The Meaning of Abstract
Eray Ozkural exa
I'm surprised that some people take "abstract" as a vague common-sense
concept. To me, it has a precise technical meaning: lossy compression.
A program is abstract, because it *loses* the architectural details of
a computation, and it is concise. A blueprint of a house is abstract
because it *loses* the architectural and material properties of an
actual house, and it is concise.
"Instantiation" is most certainly a Platonist word which includes
counter-factuals in its meaning. An "abstract" entity represents
another entity in a purposeful way, it is a sign that points to
another object or sign.
Maybe we should all study semiotics instead of Platonist computer
science and mathematics! Maybe that is how one truly becomes a
hard-core materialist!
Regards,
PS: Dan, have I gotten materialist enough now?
--
Eray Ozkural
Jesse F. Hughes
> Greetings,
>
> I'm surprised that some people take "abstract" as a vague
> common-sense concept. To me, it has a precise technical meaning:
> lossy compression.
That's a very narrow meaning that has little currency in everyday
speech.
Your meaning fails to accommodate almost any of the uses of "abstract"
in mathematics, for instance. Algebra is not a "lossy compression" of
number theory or whatever else you might want to plug in.
It's no wonder that you're a fan of Chaitin. You're more than willing
to spread horribly sloppy attempts at thought.
--
"No feeling sympathy for mathematicians who start marching with signs
like 'Will work for food' in the future... I will not show mercy
going forward. I was trained as a soldier in the United States Army
after all... We play to win." --James Harris, feel his wrath!
Stephen Harris
"Eray Ozkural exa" <er...@bilkent.edu.tr> wrote in message
news:fa69ae35.04060...@posting.google.com...
> Greetings,
>
> I'm surprised that some people take "abstract" as a vague common-sense
> concept. To me, it has a precise technical meaning: lossy compression.
> counter-factuals in its meaning. An "abstract" entity represents
> another entity in a purposeful way, it is a sign that points to
> another object or sign.
>
> Maybe we should all study semiotics instead of Platonist computer
> science and mathematics! Maybe that is how one truly becomes a
> hard-core materialist!
>
Let us take the word "eternity" ... is that both a sign and an abstract
entity?
The answer is no because your definition above is wrong.
Curt Welch
> er...@bilkent.edu.tr (Eray Ozkural exa) writes:
>
> > Greetings,
> >
> > I'm surprised that some people take "abstract" as a vague
> > common-sense concept. To me, it has a precise technical meaning:
> > lossy compression.
>
> That's a very narrow meaning that has little currency in everyday
> speech.
>
> Your meaning fails to accommodate almost any of the uses of "abstract"
> in mathematics, for instance. Algebra is not a "lossy compression" of
> number theory or whatever else you might want to plug in.
>
> It's no wonder that you're a fan of Chaitin. You're more than willing
> to spread horribly sloppy attempts at thought.
"lossy compression" is far from "horribly sloppy". It's obviously correct.
Anyone that doesn't understand the reference doesn't understand
abstraction.
You can also think of abstraction as pattern matching or signal filtering.
It's just different ways to talk about the same thing.
--
Curt Welch http://CurtWelch.Com/
cu...@kcwc.com Webmaster for http://NewsReader.Com/
patty
Well its not so very totally obviously correct to me; although i must
admit that i flashed as such when i first read it. Thing is that an
abstraction is not located in time and space which is the most usual
criteria of something that is not physical. However, a compressed set
of marks is definitely located in time and space.
patty
Lester Zick
in comp.ai.philosophy wrote:
Hi Eray - I didn't know that "instantiation" comes from Plato. The
word comes from Latin and means to provide a concrete instance of.
But I still don't like it because people seem to use it to indicate
the the agency of physical realization of an abstract rather than a
mere concrete instance or example described by an abstract.
I'm not sure your telling us a lot either by defining "abstract" as
lossy compression. These aspects of an abstract may be true without
being definitive. When I inhale air there is lossy compression too.
I think you'd be better off just analyzing predicates in relation to
one another and recognizing that predicates are abstractions because
they're objective and they're objective for reasons that aren't plain
but that relate to the mechanics of differences between differences.
In point of fact the abstraction comes from differential mechanics in
which the lossy part of the abstraction refers to the loss of identity
as the result of taking differences between concrete things out there.
Regards - Lester
Jesse F. Hughes
> je...@phiwumbda.org (Jesse F. Hughes) wrote:
>> er...@bilkent.edu.tr (Eray Ozkural exa) writes:
>>
>> > Greetings,
>> >
>> > I'm surprised that some people take "abstract" as a vague
>> > common-sense concept. To me, it has a precise technical meaning:
>> > lossy compression.
>>
>> That's a very narrow meaning that has little currency in everyday
>> speech.
>>
>> Your meaning fails to accommodate almost any of the uses of "abstract"
>> in mathematics, for instance. Algebra is not a "lossy compression" of
>> number theory or whatever else you might want to plug in.
>>
>> It's no wonder that you're a fan of Chaitin. You're more than willing
>> to spread horribly sloppy attempts at thought.
>
> "lossy compression" is far from "horribly sloppy". It's obviously
> correct. Anyone that doesn't understand the reference doesn't
> understand abstraction.
I guess I don't understand abstraction.
Is "abstract" supposed to be a synonym for "lossy compression"? If
so, do you regard an mp3 ripped from a CD as an abstraction of the
original wave file?
> You can also think of abstraction as pattern matching or signal
> filtering. It's just different ways to talk about the same thing.
These uses don't capture the full meaning of abstraction in either
mathematics or philosophy, near as I can figure.
I don't see that, for instance, algebra has arisen via any of the
notions of "lossy compression", "pattern matching" or "signal
filtering" from the various concrete mathematical structures which
preceded it.
The act of filtering a signal or lossily compressing a source is very
different to my mind than the mathematical (or philosophical) act of
abstraction. Evidently, what is "obviously correct" in your view is
"horribly sloppy" in mine.
--
"Not all features that are found on the Security tab are designed to
help make your documents and files more secure." --Microsoft on Office
security features (after it was pointed out by a third party that a
certain password setting is easily bypassed.)
Nathan Funk
>
> Is "abstract" supposed to be a synonym for "lossy compression"? If
> so, do you regard an mp3 ripped from a CD as an abstraction of the
> original wave file?
I would definitely understand an mp3 as an abstract representation of
the raw waveform data contained on a CD. Do you have a better example
for the point you are trying to make?
> The act of filtering a signal or lossily compressing a source is very
> different to my mind than the mathematical (or philosophical) act of
> abstraction. Evidently, what is "obviously correct" in your view is
> "horribly sloppy" in mine.
I wouldn't go as far as to say it's obviously correct, however it "lossy
compression" is vaguely how I understand abstraction as well.
A few different questions: Does it have to be lossy? Can't abstraction
be lossless as well? If it is lossless, is it abstraction or merely a
different form representation? How can we universally quantify the level
of abstraction?
Nathan
Curt Welch
> Curt Welch wrote:
> > "lossy compression" is far from "horribly sloppy". It's obviously
> > correct. Anyone that doesn't understand the reference doesn't
> > understand abstraction.
> >
> > You can also think of abstraction as pattern matching or signal
> > filtering. It's just different ways to talk about the same thing.
> >
>
> Well its not so very totally obviously correct to me; although i must
> admit that i flashed as such when i first read it. Thing is that an
> abstraction is not located in time and space which is the most usual
> criteria of something that is not physical. However, a compressed set
> of marks is definitely located in time and space.
The belief that an abstraction is not located in time and space is just a
false, and useless belief.
We have the power to make stuff up that has never existed by combining
features from real things together in ways that we have never seen before.
Such as a flying elephant. Our power to generate these ideas are limitless.
But, does the idea of a flying elephant help you understand yourself or the
universe? Is it "real"? Well, the concept is real beacuse I just created
it for this post. And we all understand the concept so it's real in that
sense even if there are no flying elephants in the world.
Is the belief that an abstraction is not located in time and space a flying
elephant or is it real? It is in fact just a flying elephant.
Show me an abstraction that doesn't exist in time and space.
You can't do it because anything you can show me, by defintion has to exist
in time and space. All you can do is talk to me about flying elephants and
say they are real and tell me that you believe they are real. To say that
you believe something is real is to believe that you will one day see one,
but that you just haven't seen one yet. To say that something is real but
can never been seen, is your perosonal choice to believe that things which
can never be senseed, are in fact real. And if you belive that, then I
assume you believe not only that abstractions don't exist in time and
space, but that flying elephants are also just as real.
Our culture accepts fairly easilly the idea that things we can't sense are
real. This happens for many reasons. Mostly it happens because we put a
large amount of faith in second-hand knowledge. Bob said he saw a flying
elephant, and I trust Bob, so even though I've never seen one, I too
believe they are real. Other people said they have seen time distort
according to Einstein's predictions so I believe it is real even though I
have never seen it happen.
Everyone knows that concepts exist in a dimension outside of time and space
so everyone believes it. Everyone is wrong. Or, more accurately, they are
just talking about flying elephants because it's doesn't hurt to do so -
until you try to figure out how the brain works and what consciousness is.
The other reason we accept the idea so easily is the fact that these things
happen inside our head - where we can sense them in our thoughts - but
where we can not tie them to our other sensory inputs from the physcial
world. When we see an action, and can tie that to a sound, and a smell,
and a taste, and touch, then we know the thing we sensed exists in the
world of all those senses.
But when we sense our thoughts, we see no connections to our other senses -
we can not hear, see, taste, or feel, our own thoughts. That makes us
believe our thoughts exist in a separte world from the world of the other 5
senses. But, we have learned enough about the brain to know this is not
true. Eerything we think about is linked to brain activity, which we can
sense in physical world - even if most of us have never gotten the chance
to do that.
What we in fact have is a brain full of pattern recognition hardware. We
have pattern recognition hardware that can detect elephants, and hardware
which can detect flying things. And if we ever saw a flying elephant, we
would be able to detect it in a heartbeat. We would know what we were
looking at.
And we have trained our pattern recognition hardware for detecting
"elephant" to also respond to the word "elephant". But we do not confuse
the real thing for the word because we know that "elepahant in the context
of the hearing the word elephant" is different than "elepahnt in the
context of seeing a large animal".
It's the functioning of this pattern matching hardware which is creating
abstractions. It's a lossy compression system for responding to some
aspects of the data, and ignoring other aspects. Our pattern matching
hardware is the defintion of all abstractions we know about. The
abstraction "exists" in the form of the hardware which produces it, which
is very real and physical and exists in time and space, and in the outputs
it generates each time it is used, which is also something that is very
real in time and space.
The belief that an abstraction could be "timeless" is the belief that these
pattern matching machines could exist forever, and could exist without a
physcial form - both which are impossible - i.e., flying elephants.
Curt Welch
> cu...@kcwc.com (Curt Welch) writes:
>
> > je...@phiwumbda.org (Jesse F. Hughes) wrote:
> >> er...@bilkent.edu.tr (Eray Ozkural exa) writes:
> >>
> >> > Greetings,
> >> >
> >> > I'm surprised that some people take "abstract" as a vague
> >> > common-sense concept. To me, it has a precise technical meaning:
> >> > lossy compression.
> >>
> >> That's a very narrow meaning that has little currency in everyday
> >> speech.
> >>
> >> Your meaning fails to accommodate almost any of the uses of "abstract"
> >> in mathematics, for instance. Algebra is not a "lossy compression" of
> >> number theory or whatever else you might want to plug in.
> >>
> >> It's no wonder that you're a fan of Chaitin. You're more than willing
> >> to spread horribly sloppy attempts at thought.
> >
> > "lossy compression" is far from "horribly sloppy". It's obviously
> > correct. Anyone that doesn't understand the reference doesn't
> > understand abstraction.
>
> I guess I don't understand abstraction.
>
> Is "abstract" supposed to be a synonym for "lossy compression"? If
> so, do you regard an mp3 ripped from a CD as an abstraction of the
> original wave file?
Of course it is. Is there not a large set of CDs that all have different
data values stored on them, yet they would all be abstractly represented by
the same mp3 file?
To make it easier to see, think of a lossy compression algorithm which is
easy to reverse. Just sum the values of all the bytes in a file together
using 8 bit math so that you end up with a single 8 bit check sum value to
represent the file with.
Are there not an infinite numbner of files which are abstractly represented
with the same 8 bit sum? Am I not able to talk about all files with an 8
bit check sum value of 107 as an abstract set of files?
> > You can also think of abstraction as pattern matching or signal
> > filtering. It's just different ways to talk about the same thing.
>
> These uses don't capture the full meaning of abstraction in either
> mathematics or philosophy, near as I can figure.
They do, but you just don't see it yet.
> I don't see that, for instance, algebra has arisen via any of the
> notions of "lossy compression", "pattern matching" or "signal
> filtering" from the various concrete mathematical structures which
> preceded it.
Why it arose of course is a harder idea to get a handle on but the fact
that it is created by combining abstractions together is not so hard to
grasp once you learn to look at it like that.
Do you not see that numbers are just abstactions created by combining
different properties of real things together? 3 sheep and 3 oranges share
a common abstract property which we call "3". Do you not see that all
"properties" are abstractions of real things?
> The act of filtering a signal or lossily compressing a source is very
> different to my mind than the mathematical (or philosophical) act of
> abstraction. Evidently, what is "obviously correct" in your view is
> "horribly sloppy" in mine.
Sure, I know we look at it very differently. I'm just trying to show you
how it's possible to adjust your set of abstractions about the world to
make it fit my set of abstractions and about how the world becomes much
simpler when you do that (i.e., many things which seemed unrelated are
actually very much related).
Ioannis
news:c9ir97$n5q$1...@tabloid.srv.ualberta.ca...
[snip]
> A few different questions: Does it have to be lossy? Can't abstraction
> be lossless as well? If it is lossless, is it abstraction or merely a
> different form representation? How can we universally quantify the level
> of abstraction?
One possible way to visualize whatever the original poster means by
"abstraction", would be to look at some of the evolved computer languages
that were (and are) used to represent various objects "abstractly".
In most (if not all) of these languages, the notion of "pointer" is of
fundamental importance.
Imo, the most natural way to define "abstraction" for a certain set of data,
is to provide a "pointer" to the data.
In such models, "abstraction" doesn't have to be lossy and it appears to be
a different form of representation of the same object.
In away, the Index of an encyclopedia (or book) seems to me like an
"abstraction" of the actual contents.
> Nathan
--
Ioannis Galidakis
http://users.forthnet.gr/ath/jgal/
------------------------------------------
Eventually, _everything_ is understandable
Jesse F. Hughes
>> I guess I don't understand abstraction.
>>
>> Is "abstract" supposed to be a synonym for "lossy compression"? If
>> so, do you regard an mp3 ripped from a CD as an abstraction of the
>> original wave file?
>
> I would definitely understand an mp3 as an abstract representation of
> the raw waveform data contained on a CD. Do you have a better example
> for the point you are trying to make?
Most of us don't regard the relation between mp3 and wave file as the
same as the relation between the concept of triangle and *all* of the
particular three-sided figures with which we are familiar[1].
That x is created by lossy compression from a source y puts x in a
relationship with a particular y. When we abstract, we forget certain
details of each in a collection of instances, not the details of just
one particular instance.
But even this forgetting is not really the defining feature of
abstraction. Consider the operation that takes algebras to their
underlying sets by "forgetting" the algebraic structure. This
operation loses some data, just like lossy compression, but it is not
an act that most of us would call abstraction. One doesn't abstract
algebras to reach a notion of set[2]. The fact that data is lost or
forgotten doesn't capture the notion of abstraction, although it may
be a feature of abstraction (similarly, there is nothing about
compression as far as the forgetting is concerned).
