lived
hundred more years" which are both true.

--
Helmut Richter

[Peter T. Daniels]

I showed with this one simple example that your opinion is unsupportable.

[Peter Olcott]

On 2/3/2015 9:47 AM, Helmut Richter wrote:
> Am 03.02.2015 um 16:13 schrieb Peter Olcott:
>
>> http://plato.stanford.edu/entries/montague-semantics/
>> There is in my opinion no important theoretical difference between
>> natural languages and the artificial languages of logicians; indeed I
>> consider it possible to comprehend the syntax and semantics of both
>> kinds of languages with a single natural and mathematically precise
>> theory. (Montague 1970c, 222)
>
> To a large extent, this is true.
>
> But one should not expect that the rules underlying a natural language
> and the rules underlying mathematical logic are the same.
>
> Example: When a Bavarian (bar) says something, here with (de) and (en)
> literal translation, it may mean the exact opposite of what a logician
> would understand:
>
> bar: Bei uns hod no nia koaner koan Hunger ned glittn.
> de: Bei uns hat noch nie keiner keinen Hunger nicht gelitten.
> en: With us has still never no-one no hunger not suffered.
>
> The meaning is "at our place no-one ever had to go hungry at any time"
> and not the opposite, as the four negations cancelling each other
> would suggest.
>

Thus an idiomatic meaning that is erroneous when measured against the purely
compositional literal meaning.

> In many languages, the phrases "before he comes" and before he does
> not come" mean exactly the same.
>
> Some people would call that illogical, I'd call it logical according
> to a logic with rules specific to these languages -- rules that happen
> to disagree with common logical rules.

Simply idiomatic meanings such that compositionality is superseded by
the idiomatic meaning.
It is like one forms a code such that whenever one says: "I am going to
the store right now"
they really mean {I am going to eat a bologna sandwich in a few minutes}.

>
> According to the latter, we have the principle "ex falso quodlibet"
> which allows to utter sentences like "after each return from Mongolia,
> I died withing two weeks" and "after each return from Mongolia, I
> lived hundred more years" which are both true.
>

Not when one is being 100% literally precise and compositionality applies.

[Peter Olcott]

http://en.wikipedia.org/wiki/Formal_semantics_%28linguistics%29
It is not my opinion, it is the opinion of the originator of formal
semantics
within linguistics, Richard Montague.

[DKleinecke]

No one but you takes Montague seriously as linguistics. Perhaps
there are logicians that take him seriously. There is good reason to
believe that Montague himself did not take the linguistic aspects of
his formal system seriously. He was inventing a new kind of logical
research and using an analogy with natural language to justify it.

I fear the real difficulty here is that formal semantics of any sort
are too different from the pragmatic and idiosyncratic semantics of
human languages to be of any use in linguistics.

[Peter Olcott]

Barbara Partee took him very seriously and was his liaison to the linguistic
community, and number of which took him very seriously.

> Perhaps
> there are logicians that take him seriously. There is good reason to
> believe that Montague himself did not take the linguistic aspects of
> his formal system seriously. He was inventing a new kind of logical
> research and using an analogy with natural language to justify it.
>
> I fear the real difficulty here is that formal semantics of any sort
> are too different from the pragmatic and idiosyncratic semantics of
> human languages to be of any use in linguistics.

In seems to me that many of the linguistic idiosyncrasies are merely
subtle errors. I could be totally off-base here, I consider idiomatic
meanings an error within the goal of providing effective communication.

I make this assessment because idiomatic meanings only make
communication unnecessarily difficult, and more complex than essentially
necessary.

Many of the other idiosyncrasies of natural language may also make
effective
communication more complex than necessary, too.

[Peter T. Daniels]

Montague was not working within linguistics. He was a logician. My college
logic class (in the Mathematics Department, but recommended for linguistics
majors) used his textbook (Kalish and Montague).

(BTW since you just asserted that that _is_ your opinion _and_ it is _not_
your opinion, you just asserted everything and nothing.)

[Peter T. Daniels]

On Tuesday, February 3, 2015 at 1:11:42 PM UTC-5, Peter Olcott wrote:
> On 2/3/2015 11:28 AM, DKleinecke wrote:

> > No one but you takes Montague seriously as linguistics.
>
> Barbara Partee took him very seriously and was his liaison to the linguistic
> community, and number of which took him very seriously.

And then stoppoed taking him seriously.

> > Perhaps
> > there are logicians that take him seriously. There is good reason to
> > believe that Montague himself did not take the linguistic aspects of
> > his formal system seriously. He was inventing a new kind of logical
> > research and using an analogy with natural language to justify it.
> >
> > I fear the real difficulty here is that formal semantics of any sort
> > are too different from the pragmatic and idiosyncratic semantics of
> > human languages to be of any use in linguistics.
>
> In seems to me that many of the linguistic idiosyncrasies are merely
> subtle errors. I could be totally off-base here, I consider idiomatic
> meanings an error within the goal of providing effective communication.

Yes, you are totally off base.

> I make this assessment because idiomatic meanings only make
> communication unnecessarily difficult, and more complex than essentially
> necessary.
>
> Many of the other idiosyncrasies of natural language may also make
> effective
> communication more complex than necessary, too.

Redundancy is highly necessary. Long before there was Montague, there were
Shannon and Weaver.

[Peter Olcott]

No it is a verbatim word-for-word quote from Montague, along with
exactly when and where he said it.

[Peter Olcott]

On 2/3/2015 3:29 PM, Peter T. Daniels wrote:
> On Tuesday, February 3, 2015 at 1:11:42 PM UTC-5, Peter Olcott wrote:
>> On 2/3/2015 11:28 AM, DKleinecke wrote:
>>> No one but you takes Montague seriously as linguistics.
>> Barbara Partee took him very seriously and was his liaison to the linguistic
>> community, and number of which took him very seriously.
> And then stoppoed taking him seriously.

This work is dated 2014 and heavily refers to Montague.
file:///C:/Users/Peter%20Olcott/Downloads/LACL14.pdf

>>> Perhaps
>>> there are logicians that take him seriously. There is good reason to
>>> believe that Montague himself did not take the linguistic aspects of
>>> his formal system seriously. He was inventing a new kind of logical
>>> research and using an analogy with natural language to justify it.
>>>
>>> I fear the real difficulty here is that formal semantics of any sort
>>> are too different from the pragmatic and idiosyncratic semantics of
>>> human languages to be of any use in linguistics.
>> In seems to me that many of the linguistic idiosyncrasies are merely
>> subtle errors. I could be totally off-base here, I consider idiomatic
>> meanings an error within the goal of providing effective communication.
> Yes, you are totally off base.

<sarcasm>
So idioms are helpful making effective communication simpler and easier...
</sarcasm>

>> I make this assessment because idiomatic meanings only make
>> communication unnecessarily difficult, and more complex than essentially
>> necessary.
>>
>> Many of the other idiosyncrasies of natural language may also make
>> effective
>> communication more complex than necessary, too.
> Redundancy is highly necessary. Long before there was Montague, there were
> Shannon and Weaver.

Yes redundancy is helpful, because human attention is limited it is best to
say the same thing several different ways to make sure that attention did
not drift off-track and miss some crucial aspect.

[Peter T. Daniels]

On Tuesday, February 3, 2015 at 5:17:35 PM UTC-5, Peter Olcott wrote:
> On 2/3/2015 3:29 PM, Peter T. Daniels wrote:
> > On Tuesday, February 3, 2015 at 1:11:42 PM UTC-5, Peter Olcott wrote:
> >> On 2/3/2015 11:28 AM, DKleinecke wrote:

> >>> No one but you takes Montague seriously as linguistics.
> >> Barbara Partee took him very seriously and was his liaison to the linguistic
> >> community, and number of which took him very seriously.
> > And then stoppoed taking him seriously.
>
> This work is dated 2014 and heavily refers to Montague.
> file:///C:/Users/Peter%20Olcott/Downloads/LACL14.pdf

It's not clear what an address on your hard drive is supposed to tell anyone.