Anyway, if you want to say that an mp3 is an abstraction of a
particular wave file, well, you can have that reductio ad absurdum.
Footnotes:
[1] I'm not taking the position that the notion of triangle is really
somehow abstracted from our experience with particular
three-sided-figures here.
[2] I can't say why this *isn't* a good example of abstraction, but I
think one would find that most philosophers of mathematics wouldn't
call this forgetful functor an instance of abstraction.
--
"So, at this time, I'd like to assure you that I am not interested in
making sure mathematicians worldwide get fired. I've rethought my
desire to go to Congress and try to get funding for mathematicians
cut." -- James Harris is a reasonable man. Whew!
Jesse F. Hughes
Most of us distinguish between quotienting by an equivalence relation
and abstracting.
To repeat the example I posted a moment ago: fix a signature S and
consider the functor which takes S-algebras to their underlying set.
Just like your example, for each set, there are a slew of S-algebras
which are mapped to that set. I would not think that this forgetful
functor is an example of abstracting.
I think it's fairly evident that you and I are using the word
abstraction in different senses.
[...]
>
>> > You can also think of abstraction as pattern matching or signal
>> > filtering. It's just different ways to talk about the same thing.
>>
>> These uses don't capture the full meaning of abstraction in either
>> mathematics or philosophy, near as I can figure.
>
> They do, but you just don't see it yet.
Don't patronize.
>> I don't see that, for instance, algebra has arisen via any of the
>> notions of "lossy compression", "pattern matching" or "signal
>> filtering" from the various concrete mathematical structures which
>> preceded it.
>
> Why it arose of course is a harder idea to get a handle on but the fact
> that it is created by combining abstractions together is not so hard to
> grasp once you learn to look at it like that.
>
> Do you not see that numbers are just abstactions created by combining
> different properties of real things together? 3 sheep and 3 oranges share
> a common abstract property which we call "3". Do you not see that all
> "properties" are abstractions of real things?
Oh, no, no, no. I am not walking down this path with you. I can
already predict that a discussion of the origins of the concept of
number will be somewhat more painful than licking razor blades.
Thanks, but no thanks.
>> The act of filtering a signal or lossily compressing a source is very
>> different to my mind than the mathematical (or philosophical) act of
>> abstraction. Evidently, what is "obviously correct" in your view is
>> "horribly sloppy" in mine.
>
> Sure, I know we look at it very differently. I'm just trying to show you
> how it's possible to adjust your set of abstractions about the world to
> make it fit my set of abstractions and about how the world becomes much
> simpler when you do that (i.e., many things which seemed unrelated are
> actually very much related).
The only thing common to all of these notions is that in each, certain
information is lost. Big deal. That's not all that deep and it
doesn't seem to capture the essence of abstraction.
In particular, the actual acts of compression or signal filtering take
a single input to produce a new, "simpler" output. That's different
than the abstraction from certain particulars of individuals in a set
to arrive at a new concept in which each individual is an instance.
These are just different conceptual acts. I believe that this is also
different than pattern matching, in which one finds common features
rather than finding a broader notion the includes each of the
particulars.
--
Jesse F. Hughes
"History will hate you and love me. I'm the misunderstood and
persecuted genius. You're the assholes." -- James Harris
Jesse F. Hughes
> cu...@kcwc.com (Curt Welch) writes:
>
>>
>> Of course it is. Is there not a large set of CDs that all have
>> different data values stored on them, yet they would all be
>> abstractly represented by the same mp3 file?
>
> Most of us distinguish between quotienting by an equivalence relation
> and abstracting.
>
> To repeat the example I posted a moment ago: fix a signature S and
> consider the functor which takes S-algebras to their underlying set.
> Just like your example, for each set, there are a slew of S-algebras
> which are mapped to that set. I would not think that this forgetful
> functor is an example of abstracting.
Stupid!
A simpler example.
Take the map N -> N taking 0 to 0 and n+1 to n. This is a lossy
compression scheme. Would you *really* want to call it an
abstraction?
--
Jesse F. Hughes
"Even I, who know beyond doubt that my death will be caused by a silly
girl, will not hesitate when that girl passes by." -- Merlin, as
reported by John Steinbeck.
David B. Held
news:fa69ae35.04060...@posting.google.com...
>
> common-sense concept. To me, it has a precise technical
> meaning: lossy compression.
I have to agree with Mr. Hughes. If you find that to be the
essential feature of "abstraction", then I fear you haven't
abstracted the notion properly. ;) I think a better meaning is:
to identify the necessary features of a concept. "Lossy
compression" is insufficient to describe this version because
it often includes features that are *not* essential to the
abstraction of an instance. If I show you a picture of a girl,
and you produce a JPEG image of it and present it to me as
"an abstraction", I'll say you're full of it. Now, if I ask a 5 year
old to draw what he sees in the picture, he will most likely
draw a stick figure of a girl, identifying what he believes to be
the essential features of the image. And I'm fairly certain that
a large survey with our two "abstractions" will show that the
stick figure is a much better abstraction than the JPEG. Why?
Because the JPEG contains too much information. It takes
out redundant information, but only on a local level. The JPEG
compression algorithm necessarily cannot identify the information
contained in the picture that depends on having human knowledge
and experience. And that is why it does not produce a useful
abstraction. The same can be said of the MP3 file. An MP3 is
not a very good abstraction of the original. But a piano rendition
with a monophonic melody and simple chord accompaniment
*would* be what many musicians regard as an "abstraction" of
the original song.
> A program is abstract, because it *loses* the architectural details
> of a computation, and it is concise.
Actually, programs aren't very abstract at all. They tend to be rather
concrete engineering solutions to practical problems. And a review
of www.dictionary.com shows that "concrete" and "practical" are
practically antonyms of "abstract".
> A blueprint of a house is abstract because it *loses* the
> architectural and material properties of an actual house, and it is
> concise.
> [...]
A blueprint is certainly more abstract than an actual house, but not
because of what it is missing. It is abstraction because of what it
has, which is the essential features of the house. Whether it
specifies the placement of all the electrical outlets or not is of no
concern. Whether it conveys the proper layout of rooms and floors
is. If I created a blueprint that located all the electrical outlets in the
house but not any of the walls, according to your definition, it is an
"abstraction" of the house, because I have *lost* detail, and created
a "concise" version of the house. But merely losing detail is not the
key feature. The key feature is identifying the details that we use to
classify the concrete instances. Abstraction is a notion concerned
with classes and types, not details and data representation. Such
a definition is not sufficiently "abstract".
Dave
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.691 / Virus Database: 452 - Release Date: 5/26/2004
Curt Welch
> A simpler example.
Good. Your other example was beyond me. I'm posting from the ai group,
not the math group. I really don't know why this thread was cross posted
to this set of groups.
> Take the map N -> N taking 0 to 0 and n+1 to n. This is a lossy
> compression scheme. Would you *really* want to call it an
> abstraction?
Sure, it's an abstraction. Just not a very useful one. But I agree, it's
very unlikely I would ever use the world "abstraction" to describe it
outside the conext of this type of discussion.
In AI work, ideas of "abstraction" are a common themem. Understanding
abstraction is key to understanding what the brain does. When you explore
these things, you start to see how all these different ideas (and more -
like data clasification) are closely realated and very likely, only one
problem to the brain hardware.
When I first posted to this thread, I didn't notice the strange cross
posting and thought the AI context was assumed. So I wasn't trying to
debate the correct usage of the word "abstraction" but to simply debate the
concept in the context of building smart machines.
ZZBunker
> Greetings,
>
> I'm surprised that some people take "abstract" as a vague common-sense
> concept. To me, it has a precise technical meaning: lossy compression.
>
> A program is abstract, because it *loses* the architectural details of
> a computation, and it is concise. A blueprint of a house is abstract
> because it *loses* the architectural and material properties of an
> actual house, and it is concise.
No it hasn't. Since real time programs
lose no architectural details.
since otherwise Internet wouldn't exist.
And neither blueprints or houses or science would
exist if real-time programs didn't exist.
A. G. McDowell
Ozkural exa <er...@bilkent.edu.tr> writes
>Greetings,
>
>I'm surprised that some people take "abstract" as a vague common-sense
>concept. To me, it has a precise technical meaning: lossy compression.
>
that suggests itself to me is from program-proving. See e.g. "the b-
method" by steve schneider, chapter 14.
He describes how an implementation T1 can be proved to implement
(actually, refine) a more abstract machine T, so that we can take an
argument that the T-machine satisfies some requirements and turn it into
an argument that the T1-machine also satisfies some requirements. What
we require is some link between the possible sequence of states of T1
and that of T. It turns out that it is not useful to try and link the
two with a function, going in either direction: T1 may be allowed some
non-determinism, so that more than one possible sequence of T1-states is
equivalent to a single sequence of T-states. The behaviour we require of
T may be satisfied by a T-machine that has some non-determinism not
present in T1, so it may be perfectly OK for more than one sequence of
T-states to map to a single sequence of T1-states.
(While I would accept that an injective function might be described as a
lossy compression, I would not normally expect a compression scheme to
sometimes map the same input to different outputs, or equivalently to
map a single input to a set of possible outputs, which is what can
happen here).
To link the machines, we need the more general notion of a predicate,
which takes a sequence of T1-states and a sequence of T-states and says
whether the two are equivalent. For the predicate to allow us to show
that T1 satisfies its requirements, there are specific conditions on it
(proof obligations). Schneider takes a chapter to explain them, so I'm
not going to try very hard, but basically you need to show that the
predicate links T1 and T closely enough that any reachable sequence in
T1 has a matching (predicate(T, T1) = true) sequence in T, and that
since the T-sequence is a match, if it satisfies the requirements then
the T1-sequence must also satisfy the requirements. This allows you to
work out that a proof that T is correct must translate to a proof that
T1 can exist and is and correct.
--
A. G. McDowell
Jesse F. Hughes
> je...@phiwumbda.org (Jesse F. Hughes) wrote:
>
>> A simpler example.
>
> Good. Your other example was beyond me. I'm posting from the ai group,
> not the math group. I really don't know why this thread was cross posted
> to this set of groups.
>
>> Take the map N -> N taking 0 to 0 and n+1 to n. This is a lossy
>> compression scheme. Would you *really* want to call it an
>> abstraction?
>
> Sure, it's an abstraction. Just not a very useful one. But I agree, it's
> very unlikely I would ever use the world "abstraction" to describe it
> outside the conext of this type of discussion.
>
> In AI work, ideas of "abstraction" are a common themem. Understanding
> abstraction is key to understanding what the brain does.
I am doubtful that the brain's facility of abstraction has much to do
with the metaphors which you claimed were obviously correct, so I'm
not too sure that you have grasped that key to understanding yet.
But I will not debate that point, since it is possible that your
notion of abstraction is simply a technical term distinct from the
common usage.
> When you explore these things, you start to see how all these
> different ideas (and more - like data clasification) are closely
> realated and very likely, only one problem to the brain hardware.
[...]
--
Jesse Hughes
"Well, if I can get [my proof of FLT accepted], then I hopefully get a
book deal down the road, and maybe I get to go on 'Oprah'."
James Harris, on the rewards of mathematical endeavours.
Herman Jurjus
"Jesse F. Hughes" <je...@phiwumbda.org> wrote in message news:87y8n7c...@phiwumbda.org...
> Nathan Funk <noe...@noemail.com> writes:
>
> >> I guess I don't understand abstraction.
> >>
> >> Is "abstract" supposed to be a synonym for "lossy compression"? If
> >> so, do you regard an mp3 ripped from a CD as an abstraction of the
> >> original wave file?
> >
> > I would definitely understand an mp3 as an abstract representation of
> > the raw waveform data contained on a CD. Do you have a better example
> > for the point you are trying to make?
>
> Most of us don't regard the relation between mp3 and wave file as the
> same as the relation between the concept of triangle and *all* of the
> particular three-sided figures with which we are familiar[1].
Isn't that called "generalization", rather than abstraction?
BTW, the common sense meanings of 'group', 'set', 'class'
'number', 'sequence', 'field', 'contradiction' are also rather
different from their formal mathematical counterparts.
One wonders how sloppy, irrelevant, or plainly wrong
mathematics must look in the eyes of a non-mathie...
;-)
Cheers,
Herman Jurjus
Jesse F. Hughes
> "Jesse F. Hughes" <je...@phiwumbda.org> wrote in message news:87y8n7c...@phiwumbda.org...
>> Nathan Funk <noe...@noemail.com> writes:
>>
>> >> I guess I don't understand abstraction.
>> >>
>> >> Is "abstract" supposed to be a synonym for "lossy compression"? If
>> >> so, do you regard an mp3 ripped from a CD as an abstraction of the
>> >> original wave file?
>> >
>> > I would definitely understand an mp3 as an abstract representation of
>> > the raw waveform data contained on a CD. Do you have a better example
>> > for the point you are trying to make?
>>
>> Most of us don't regard the relation between mp3 and wave file as the
>> same as the relation between the concept of triangle and *all* of the
>> particular three-sided figures with which we are familiar[1].
>
> Isn't that called "generalization", rather than abstraction?
Maybe so. Bad example, I guess.
>
> BTW, the common sense meanings of 'group', 'set', 'class'
> 'number', 'sequence', 'field', 'contradiction' are also rather
> different from their formal mathematical counterparts.
> One wonders how sloppy, irrelevant, or plainly wrong
> mathematics must look in the eyes of a non-mathie...
But mathematicians don't claim that they've provided the "correct"
technical definition for the common-sense usage. If one wants to use
the term "abstraction" in a certain technical sense, then there is no
controversy. If they claim, on the other hand, that this technical
sense clarifies the common-sense usage, then that's a different
matter.
I've never heard a mathematician argue that Farmer Joe Bob's use of
the term "field" is just a vague approximation of the correct and
precise definition found in mathematics[1].
But lately, folks have responded to my intended reductio ad absurdums
by agreeing with them, so maybe I should offer that example with
trepidation.
Footnotes:
[1] Maybe some folk would argue that "set", "class" or "sequence" are
precisifications[2] of the common sense notions. In fact, the claim
seems fairly plausible for "sequence" in particular.
[2] A technical term of art. Honest.
--
Jesse Hughes
"If you really think there's a bug you should report a bug. Maybe
you're not using it properly... It turns out Luddites don't know how
to use software properly, so you should look into that." -- Bill Gates
Stephen Harris
"Eray Ozkural exa" <er...@bilkent.edu.tr> wrote in message
news:fa69ae35.04060...@posting.google.com...
>
> I'm surprised that some people take "abstract" as a vague common-sense
> concept. To me, it has a precise technical meaning: lossy compression.
> "Instantiation" is most certainly a Platonist word which includes
> counter-factuals in its meaning. An "abstract" entity represents
> another entity in a purposeful way, it is a sign that points to
> another object or sign.
>
> Maybe we should all study semiotics instead of Platonist computer
> science and mathematics! Maybe that is how one truly becomes a
> hard-core materialist!
>
I am a bit uncertain that all traces of Platonism are excluded in the
explantion provided below or that "abstract" is treated as "vague
common-sense",
but the comprehension compression is certainly loses nothing to the Tao*.
Adding Abstract to Formal and Content Schemata:
Results of Recent Work in Peircean Semiotics
John W. Oller, Jr.
http://members.door.net/arisbe/menu/library/aboutcsp/oller/schemata.htm
"Abstract schemata, by contrast, concern everything that is contained
within the meaning or definition of a symbol (including propositions,
arguments, and discourses). They take all that possibly could be (as
known through the symbols used) and relate it to whatever must be
(provided the symbols are used truly).
*As a result, the abstract level
of the symbol (as Peirce showed) reaches from outside of time and space
into the material world and yet is itself neither temporal nor spatial
in its compass.* It comes nearer to our ideas of eternity, infinity,
continuity, and universality than to anything known through our senses
in the material world. It involves what Peirce called a 'thirdness'
beyond the duality of opposing forces clashing in the material realm."
Industrial strength supervening,
Stephen
David C. Ullrich
>je...@phiwumbda.org (Jesse F. Hughes) wrote:
>
>> A simpler example.
>
>Good. Your other example was beyond me. I'm posting from the ai group,
>not the math group. I really don't know why this thread was cross posted
>to this set of groups.
>
>> Take the map N -> N taking 0 to 0 and n+1 to n. This is a lossy
>> compression scheme. Would you *really* want to call it an
>> abstraction?
>
>Sure, it's an abstraction.
So 41 is an "abstraction" derived from 42?
That's nothing at all like "abstraction" as we know it - if
this is an abstraction of that then any particular that
should be an instance of this. Is 42 an instance of 41?
(42 is a special case of 41 somehow?)
>Just not a very useful one. But I agree, it's
>very unlikely I would ever use the world "abstraction" to describe it
>outside the conext of this type of discussion.
Glad to hear _that_, at least. I can't imagine why 41 would
count as an abstraction of 42 in _any_ context.
>In AI work, ideas of "abstraction" are a common themem. Understanding
>abstraction is key to understanding what the brain does. When you explore
>these things, you start to see how all these different ideas (and more -
>like data clasification) are closely realated and very likely, only one
>problem to the brain hardware.
That's fascinating, but it doesn't give me any hint why it is that
in AI it's appropriate to consider 42 a special case of 41.
>When I first posted to this thread, I didn't notice the strange cross
>posting and thought the AI context was assumed. So I wasn't trying to
>debate the correct usage of the word "abstraction"
Which is what the thread is about.
>but to simply debate the
>concept in the context of building smart machines.
************************
David C. Ullrich
Herman Rubin
>I'm surprised that some people take "abstract" as a vague common-sense
>concept. To me, it has a precise technical meaning: lossy compression.
>A program is abstract, because it *loses* the architectural details of
>a computation, and it is concise. A blueprint of a house is abstract
>because it *loses* the architectural and material properties of an
>actual house, and it is concise.
You are confusing "abstract" with "abstraction". None
of these is abstract, although they are abstractions.
When understood, an abstract idea is a mental structure,
which may or may not arise by the process of abstraction.
It is likely to be better understood if it does not.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hru...@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558
Eray Ozkural exa
> > I guess I don't understand abstraction.
> >
> > Is "abstract" supposed to be a synonym for "lossy compression"? If
> > so, do you regard an mp3 ripped from a CD as an abstraction of the
> > original wave file?
>
> I would definitely understand an mp3 as an abstract representation of
> the raw waveform data contained on a CD. Do you have a better example
> for the point you are trying to make?
Whether such an example exists is of interest for me, as well.
> > The act of filtering a signal or lossily compressing a source is very
> > different to my mind than the mathematical (or philosophical) act of
> > abstraction. Evidently, what is "obviously correct" in your view is
> > "horribly sloppy" in mine.
> I wouldn't go as far as to say it's obviously correct, however it "lossy
> compression" is vaguely how I understand abstraction as well.
Yes, I am trying to put it in mechanical terms so that we can build an
AI that abstracts.
> A few different questions: Does it have to be lossy? Can't abstraction
> be lossless as well? If it is lossless, is it abstraction or merely a
> different form representation? How can we universally quantify the level
> of abstraction?
In my opinion, it has to be lossy for it to be abstract. If it is not
lossless, it is simply an algorithmic copy of the information (in
requiring a decompressor program), ie. a full representation of the
object. That is also quite important, as shown in "Information
Distance" of Charles Bennett et al.
You have stated an important question: lossy in what sense? I am
working on this question, and I have a few mathematical ideas, but I
do not know how well they will turn out. Immature as they are, I'd
rather not speak of them. References to theoretical work would be
appreciated.
Best Regards,
--
Eray Ozkural
patty
An ideal circle. Take an line segment and keeping the segment in the
same plane, rotate it around one end, the shape that is made by the
opposite end is the shape of an ideal circle. It is interesting to note
that we can imagine and\or recognize the shape without knowing the
procedure.
That shape (note the singular) is not located in time and space. If it
is located there, please provide me the coordinates.
> You can't do it because anything you can show me, by defintion has to exist
> in time and space. All you can do is talk to me about flying elephants and
> say they are real and tell me that you believe they are real.
I have no problem with calling a ideal circle, not real.
> To say that
> you believe something is real is to believe that you will one day see one,
> but that you just haven't seen one yet. To say that something is real but
> can never been seen, is your perosonal choice to believe that things which
> can never be senseed, are in fact real. And if you belive that, then I
> assume you believe not only that abstractions don't exist in time and
> space, but that flying elephants are also just as real.
>
That last sentence does not follow, but then the whole paragraph is
based upon the erroneous assumption that i believe that a ideal circle
is real in the sense that it can be seen or sensed.
Although the ideal circle cannot be sensed, we can manifest approximate
instances of it that can be sensed. That is why we call the ideal
circle an abstraction and why we make a special slot for it in our
ontology. It exists in our ontology (in our culture and language) for
the purpose of discussions, but it does not exist in time and space.
> Our culture accepts fairly easilly the idea that things we can't sense are
> real. This happens for many reasons. Mostly it happens because we put a
> large amount of faith in second-hand knowledge. Bob said he saw a flying
> elephant, and I trust Bob, so even though I've never seen one, I too
> believe they are real. Other people said they have seen time distort
> according to Einstein's predictions so I believe it is real even though I
> have never seen it happen.
>
There is a big difference between a fictional thing and a abstract
thing. Both exist in our ontology. Neither of them can be found in
time and space. Neither of them can be sensed. Both of them can be
imagined.
> Everyone knows that concepts exist in a dimension outside of time and space
> so everyone believes it. Everyone is wrong. Or, more accurately, they are
> just talking about flying elephants because it's doesn't hurt to do so -
> until you try to figure out how the brain works and what consciousness is.
>
> The other reason we accept the idea so easily is the fact that these things
> happen inside our head - where we can sense them in our thoughts - but
> where we can not tie them to our other sensory inputs from the physcial
> world. When we see an action, and can tie that to a sound, and a smell,
> and a taste, and touch, then we know the thing we sensed exists in the
> world of all those senses.
>
Yes it is reasonable to call those things that we sense and that others
can sense as well as "real". Those things that we make up as fictions
or that we just imagine are then termed "not real". I would call them
"imaginary". They don't exist in some other dimension, they are
artifacts of the behavior of our brains and are signs that we stick in
our culture. We can play a language game about them and discuss them
and if there are effective procedures for reproducing instances of them
we can call them "ideals" or "abstractions" to differentiate them from
fictions which are related to no such effective procedures. I would
think that even you would agree that all of that is fairly
uncontroversial. Do you not?
> But when we sense our thoughts, we see no connections to our other senses -
> we can not hear, see, taste, or feel, our own thoughts. That makes us
> believe our thoughts exist in a separte world from the world of the other 5
> senses. But, we have learned enough about the brain to know this is not
> true. Eerything we think about is linked to brain activity, which we can
> sense in physical world - even if most of us have never gotten the chance
> to do that.
>
> What we in fact have is a brain full of pattern recognition hardware. We
> have pattern recognition hardware that can detect elephants, and hardware
> which can detect flying things. And if we ever saw a flying elephant, we
> would be able to detect it in a heartbeat. We would know what we were
> looking at.
>
> And we have trained our pattern recognition hardware for detecting
> "elephant" to also respond to the word "elephant". But we do not confuse
> the real thing for the word because we know that "elepahant in the context
> of the hearing the word elephant" is different than "elepahnt in the
> context of seeing a large animal".
>
I have no major problems with those last three paragraphs.
> It's the functioning of this pattern matching hardware which is creating
> abstractions.
Ok.
> It's a lossy compression system for responding to some
> aspects of the data, and ignoring other aspects. Our pattern matching
> hardware is the defintion of all abstractions we know about. The
> abstraction "exists" in the form of the hardware which produces it, which
> is very real and physical and exists in time and space, and in the outputs
> it generates each time it is used, which is also something that is very
> real in time and space.
>
If we base our ontology on process rather than substance then we could
say that the process of our brains imagining an ideal (of abstracting)
is a *subclass* of the process of making a lossy compression of marks.
It is merely a *sub* class because there are compressions that do not
have the particular features that are necessary for us to call it the
process of constructing an abstraction, as others in this thread have
pointed out.
> The belief that an abstraction could be "timeless" is the belief that these
> pattern matching machines could exist forever, and could exist without a
> physcial form - both which are impossible - i.e., flying elephants.
>
There is an important distinction between saying that something exists
forever and saying that for it time is irrelevant.
I would like to use this example an an excuse to show how semiotics can
help clarify some of these issues for us. Please refer to the diagram:
<http://www.icyberspace.net/patty/diagrams/SEMIOT.JPG>
Now let's discuss where the object of the sign that would be marked with
the letters "circle" exists ? The diagram indicates that all objects
exist in a plane labeled "environment, language, culture". Well we
already eliminated the possibility that it is a sensual object in the
"real" environment, so it must be an object in language and\or culture.
Our friends from the EAB might say that it exists as contingencies in
our language and our culture. And I would agree with them; that is
most reasonable place for us to place it.
You, on the other hand, might argue that it is an object in the squiggly
line (which stands for cause and effect connections) that are drawn
inside the interpreters which normally are found inside peoples heads.
Let's see if that would work. Does the ideal circle go away if your
head ceases to exist? I think not. Does the ideal circle go away if
all such human heads cease to exist? I think not, for some alien
creature could study the markings of our culture and recreate that
concept from those. So it is in that culture that this object does
exist. I rest my case.
Do you see the use of semiotics now?
patty
Curt Welch
> You, on the other hand, might argue that it is an object in the squiggly
> line (which stands for cause and effect connections) that are drawn
> inside the interpreters which normally are found inside peoples heads.
> Let's see if that would work. Does the ideal circle go away if your
> head ceases to exist? I think not. Does the ideal circle go away if
> all such human heads cease to exist?
Yes.
> I think not, for some alien
> creature could study the markings of our culture and recreate that
> concept from those.
Then it comes back at that time. Where was it between those times?
> So it is in that culture that this object does
> exist. I rest my case.
>
> Do you see the use of semiotics now?
No. It seems to be blinding you to some simple facts of reality.
Do you belive in ghosts as well?
The only reason you think the idea of a perfect circle exists outside of
time and space is because you have chosen to believe it to be so. There is
nothing you can sense that tells you this is true. You have just made a
personal decison to believe in a ghost.
Sounds like fun. Let me try it. Ok, fugrats exist outside of time and
space. Even if you kill me, fugrats will continue to exist! You can not
kill the fugrat because even if you kill everyone that has read this
message some alien could come and learn about the glorious fugrat!
So did fugrats exist before now? Do they exist outside of time and space
now that I have chosen to believe they do?
If we were to erase all copies of the fugrat idea from the universe, do you
think it would ever come back again? Do think if it came back again that
it would have anything to do with my fugrat?
Do you think there is something special about your fugrat (the perfect
circle) that makes it better than my fugrats so your fugrat can exist
outside of space and time whereas my fugrat can't?
Or do you have a problem that I haven't defined what a fugrat is yet?
It's a configuration of 24 objects of sizes from 1 to 24 where the 24
object is exactly 24 times the size of the 1 object. They are lined up in a
row where the order of the objects in the set runs from one end to the
other like this:
22 24 14 15 1 7 20 8 17 9 11 23 19 12 13 3 4 5 6 10 2 16 18 21
So has the idea of a fugrat always existed? Does it now exist outside of
space and time now that I have defined it?
patty
> patty <pat...@SPAMicyberspace.net> wrote:
>
>
>>You, on the other hand, might argue that it is an object in the squiggly
>>line (which stands for cause and effect connections) that are drawn
>>inside the interpreters which normally are found inside peoples heads.
>>Let's see if that would work. Does the ideal circle go away if your
>>head ceases to exist? I think not. Does the ideal circle go away if
>>all such human heads cease to exist?
>
>
> Yes.
>
>
>>I think not, for some alien
>>creature could study the markings of our culture and recreate that
>>concept from those.
>
>
> Then it comes back at that time. Where was it between those times?
>
>
>>So it is in that culture that this object does
>>exist. I rest my case.
>>
>>Do you see the use of semiotics now?
>
>
> No. It seems to be blinding you to some simple facts of reality.
>
> Do you belive in ghosts as well?
>
> The only reason you think the idea of a perfect circle exists outside of
> time and space is because you have chosen to believe it to be so.
I have never said that a circle exists outside of time and space. You
keep foisting these words upon me. For one thing the word "exists" is
problematic, and for another thing time is irrelivant to the circle. To
talk of time space coordinates of such an ideal circle would be a
catagory error.
> There is
> nothing you can sense that tells you this is true. You have just made a
> personal decison to believe in a ghost.
>
Well in a sense, yes. I have made a choice to talk about it in this
manner and to look at the situation from this perspective. There is a
sense in which we are both right. I am saying that the circle owes it's
persistance to the marks in our culture (and it does) ... you are saying
that without the brain processess those marks remain just marks (and
they do). Why must we think only in coins of one side ? We know that
there is a feedback loop working here, let us use it.
> Sounds like fun. Let me try it. Ok, fugrats exist outside of time and
> space. Even if you kill me, fugrats will continue to exist! You can not
> kill the fugrat because even if you kill everyone that has read this
> message some alien could come and learn about the glorious fugrat!
>
> So did fugrats exist before now? Do they exist outside of time and space
> now that I have chosen to believe they do?
>
> If we were to erase all copies of the fugrat idea from the universe, do you
> think it would ever come back again? Do think if it came back again that
> it would have anything to do with my fugrat?
>
> Do you think there is something special about your fugrat (the perfect
> circle) that makes it better than my fugrats so your fugrat can exist
> outside of space and time whereas my fugrat can't?
>
> Or do you have a problem that I haven't defined what a fugrat is yet?
>
> It's a configuration of 24 objects of sizes from 1 to 24 where the 24
> object is exactly 24 times the size of the 1 object. They are lined up in a
> row where the order of the objects in the set runs from one end to the
> other like this:
>
> 22 24 14 15 1 7 20 8 17 9 11 23 19 12 13 3 4 5 6 10 2 16 18 21
>
> So has the idea of a fugrat always existed? Does it now exist outside of
> space and time now that I have defined it?
>
Well i doubt that your fugrat will stick in our culture ... though it
will perculate to many clients today and it will remain on google groups
for many years. But there is one thing about the circle which makes it
so very persistent and distinguishes itself from your fugrat. It is
useful. It is easy to recreate. I would even bet that it would be
recreated in most technological cultures regargless of how alien. Your
fugrat is just a doodle in the sand that will soon be buried in a flood
of useless usenet jibberish.
The real question here is one of semantics. To what does the word
"circle" refer ? Does it refer collectively to all the brain process
that have ever comprehended it - however vague and diverse ? Or does
it refer to what caused the persistant usage of the concept in our
culture ? When i refer to a penny am i refering to the head side or
the tail side?
patty
Curt Welch
> The real question here is one of semantics.
Yes, I think the difference in our views is all a mater of semantics.