[Peter Olcott]

It usually provides a web link. Here is the 2014 paper with extensive
reference to Montague.
Monotonicity Reasoning in Formal Semantics Based on Modern Type Theories

[Yusuf B Gursey]

https://www.researchgate.net/publication/265651913_Monotonicity_Reasoning_in_Formal_Semantics_Based_on_Modern_Type_Theories

[Peter T. Daniels]

I don't know why Yusuf bothered to provide the link you failed to, and I'm certainly not going to sign up for some intrusive social network to look at
an abstract, but the title, the affiliations of the authors, and the "outline"
give no hint that it has anything whatsoever to do witb linguistics or with
language.

[Peter Olcott]

This is a link to the full text and it was on Yusuf's link
http://www.researchgate.net/publication/265651913_Monotonicity_Reasoning_in_Formal_Semantics_Based_on_Modern_Type_Theories


It is not merely a social network it is a network of research scientists
to join you must have at least one significant publication. I used my
issued patent.

[Peter T. Daniels]

It still doesn't have anything to do with human language, and I'm still not
going to sign up for it.

[Peter Olcott]

It is research scientists of every field.
So it seems that you are one of the kind of people that presumes that
if it is based on math then it is not relevant to language.

That is OK scientific studies have shown a sharp dichotomy between
analytical versus communicative abilities. To the same extent that one
is relatively good at one, they are relatively poor at the other in a
mutually sort of exclusive way.

[DKleinecke]

There is another way to read the article without signing up for
anything and I have read the article. I don't really care enough
about formal semantics to have an opinion about the value of the
article but it is easy to see that it politely acknowledges Montague
and proceeds to develop a computational scheme quite different than his.

I wouldn't describe it as based on Montague or about Montague
grammar or even about natural language. It describes a computer-
based approach to whatever it is that formal semanticists think
they are doing.

As a card carrying mathematician I would take the position that,
while anything is possible, so far there have been no useful
applications of mathematics or mathematically-based techniques to
linguistics. Chomsky started with veneer of what looked like
mathematics but he had already given up mathematical type reasoning
in "Aspects".

[Peter Olcott]

It explicitly states that it is about formalized natural language.

> It describes a computer-
> based approach to whatever it is that formal semanticists think
> they are doing.
>
> As a card carrying mathematician I would take the position that,
> while anything is possible, so far there have been no useful
> applications of mathematics or mathematically-based techniques to
> linguistics. Chomsky started with veneer of what looked like
> mathematics but he had already given up mathematical type reasoning
> in "Aspects".

Chomsky was a very great contributor to Linguistics, yet even
Einstein was totally incorrect about quantum mechanics in physics.

It looks like the sharp (mutually exclusive) dichotomy between
communicative ability and analytical ability leaves almost no one
with sufficient skills in both to effectively evaluate the mathematical
formalization of natural language.

It seems that (for the most part) the best that we can do is get someone
that is relatively good at one of these and somewhat of a crank (mostly
incompetent) in the other.

Thus if Montague was correct:

There is in my opinion no important theoretical difference between
natural languages and the artificial languages of logicians; indeed I
consider it possible to comprehend the syntax and semantics of both
kinds of languages with a single natural and mathematically precise
theory. (Montague 1970c, 222)

There is close to no one that can double check him. I myself can directly
see that he was entirely correct at this, yet lack the communicative skill
to directly prove this to others.

When I must effectively communicate subtle nuances of very complex
technical ideas it takes me very many years of hit-and-miss until I get
any hits at all.

The exception to my communication limitations is when I can directly
encode these subtle nuances of meaning in a programming language.
In these cases my communication is 100% effective for everyone knowing
this programming language.

[DKleinecke]

On Wednesday, February 4, 2015 at 11:16:39 AM UTC-8, Peter Olcott wrote:

Thinking about my last comment I think it might be useful to state
what I think Chomsky proposed (very clumsily) in "Syntactic Structures".

He assumed the existence of a set called the set of all sentences in
the language (at this time he had not addressed the question of
different languages and effectively spoke as though there was one
language - English). He visualized a sentence as a string of words
and a word as a string of letters. He then proposed a way to build a
method of generating this set - in the sense that, for any sentence
in the set, it will appear in a finite number of steps.

The method had two parts - phrase structure and transformations. The
phrase structure (context-free) generated a set called kernel sentences
and the transformations generated more sentences by operating on the
kernel sentences.

I believe that when SS was written he imagined that creating a grammar
(phrase structure plus transformations) was relatively easy and that he
was aware that multiple grammars could generate the same set. Hence, he
felt - at that time - that selecting a "best" grammar was an important
task.

This relatively simple model of language failed before "Aspects of
Syntax" was written. I believe that no technical explanation has ever
been formulated as to why it failed - many "political" explanations
have been written.

Since that fiasco nothing quite like it has ever been attempted.
applied to

[DKleinecke]

On Wednesday, February 4, 2015 at 11:16:39 AM UTC-8, Peter Olcott wrote:

> This work is dated 2014 and heavily refers to Montague.

> Monotonicity Reasoning in Formal Semantics Based on Modern Type Theories

> > I wouldn't describe it as based on Montague or about Montague
> > grammar or even about natural language.
>
> It explicitly states that it is about formalized natural language.

That it calls what it is doing "formalized natural language" does
not mean it pertains to natural language. As nearly as I can tell
it is a proposal for a way to make valid inferences in a computer
data base. If I am right it is a contribution of some interest to
society as a whole - but it has nothing to do with linguistics
which is the study of how natural languages work.

> It looks like the sharp (mutually exclusive) dichotomy between
> communicative ability and analytical ability leaves almost no one
> with sufficient skills in both to effectively evaluate the mathematical
> formalization of natural language.

I would evaluate it as a waste of time.

> Thus if Montague was correct:
> There is in my opinion no important theoretical difference between
> natural languages and the artificial languages of logicians; indeed I
> consider it possible to comprehend the syntax and semantics of both
> kinds of languages with a single natural and mathematically precise
> theory. (Montague 1970c, 222)

But he is wrong - seriously wrong and I think he knew he was wrong when
he wrote that. I knew Dick Montague when we were both at Berkeley and
he then showed no knowledge of or interest in linguistics. Since this
the middle of the wars about generative semantics I think he probably
stayed uninformed about linguistics and, in the paragraph quoted, was
speaking from a naive position of ignorance about natural language.

He was not the only logician who was this naive. I call your attention
to Frederick B. Thompson (of Cal Tech, recently deceased) an interesting
contemporary of Chomsky whom Chomsky's biographers appear to have
ignored. His DEACON project was based on a complete misunderstanding
of the complexity of natural language. His obituary seems to say that
he spend the rest of his life trying to make his ideas work. So far
as I can tell nobody has ever paid any attention to him.

Thompson and Montague were both students of Tarski. Thompson was the
star pupil. Montague carried a spear.

[Peter Olcott]

Yet since I can directly see first-hand how his system could be extended
to prove his point beyond all possible doubt I know that you and everyone
that agrees with you must be wrong. Maybe eventually I will, after countless
attempts, find a way to communicate this understanding so that others
could see this too.

[Peter T. Daniels]

On Wednesday, February 4, 2015 at 2:16:39 PM UTC-5, Peter Olcott wrote:

> Thus if Montague was correct:
> There is in my opinion no important theoretical difference between
> natural languages and the artificial languages of logicians; indeed I
> consider it possible to comprehend the syntax and semantics of both
> kinds of languages with a single natural and mathematically precise
> theory. (Montague 1970c, 222)

I was going to simply say "But he was not correct," but David beat me to it,
and with inside knowledge, no less.