> To what does the word
> "circle" refer ? Does it refer collectively to all the brain process
> that have ever comprehended it - however vague and diverse ? Or does
> it refer to what caused the persistant usage of the concept in our
> culture ? When i refer to a penny am i refering to the head side or
> the tail side?
Well, this is what I'm getting at. In the real world, we have words, and
we have objects. And that's all we have. And the words and objects exist,
or they don't exist, and there's just no question about that.
We have the power to write words like "perfect circle", that functions in
the langauge as if it were refering to a real object, but in fact, it is
refering to something that can't be real - that can't exist. Perfect
circles don't exist. Only the words that describe them exist.
Just because we have created words, doesn't me we have created a "perfect
circle". We have only created a description. We have created the idea of
a perfect circle.
You said the word exist was problematical. It's not if you use it like I
use it.
perfect circles don't exist. They are imaginary. Only the description of
them in our langauge exists. You only get confused over the correct usage
of "exist" if you confuse the reference (which does exist) with the object
of the reference which, for perfect circle, does not exist.
Everything that exists has a location in space and time.
patty
> patty <pat...@SPAMicyberspace.net> wrote:
>
>
>>The real question here is one of semantics.
>
>
> Yes, I think the difference in our views is all a mater of semantics.
>
>
>>To what does the word
>>"circle" refer ? Does it refer collectively to all the brain process
>>that have ever comprehended it - however vague and diverse ? Or does
>>it refer to what caused the persistant usage of the concept in our
>>culture ? When i refer to a penny am i refering to the head side or
>>the tail side?
>
>
> Well, this is what I'm getting at. In the real world, we have words, and
> we have objects. And that's all we have.
What about the agents (interpreters) who map the words to the objects?
> And the words and objects exist,
> or they don't exist, and there's just no question about that.
>
It's not really that simple. What about mental states? Do intentions,
wants, desires, beliefs exist in your so called "real world" ? What
about patterns? What about the price of beans in China today? What is
the space time cordinates of that value? What about 3 ? Where is it
located?
> We have the power to write words like "perfect circle", that functions in
> the langauge as if it were refering to a real object, but in fact, it is
> refering to something that can't be real - that can't exist. Perfect
> circles don't exist. Only the words that describe them exist.
>
Well in a way that's what i have been saying all along. The word
"circle" refers to the contingencies in our environment that cause us to
use words to describe it. There just is no way that we are talking
about a fiction here. The circle is far to useful for that to be the
case.
> Just because we have created words, doesn't me we have created a "perfect
> circle". We have only created a description. We have created the idea of
> a perfect circle.
>
I have no problem with saying it that way. Note your usage of the word
"idea". A circle is an ideal. If you want to disallow ideals in your
ontology, then that is fine with me, but i think i can function with
those in my ontology and not get confused. Check out the SUMO IEEE
ontology; this is not just me. I do try to be careful and not reify
ideals too much because therein lies folly; but there is a long history
of playing the language game with these ideals to useful ends; even, i
might add, doing logic with them. I am not prepared to abandon that
tradition and reduce everything to the simplicity of just those things
for which i can assign space time coordinates.
> You said the word exist was problematical. It's not if you use it like I
> use it.
>
> perfect circles don't exist. They are imaginary. Only the description of
> them in our langauge exists. You only get confused over the correct usage
> of "exist" if you confuse the reference (which does exist) with the object
> of the reference which, for perfect circle, does not exist.
>
Well something exists that caused this word to be used so much; that is
what the word refers to. "quigunk" is an example of a word for which no
object exists. Is the referent of "circle" anything like the referent
of "quigunk" ?
> Everything that exists has a location in space and time.
>
No, that is a subset of those things for which the word "exists" can be
usefully applied.
patty
Lester Zick
comp.ai.philosophy wrote:
Existence has more than one meaning. The description of perfect
circles exists in language because their existence is real in the
description. Whether perfect circles have material realizations is
another issue. There are probably about as many examples of perfect
circles that exist in physical terms as there are perfect computers or
perfect anything. But the problem isn't trivial and if you eliminate
the idea of perfect, the issue of what is and isn't materially
realizable remains.
Regards - Lester
Will Twentyman
> Greetings,
>
> I'm surprised that some people take "abstract" as a vague common-sense
> concept. To me, it has a precise technical meaning: lossy compression.
This does not agree with how I would compare College Algebra with
Abstract Algebra (the most obvious use of the word in mathematics).
Both of them deal with operations on sets, and both of them are
concerned with the properties of these operations and sets. The
differences lie in the focus. While College Algebra is focused on
specific manipulations, such as u-substitutions, solving equations,
etc., Abstract Algebra is focused on properties of the operation(s).
The distinction appears to be in how much we care about the individual
elements of the set. In both cases, the various properties of the
operation(s) on the set are important (distributive, commutative,
associative, etc.) I would consider College Algebra an abstraction of
Arithmatic, and Abstract Algebra as an abstraction of College Algebra.
The abstraction occurs in the form of a shift of focus. First, from
knowing the numbers to not always knowing the numbers, then from caring
about any of the numbers to frequently caring about only a few.
Based on that, Abstract algebra is more abstract than College Algebra,
which is more abstract than Arithmatic. I see no evidence of
compression, much less lossy compression.
> A program is abstract, because it *loses* the architectural details of
> a computation, and it is concise. A blueprint of a house is abstract
> because it *loses* the architectural and material properties of an
> actual house, and it is concise.
Some programs may represent an abstraction of a real-life scenario, if
the program is a model. The program itself is not abstract. It is, at
times, annoyingly precise, especially when it involves an extra
semi-colon between the statement of a for loop and the "body" of the for
loop.
> "Instantiation" is most certainly a Platonist word which includes
> counter-factuals in its meaning. An "abstract" entity represents
> another entity in a purposeful way, it is a sign that points to
> another object or sign.
I don't think Instantiation shows up in mathematics at all. I could be
wrong, of course.
The word "car" is the name of an abstract concept. When I think of car,
I do not think of any particular car, but of several images of cars and
the characteristics that must be satisfied to be considered a car.
Along with this are certain filters to distinguish between a car and a
truck, SUV, airplane, etc.
I think abstract is a relative term, also. A stack as an abstract data
type compared to a stack implemented in C++ (or other language) would
serve. The ADT for the stack is the abstract version, while the
implementation, with all its code, is the concrete version. Yet an ADT
specification is still concrete compared to "something to hold objects
satisfying LIFO".
"Lossy Compression" may be applicable to some of these examples, but it
doesn't seem to adequately describe what it means to be abstract, or
even to be an abstraction.
--
Will Twentyman
email: wtwentyman at copper dot net
Curt Welch
> Curt Welch wrote:
>
> > patty <pat...@SPAMicyberspace.net> wrote:
> >
> >
> >>The real question here is one of semantics.
> >
> >
> > Yes, I think the difference in our views is all a mater of semantics.
> >
> >
> >>To what does the word
> >>"circle" refer ? Does it refer collectively to all the brain process
> >>that have ever comprehended it - however vague and diverse ? Or does
> >>it refer to what caused the persistant usage of the concept in our
> >>culture ? When i refer to a penny am i refering to the head side or
> >>the tail side?
> >
> >
> > Well, this is what I'm getting at. In the real world, we have words,
> > and we have objects. And that's all we have.
>
> What about the agents (interpreters) who map the words to the objects?
Those are objects too.
> > And the words and objects exist,
> > or they don't exist, and there's just no question about that.
> >
>
> It's not really that simple. What about mental states?
What about the state of a pile of rocks? How is it different?
> Do intentions,
> wants, desires, beliefs exist in your so called "real world" ?
The words exist. But like "perfect circle" the thing they imply exists
doesn't exist. What they are used for however is to understand and describe
behavior, which does exist.
When I tie a weight to the end of a rope and spin it in a circle, I'm using
the idea of a "perfect circle" to help undertand the path of the weight at
the end of the rope. The weight does not move in a perfect circle, but
yet, it's LIKE a perfect circle - meaning it shares features in common with
this thing called a perfect circle which doesn't actually exist - only the
description of it exists. The idea of a perfect circle is useful as a tool
for understanding other real things in the world. If we study the
properties of a perfect circle, we can gain understanding about the real
things in the world which are LIKE a perfect circle.
The idea of "wants" is like the idea of a perfect circle. Wants are not
real, but the the idea is useful to help us understand real things - our
behavior.
> What
> about patterns? What about the price of beans in China today? What is
> the space time cordinates of that value? What about 3 ? Where is it
> located?
You really are having a hard time with this aren't you?
> > We have the power to write words like "perfect circle", that functions
> > in the langauge as if it were refering to a real object, but in fact,
> > it is refering to something that can't be real - that can't exist.
> > Perfect circles don't exist. Only the words that describe them exist.
> >
>
> Well in a way that's what i have been saying all along.
Yes, I think it is. I think we are saying and thinking bascially the same
thing. I think the idea of "existence" and "real" is where we drift apart
on undersanding.
> The word
> "circle" refers to the contingencies in our environment that cause us to
> use words to describe it. There just is no way that we are talking
> about a fiction here. The circle is far to useful for that to be the
> case.
Exactly. We extract properties of real things, and talk about the
properties in new ways. We can extract the idea of wings, and elephants,
and put these two properties together to create a winged elephant that can
fly yet we have never seen one. And we can extract the entire field of
math from what the brain sees in the real world, but what we manipulate in
math are these extracted properties, not the real things they came from.
> > Just because we have created words, doesn't me we have created a
> > "perfect circle". We have only created a description. We have created
> > the idea of a perfect circle.
> >
>
> I have no problem with saying it that way. Note your usage of the word
> "idea". A circle is an ideal. If you want to disallow ideals in your
> ontology, then that is fine with me, but i think i can function with
> those in my ontology and not get confused.
I don't know what an ontology is so I can't use that idea to discuss this
with you.
But, I have no problem with the idea of a perfect circle. I of course use
it all the time. I just happen to know it's a description which maps to an
imaginary object - i.e. an object that can't exist.
Maybe your problem with my langauge is that you start with the assumption
that all objects referenced by langauge do exist simply because we have a
word for them? So you define existence based on what we can describe
instead of what we can sense?
I describe existence based on what we can sense.
We can sense circular objects, but we could not sense a perfect circle.
We learn to first describe those things which we can sense, and then use
that lanague to describe things which we have never sensed, and in some
cases, like the perfect circle, can never sense.
> Check out the SUMO IEEE
> ontology; this is not just me. I do try to be careful and not reify
> ideals too much because therein lies folly; but there is a long history
> of playing the language game with these ideals to useful ends; even, i
> might add, doing logic with them. I am not prepared to abandon that
> tradition and reduce everything to the simplicity of just those things
> for which i can assign space time coordinates.
>
> > You said the word exist was problematical. It's not if you use it like
> > I use it.
> >
> > perfect circles don't exist. They are imaginary. Only the description
> > of them in our langauge exists. You only get confused over the correct
> > usage of "exist" if you confuse the reference (which does exist) with
> > the object of the reference which, for perfect circle, does not exist.
> >
>
> Well something exists that caused this word to be used so much;
It's a an abstract property which can be observed in all real things that
are circular in nature.
The brain has hardware which abstracts properties from the senses and the
"property" is only real in the sense that we have pattern matching hardware
which activates whenever the hardware is exposed to a sensory signal which
has that property. It's this basic low level data clasification hardware
which allows us at the higher levels to create ideas like "perfect circle"
and talk about it and use it help us understand the world.
> that is
> what the word refers to. "quigunk" is an example of a word for which no
> object exists. Is the referent of "circle" anything like the referent
> of "quigunk" ?
>
> > Everything that exists has a location in space and time.
> >
>
> No, that is a subset of those things for which the word "exists" can be
> usefully applied.
Yeah, see, we are just using the word "exist" differently. And probably
other associated words like "real".
"wants" are not real to me. The idea of "wants" is what is real. And the
idea of "wants" exists only as language behavior which helps us to
understand the real things in the universe - like human behavior.
Eray Ozkural exa
>
> I'm surprised that some people take "abstract" as a vague common-sense
> concept. To me, it has a precise technical meaning: lossy compression.
>
> a computation, and it is concise. A blueprint of a house is abstract
> because it *loses* the architectural and material properties of an
> actual house, and it is concise.
Maybe, I should state my question in a more mathematical way.
Suppose that we have a meta-mathematical theory of theorem proving. In
this system, there are entities like proofs, which are derivations in
a formal axiomatic system. Now, suppose that some of the proofs we
would like to observe are practically impossible to understand due to
their excessive length. [+] We would like to construct a complementary
system that will summarize these proofs in another language [*],
giving us proof sketches, which will be comprehensible to
mathematicians. What is the best way to formalize this process of
abstraction?
References to work with similar goals would be much appreciated.
Best Regards,
--
Eray Ozkural
[+] At this point, it could be argued that it is not necessary to
"observe" anything. The very purpose of such a device is generating
proofs that we can rely on. However, bear with me, and consider that
we would like to somehow "visualize" a proof for studying these
proofs.
[*] By this I mean an artificial language, but I suspect it could as
well be English.
patty
> patty <pat...@SPAMicyberspace.net> wrote:
>
>>I have no problem with saying it that way. Note your usage of the word
>>"idea". A circle is an ideal. If you want to disallow ideals in your
>>ontology, then that is fine with me, but i think i can function with
>>those in my ontology and not get confused.
>
>
> I don't know what an ontology is so I can't use that idea to discuss this
> with you.
>
Formal ontologies can be very useful; they can eliminate most of the
kinds of misunderstandings that have been happening in this train.
Check out the one that the IEEE is working on:
<http://ontology.teknowledge.com/>
> But, I have no problem with the idea of a perfect circle. I of course use
> it all the time. I just happen to know it's a description which maps to an
> imaginary object - i.e. an object that can't exist.
>
> Maybe your problem with my langauge is that you start with the assumption
> that all objects referenced by langauge do exist simply because we have a
> word for them? So you define existence based on what we can describe
> instead of what we can sense?
>
Well there is some precedence for that. Pat Hayes
<http://www.ihmc.us/users/user.php?UserID=42> who is a respected
authority in this area once said something to the effect that if you
don't want Unicorns to exist, then don't talk about them. Whenever we
use language we invent worlds (formal semantics calls these
interpretations) and we talk of whether a thing exists in a particular
world or not.
> I describe existence based on what we can sense.
If you want to deal with only the world of things with time\space
coordinates, with things that can be sensed and measured, then that is
fine with me. There are lots of terms for just that world ... my
favorite is "the natural world". From now on (unless otherwise
specified by you) i will take you term "exists x" as meaning that x
exists in what i call the "natural world". However, when i communicate
with most other people, I think that reduction would be an inconvenience.
>
> We can sense circular objects, but we could not sense a perfect circle.
>
> We learn to first describe those things which we can sense, and then use
> that lanague to describe things which we have never sensed, and in some
> cases, like the perfect circle, can never sense.
>
Fine.
>>>
>>>perfect circles don't exist. They are imaginary. Only the description
>>>of them in our langauge exists. You only get confused over the correct
>>>usage of "exist" if you confuse the reference (which does exist) with
>>>the object of the reference which, for perfect circle, does not exist.
>>>
>>
>>Well something exists that caused this word to be used so much;
>
>
> It's a an abstract property which can be observed in all real things that
> are circular in nature.
>
> The brain has hardware which abstracts properties from the senses and the
> "property" is only real in the sense that we have pattern matching hardware
> which activates whenever the hardware is exposed to a sensory signal which
> has that property. It's this basic low level data clasification hardware
> which allows us at the higher levels to create ideas like "perfect circle"
> and talk about it and use it help us understand the world.
>
That is fine with me; however i doubt that most of us came to the idea
of a circle by inductive pattern matching. But that is a whole
different topic.
>>>Everything that exists has a location in space and time.
>>>
>>
>>No, that is a subset of those things for which the word "exists" can be
>>usefully applied.
>
>
> Yeah, see, we are just using the word "exist" differently. And probably
> other associated words like "real".
>
Yes i think we are in violent agreement on that :)
> "wants" are not real to me. The idea of "wants" is what is real. And the
> idea of "wants" exists only as language behavior which helps us to
> understand the real things in the universe - like human behavior.
>
Whether "wants" and other mental states exist in the natural world is,
as you now doubt know, a hotly debated topic on this forum. There is no
doubt in my mind that these terms can be mapped to objects in some
interpretation. As such they are useful tokens in a language game. Is
that interpretation (world) scientific and extensional. I think not.
Should we avoid these terms when trying to understand human psychology?
Luckily i do not need to answer that question because it is not my
field of study. Should we avoid attributing these states to our AI
mechanisms? I don't know Curt, what do you think?
patty, sitting on a fence.
Neil W Rickert
Hash: SHA1
er...@bilkent.edu.tr (Eray Ozkural exa) writes:
>er...@bilkent.edu.tr (Eray Ozkural exa) wrote in message news:<fa69ae35.04060...@posting.google.com>...
>> Greetings,
>> I'm surprised that some people take "abstract" as a vague common-sense
>> concept. To me, it has a precise technical meaning: lossy compression.
>> A program is abstract, because it *loses* the architectural details of
>> a computation, and it is concise. A blueprint of a house is abstract
>> because it *loses* the architectural and material properties of an
>> actual house, and it is concise.
The trouble is that you have confused the technical term "abstract"
from object oriented programming, with the ordinary word "abstract".
>Maybe, I should state my question in a more mathematical way.
>Suppose that we have a meta-mathematical theory of theorem proving. In
>this system, there are entities like proofs, which are derivations in
>a formal axiomatic system. Now, suppose that some of the proofs we
>would like to observe are practically impossible to understand due to
>their excessive length. [+] We would like to construct a complementary
>system that will summarize these proofs in another language [*],
>giving us proof sketches, which will be comprehensible to
>mathematicians. What is the best way to formalize this process of
>abstraction?
It's a process of summarization, not of abstraction. And you probably
cannot formalize it.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (SunOS)
iD8DBQFAv9s2vmGe70vHPUMRAj4EAKC2+B/dg3BC2Gv0oRg4k2D0FZZ7wgCfYptF
k1GWyCkdzNdqMYEs8Rr2DVk=
=pXYr
-----END PGP SIGNATURE-----
fishfry
Many years ago I was in the audience at a debate on artificial
intelligence. Someone made the point that a good test of intelligence
would be to input a novel, and have the computer produce a decent
summarization. The point being that summarizing a novel involves
understanding the meaning of the novel, whatever we mean by
"understanding" and "meaning," and is qualitatively different than any
brute-force analysis of the symbols of the novel.
Your challenge seems similar -- to abstract from a formal proof, a
"sketch" of a proof that is (1) logically rigorous even with all the
tiny details missing, and (2) "comprehensible" by mathematicians -- a
term that takes into account the backgrounds, abilities, and
psychological dispositions of a group of typical professional
mathematicians.
I don't think humans are obsolete quite yet.
Curt Welch
> Whether "wants" and other mental states exist in the natural world is,
> as you now doubt know, a hotly debated topic on this forum. There is no
> doubt in my mind that these terms can be mapped to objects in some
> interpretation. As such they are useful tokens in a language game. Is
> that interpretation (world) scientific and extensional. I think not.
> Should we avoid these terms when trying to understand human psychology?
> Luckily i do not need to answer that question because it is not my
> field of study. Should we avoid attributing these states to our AI
> mechanisms? I don't know Curt, what do you think?
>
> patty, sitting on a fence.
I would not try to build "wants" into my machine, but I would try to build
a machine which wants. If I built it correctly, I would expect to see it
act in a way that tells me it wants stuff. I would expect my food seeking
AI to show clear signs that it wants food for example. If it was put in a
test environment where it had to collect tokens to get food, I would expect
it to develop clear signs of wanting tokens as well.
David B. Held
news:fa69ae35.0406...@posting.google.com...
> er...@bilkent.edu.tr (Eray Ozkural exa) wrote in message
news:<fa69ae35.04060...@posting.google.com>...
> Suppose that we have a meta-mathematical theory of theorem
> proving. In this system, there are entities like proofs, which are
> derivations in a formal axiomatic system. Now, suppose that some
> of the proofs we would like to observe are practically impossible
> to understand due to their excessive length.
Are proofs difficult to understand because of their length, or because
they don't relate to something we understand?
> [+] We would like to construct a complementary system that will
> summarize these proofs in another language [*], giving us proof
> sketches, which will be comprehensible to mathematicians. What
> is the best way to formalize this process of abstraction?
It sounds to me like you want a formal system translator. The
problem is that the "target language" is, in fact, a "moving target".
What constitutes "comprehensible to mathematicians" depends
on which mathematicians you mean, and the state of the art. In fact,
I would say that such a system would "understand" the theorems
as well as or better than the client mathematicians. That's because
in order to construct a "summary proof" that is comprehensible to
mathematicians, the translator will need to have knowledge of a
similar set of experiences and facts as the mathematicians. A
good teacher is not one that can explain a concept to her peers,
but to novices (or, as I like to say, to 5 year olds). That's because
the teacher is relying on the experience of the audience, which
happens to be quite different from her own.
Furthermore, the translator will have to understand the proof well
enough to perform the mapping. But in order to understand the
proofs, the translator will have to itself have a lot of experience in
the source domain. Essentially, the translator would have to be a
first-rate expert system.
Can such a system be formalized? Possibly, but one wonders
whether the process of formalization would enable the knowledge
engineers themselves to perform the target translation. After all,
it is rare to find an expert system that performs duties considerably
different from the experience of the experts who were consulted to
design it.
I would say that all across the board, science is reaching the "limits
of reductionism". Such formalizations rely on the reductionist
premise that understanding the parts of abstraction will enable us
to create building blocks for a theorem translator that will be greater
than the sum of the parts. But more and more we are finding that
the sum of the parts is much less than the emergent whole. So I
fear that such things as "understanding theorems" will not be
tackled reductionistically. If it were possible, why wouldn't we have
significant progress already, when we have made significant
progress in other fields that are amenable to reductionism?
Such a translator touches on issues of epistemology and the
like that we obviously haven't gotten down to an engineering
practice yet. Like anything, however, I think insight would be
gained by starting with a trivial such system and iterating it
until you identify the step in which there is a "phase change"
and the complexity of the system's behavior runs away at a
rate that is difficult to follow. We sure aren't going to understand
mathematician's brains by looking at individual neurons, so
why would we take that approach with an automated theorem
translator?
Curt Welch
> Curt Welch wrote:
> > I don't know what an ontology is so I can't use that idea to discuss
> > this with you.
> >
>
> Formal ontologies can be very useful; they can eliminate most of the
> kinds of misunderstandings that have been happening in this train.
> Check out the one that the IEEE is working on:
> <http://ontology.teknowledge.com/>
IEEE is actually IEEE. surprise surprise.
I suspect they actually create and/or support the kinds of misunderstanding
that have been happening here. But I would have to study them more to be
able to debate that idea.
David Longley
<pat...@SPAMicyberspace.net> writes
The problem with McCarthy (and possibly Hayes too, but I'll leave it at
McCarthy as I've had exchanges with him and not Hayes), is that the idea
of coding "common sense" knowledge goes rather against that. I've
suggested where that went wrong (ie in turning to Carnap and not
understanding Quine). The fact that people can have skills in one area
and make what amount to major faux pas in another, and be unable to see
this, or readily change this behaviour, is something I have,
paradoxically perhaps, gone to some lengths to explicate here.
>
>> I describe existence based on what we can sense.
>
>If you want to deal with only the world of things with time\space
>coordinates, with things that can be sensed and measured, then that is
>fine with me. There are lots of terms for just that world ... my
>favorite is "the natural world". From now on (unless otherwise
>specified by you) i will take you term "exists x" as meaning that x
>exists in what i call the "natural world". However, when i communicate
>with most other people, I think that reduction would be an inconvenience.
It isn't a *reduction* though. That was the point of "Two Dogmas of
Empiricism" and it has been central to what I (and Glen) have been
telling you and others here (although in Glen's case he doesn't need the
Quinean philosophy as he tends to refer to the empirical work directly -
something Quine himself advocates). When you resort to intensional
locutions you simply behave in ways which we know are demonstrably
indeterminate. Why you don't take the advice you just cited Hayes as
giving is just more grist to my mill.
But it *is* in your field of application/behaviour. Like the rest of us
you can tell yourself whatever you like. You can "assume" that you are
"understood", you can "think that" you "know" what others "mean". You
can "attribute" and "infer", you can "intend", and "hope" or "feel that"
you are right, and all that time you can "believe" what you "want". But
the chances are, a lot of the time, you just won't know what you are
doing or be able to say what you are talking about if you stand back and
try to analyse any of it. You'll just see that you are going round and
round in intensional circles. When you come to program, or do anything
practical, all of that talk will really just be dropped and you'll see
you can do without it. In fact, to do what you need to "effectively",
you really have no choice.
Given that, what do you think you might be doing when you resort to folk
psychological locutions? (might you just be behaving in conventional
ways? Try seeing those locutions as concatenated phonemes, ie as
behavious - mands and tacts).
> Should we avoid attributing these states to our AI mechanisms? I
>don't know Curt, what do you think?
>
>patty, sitting on a fence.
--
David Longley
patty
> In article <6JQvc.41854$eY2.28042@attbi_s02>, patty
> <pat...@SPAMicyberspace.net> writes
>>
>> <http://www.ihmc.us/users/user.php?UserID=42> who is a respected
>> authority in this area once said something to the effect that if you
>> don't want Unicorns to exist, then don't talk about them. Whenever we
>> use language we invent worlds (formal semantics calls these
>> interpretations) and we talk of whether a thing exists in a particular
>> world or not.
>
>
> The problem with McCarthy (and possibly Hayes too, but I'll leave it at
> McCarthy as I've had exchanges with him and not Hayes), is that the idea
> of coding "common sense" knowledge goes rather against that. I've
> suggested where that went wrong (ie in turning to Carnap and not
> understanding Quine). The fact that people can have skills in one area
> and make what amount to major faux pas in another, and be unable to see
> this, or readily change this behaviour, is something I have,
> paradoxically perhaps, gone to some lengths to explicate here.
>
If coding this "common sense" was really feasible, and if you could get
everybody to use the same coding, then the world would be a different
place; and if pigs had winds we would see them make fly-bys. Look at
all the problems and years of dilog that has been going on over on the
IEEE SUO ontology mailing list - and mostly those are professional
people with a practical goal in mind. But I have no idea what you are
getting at in your paragraph above.
>> If you want to deal with only the world of things with time\space
>> coordinates, with things that can be sensed and measured, then that is
>> fine with me. There are lots of terms for just that world ... my
>> favorite is "the natural world". From now on (unless otherwise
>> specified by you) i will take you term "exists x" as meaning that x
>> exists in what i call the "natural world". However, when i
>> communicate with most other people, I think that reduction would be an
>> inconvenience.
>
>
> It isn't a *reduction* though. That was the point of "Two Dogmas of
> Empiricism" and it has been central to what I (and Glen) have been
> telling you and others here (although in Glen's case he doesn't need the
> Quinean philosophy as he tends to refer to the empirical work directly -
> something Quine himself advocates). When you resort to intensional
> locutions you simply behave in ways which we know are demonstrably
> indeterminate. Why you don't take the advice you just cited Hayes as
> giving is just more grist to my mill.
>
Maybe it's because i like unicorns.
>> Whether "wants" and other mental states exist in the natural world is,
>> as you now doubt know, a hotly debated topic on this forum. There is
>> no doubt in my mind that these terms can be mapped to objects in some
>> interpretation. As such they are useful tokens in a language game.
>> Is that interpretation (world) scientific and extensional. I think
>> not. Should we avoid these terms when trying to understand human
>> psychology? Luckily i do not need to answer that question because it
>> is not my field of study.
>
>
> But it *is* in your field of application/behaviour. Like the rest of us
> you can tell yourself whatever you like. You can "assume" that you are
> "understood", you can "think that" you "know" what others "mean". You
> can "attribute" and "infer", you can "intend", and "hope" or "feel that"
> you are right, and all that time you can "believe" what you "want". But
> the chances are, a lot of the time, you just won't know what you are
> doing or be able to say what you are talking about if you stand back and
> try to analyse any of it. You'll just see that you are going round and
> round in intensional circles. When you come to program, or do anything
> practical, all of that talk will really just be dropped and you'll see
> you can do without it. In fact, to do what you need to "effectively",
> you really have no choice.
>
Actually i rather "like" that paragraph; it is good writing. It made me
stop and think about it. Certainly what goes on in usenet and in most
of human discourse has little practical import and just goes in endless
circles. Makes one wonder what game is being spun. Why shouldn't we
all just *stop it* ! Let us also stop reinforcing our children for
using intensions. Why should my child be concerned with what i am
thinking; when he could not ever "know that" i am "thinking that" his
life is heading toward a disaster. And when he bips me on my head why
should i be "concerned that" he doesn't "know that" it makes me "feel
like" crying. These are all so "demonstrably indeterminate", let us now
just excise them from our behavioral repertory.
Or not! Perhaps what we do here is build the very fabric which
inevitably is experienced as consciousness.
But of course we don't need that either :(
> Given that, what do you think you might be doing when you resort to folk
> psychological locutions? (might you just be behaving in conventional
> ways? Try seeing those locutions as concatenated phonemes, ie as
> behavious - mands and tacts).
>
For those (like me) unfamiliar with "mands and tacts" here is a passage
from Psych 1AA3 Child Development:
""
Skinner distinguishes two broad classes of multi-word utterances which
he calls mands and tacts. A mand is an utterance involving a request:
such as "Give cookie", or "want milk". A tact is an utterance that
simply makes a statement or observation about the word. "Red ball",
"big doggie" : are statements of this sort.
Skinner says a mand is reinforced when the request it involves is
granted. A child says "want cookie", and is given a cookie. That
reinforcement makes the utterance more likely to occur again, just like
any reinforced response. A tact is reinforced by the general response
of the listener. if the child says "big doggie", her mother may say
"yes, that's right!" in a tone that the child has come to associate with
positive reinforcement.
""
I guess Skinner did put his baby in an operant chamber.
patty
Glen M. Sizemore
from Psych 1AA3 Child Development:
""
Skinner distinguishes two broad classes of multi-word utterances which
he calls mands and tacts.
GS: Wrong. Mands and tacts are not necessarily "multi-word utterances."
Indeed, they need not be utterances at all. For example, gesturing to a
waiter is a mand.
P: A mand is an utterance involving a request:
such as "Give cookie", or "want milk". A tact is an utterance that
simply makes a statement or observation about the word. "Red ball",
"big doggie" : are statements of this sort.
GS: Technically, mands are verbal responses whose probability depends
largely on specific conditions of deprivation and satiation and aversive
stimuli. Tacts are verbal responses that are under stimulus control of some
feature of the world, and are relatively free of control by specific
motivational variables.
P: Skinner says a mand is reinforced when the request it involves is
granted. A child says "want cookie", and is given a cookie. That
reinforcement makes the utterance more likely to occur again, just like
any reinforced response. A tact is reinforced by the general response
of the listener. if the child says "big doggie", her mother may say
"yes, that's right!" in a tone that the child has come to associate with
positive reinforcement.