[Peter Olcott]

[DKleinecke]

On Wednesday, February 4, 2015 at 2:17:51 PM UTC-8, Peter Olcott wrote:

> >> Thus if Montague was correct:

> > But he is wrong - seriously wrong and I think he knew he was wrong when
> > he wrote that.

> Yet since I can directly see first-hand how his system could be extended
> to prove his point beyond all possible doubt I know that you and everyone
> that agrees with you must be wrong. Maybe eventually I will, after countless
> attempts, find a way to communicate this understanding so that others
> could see this too.

You do realize that this sounds like you are insane ?

[Peter Olcott]

I only thought of it as something that I know intuitively that I haven't
been
able to express in words yet.

I also know that ignorance can only be perceived by ignorant minds as
disagreement because the direct perception of the ignorance itself requires
contrasting the missing knowledge against the current lack of this missing
knowledge.

I coined the term [ignorance squared] decades ago to account for this
concept.
It literally means that one is ignorant of their own ignorance, they
don't know
that they don't know, thus they presume that they do know.

Other people have used this term incorrectly since I derived it:
http://forum.quoteland.com/eve/forums/a/tpc/f/99191541/m/6991970196

Not knowing what one does not know is merely ignorance, not knowing
that one does not know is ignorance of ignorance, thus ignorance squared.

[Helmut Richter]

Am 03.02.2015 um 17:16 schrieb Peter Olcott:

> On 2/3/2015 9:47 AM, Helmut Richter wrote:
>> Am 03.02.2015 um 16:13 schrieb Peter Olcott:
>>
>>> http://plato.stanford.edu/entries/montague-semantics/
>>> There is in my opinion no important theoretical difference between
>>> natural languages and the artificial languages of logicians; indeed I
>>> consider it possible to comprehend the syntax and semantics of both
>>> kinds of languages with a single natural and mathematically precise
>>> theory. (Montague 1970c, 222)
>>
>> To a large extent, this is true.
>>
>> But one should not expect that the rules underlying a natural language
>> and the rules underlying mathematical logic are the same.
>>
>> Example: When a Bavarian (bar) says something, here with (de) and (en)
>> literal translation, it may mean the exact opposite of what a logician
>> would understand:
>>
>> bar: Bei uns hod no nia koaner koan Hunger ned glittn.
>> de: Bei uns hat noch nie keiner keinen Hunger nicht gelitten.
>> en: With us has still never no-one no hunger not suffered.
>>
>> The meaning is "at our place no-one ever had to go hungry at any time"
>> and not the opposite, as the four negations cancelling each other
>> would suggest.
>>
>
> Thus an idiomatic meaning that is erroneous when measured against the
> purely
> compositional literal meaning.

No, it is a purely compositional literal meaning but the rules of
composition are different.

In such languages (e.g. today's Bavarian language or German of Luther's
time), a sentence is negated by *one or more* negations. In other
languages (e.g. today's standard German), more than one negation on the
same level is ungrammatical, and if used all the same, can be ambiguous.
In still another language (e.g. classical propositional calculus), any
two negations on the same level cancel each other, that is, ¬¬A <=> A.
In still another language (e.g. intuitionistic propositional calculus),
¬¬A ("it cannot be proved that it cannot be proved that A is true") is
weaker than A ("it can be proved that A is true").

So there is not *one* set of rules to assign semantics but there is one
per language. This is most blatant for negations but holds for many
other features of language. This does *not* contradict compositionality,
it does only forbid that compositional rules be regarded as universal.
Either one has to accept that the rules of one language consist not only
of axioms but rather of inference rules as well, or -- what I find the
better solution -- one has to carefully distinguish logic (logical
operators, quantifiers, equality, ...) in the object language from logic
in the metalanguage.

>> According to the latter, we have the principle "ex falso quodlibet"
>> which allows to utter sentences like "after each return from Mongolia,
>> I died withing two weeks" and "after each return from Mongolia, I
>> lived hundred more years" which are both true.
>>
>
> Not when one is being 100% literally precise and compositionality applies.

Maybe I am spoiled by having done too much maths and logic in my life.
For me, each sentence of the form "for all R (R is one of my returns
from Mongolia => Pred(R))" is true irrespective of what predicate "Pred"
is if I never returned from Mongolia. For the simple reason that every
formula A=>B is true if A is false (assuming classical logic). Your
understanding of "for all x" in language differs from mine: it means
"there is at least one x and for each of them" which is different. This
is why I explicitly started my asserting with the words "according to
the latter", referring to the rules of logic, not the rules of some
language. Needless to say that the rules of language are defined vaguely
enough that there can be misunderstandings.

I apologise for triggering a discussion about the compositionality
principle as a whole, which was not my intention. It seems to be a
highly emotional topic in linguistics.

--
Helmut Richter

[Peter Olcott]

http://en.wikipedia.org/wiki/Principle_of_compositionality
No most often if not always idiomatic meanings are simply macro
substitutions,
substituting an entirely different meaning for a string of words.
http://www.smart-words.org/quotes-sayings/idioms-meaning.html

>
> In such languages (e.g. today's Bavarian language or German of
> Luther's time), a sentence is negated by *one or more* negations. In
> other languages (e.g. today's standard German), more than one negation
> on the same level is ungrammatical, and if used all the same, can be
> ambiguous. In still another language (e.g. classical propositional
> calculus), any two negations on the same level cancel each other, that
> is, ¬¬A <=> A. In still another language (e.g. intuitionistic
> propositional calculus), ¬¬A ("it cannot be proved that it cannot be
> proved that A is true") is weaker than A ("it can be proved that A is
> true").
>
> So there is not *one* set of rules to assign semantics but there is
> one per language.

Ultimately there is a single underlying language of pure semantics that
all other
languages map to and from.

It seems to me that compositionality may be the single most important aspect
of all of linguistics.

[Peter T. Daniels]

On Thursday, February 5, 2015 at 7:58:21 PM UTC-5, Peter Olcott wrote:

> Ultimately there is a single underlying language of pure semantics that
> all other languages map to and from.

You are Chomsky, and I claim my five pounds.

[Peter Olcott]

Is that a complement ?

[Peter T. Daniels]

A complement of or to what?

[Peter Olcott]

A compliment as opposed to and contrast with an insult.
(I did not realize that the two words were spelled differently).

[Helmut Richter]

Am 06.02.2015 um 01:58 schrieb Peter Olcott:

> No most often if not always idiomatic meanings are simply macro
> substitutions,
> substituting an entirely different meaning for a string of words.

I am afraid I am unable to make myself clear. One more attempt:

1. I did not intend to express an opinion about compositionality. Maybe
I'll do that briefly in this thread after this contribution. I assume
that our opinions will diverge less than what you expect.

2. I share your view about how idiomatic meanings can be fitted into the
framework of compositionality. Hence, the existence of idioms does not
contradict compositionality but rather only contradicts the idea that
the compositional mapping of a string to its meaning is unambiguous.
Quite the contrary: it is nearly always ambiguous, and the more so when
idioms are involved. I assume that you would agree in this point as well.

3. Where we disagree: If in some language the usage of more than one
negation particle for a logically single negation is allowed or in some
cases mandatory, it is not so that each occurrence of a multiple
negation constitutes an idiom. Rather, the mapping of a string to its
meaning follows other rules than in other languages or in mathematical
logic, with no idioms underway.

--
Helmut Richter

[Helmut Richter]

(Not considering the spelling error.) This is why I wrote:

I apologise for triggering a discussion about the compositionality
principle as a whole, which was not my intention. It seems to be a
highly emotional topic in linguistics.