""
GS: Basically correct.
P: I guess Skinner did put his baby in an operant chamber.
GS: Don't be an idiot - if that is at all possible.
"patty" <pat...@SPAMicyberspace.net> wrote in message
news:x_0wc.5537$Sw.3130@attbi_s51...
JXStern
wrote:
> Pat Hayes
><http://www.ihmc.us/users/user.php?UserID=42> who is a respected
>authority in this area once said something to the effect that if you
>don't want Unicorns to exist, then don't talk about them. Whenever we
>use language we invent worlds (formal semantics calls these
>interpretations) and we talk of whether a thing exists in a particular
>world or not.
I've always like Pat Hayes, thanks for this link, looks like he's
built up quite an operation down there in Florida.
If you have a reference to any published work along the line you
describe here, about inventing worlds, formal semantics calling these
interpretations, I'd appreciate a pointer.
> Luckily i do not need to answer that question because it is not my
>field of study. Should we avoid attributing these states to our AI
>mechanisms?
Eliminativism, folk theories, attribution, reification, reduction,
preserving the phenomenon, ...
Fooey.
Even if and when a full reduction is available, it's often useful to
retain folk theories and terms, like "chair" and "red" and "cat".
Attributing state to mechanism is more reductive than folk, anyway.
Clock mechanisms have state, so do computers. The only way to *avoid*
attributing state to intelligent mechanisms, artificial or real, is to
throw intelligence somehow into a different category.
It is still a useful activity to discover what reductions are
available, as they should be ultimately explanatory and predictive,
but this seldom will eliminate the convenience of higher-level terms
of the special sciences or folk theories.
I've been reading the extended introduction in a book I mentioned here
a few weeks ago, "The Structure of Scientific Theories", Frederick
Suppe, Illinois Books 1977, ISBN 0-25200634-8, Q174.S87, talking about
the "Received View" of scientific theories before and after things
changed circa 1970. This is philosophy of science, but it parallels
exactly - or perhaps dominates - the similar discussion in analytic
philosophy about the same time. It seems that Suppe is mapping this
out in excruciating detail - it's a painful read, but very worthwhile.
I suggest this book be near the top of the reading list for any AI
program with aspirations to theory, even if this is more meta-theory.
J.
patty
> On Fri, 04 Jun 2004 02:00:34 GMT, patty <pat...@SPAMicyberspace.net>
> wrote:
>
>>Pat Hayes
>><http://www.ihmc.us/users/user.php?UserID=42> who is a respected
>>authority in this area once said something to the effect that if you
>>don't want Unicorns to exist, then don't talk about them. Whenever we
>>use language we invent worlds (formal semantics calls these
>>interpretations) and we talk of whether a thing exists in a particular
>>world or not.
>
>
> I've always like Pat Hayes, thanks for this link, looks like he's
> built up quite an operation down there in Florida.
>
> If you have a reference to any published work along the line you
> describe here, about inventing worlds, formal semantics calling these
> interpretations, I'd appreciate a pointer.
>
>
Well of course there is this major work
<http://www.w3.org/TR/rdf-mt/>
Also look up "common logic"
<http://www.google.com/search?q=%22common+logic%22+%22pat+hayes%22>
Incidentally do not attribute my words about inventing worlds to Pat
Hayes, they are but my flaky paraphrase of an understanding that i
picked up in my meanderings. In fact i cannot even find where Pat said
what i said he did .. all that i can find is where other people
attributed that saying to him ... but note that he did not deny saying
it. And if you want to see what Pat has been up to of late, google
would be your best bet; he is all over the place.
patty
patty
<http://www.cartage.org.lb/en/themes/Biographies/MainBiographies/S/Skinner/1.html>
<http://www.coedu.usf.edu/cybertutorial/images/babyinbox.jpg>
... but of course you knew that.
patty
David Longley
<pat...@SPAMicyberspace.net> writes
Well, I don't know what *that* paragraph is all about for what that's
worth. If you don't understand what I said above I reckon it's further
reason that you should read the later sections of "Fragments".
>
>
>>> If you want to deal with only the world of things with time\space
>>>coordinates, with things that can be sensed and measured, then that
>>>is fine with me. There are lots of terms for just that world ... my
>>>favorite is "the natural world". From now on (unless otherwise
>>>specified by you) i will take you term "exists x" as meaning that x
>>>exists in what i call the "natural world". However, when i
>>>communicate with most other people, I think that reduction would be an inconvenience.
>> It isn't a *reduction* though. That was the point of "Two Dogmas of
>>Empiricism" and it has been central to what I (and Glen) have been
>>telling you and others here (although in Glen's case he doesn't need
>>the Quinean philosophy as he tends to refer to the empirical work
>>directly - something Quine himself advocates). When you resort to
>>intensional locutions you simply behave in ways which we know are
>>demonstrably indeterminate. Why you don't take the advice you just
>>cited Hayes as giving is just more grist to my mill.
>>
>
>Maybe it's because i like unicorns.
We all "like" putting fragments of behaviours together. It's just that a
lot of what we emit isn't useful.
>
>
>>> Whether "wants" and other mental states exist in the natural world
>>>is, as you now doubt know, a hotly debated topic on this forum.
>>>There is no doubt in my mind that these terms can be mapped to
>>>objects in some interpretation. As such they are useful tokens in a
>>>language game. Is that interpretation (world) scientific and
>>>extensional. I think not. Should we avoid these terms when trying
>>>to understand human psychology? Luckily i do not need to answer that
>>>question because it is not my field of study.
>> But it *is* in your field of application/behaviour. Like the rest
>>of us you can tell yourself whatever you like. You can "assume" that
>>you are "understood", you can "think that" you "know" what others
>>"mean". You can "attribute" and "infer", you can "intend", and "hope"
>>or "feel that" you are right, and all that time you can "believe"
>>what you "want". But the chances are, a lot of the time, you just
>>won't know what you are doing or be able to say what you are talking
>>about if you stand back and try to analyse any of it. You'll just see
>>that you are going round and round in intensional circles. When you
>>come to program, or do anything practical, all of that talk will
>>really just be dropped and you'll see you can do without it. In fact,
>>to do what you need to "effectively", you really have no choice.
>>
>
>Actually i rather "like" that paragraph; it is good writing. It made
>me stop and think about it. Certainly what goes on in usenet and in
>most of human discourse has little practical import and just goes in
>endless circles. Makes one wonder what game is being spun.
There are all sorts of reasons. I suspect a lot that post here couldn't
care less about "AI". They, like nearly half the rest of the UK and US
population seem to go to university these days, and there's a whole new
industry churning out books and other intellectual paraphernalia which
collectively plays a part in what amounts to little more than
entertainment. It's a cultural change. I guess there's no harm in it,
but when it's done at the expense of doing something which contributes
to our other welfare, I reckon we have a problem. That's partly what I'm
concerned about. Dennett, Crick, Damasio, Edelman, Fodor, Minsky..
there's no end of these celebrities, have a market and I don't begrudge
them that. I just wish they didn't make out that they were educating
people. Most of them mislead people and I've said why.
> Why shouldn't we all just *stop it* ! Let us also stop reinforcing
>our children for using intensions.
Some parents do stop bullshitting their children. They teach them not to
take that "psychology" too seriously. "it's only psychological" etc.
> Why should my child be concerned with what i am thinking; when he
>could not ever "know that" i am "thinking that" his life is heading
>toward a disaster.
There is behaviour and there is "thinking". Caring about what people do
and don't do is one thing - speculating about their "thinking" seems a
bit of a waste of time.
> And when he bips me on my head why should i be "concerned that" he
>doesn't "know that" it makes me "feel like" crying. These are all so
>"demonstrably indeterminate", let us now just excise them from our
>behavioral repertory.
You're mixing behaviour with quite inconsequential talk. Can you not see
the difference and how focussing more on behaviour and less on drivel
can actually make you more useful and helpful as a role model and
mentor?
>
>Or not! Perhaps what we do here is build the very fabric which
>inevitably is experienced as consciousness.
>
>But of course we don't need that either :(
You're being silly and vague. Try to be more discerning ;-)
>
>> Given that, what do you think you might be doing when you resort to
>>folk psychological locutions? (might you just be behaving in
>>conventional ways? Try seeing those locutions as concatenated
>>phonemes, ie as behavious - mands and tacts).
>>
>
>For those (like me) unfamiliar with "mands and tacts" here is a passage
>from Psych 1AA3 Child Development:
>
snip - the point is to see these as behaviours "give_cookie",
"big-doggie".
--
David Longley
patty
> patty <pat...@SPAMicyberspace.net> wrote:
>
>>Curt Welch wrote:
>
>
>
>>>I don't know what an ontology is so I can't use that idea to discuss
>>>this with you.
>>>
>>
>>Formal ontologies can be very useful; they can eliminate most of the
>>kinds of misunderstandings that have been happening in this train.
>>Check out the one that the IEEE is working on:
>><http://ontology.teknowledge.com/>
>
>
> IEEE is actually IEEE. surprise surprise.
>
Kurt, you doubted me :(
> I suspect they actually create and/or support the kinds of misunderstanding
> that have been happening here.
Why would you say a thing like that? Their intention is to eliminate
misunderstandings; albeit more for formal systems than for these kind of
flaky discussions. But i don't see why we couldn't use that kind of
ontology to disambiguate some of our word usage. For example i could
say that your "real world" is a collections of Physical objects
according to the following URL:
<http://128.136.11.20:8080/sigma/skb.jsp?req=skb_sr&skb=SUMO_skb&term=Physical>
Now their discussions may never transcend their political squabbling
long enough for them to settle on a ontology; but auditing their
discussions is certainly interesting. There are a lot of heavy hitters
there, Hayes (who stormed out in disgust), Sowa, etc. You can join in
the discussion and even have voting powers and have a say in determing
the standard. This is supposed to be an open project.
<http://grouper.ieee.org/groups/suo/email/thrd1.html>
patty
JXStern
wrote:
>Incidentally do not attribute my words about inventing worlds to Pat
>Hayes, they are but my flaky paraphrase of an understanding that i
>picked up in my meanderings. In fact i cannot even find where Pat said
>what i said he did .. all that i can find is where other people
>attributed that saying to him ...
The unicorns bit is good, but I like the "inventing world" phrase even
better. I don't care who said it, but would like to see some full
article or book where the idea is developed or used. Again, if you
have a reference for either or both of those, I'd appreciate it - I
gather you didn't refer me to w3 or common logic for those.
(the idea of the common logic thing, that's another topic entirely!
have you ever read Lewis Carroll's "Achilles and the Tortoise",
published in the philosophy journal Mind in 1895?)
http://www.mathacademy.com/pr/prime/articles/carroll/index.asp
(btw, I disagree entirely with the analysis given on this page, just
stick to the Carroll stuff and work the problem yourself, I suggest)
J.
Curt Welch
> Curt Welch wrote:
>
> > patty <pat...@SPAMicyberspace.net> wrote:
> >
> >>Curt Welch wrote:
> >
> >
> >
> >>>I don't know what an ontology is so I can't use that idea to discuss
> >>>this with you.
> >>>
> >>
> >>Formal ontologies can be very useful; they can eliminate most of the
> >>kinds of misunderstandings that have been happening in this train.
> >>Check out the one that the IEEE is working on:
> >><http://ontology.teknowledge.com/>
> >
> >
> > IEEE is actually IEEE. surprise surprise.
> >
>
> Kurt, you doubted me :(
I just assumed that "ontology" was a linguistic idea that wouldn't be
connected with the IEEE. So when I saw IEEE in your post I assumed it was
something else.
I've known people have worked on building knowledge databases for a long
time. I just never realized the term ontology was connected with that
concept.
> > I suspect they actually create and/or support the kinds of
> > misunderstanding that have been happening here.
>
> Why would you say a thing like that?
People don't know how the brain works. They don't understand what
consciousness is and they don't understand what the mind is. They don't
understant what they are. Beacuse of all this, they just made stuff up
when needed to fill in all the holes they couldn't otherwise explain. This
has left us with a language which is flawed from the ground up. If you
attempt to formalize a flawed langauge, you only perpetuate, and justify,
the mistakes inherent in our understanding of reality expresed by the
langauge.
On top of that, the knowledge in our heads is 100% fuzzy. It can't be
formalized in a system built on the assumption that "knowledge" is not
fuzzy. The whole approach of formalizing knoweldge is built on the false
assumption that removing the fuzziness from our langauge somehow makes it
better. It does help us to communicate facts, but it prevents us from
communicating knowledge.
The misunderstanding we have in these discussions is the basis for how we
create knowledge. If you remove the misunderstanding, you have removed the
creation of knowledge, and in the end, you have killed the very thing you
were trying to capture.
ray scanlon
> Whether "wants" and other mental states exist in the natural world is,
> as you now doubt know, a hotly debated topic on this forum. There is no
> doubt in my mind that these terms can be mapped to objects in some
> interpretation. As such they are useful tokens in a language game. Is
> that interpretation (world) scientific and extensional. I think not.
> Should we avoid these terms when trying to understand human psychology?
> Luckily i do not need to answer that question because it is not my
> field of study. Should we avoid attributing these states to our AI
> mechanisms? I don't know Curt, what do you think?
>
> patty, sitting on a fence.
Since the oblique advocation and the strenous denial of mental states
seems to be the main subject on c.a.p., whatever shall we talk about
if they are denied. We shall have the sound of one hand clapping.
If artificial intelligence means anything at all, it suggests the
contrivance of a device that shall embody a soul (mind, spirit,
essence, consciousness, awareness, intellect, mentality, self,
individuality, persona, personality, executive function, conscious
mental field, self-awareness, sentience).
Why else should McCarthy call for "artificial intelligence"? Why not
just "robotry".
Ray
John Casey
"patty" <pat...@SPAMicyberspace.net> wrote in message
> Curt Welch wrote:
[...]
>
> > "wants" are not real to me. The idea of "wants" is what is real. And
the
> > idea of "wants" exists only as language behavior which helps us to
> > understand the real things in the universe - like human behavior.
> >
>
> Whether "wants" and other mental states exist in the natural world is,
> as you now doubt know, a hotly debated topic on this forum. There is no
> doubt in my mind that these terms can be mapped to objects in some
> interpretation. As such they are useful tokens in a language game. Is
> that interpretation (world) scientific and extensional. I think not.
> Should we avoid these terms when trying to understand human psychology?
> Luckily i do not need to answer that question because it is not my
> field of study. Should we avoid attributing these states to our AI
> mechanisms? I don't know Curt, what do you think?
>
> patty, sitting on a fence.
Why would you avoid attributing these states to our AI mechanisms?
They serve us well in everyday life. Surely you would want to
communicate with your AI system with the same ease you communicate
with your fellow human being? Unless you want to be a social misfit
with autistic tendencies by treating others like wind-up toys.
Mental terms must have some functional use. Why else would we have them?
Evolution has equipped us with an intuitive psychology that enables us
to guess with varying degrees of success the intentions of others. That
is, to guess what they are likely to do next. We don't have the time to
do a radical evidential behavioural analysis using the latest statistical
techniques when we have to interact with each other in real time.
My view is that "mental states" are physical configurations in the brain
that can be triggered by the things they represent via our sensory system
(including our internal sensory system).
They may be the result of "past histories" both personal and evolutionary.
But they exist in some physical form in order to have an effect. Exactly
what form they take in a brain is up for discovery. In the case of AI we
can decide ourselves what physical form they should take.