With "emotional" I meant that there are only allies and enemies, with
Chomsky being either the guru or the villain. As his work has
mathematical background, mathematics is either the universal tool or the
abomination of desolation. However, I think we can have a much less
emotional access to these questions.

For millennia, people have heard and understood sentences they had never
heard before, and their intuitive and correct understanding was based on
the fact that they were able to analyze the sentence structure and
sythesize their understanding of the meaning by considering the meaning
of the constituents and their knowledge how the sentence construction
combines the meanings of the constituents. They did not need Frege's or
Montague's work to do so. So, Frege and Montague did nothing but express
the ideas that grammarians always have had, but in another language, to
wit the language of mathematics. This may be useful or not, but it is
not evil.

At least for centuries, grammarians have analyzed sentences, giving the
constituents names like "subject", "object", "adverbial", and others.
They did not need Chomsky's work to do so. So, Chomsky did nothing but
express the ideas that grammarians always have had, but in another
language, to wit the language of mathematics. This may be useful or not,
but it is not evil.

The language of mathematics is useful when, by its greater formality,
irregularities and ambiguities in the object that is modelled, to wit
human languages, are more easily detected. I take that as an advantage:
one is forced to refine the model so that it comprehends more and more
of these irregularities, thus leading to more insight. On the other
hand, this can also serve as a weapon against the formal models, as they
do not normally match the actual complexity of language. But neither
does grammar in flowery plain language, only is there hardly a chance to
pinpoint the fallacies as precisely as is possible in a mathematical model.

Where I remain sceptical, is about the abuse of formalism to corroborate
models of how the human brain works when processing language. I feel
this is a confusion of the two meanings of "generative". I think I have
already commented on this topic.

--
Helmut Richter

[Peter Olcott]

What I mean when I say that idioms are not compositional is that most
often the meaning of the expression has nothing at all to do with any of
the meanings of any of the words in the idiom.

See these concrete examples:
http://www.smart-words.org/quotes-sayings/idioms-meaning.html

You are [barking up the wrong tree] if you think otherwise. Thus you
must go [back to the drawing board].

[António Marques]

Helmut Richter <hh...@web.de> wrote:
> 2. I share your view about how idiomatic meanings can be fitted into the
> framework of compositionality. Hence, the existence of idioms does not
> contradict compositionality but rather only contradicts the idea that the
> compositional mapping of a string to its meaning is unambiguous. Quite
> the contrary: it is nearly always ambiguous, and the more so when idioms
> are involved. I assume that you would agree in this point as well.

Last time I checked, Peter was only interested in a a subset of natural
language (that he believes exists) that is 'well formed' enough to be
described in Montagovian terms. His goal is to build a machine that can
actually understand such texts, as in acquiring their actual meanings
(which is a separate problem with separate assumptions), and that way learn
by reading (which again is a separate thing).
--
Sent from one of my newsreaders

[António Marques]

Peter Olcott <OCR4Screen> wrote:
> On 2/6/2015 7:37 AM, Peter T. Daniels wrote:
>> On Thursday, February 5, 2015 at 11:38:10 PM UTC-5, Peter Olcott wrote:
>>> On 2/5/2015 10:04 PM, Peter T. Daniels wrote:
>>>> On Thursday, February 5, 2015 at 7:58:21 PM UTC-5, Peter Olcott wrote:
>>>>> Ultimately there is a single underlying language of pure semantics that
>>>>> all other languages map to and from.
>>>> You are Chomsky, and I claim my five pounds.
>>> Is that a complement ?
>> A complement of or to what?
>
> A compliment as opposed to and contrast with an insult.

SEE how your mind works? It is neither. Your sentence quoted above sums up
what at some point was Chomsky's position. (Of course, Chomsky has never
stood still.)

[António Marques]

You should rather proceed to extend the system and present some results.
Proof-of-concept, or something. Of course, if your point is that you'd need
the help of others to do it, then that's not an option, but it's still the
only way to go on about it, since trying to simply convince people has
proved fruitless - you claim linguists don't understand math and
mathematicians don't understand language, but there are folks here with an
extensive understanding of, and training in, both. And they're not saying
anything different than the ones only versed in one.

[Helmut Richter]

Am 06.02.2015 um 21:01 schrieb António Marques:

> Helmut Richter <hh...@web.de> wrote:
>> 2. I share your view about how idiomatic meanings can be fitted into the
>> framework of compositionality. Hence, the existence of idioms does not
>> contradict compositionality but rather only contradicts the idea that the
>> compositional mapping of a string to its meaning is unambiguous. Quite
>> the contrary: it is nearly always ambiguous, and the more so when idioms
>> are involved. I assume that you would agree in this point as well.

> Last time I checked, Peter was only interested in a a subset of natural
> language (that he believes exists) that is 'well formed' enough to be
> described in Montagovian terms.

In the terms of my paragraph cited above, 'well formed enough' means
*unamguously* described in Montagovian terms. An idiom which is known
both to the speaker and the interlocutor introduces an ambiguity, e.g.

1. The dog barks up the wrong tree (that is, another than the one where
the cat is sitting), by the meaning of the vocabulary items "bark",
"tree", ...

2. He who calls me a Chomskyan barks up the wrong tree (that is, fails
to understand to what limited extent I consider Chomsky's ideas cogent),
by the meaning of the vocabulary item "bark up the wrong tree".

The resolution, as always, is by commonsense knowledge.

> His goal is to build a machine that can
> actually understand such texts, as in acquiring their actual meanings
> (which is a separate problem with separate assumptions), and that way learn
> by reading (which again is a separate thing).

So, one prerequisite of that machine is enough commonsense knowledge,
which is required not only for idioms but also for idiom-free sentences
that were devised as unambiguous by the speaker but are not -- which is
more or less every real-world sentence of at least medium complexity.

When I say "I have at home a translation of Tolstoy's 'War and peace'",
commonsense knowledge tells that I have not a translation process but
its outcome, not the translation outcome but a write-up thereof, not the
original write-up but a copy of a printed edition thereof. In another
language, the three words may be different.

--
Helmut Richter

[Peter T. Daniels]

That depends entirely on your opinion of Chomsky. If you accept that all
languages can be accounted for as if they were English, then I suppose you
would take it as a compliment.

[Peter T. Daniels]

On Friday, February 6, 2015 at 1:20:03 PM UTC-5, Helmut Richter wrote:

> At least for centuries, grammarians have analyzed sentences, giving the
> constituents names like "subject", "object", "adverbial", and others.
> They did not need Chomsky's work to do so. So, Chomsky did nothing but
> express the ideas that grammarians always have had, but in another
> language, to wit the language of mathematics. This may be useful or not,
> but it is not evil.

But Cnomsky discarded all such labels -- instead of subjects and objects,
he had only [NP [VP NP]].

[Peter Olcott]

I like the way that formalisms are able to provide complete clarity and
totally
eliminate all ambiguity. Also only by formalizing natural language will
we ever
be able to create machines with the capabilities of the human mind.

[Helmut Richter]

Am 06.02.2015 um 21:58 schrieb Peter Olcott:

> On 2/6/2015 12:19 PM, Helmut Richter wrote:

>> [...]

>> The language of mathematics is useful when, by its greater formality,
>> irregularities and ambiguities in the object that is modelled, to wit
>> human languages, are more easily detected. I take that as an
>> advantage: one is forced to refine the model so that it comprehends
>> more and more of these irregularities, thus leading to more insight.
>> On the other hand, this can also serve as a weapon against the formal
>> models, as they do not normally match the actual complexity of
>> language. But neither does grammar in flowery plain language, only is
>> there hardly a chance to pinpoint the fallacies as precisely as is
>> possible in a mathematical model.

>> [...]

> I like the way that formalisms are able to provide complete clarity and
> totally
> eliminate all ambiguity.

If used correctly, they eliminate all ambiguity in the meta-language
(the language in which we talk about language) but not in the object
language (the language about which we talk). This is what I meant with
the paragraph cited above.