--
John
David Longley
<JXSternC...@gte.net> writes
>On Fri, 04 Jun 2004 18:00:36 GMT, patty <pat...@SPAMicyberspace.net>
>wrote:
>>Incidentally do not attribute my words about inventing worlds to Pat
>>Hayes, they are but my flaky paraphrase of an understanding that i
>>picked up in my meanderings. In fact i cannot even find where Pat said
>>what i said he did .. all that i can find is where other people
>>attributed that saying to him ...
>
>The unicorns bit is good, but I like the "inventing world" phrase even
>better. I don't care who said it, but would like to see some full
>article or book where the idea is developed or used. Again, if you
>have a reference for either or both of those, I'd appreciate it - I
>gather you didn't refer me to w3 or common logic for those.
If you really want to entertain yourself this way, why not look into
Nelson Goodman's work? "Ways of Worldmaking" etc. He took his early cue
from Carnap's "Aufbau". Note that Carnap was never happy about having it
translated into English (although it is now available). You will find
Fodor and Dennett in what Carnap abandoned as far back as 1928!
Then there's all that possible worlds semantics/metaphysics (nonsense)
from David Lewis and chums.
--
David Longley
patty
I tend to agree on all points. Thank you for taking the time to write
them down so clearly :) Whatever AI is, it certainly will not have the
same psychology as we do. We can deal with it a bit more objectively
than we deal with our own behavior. We can clearly give it a goal. We
can assertain when it chooses its own goal. We can, as you say, decide
what physical form that goal will take. For engineering purposes, we do
not need to equivocate at all.
I just got off the fence.
patty
Glen M. Sizemore
html
http://www.coedu.usf.edu/cybertutorial/images/babyinbox.jpg
What is the purpose of these URL's? The first is incorrect - she was not
"raised" in the Air Crib, the AC was basically a climate controlled
playpen/crib. The second is merely a picture of Debbie, in the AC, published
in Ladies Home Journal (and was a PR disaster).
In any event, Skinner did not raise his daughter in a box, and did not use
the AC for experiments (though Skinner described an "experiment" where he
reinforced arm movements in one of the infant girls by turning on a lamp,
but she wasn't in the AC, he was holding her).
This sort of innuendo does nobody any good, except for those that wish to
propagate the urban myths about Debbie Buzan (nee Skinner). She was forced,
as late as this year, to write a lengthy reply to a book about famous
psychologists that had just been published where the old myth was repeated
that she sued Skinner as an adult, was institutionalized, and later
committed suicide in a bowling alley in Billings, Montana. She replied, in
part, and somewhat persuasively, that she was certain she wasn't dead, and
pretty sure that she hadn't gone crazy. She also wrote that Skinner was a
devoted and loving father, deeply missed by both she and Julie. Julie Vargas
is a psychologist in West Virginia, and Debbie is a successful artist living
in London.
"patty" <pat...@SPAMicyberspace.net> wrote in message
news:RB3wc.40822$pt3.35514@attbi_s03...
patty
Look, i don't know the facts of the matter. I found those URL's from
Google in a minutes time. If you do know the facts, why not point to an
article which presents them. This started because you called me an
idiot after i guessed (albeit from very sketchy data) that Skinner put
his baby in an operant chamber. As far as i can tell my guess was
correct. Now justify calling me an idiot. Or better yet, don't bother;
it's not important; rather just tell me what problem you may have with
the assumption below:
If we can empirically establish that a public
behavior of checking a box is correlated with
a follow up behavior that is described next to
the box; then we can assume that the person had
a disposition toward that follow up behavior.
We call that disposition an intention.
Example:
[X] I will not answer any more posts in which I am called an idiot.
The box is checked, you can observer my future usenet behavior yourself.
patty
Bjorn Reese
> admit that i flashed as such when i first read it. Thing is that an
> abstraction is not located in time and space which is the most usual
> criteria of something that is not physical. However, a compressed set
That is but one explanation. See:
http://plato.stanford.edu/entries/abstract-objects/
--
mail1dotstofanetdotdk
patty
After a moment's reflection I would like to take a stab at answering
this question the way i would have expected it to be answered by a
behaviorist:
""
Make whatever assumptions you like. However your example cannot be
taken as vindication of Dennett's Intentional Stance as making the
assumption in your example will not *improve* your abilities to predict
the follow up behavior in the least.
""
And because nothing in life is free, one might expect that such a
hypothetical behaviorist would have said something to solidify his
jollies ... perhaps something discreet like:
"Back to the drawing boards, eh patty?"
patty
Lester Zick
Patty, let me just comment that there is more to scientific progress
than is dreamed of by students of science who are taught to see the
subject in very simplistic terms of unified knowledge, successful
results and the huge impact they've had on human progress.
Like all professions scientists can be nasty, brutal, and very short
sighted in terms of infighting and partisan bickering. Most students
view science as noble profession aloof and above the fray. And it is
in terms of end product. But its practioners are often petty,devisive,
and frankly manipulative when it comes to defending their own ideas.
Regards - Lester
patty
> On Fri, 04 Jun 2004 18:00:36 GMT, patty <pat...@SPAMicyberspace.net>
> wrote:
>
>>Incidentally do not attribute my words about inventing worlds to Pat
>>Hayes, they are but my flaky paraphrase of an understanding that i
>>picked up in my meanderings. In fact i cannot even find where Pat said
>>what i said he did .. all that i can find is where other people
>>attributed that saying to him ...
>
>
> The unicorns bit is good, but I like the "inventing world" phrase even
> better. I don't care who said it, but would like to see some full
> article or book where the idea is developed or used. Again, if you
> have a reference for either or both of those, I'd appreciate it - I
> gather you didn't refer me to w3 or common logic for those.
>
Well i don't have a reference, but i still claim that this is nothing
new, and is just a standard way of thinking when doing formal semantics.
Perhaps you will find your reference somewhere in a paper on formal
semantics. The RDF Semantics document which i referenced is an example
of this ... just note how Pat defined the interpretation and how he drew
the diagram.
I think that when we try to understand what a person is saying we
frequently concoct a imaginary world which contains the assumptions
which we have attributed to the speaker and comprehend his words in
relationship to that world, rather than our own. However those of us
who do not use that strategy more frequently can make no sense
whatsoever of what others are talking about.
> (the idea of the common logic thing, that's another topic entirely!
> have you ever read Lewis Carroll's "Achilles and the Tortoise",
> published in the philosophy journal Mind in 1895?)
>
> http://www.mathacademy.com/pr/prime/articles/carroll/index.asp
>
> (btw, I disagree entirely with the analysis given on this page, just
> stick to the Carroll stuff and work the problem yourself, I suggest)
>
Well in regards to the last paragraph where he says ...
""
The formalist solution, while effective, has its own philosophical
drawbacks. Not the least of these is that, by reducing logic to
uninterpreted symbols, all semantic content is removed from the
conclusions of formal logic. In other words, what we would ordinarily
consider meaning is lost. How to restore meaning to systems of inference
while still avoiding difficulties such as Carroll’s Paradox remains a
thorny question for philosophers of mathematics.
""
I would just like to note that when the semantic content is restored to
the symbols, some of the formal assumptions (usually the law of the
excluded middle) become invalid. Which, i claim, is why classical logic
rarely works in real life.
patty
David Longley
<lester...@worldnet.att.net> writes
Scientists are ruthless when they encounter, or discover deception and
misrepresentation. That's because to scientists it's anathema. Science
is the pursuit of truth (Quine 1992).
W.V.O Quine "Pursuit of Truth", Harvard University Press, 1992.
--
David Longley
David Longley
<pat...@SPAMicyberspace.net> writes
I don't know what this "real life" is that *you* keep alluding to. From
the sound of it, I'd go as far as saying that you don't know what you're
talking about, and that you have no hope of ever telling us either if
you persist in talking this way. Our economic and general physical
well-being is critically dependent on the extensional stance. I don't
know what you mean by "classic logic" either, but if you mean modern,
post-Fregian logic, we depend on it. To a very large extent "classic"
logic runs our "real life", it certainly runs the Internet!
--
David Longley
Eray Ozkural exa
"A. G. McDowell" <mcdo...@nospam.co.uk> wrote in message news:<VuNcaIAm...@mcdowella.demon.co.uk>...
> In article <fa69ae35.04060...@posting.google.com>, Eray
> Ozkural exa <er...@bilkent.edu.tr> writes
> >Greetings,
> >
> >I'm surprised that some people take "abstract" as a vague common-sense
> >concept. To me, it has a precise technical meaning: lossy compression.
> >
> There are precise notions of abstraction in Computer Science. The one
> that suggests itself to me is from program-proving. See e.g. "the b-
> method" by steve schneider, chapter 14.
>
> He describes how an implementation T1 can be proved to implement
> (actually, refine) a more abstract machine T, so that we can take an
> argument that the T-machine satisfies some requirements and turn it into
> an argument that the T1-machine also satisfies some requirements. What
> we require is some link between the possible sequence of states of T1
> and that of T. It turns out that it is not useful to try and link the
> two with a function, going in either direction: T1 may be allowed some
> non-determinism, so that more than one possible sequence of T1-states is
> equivalent to a single sequence of T-states. The behaviour we require of
> T may be satisfied by a T-machine that has some non-determinism not
> present in T1, so it may be perfectly OK for more than one sequence of
> T-states to map to a single sequence of T1-states.
>
> (While I would accept that an injective function might be described as a
> lossy compression, I would not normally expect a compression scheme to
> sometimes map the same input to different outputs, or equivalently to
> map a single input to a set of possible outputs, which is what can
> happen here).
>
> To link the machines, we need the more general notion of a predicate,
> which takes a sequence of T1-states and a sequence of T-states and says
> whether the two are equivalent. For the predicate to allow us to show
> that T1 satisfies its requirements, there are specific conditions on it
> (proof obligations). Schneider takes a chapter to explain them, so I'm
> not going to try very hard, but basically you need to show that the
> predicate links T1 and T closely enough that any reachable sequence in
> T1 has a matching (predicate(T, T1) = true) sequence in T, and that
> since the T-sequence is a match, if it satisfies the requirements then
> the T1-sequence must also satisfy the requirements. This allows you to
> work out that a proof that T is correct must translate to a proof that
> T1 can exist and is and correct.
Thanks for this excellent post. This was precisely the kind of reply I
sought by cross-posting to comp.theory. I will meditate on your reply
and if I see the need to make any additional points, I will make them.
However, I estimate that I will probably have to remain silent in view
of your most remarkable exposition!
Cordially,
--
Eray Ozkural
phd candidate, bilkent university cs dept., ankara
JXStern
<Da...@longley.demon.co.uk> wrote:
>If you really want to entertain yourself this way, why not look into
>Nelson Goodman's work? "Ways of Worldmaking" etc. He took his early cue
>from Carnap's "Aufbau". Note that Carnap was never happy about having it
>translated into English (although it is now available). You will find
>Fodor and Dennett in what Carnap abandoned as far back as 1928!
>
>Then there's all that possible worlds semantics/metaphysics (nonsense)
>from David Lewis and chums.
David, thanks very much, these are two good suggestions. The issue is
a little bit peripheral for me, but both Goodman and "Aufbau" may at
least make good reference points. I should have Aufbau on the shelf
as the originator (?) of methodological solipsism, tho how close that
comes to Fodor's version, I need to see. Carnap is just so difficult.
Aufbau 2003 printing
http://search.barnesandnoble.com/textbooks/booksearch/isbnInquiry.asp?isbn=0812695232
Remember, Wittgenstein's TLP preceeded Aufbau by six years, yet Carnap
was still doing something akin to logical atomism and/or logicism.
And Turing 1936 has to be seen as huge progress in these matters,
well, I view it that way. Quine writing in 1953 about Carnap and not
mentioning Turing, is just amazing to me, but then, until we all had
experience with computers, say at least through the 1950s (and at
least until AI was invented as a term and a project), if not through
the 1990s, I guess it just was that easy to ignore computation as a
philosophical topic.
I share Quine's skepticism on the value of modal logics as philosophy.
David Lewis seems to have one take on such matters, but it's just a
technical twist.
I like patty's formulation better. I know there is literature in
various fields about fictional worlds we invent and use, but by far
the majority of semantic and semiotic materials, like most logicist
and philosophical ones, worry primarily about how we relate to The
Real World, and consider hypothetical ones difficult at best, aberrant
or nil at worst.
I take Wittgenstein's linguistic turn and language game ideas to be
far better matches for any kind of theories for computation and any
computational theories of mind, but the trick is not to get so lost in
the arbitrary, skeptical view of language that semantics becomes
totally impossible. Even Wittgenstein lost the reins on these matters
in his later writing, which is why everyone turned away from him - as
much the same thing happened to Quine. I still say (mounting my
soapbox and banging a drum) that Turing shows the way to rehabilitate
and complete these analytic viewpoints.
J.
patty
topic above]
patty wrote:
Ok, here is another take only this time we will require the scientist to
make a real prediction. The subject can read Chinese and the
description next to the box is written in Chinese. The cognitive
scientist can read Chinese; but the behavioral scientist cannot.
Obviously the cognitive scientist will be able to attribute the
intentional state to the subject and make an accurate prediction. The
behaviorist who has access to the same recorded behavior (the checking
of the box) will not be able to make an accurate prediction.
Does that example vindicate Dennett's stance?
patty
patty
Real life is what humans experience between birth and death. It is the
events that happen inside their "experiential bubble".
<http://home.tiscali.nl/boynalechmipo/index.htm>
> Our economic and general physical
> well-being is critically dependent on the extensional stance. I don't
> know what you mean by "classic logic" either, but if you mean modern,
> post-Fregian logic, we depend on it. To a very large extent "classic"
> logic runs our "real life", it certainly runs the Internet!
I take classical logic to be the kind of logic discussed on the page
about Carroll’s Paradox, see above. As noted on that page, the symbols
that are processed are devoid of the content that we find in real signs.
Yet when we try to apply classical logic we must still assume that a
strong identity on these symbols are possible and use that to exclude
possibilities (LEM) which actually exist in the world around us. Fact
is, this logic only works in toy domains and in academic games. It has
been hyped in our culture far in excess of any hype associated with AI.
Were it never invented by man, our science and technology would be
just as advanced as they are today.
Our world runs on physical mechanism and we depend on them. Computers,
for the most part, run on if-then statements, not the modus ponens of
classical logic. In many cases we utilize statistical analysis. These
statical heuristics do not gain their utility from anything that could
be called "extensional" using your definition of that term.
patty
Lester Zick
<Da...@longley.demon.co.uk> in comp.ai.philosophy wrote:
>In article <40c23d33...@netnews.att.net>, Lester Zick
><lester...@worldnet.att.net> writes
>>On Sat, 05 Jun 2004 20:03:53 GMT, patty <pat...@SPAMicyberspace.net>
>>in comp.ai.philosophy wrote:
[. . .]
>>Patty, let me just comment that there is more to scientific progress
>>than is dreamed of by students of science who are taught to see the
>>subject in very simplistic terms of unified knowledge, successful
>>results and the huge impact they've had on human progress.
>>
>>Like all professions scientists can be nasty, brutal, and very short
>>sighted in terms of infighting and partisan bickering. Most students
>>view science as noble profession aloof and above the fray. And it is
>>in terms of end product. But its practioners are often petty,devisive,
>>and frankly manipulative when it comes to defending their own ideas.
>>
>>Regards - Lester
>>
>
>Scientists are ruthless when they encounter, or discover deception and
>misrepresentation. That's because to scientists it's anathema. Science
>is the pursuit of truth (Quine 1992).
>
>W.V.O Quine "Pursuit of Truth", Harvard University Press, 1992.
I know what science is, David, and I know what scientists are. Only
occasionally do the twain meet and never in your instantiation.