Ambiguities in the object language simply exist; they cannot be avoided
except if we switch to an artificial Newspeak -- and that will not work.
As human language always has a fair amount of ambiguity, it seems to be
a feature rather than a bug.

--
Helmut Richter

[Peter Olcott]

On 2/6/2015 2:01 PM, António Marques wrote:
> Helmut Richter <hh...@web.de> wrote:
>> 2. I share your view about how idiomatic meanings can be fitted into the
>> framework of compositionality. Hence, the existence of idioms does not
>> contradict compositionality but rather only contradicts the idea that the
>> compositional mapping of a string to its meaning is unambiguous. Quite
>> the contrary: it is nearly always ambiguous, and the more so when idioms
>> are involved. I assume that you would agree in this point as well.
> Last time I checked, Peter was only interested in a a subset of natural
> language (that he believes exists) that is 'well formed' enough to be
> described in Montagovian terms.

Not at all.

> His goal is to build a machine that can
> actually understand such texts, as in acquiring their actual meanings
> (which is a separate problem with separate assumptions), and that way learn
> by reading (which again is a separate thing).

It is the deep meaning specified by the compositionality of how the
meanings of
words connect together in sentences and discourse context that only formal
semantics could possibly sufficiently capture such that machines could
attain an
understanding of any subject matter equal or better than that humans.

The AI guys don't seem to generally know this yet.

[Peter Olcott]

On 2/6/2015 2:01 PM, António Marques wrote:

That is good if you stand still too long you can get blood clots in your
legs
that can kill you.

[Peter Olcott]

On 2/6/2015 2:01 PM, António Marques wrote:

> Peter Olcott <OCR4Screen> wrote:
>> On 2/4/2015 10:19 PM, Peter T. Daniels wrote:
>>> On Wednesday, February 4, 2015 at 2:16:39 PM UTC-5, Peter Olcott wrote:
>>>
>>>> Thus if Montague was correct:
>>>> There is in my opinion no important theoretical difference between
>>>> natural languages and the artificial languages of logicians; indeed I
>>>> consider it possible to comprehend the syntax and semantics of both
>>>> kinds of languages with a single natural and mathematically precise
>>>> theory. (Montague 1970c, 222)
>>> I was going to simply say "But he was not correct," but David beat me to it,
>>> and with inside knowledge, no less.
>> Yet since I can directly see first-hand how his system could be extended
>> to prove his point beyond all possible doubt I know that you and everyone
>> that agrees with you must be wrong. Maybe eventually I will, after countless
>> attempts, find a way to communicate this understanding so that others
>> could see this too.
> You should rather proceed to extend the system and present some results.

I would really love to do that full time as a profession, yet I have no
time for that.

> Proof-of-concept, or something. Of course, if your point is that you'd need
> the help of others to do it, then that's not an option, but it's still the
> only way to go on about it, since trying to simply convince people has
> proved fruitless

Yes I am finally making progress on my work on the computer science
halting problem,
only because I have begun to finally started to express these ideas in a
way that they
could be directly evaluated: predicate logic.

> - you claim linguists don't understand math and
> mathematicians don't understand language, but there are folks here with an
> extensive understanding of, and training in, both. And they're not saying
> anything different than the ones only versed in one.

They are saying the formalizing natural language is impossible?

[Peter Olcott]

On 2/6/2015 2:27 PM, Helmut Richter wrote:
> Am 06.02.2015 um 21:01 schrieb António Marques:
>
>> Helmut Richter <hh...@web.de> wrote:
>>> 2. I share your view about how idiomatic meanings can be fitted into
>>> the
>>> framework of compositionality. Hence, the existence of idioms does not
>>> contradict compositionality but rather only contradicts the idea
>>> that the
>>> compositional mapping of a string to its meaning is unambiguous. Quite
>>> the contrary: it is nearly always ambiguous, and the more so when
>>> idioms
>>> are involved. I assume that you would agree in this point as well.
>
>> Last time I checked, Peter was only interested in a a subset of natural
>> language (that he believes exists) that is 'well formed' enough to be
>> described in Montagovian terms.
>
> In the terms of my paragraph cited above, 'well formed enough' means
> *unamguously* described in Montagovian terms. An idiom which is known
> both to the speaker and the interlocutor introduces an ambiguity, e.g.
>
> 1. The dog barks up the wrong tree

The English idiom about [barking up the wrong tree]
has nothing to do with dogs, barking, or trees.

[Peter Olcott]

No the universal semantic language is enormously simpler than any
natural language.
I have provided tid bits of examples of this from time to time.

[Peter Olcott]

Whether you call them by name or their relative position on a di-graph
makes no difference as long as you have some point of reference.

[Helmut Richter]

Am 07.02.2015 um 04:00 schrieb Peter Olcott:

>> 1. The dog barks up the wrong tree
>
> The English idiom about [barking up the wrong tree]
> has nothing to do with dogs, barking, or trees.

You would have noticed that I am familiar with it, had you read my next
line before answering.

--
Helmut Richter

[Peter Olcott]

I could not tell that you were familiar with it based on what you said.
Based on what you said I thought that it may be possible that you
were totally unfamiliar with the completely non-compositional meaning
of that idiom and were referring to a compositional meaning.

Normally one never inserts the concept or word of dog into any use
of this idiom. Normally one is referring to a human being in the use of
this idiom, and never referring to any dog.

[DKleinecke]

Practically speaking "formalizing natural language is impossible" at
this time because we do not understand natural language well enough
to formalize it. There is no likelihood that neurologists will reach
an adequate understanding of how the brain works in the foreseeable
future. And, if we did reach nough understanding, it seems most probable
that formalization would have to be different for different languages.
There being no such thing as "natural language" only a collection of
natural languages - all different. I suspect that for a language
understanding in detail enough for formulization each idiolect would
be different.

Chomsky is aware of what I just expounded. His solution is to postulate
a super-language from which actual languages are formed by selecting
the values of specific parameters. Many of us think that he has drifted
too far from reality to be significant any longer.

[Peter Olcott]

On 2/7/2015 11:12 AM, DKleinecke wrote:
> On Friday, February 6, 2015 at 6:52:54 PM UTC-8, Peter Olcott wrote:
>> On 2/6/2015 2:01 PM, António Marques wrote:
>>
>>> - you claim linguists don't understand math and
>>> mathematicians don't understand language, but there are folks here with an
>>> extensive understanding of, and training in, both. And they're not saying
>>> anything different than the ones only versed in one.
>> They are saying the formalizing natural language is impossible?
> Practically speaking "formalizing natural language is impossible" at
> this time because we do not understand natural language well enough
> to formalize it.

That does not really meet the common understanding of the meaning of the
term [impossible].

> There is no likelihood that neurologists will reach
> an adequate understanding of how the brain works in the foreseeable
> future.

This is 100% totally moot.

> And, if we did reach nough understanding, it seems most probable
> that formalization would have to be different for different languages.

Not at all. A single acyclic digraph that directly connects every pure
concept to every other pure concept using unique integer handles for each
unique concept can do the job. All of the syntax of this whole system is
merely different types of connections between these integer handles.

Something like this system provides the types of connections (edges) and
the types of concepts (nodes) within the digraph:

Kurt Gödel (1944)
By the theory of simple types I mean the doctrine which says that the
objects of thought (or, in another interpretation, the symbolic
expressions) are divided into types, namely: individuals, properties of
individuals, relations between individuals, properties of such
relations, etc. (with a similar hierarchy for extensions), and that
sentences of the form: " a has the property φ ", " b bears the relation
R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
fitting together.

> There being no such thing as "natural language" only a collection of
> natural languages - all different. I suspect that for a language
> understanding in detail enough for formulization each idiolect would
> be different.
>
> Chomsky is aware of what I just expounded. His solution is to postulate
> a super-language from which actual languages are formed by selecting
> the values of specific parameters. Many of us think that he has drifted
> too far from reality to be significant any longer.

Well if he thinks it is fundamentally based on English he is whacked out,
otherwise its seems that we may be in at least somewhat of agreement.

[Helmut Richter]

What is so evil: that he used other labels, or that he abbreviated them
in order that the formula-like grammar rules fit on one line?

These terms are used in different ways anyway. In grammars written in
English, objects are mostly considered part of the predicate; in
grammars written in German, they are a separate clause constituent
besides subject and predicate. And there are many more details that make
a definition of terms necessary instead of relying on the terminology we
are acquainted with from old grammars of Greek and Latin -- whose terms
do not always fit for features of other languages.

--
Helmut Richter

[Peter T. Daniels]

On Sunday, February 8, 2015 at 3:43:38 AM UTC-5, Helmut Richter wrote:
> Am 06.02.2015 um 21:57 schrieb Peter T. Daniels:
> > On Friday, February 6, 2015 at 1:20:03 PM UTC-5, Helmut Richter wrote:
> >
> >> At least for centuries, grammarians have analyzed sentences, giving the
> >> constituents names like "subject", "object", "adverbial", and others.
> >> They did not need Chomsky's work to do so. So, Chomsky did nothing but
> >> express the ideas that grammarians always have had, but in another
> >> language, to wit the language of mathematics. This may be useful or not,
> >> but it is not evil.
> >
> > But Cnomsky discarded all such labels -- instead of subjects and objects,
> > he had only [NP [VP NP]].
>
> What is so evil: that he used other labels, or that he abbreviated them
> in order that the formula-like grammar rules fit on one line?

That he thought human language could be analyzed purely formally -- that is,
in the syntax -- and not at all functionally -- that is, in the semantics
(or "grammatical relations"). The concepts of subject, object, predicate,
modifier simply do not exist in Chomskyism.

> These terms are used in different ways anyway. In grammars written in
> English, objects are mostly considered part of the predicate; in
> grammars written in German, they are a separate clause constituent
> besides subject and predicate. And there are many more details that make
> a definition of terms necessary instead of relying on the terminology we
> are acquainted with from old grammars of Greek and Latin -- whose terms
> do not always fit for features of other languages.

In German you can't ignore subjects and objects formally because you have
noun cases.

[Mścisław Wojna-Bojewski]

On Sunday, February 1, 2015 at 8:26:56 PM UTC+2, Peter T. Daniels wrote:
> On Sunday, February 1, 2015 at 12:26:05 PM UTC-5, Peter Olcott wrote:
> > On 2/1/2015 10:47 AM, DKleinecke wrote:
>
> > > "What time is it (yes or no)?" would be called "an unfair question"
> > > in normal, nontechnical parlance. Do we need a different term?
> >
> > If we want to formalize our semantic intuitions, then yes we need this
> > additional
> > term and its related concept. For example we know that there is something
> > wrong with the following sentence: Colorless green ideas sleep furiously.
> > Yet it seems that we lack a sufficient formalism to point out exactly
> > what is
> > incorrect about this sentence.
>
> Chomsky's ENTIRE ENTERPRISE is based on the simple fact that THERE IS NOTHING
> WRONG FORMALLY with that sentence, whereas "Furiously sleep ideas green colorless"
> IS formally impossible.
>
> If you don't understand that, then you are _certainly_ in no position to be
> investigating formal semantics, whether Montague's version or anyone else's.


"Colorless green ideas sleep furiously" is perfectly easy to interpret. "Old-hat and uninspiring ecologist ideas are having no real effect on political decision-making, while they are fiercely debated by the chattering classes."

[DKleinecke]

On Saturday, February 7, 2015 at 12:10:07 PM UTC-8, Peter Olcott wrote:
>
> Kurt Gödel (1944)
> By the theory of simple types I mean the doctrine which says that the
> objects of thought (or, in another interpretation, the symbolic
> expressions) are divided into types, namely: individuals, properties of
> individuals, relations between individuals, properties of such
> relations, etc. (with a similar hierarchy for extensions), and that
> sentences of the form: " a has the property φ ", " b bears the relation
> R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
> fitting together.

KG says clearly that he is expounding a doctrine about the objects of
thought. He has not in any sense proved it - he has postulated it.

I observe a recent book about "objects of thought" (by Tim Crane). Since
you are deeply into these objects I assume you keep up with the
literature and do not rely solely on (nearly) century-old studies.
Do you keep up on the literature?

The Wikipedia article on type theory does not mention KG but does
ennumerate a dozen different type theories. It looks to me like KG
is simply assuming Russell's theory which appears, to my eye, to be
today treated as a primitive predecessor of modern theories.

But the point is - this is an assumption - there is no proof. These
ideas are falsifiable if neurology tells us that they do not apply.
I find extreme difficulty in trying to match the mechanisms of vision
to any of these logical theories. Vision is the only case where
neurology has advanced far enough to give us a hint of what is going
on. And, it seems to me, vision falsifies the logical theories.

> > There being no such thing as "natural language" only a collection of
> > natural languages - all different. I suspect that for a language
> > understanding in detail enough for formulization each idiolect would
> > be different.
> >
> > Chomsky is aware of what I just expounded. His solution is to postulate
> > a super-language from which actual languages are formed by selecting
> > the values of specific parameters. Many of us think that he has drifted
> > too far from reality to be significant any longer.
>
> Well if he thinks it is fundamentally based on English he is whacked out,
> otherwise its seems that we may be in at least somewhat of agreement.

At one point in time Chomsky seems to have thought that it was only
necessary to completely understand one language and all the others
would follow as corollaries. English was convenient. It could just as
well have been Japanese - native speakers were required. These days,
and for many years, Chomsky has encouraged explorations into multiple
languages.

[António Marques]

Peter Olcott <OCR4Screen> wrote:
> On 2/7/2015 3:24 AM, Helmut Richter wrote:
>> Am 07.02.2015 um 04:00 schrieb Peter Olcott:
>>
>>>> 1. The dog barks up the wrong tree
>>>
>>> The English idiom about [barking up the wrong tree]
>>> has nothing to do with dogs, barking, or trees.
>>
>> You would have noticed that I am familiar with it, had you read my >
>> next line before answering.
>>
>
> I could not tell that you were familiar with it based on what you said.
> Based on what you said I thought that it may be possible that you
> were totally unfamiliar with the completely non-compositional meaning
> of that idiom and were referring to a compositional meaning.

What's non-compositional about it?

[António Marques]

Peter Olcott <OCR4Screen> wrote:
> On 2/6/2015 2:01 PM, António Marques wrote:
>> Helmut Richter <hh...@web.de> wrote:
>>> 2. I share your view about how idiomatic meanings can be fitted into the
>>> framework of compositionality. Hence, the existence of idioms does not
>>> contradict compositionality but rather only contradicts the idea that the
>>> compositional mapping of a string to its meaning is unambiguous. Quite
>>> the contrary: it is nearly always ambiguous, and the more so when idioms
>>> are involved. I assume that you would agree in this point as well.
>> Last time I checked, Peter was only interested in a a subset of natural
>> language (that he believes exists) that is 'well formed' enough to be
>> described in Montagovian terms.
>
> Not at all.

'last time I checked'

[Helmut Richter]

When analyzing a sentence syntactically, one can do so by the function
of the constituents (subject, object, ...) or by their form (noun
phrase, prepositional phrase, ...). In many cases, one can infer the
function from the form, e.g that a noun phrase in some position relative
to the verb can only be an object. In other cases, the form does not
suffice, e.g. for deciding whether a prepositional phrase has the
function of an object or of an adverbial. I would have expected that a
formal grammar first analyzes according to function, and then the
functional constituents according to form until one reaches words (or
even mere morphemes) as last atoms that have only form whereas their
function is determined by their syntactic context. I see no reason why
one could not include *both* levels in such a formal model, and I do not
know why Chomsky has skipped the first step.

Although the function aspect sounds "semantic", all this is pure syntax
and does not require semantics -- that is, up to here, the "colourless
green ideas" are perfectly fine, and only semantics will reveal that
they are nonsense.

I find it a fascinating question to learn how far this kind of syntactic
analysis can get, and at what point semantics is necessarily involved.
It is known that in the brain the two run simultaneously: so-called
garden-path sentences would not work if the partially recognized
semantics were not able to confuse the syntanctic processing.
Nonetheless, the method to concentrate on syntax in order to learn which
structures are purely syntactic can bring new insight. I am not well
enough acquainted with Chonsky's work to judge; my impression is that
his problem is not what he actually did, but that he did not clearly
enough see and respect the limitations of his model.

Some of his ideas were successful in artificial computer languages
because these were designed with a syntax that is a one-to-one mapping
of the desired semantics. This is obviously not so with natural language
so that the experiences with artificial languages cannot be transferred.

--
Helmut Richter

[Peter Olcott]

[Peter Olcott]

On 2/8/2015 12:42 PM, DKleinecke wrote:
> On Saturday, February 7, 2015 at 12:10:07 PM UTC-8, Peter Olcott wrote:
>> Kurt Gödel (1944)
>> By the theory of simple types I mean the doctrine which says that the
>> objects of thought (or, in another interpretation, the symbolic
>> expressions) are divided into types, namely: individuals, properties of
>> individuals, relations between individuals, properties of such
>> relations, etc. (with a similar hierarchy for extensions), and that
>> sentences of the form: " a has the property φ ", " b bears the relation
>> R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
>> fitting together.
> KG says clearly that he is expounding a doctrine about the objects of
> thought. He has not in any sense proved it - he has postulated it.

I can see that it is inherently self-evidently true and this truth can
be proven
entirely on the basis of the meaning of his words.

Have you ever uncorked the color of your car?
How about washing a little ineptitude in the airplane parts?

It is possible to form semantically incorrect sentences using
correct syntax and word meanings that do not combine together
coherently.

[Peter Olcott]

It is *not* referring to any of these meanings: {dogs, barking, up, trees}.

[Peter Olcott]

On 2/8/2015 3:55 PM, António Marques wrote:

I am looking for a system such that anything that can be represented in any
language can be fully represented. I have never been looking for
anything else.

[António Marques]

That neither affirms nor contradicts what I said.

> I have never been looking for anything else.

Your interest was in making a machine understand scientific textbooks.

[António Marques]

Yes it is, only metaphorically.

[António Marques]

'Sure, that's a question I can ask.'

[Peter Olcott]

My interest has *always* been in creating a human mind with software.
The *only* reason that I have ever explored formal semantics within
linguistics is to either directly meet this goal, or to indirectly meet
this
goal by breaking this goal down into its component parts.

The key aspect of breaking this goal down into its component parts is to
somehow formalize the precise functional equivalent of the human
understanding of natural language.

The key aspect of formalizing the precise functional equivalent of the
human
understanding of natural language is to formalize the meaning of concepts
and the connections between concepts such that the [natural order] of
the essential structure of the universal set of all conceptual knowledge
could be discovered and exhaustively specified (fully elaborated) in a
rigorous
mathematical way.

The key aspect of formalizing the meaning of concepts and the connections
between concepts such that the [natural order] of the essential
structure of
the universal set of all conceptual knowledge could be discovered is finding
a system of absolute minimum complexity such that every detail of any
concept could be exhaustively specified and thus fully elaborated.

The essential framework for accomplishing this goal is a knowledge ontology
directed graph using integer handles are placeholders referring to 100%
precisely
specified concepts, such that all of these concepts are entirely created
on the
basis of their connections to other concepts within this knowledge
ontology.

The [natural order] of this knowledge ontology is defined by eliminating
redundancy in this knowledge ontology. the nodes in the digraph represent
concepts and the edges in the digraph represent connections between
concepts.

The first guess of the types of concepts(nodes) and the types of
connections between these nodes (edges) is specified below:

http://en.wikipedia.org/wiki/History_of_type_theory#G.C3.B6del_1944

Gödel 1944
Kurt Gödel in his 1944 Russell's mathematical logic gave the following
definition of the "theory of simple types" in a footnote:

By the theory of simple types I mean the doctrine which says that the
objects of thought (or, in another interpretation, the symbolic
expressions) are divided into types, namely: individuals, properties of
individuals, relations between individuals, properties of such
relations, etc. (with a similar hierarchy for extensions), and that
sentences of the form: " a has the property φ ", " b bears the relation
R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
fitting together.

Mixed types (such as classes containing individuals and classes as
elements) and therefore also transfinite types (such as the class of all
classes of finite types) are excluded. That the theory of simple types
suffices for avoiding also the epistemological paradoxes is shown by a
closer analysis of these. (Cf. Ramsey 1926 and Tarski 1935, p. 399).".[23]

Gödel 1944
Kurt Gödel in his 1944 Russell's mathematical logic gave the following
definition of the "theory of simple types" in a footnote:

By the theory of simple types I mean the doctrine which says that the
objects of thought (or, in another interpretation, the symbolic
expressions) are divided into types, namely: individuals, properties of
individuals, relations between individuals, properties of such
relations, etc. (with a similar hierarchy for extensions), and that
sentences of the form: " a has the property φ ", " b bears the relation
R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
fitting together.

Mixed types (such as classes containing individuals and classes as
elements) and therefore also transfinite types (such as the class of all
classes of finite types) are excluded. That the theory of simple types
suffices for avoiding also the epistemological paradoxes is shown by a
closer analysis of these. (Cf. Ramsey 1926 and Tarski 1935, p. 399).".[23]

[Peter Olcott]

Metaphorically then any issue of a Batman comic species
the complete meaning of the entire Holy Bible.

[Peter Olcott]

On 2/9/2015 8:32 AM, António Marques wrote:

What is the correct answer?
The correct answer must come from the intersection of the
set of hours and minutes with the set of elements: {yes, no}.

[DKleinecke]

On Sunday, February 8, 2015 at 8:15:24 PM UTC-8, Peter Olcott wrote:
> On 2/8/2015 12:42 PM, DKleinecke wrote:
> > On Saturday, February 7, 2015 at 12:10:07 PM UTC-8, Peter Olcott wrote:
> >> Kurt Gödel (1944)
> >> By the theory of simple types I mean the doctrine which says that the
> >> objects of thought (or, in another interpretation, the symbolic
> >> expressions) are divided into types, namely: individuals, properties of
> >> individuals, relations between individuals, properties of such
> >> relations, etc. (with a similar hierarchy for extensions), and that
> >> sentences of the form: " a has the property φ ", " b bears the relation
> >> R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
> >> fitting together.

> > KG says clearly that he is expounding a doctrine about the objects of
> > thought. He has not in any sense proved it - he has postulated it.
>
> I can see that it is inherently self-evidently true

Very few people share that view.

> and this truth can be proven entirely on the basis of the
> meaning of his words.

In this paragraph his words do not describe the situation well
enough for anyone else to even grasp what he is saying. For,
example, he does not explicate whether a relation can be between
more than two objects. Nor does he define "fitting together"

He describes types by ennumerating

individuals,
properties of individuals,
relations between individuals,
properties of such relations,
etc.

but it is not clear what "etc." covers.

It looks to me like there are really two ideas being mixed together.
Properties seems to have been imposed on another simpler structure
individuals,
relations between individuals,
etc.

The next step is not well defined. Is it
relations between relations between individuals
or is it
relations between two types at least one of which is a relation
between individuals ?
I assume this question could be answered by reading further into
KG's system - but it is not answered here. Nor is the answer obvious.

And if properties do not enter into relations why identify them as
types? Again KG probably has a precise definition - but not here.

> Have you ever uncorked the color of your car?
> How about washing a little ineptitude in the airplane parts?
>
> It is possible to form semantically incorrect sentences using
> correct syntax and word meanings that do not combine together
> coherently.

This is not news - colorless ideas were already sleeping.

And, as always, these questions can be given meaning

Have you ever sicked the Sierra Club on anyone?
How about overlooking a little ineptitude in my work in the
class about airplane parts?

The notion "semantically incorrect" does not seem to be well-defined.

[DKleinecke]

On Monday, February 9, 2015 at 9:35:06 AM UTC-8, Peter Olcott wrote:
>
> My interest has *always* been in creating a human mind with software.
> The *only* reason that I have ever explored formal semantics within
> linguistics is to either directly meet this goal, or to indirectly meet
> this goal by breaking this goal down into its component parts.
>
> The key aspect of breaking this goal down into its component parts is to
> somehow formalize the precise functional equivalent of the human
> understanding of natural language.
>
> The key aspect of formalizing the precise functional equivalent of the
> human understanding of natural language is to formalize the meaning of
> concepts and the connections between concepts such that the [natural

> order] of the essential structure of the universal set of all conceptual.

> knowledge could be discovered and exhaustively specified (fully elaborated)
> in a rigorous mathematical way.
>
> The key aspect of formalizing the meaning of concepts and the connections
> between concepts such that the [natural order] of the essential
> structure of the universal set of all conceptual knowledge could be
> discovered is finding a system of absolute minimum complexity such
> that every detail of any concept could be exhaustively specified and
> thus fully elaborated.
>
> The essential framework for accomplishing this goal is a knowledge
> ontology directed graph using integer handles are placeholders
> referring to 100% precisely specified concepts, such that all of these
> concepts are entirely created on the basis of their connections to other
> concepts within this knowledge ontology.
>
> The [natural order] of this knowledge ontology is defined by eliminating
> redundancy in this knowledge ontology. the nodes in the digraph represent
> concepts and the edges in the digraph represent connections between
> concepts.

At this point you have stated a coherent model of knowledge - a network
of concepts and relationships between pairs of concepts. I don't think
it is a useful model but it is well defined.

Why go further?

> The first guess of the types of concepts(nodes) and the types of
> connections between these nodes (edges) is specified below:

Why should you care about type theory at all? What does it add to
your model? Why is your model inadequate without it?

[Peter Olcott]

On 2/9/2015 12:14 PM, DKleinecke wrote:
> On Sunday, February 8, 2015 at 8:15:24 PM UTC-8, Peter Olcott wrote:
>> On 2/8/2015 12:42 PM, DKleinecke wrote:
>>> On Saturday, February 7, 2015 at 12:10:07 PM UTC-8, Peter Olcott wrote:
>>>> Kurt Gödel (1944)
>>>> By the theory of simple types I mean the doctrine which says that the
>>>> objects of thought (or, in another interpretation, the symbolic
>>>> expressions) are divided into types, namely: individuals, properties of
>>>> individuals, relations between individuals, properties of such
>>>> relations, etc. (with a similar hierarchy for extensions), and that
>>>> sentences of the form: " a has the property φ ", " b bears the relation
>>>> R to c ", etc. are meaningless, if a, b, c, R, φ are not of types
>>>> fitting together.
>>> KG says clearly that he is expounding a doctrine about the objects of
>>> thought. He has not in any sense proved it - he has postulated it.
>> I can see that it is inherently self-evidently true
> Very few people share that view.
>
>> and this truth can be proven entirely on the basis of the
>> meaning of his words.
> In this paragraph his words do not describe the situation well
> enough for anyone else to even grasp what he is saying. For,
> example, he does not explicate whether a relation can be between
> more than two objects. Nor does he define "fitting together"

// fitting together, thus incoherent
I a going to buy a can of paint and paint my house red.

// not fitting together, thus semantically incoherent
I am going to buy a can of injustice and paint my ineptitude blue.

> He describes types by ennumerating
> individuals,
> properties of individuals,
> relations between individuals,
> properties of such relations,
> etc.
> but it is not clear what "etc." covers.

He specifies the first, next and (thus implicitly) subsequent levels of
a recursion hierarchy.

>
> It looks to me like there are really two ideas being mixed together.
> Properties seems to have been imposed on another simpler structure
> individuals,
> relations between individuals,
> etc.
>
> The next step is not well defined. Is it
> relations between relations between individuals

You got that incorrectly, it is properties of relations between
individuals.
one person is greater than (relation between individuals) another person by
these properties:
1) height
2) weight

This is relations between relations:
[numerically greater than] and [physically darker than] are related to each
other in that they are both relations of increasing magnitude along
their respective dimensions: [numerical order] and [shade of color]
respectively.

> or is it
> relations between two types at least one of which is a relation
> between individuals ?
> I assume this question could be answered by reading further into
> KG's system - but it is not answered here. Nor is the answer obvious.
>
> And if properties do not enter into relations why identify them as
> types? Again KG probably has a precise definition - but not here.
>
>> Have you ever uncorked the color of your car?
>> How about washing a little ineptitude in the airplane parts?
>>
>> It is possible to form semantically incorrect sentences using
>> correct syntax and word meanings that do not combine together
>> coherently.
> This is not news - colorless ideas were already sleeping.
>
> And, as always, these questions can be given meaning
>
> Have you ever sicked the Sierra Club on anyone?
> How about overlooking a little ineptitude in my work in the
> class about airplane parts?
>
> The notion "semantically incorrect" does not seem to be well-defined.

Yet it obviously is fulfilled.

[Peter Olcott]

As I have already fully elaborated step-by-step, item-by-item,
detail-by-detail in the post that you just replied to:
To be able to recreate all of the capabilities of the human
mind in software.

I forgot to say acyclic digraph this time, (thus not a network)
also it does not seem that pairs of concepts will be defined,
but how-so-ever many naturally belong together. For example
the colors of the rainbow will not form a pair of concepts.

>> The first guess of the types of concepts(nodes) and the types of
>> connections between these nodes (edges) is specified below:
> Why should you care about type theory at all? What does it add to
> your model? Why is your model inadequate without it?
>

It seems to me that defining the types of nodes and edges of the acyclic
digraph forms the next most specific unit (thus minimal increment) of
further elaboration of the essential nature of the compositionality of the
atomic units of meaning: (comprised of the nodes and edges themselves)
within the acyclic digraph knowledge ontology.

I will repeat that , the nodes and edges within the acyclic digraph
knowledge
ontology are the atomic units of meaning.

Defining the types of nodes and edges within the acyclic digraph knowledge
ontology is the next minimal increment of further elaboration of the
[Natural
Order] of the set of all knowledge.

[DKleinecke]

Curious use of the the word "fulfilled".

I am reduced to not quoting the ancient joke about "obviously".

It isn't obvious to me and, as far as I know, anyone else.

[DKleinecke]

On Monday, February 9, 2015 at 12:10:01 PM UTC-8, Peter Olcott wrote:

> I will repeat that , the nodes and edges within the acyclic digraph
> knowledge ontology are the atomic units of meaning.

And I will repeat that that is a coherent possible model for
knowledge.

That I think it is not a useful model is a side issue.

> Defining the types of nodes and edges within the acyclic digraph knowledge
> ontology is the next minimal increment of further elaboration of the
> [Natural Order] of the set of all knowledge.

You need to offer at least a token justification for that statement.
What are going to gain from the types.

By postulating that your model is acyclic you have already avoided all
the paradoxes.