Regards - Lester
David Longley
<JXSternC...@gte.net> writes
>On Sat, 5 Jun 2004 11:25:28 +0100, David Longley
><Da...@longley.demon.co.uk> wrote:
>>If you really want to entertain yourself this way, why not look into
>>Nelson Goodman's work? "Ways of Worldmaking" etc. He took his early cue
>>from Carnap's "Aufbau". Note that Carnap was never happy about having it
>>translated into English (although it is now available). You will find
>>Fodor and Dennett in what Carnap abandoned as far back as 1928!
>>
>>Then there's all that possible worlds semantics/metaphysics (nonsense)
>>from David Lewis and chums.
>
>David, thanks very much, these are two good suggestions. The issue is
>a little bit peripheral for me, but both Goodman and "Aufbau" may at
>least make good reference points. I should have Aufbau on the shelf
>as the originator (?) of methodological solipsism, tho how close that
>comes to Fodor's version, I need to see. Carnap is just so difficult.
Originator of methodological solipsism in the analytic tradition, but
it's really Husserl's epoche (phenomenological reduction). You'll find
Carnap talking about the "heteropsychological" (which is really
"heterophenomenology" (Dennett)). All that's happened is that a few
modern american philosophers (like Dennett) have caught up with what
Husserl and Heidegger were up to in the early decades of the 20th C. he
should have learned all that through Ryle, as that's where Ryle got his
"knowing that" and "knowing how" from in my view (Ryle was favourably
disposed to Heidegger in his early days).
Ironically (for those who think behaviourism is at odds with
phenomenology) the original Husserlian phenomenology and Heideggerian
Existential phenomenology, shares quite a lot with Skinner and even
Quine once you understand it.
>
>Aufbau 2003 printing
>http://search.barnesandnoble.com/textbooks/booksearch/isbnInquiry.asp?is
>bn=0812695232
>
>Remember, Wittgenstein's TLP preceeded Aufbau by six years, yet Carnap
>was still doing something akin to logical atomism and/or logicism.
Wittgenstein resented what Carnap was doing with his philosophy as I
understand it.
>And Turing 1936 has to be seen as huge progress in these matters,
>well, I view it that way.
I suspect Wittgenstein thought little of Turing and Goedel. I think
people don't realise just how profoundly radical/revolutionary the later
Wittgenstein (along with Skinner and Quine) actually is. What we get
instead today is all this trivial "cognititivist" rubbish. Those who lap
it up don't realise how trite it all is, and how profoundly rich the
Wittgensteinian-Skinnerian-Quinean "behaviourism/empircism" is.
> Quine writing in 1953 about Carnap and not
>mentioning Turing, is just amazing to me, but then, until we all had
>experience with computers, say at least through the 1950s (and at
>least until AI was invented as a term and a project), if not through
>the 1990s, I guess it just was that easy to ignore computation as a
>philosophical topic.
I suspect Wittgenstein would have regarded it as trivial and barren,
which (by itself) it is. The notion of "effectivity" is basic to the
Tractatus. The reason why Quine doesn't make much of Turing is because
he's taken logical positivism to completion I think. There is far more
to Quine than I think you might appreciate.
>
>I share Quine's skepticism on the value of modal logics as philosophy.
>David Lewis seems to have one take on such matters, but it's just a
>technical twist.
It looks like nonsense to me.
>
>I like patty's formulation better. I know there is literature in
>various fields about fictional worlds we invent and use, but by far
>the majority of semantic and semiotic materials, like most logicist
>and philosophical ones, worry primarily about how we relate to The
>Real World, and consider hypothetical ones difficult at best, aberrant
>or nil at worst.
>
>I take Wittgenstein's linguistic turn and language game ideas to be
>far better matches for any kind of theories for computation and any
>computational theories of mind, but the trick is not to get so lost in
>the arbitrary, skeptical view of language that semantics becomes
>totally impossible. Even Wittgenstein lost the reins on these matters
>in his later writing, which is why everyone turned away from him - as
>much the same thing happened to Quine. I still say (mounting my
>soapbox and banging a drum) that Turing shows the way to rehabilitate
>and complete these analytic viewpoints.
I suspect you may not understand the later Wittgenstein, Skinner and
Quine then. Computationalism by itself is trivial and barren. The
advances that need to be made will come from replacing the intensional
idioms with behavioural explications expressed extensionally. That is
profoundly difficult, and most people in this newsgroup haven't even got
a clue about the nature of the problem. It really doesn't matter how
many people assert that the above have
"lost the reins" etc - if just one of those who made such assertions
showed me that they understood the later Wittgenstein, Skinner or Quine,
I might listen attentively. What they tend to do is just abuse because
they *don't* understand as I have said all too often here.
If you look carefully, you'll see that that is true.
--
David Longley
David Longley
David Longley
<pat...@SPAMicyberspace.net> writes
That's all bullshit. If you can't see why, you are deluded.
--
David Longley
Eray Ozkural exa
> I take Wittgenstein's linguistic turn and language game ideas to be
> far better matches for any kind of theories for computation and any
> computational theories of mind, but the trick is not to get so lost in
> the arbitrary, skeptical view of language that semantics becomes
> totally impossible.
Right.
> Even Wittgenstein lost the reins on these matters
> in his later writing, which is why everyone turned away from him - as
> much the same thing happened to Quine. I still say (mounting my
> soapbox and banging a drum) that Turing shows the way to rehabilitate
> and complete these analytic viewpoints.
In fact, Turing (and Godel's) work on undecidability can be seen as
the rigorous explanation for some of the famous philosophical
arguments of Wittgenstein and Quine. On the other hand, I think
Wittgenstein did make some sound metaphysical observations which we
might be only beginning to discover mathematically. [By this I mean
that there is a semi-Wittgensteinean interpretation of recent results
in metamathematics, perhaps at a different level than he imagined:
language game/algorithmic independence. no rules/randomness of
mathematical knowledge. Did you notice?]
Best Regards,
--
Eray Ozkural
patty
>> various fields about fictional worlds we invent and use, but by far
>> the majority of semantic and semiotic materials, like most logicist
>> and philosophical ones, worry primarily about how we relate to The
>> Real World, and consider hypothetical ones difficult at best, aberrant
>> or nil at worst.
We have a couple of definitions of this alleged "Real World" running
around now and perhaps we should distinguish between them. I say start
with what we know - our own "experiential bubble". There really is
nothing outside of that which we can talk authoritative about - and that
includes our understandings which we get from reading and writing
language. But there is another world which is populated with those
points (events) which have matches (however approximate) with other
experiential bubbles. Until a better term comes along, let us call that
the "Cultural world". The problem with that world is that none of us
can stand there - no one of us can take that perspective and speak from
it - such a person would need to *be* that culture. Historically we
have had seen persons come forward and claim that stance: Hitler - "I am
Germany", De Gaul "I am France", Longley "If you look carefully, you'll
see that that is true." etc. Accepting one of those worlds as "The"
world is a matter of faith, or politics. Where "The World of Science"
can eliminate those assumptions which we make as a matter of faith, and
those authorities which we believe as a matter of politics; it can stand
on firm testable matching points that no one can deny.
A trans human level AI that might emerge may be in a position to step
forward and take that perspective and say to us "I am your culture". If
it were an effective Internet AI which we were forced to rely upon, many
people would believe in its authority. I think we should be careful
what we wish for.
patty
David Longley
<pat...@SPAMicyberspace.net> writes
It appears to me that you are now writing muddled metaphysical nonsense.
"Snow_is_white" is true, if snow is white.
When I have said that you will find that what I have said is true, it is
an invitation or direction for you to either look to the primary sources
to see what is written, and/or, conduct the required empirical test or
experiment, to see that what is being referred to as an empirical fact,
is in fact the case.
Glen and I have both said that within the EAB research, there is
disagreement just as there is elsewhere in empirical science. But to be
able to see *what* that disagreement is about, you need to find out what
has been done.
Logic/programming by itself is just not enough, but you don't seem to
appreciate *why* it is not enough.
--
David Longley
patty
Yes, certainly it is bullshit in your world. You could have said
nothing else.
patty
JXStern
<Da...@longley.demon.co.uk> wrote:
>Originator of methodological solipsism in the analytic tradition, but
>it's really Husserl's epoche (phenomenological reduction).
Yes, that Suppe book was making some point about phenomenology as an
influence on philosophy of science circa 1900, that Mach was a
phenomenologist, the phenomenology licensed separating observations
from any underlying theory, versus Poincare's view of theoretical
terms as some flavor of instrumentalist (p 11-12). Maybe neither view
is completely convincing on its own, but Mach spent most of his later
career denying the existence of atoms because of his ill-founded views
on reduction. You do the math.
> You'll find
>Carnap talking about the "heteropsychological" (which is really
>"heterophenomenology" (Dennett)). All that's happened is that a few
>modern american philosophers (like Dennett) have caught up with what
>Husserl and Heidegger were up to in the early decades of the 20th C. he
>should have learned all that through Ryle, as that's where Ryle got his
>"knowing that" and "knowing how" from in my view (Ryle was favourably
>disposed to Heidegger in his early days).
Ryle gives me fits.
>Ironically (for those who think behaviourism is at odds with
>phenomenology) the original Husserlian phenomenology and Heideggerian
>Existential phenomenology, shares quite a lot with Skinner and even
>Quine once you understand it.
Per above, I find it quite easy to reconcile behaviorism and
phenomenology, but if the direction this takes you in is Heidegger,
it's of no further use.
>Wittgenstein resented what Carnap was doing with his philosophy as I
>understand it.
Resented the Vienna Circle including Carnap, anyway, and with good
reason, as they restored the ontological commitments Wittgenstein
considered crucial to avoid.
>>And Turing 1936 has to be seen as huge progress in these matters,
>>well, I view it that way.
>
>I suspect Wittgenstein thought little of Turing and Goedel.
It seems Wittgenstein so hated the basis and implications of Turing's
work, that he ignored it. I guess he could do this as long as
Turing's paper looked like a mathematical exercise, but once it was
reified into desktop computers (a day neither Wittgenstein nor Turing
lived to see), well, that has to seriously change the basis for
debate! I don't know offhand what if anything Wittgenstein ever said
of Godel, but it couldn't be good, since Godel was a raving platonist,
which would be totally at odds with TLP, PI, and everything else.
> I think
>people don't realise just how profoundly radical/revolutionary the later
>Wittgenstein (along with Skinner and Quine) actually is. What we get
>instead today is all this trivial "cognititivist" rubbish. Those who lap
>it up don't realise how trite it all is, and how profoundly rich the
>Wittgensteinian-Skinnerian-Quinean "behaviourism/empircism" is.
Rich like compost, indeed.
Remember, my "cognitivism" is limited to computational AI, yours is
apparently related to criminal behavior diagnosis and modification of
actual people. I would never recommend today's cognitivism for human
behavioral purposes, nor is your recommendation of Skinnerian
behaviorism as an issue for AI anything but blatant category error.
It might be an interesting exercise to expand on just how different
those categories are.
>I suspect Wittgenstein would have regarded it as trivial and barren,
>which (by itself) it is.
No, he considered it _wrong_, because Wittgenstein wanted to stay with
skepticism about rule-following.
> The notion of "effectivity" is basic to the Tractatus.
Yes, but only as the ladder you kick away. Turing's effectivity is a
different animal. I'm sure Turing thought he was refining
Wittgenstein's work, and that Wittgenstein just didn't see it that
way!
>The reason why Quine doesn't make much of Turing is because
>he's taken logical positivism to completion I think. There is far more
>to Quine than I think you might appreciate.
Read that Suppe book and get back to me (818 pages w index and refs),
or even the first thirty or so pages, on why positivism was declared
defunct thirty to forty years ago.
Now, this is a SUBTLE point, but I agree with you, that this was, at
least, throwing out the baby with the bathwater. You're still trying
to preserve both baby and bathwater, my interests are more limited.
>>I share Quine's skepticism on the value of modal logics as philosophy.
>>David Lewis seems to have one take on such matters, but it's just a
>>technical twist.
>
>It looks like nonsense to me.
That would be a concise statement I could endorse, yes.
However, there is still some value to the subjunctive mood, and there
is value to mapping out different scenarios, but this is
methodological or mechanical or something, not metaphysical.
>I suspect you may not understand the later Wittgenstein, Skinner and
>Quine then.
Maybe if you tried explaining it, then.
That's a joke, son, Google is working fine.
J.
patty
... err in some world of interpretation. A more accurate phrasing would
read: "Snow is white" is true, if snow is white in the world of your
interpretation. If you say "snow is white" to the Eskimo, who has a
vastly greater exposure to the matter, they would probably look at you
as if you were crazy.
> When I have said that you will find that what I have said is true, it is
> an invitation or direction for you to either look to the primary sources
> to see what is written, and/or, conduct the required empirical test or
> experiment, to see that what is being referred to as an empirical fact,
> is in fact the case.
>
> Glen and I have both said that within the EAB research, there is
> disagreement just as there is elsewhere in empirical science. But to be
> able to see *what* that disagreement is about, you need to find out what
> has been done.
>
> Logic/programming by itself is just not enough, but you don't seem to
> appreciate *why* it is not enough.
It's not enough because the symbols of logical mechanisms must be bound
to actual events to become useful. The system as it stands today cannot
do that, sans help from humans. But humans are not bound by the law of
identity nor the law of the excluded middle, as are the formal systems
of logic. Hence there is a mismatch. That mismatch is problematic. We
could invent a logic that would be a better match. However the academic
world has so much vested in the current state of affairs, that would
appears to be unlikely. So what happens is that we muddle through, but
the credit should go to the flexibility of humans and not to the
brittleness of logic.
patty
David Longley
<pat...@SPAMicyberspace.net> writes
Are you telling me you're an Eskimo now?!!!! Do you not see why I am
telling you that you are writing nonsense?
>> When I have said that you will find that what I have said is true, it
>>is an invitation or direction for you to either look to the primary
>>sources to see what is written, and/or, conduct the required
>>empirical test or experiment, to see that what is being referred to
>>as an empirical fact, is in fact the case.
>> Glen and I have both said that within the EAB research, there is
>>disagreement just as there is elsewhere in empirical science. But to
>>be able to see *what* that disagreement is about, you need to find
>>out what has been done.
>> Logic/programming by itself is just not enough, but you don't seem
>>to appreciate *why* it is not enough.
>
>It's not enough because the symbols of logical mechanisms must be bound
>to actual events to become useful. The system as it stands today
>cannot do that, sans help from humans. But humans are not bound by the
>law of identity nor the law of the excluded middle, as are the formal
>systems of logic. Hence there is a mismatch.
People use intensional heuristics to describe and explain their own, and
other peoples' behaviour. They get a lot wrong in the process. Our
behaviour is, in such contexts, situation specific. We turn to the
extensional stance for rule governed behaviour. Take acting or driving a
car. We follow rules to get us into the contingencies and then the
latter tend to take over. We do the same in science - the extensional
stance is used to better predict and manage behaviour. That goes for our
own behaviour too.
> That mismatch is problematic. We could invent a logic that would be
>a better match. However the academic world has so much vested in the
>current state of affairs, that would appears to be unlikely. So what
>happens is that we muddle through, but the credit should go to the
>flexibility of humans and not to the brittleness of logic.
I really don't see what's wrong with modern logic, mathematics and
science (except the intensional bits). I think you are making veiled and
inappropriate honorific references to some imagined tacit knowledge or
"LOT". The best evidence is, I have suggested, that this is a
misunderstanding. Even our so called "intuitive" judgement (or
predictions) would seem to be no more than a poor approximation of the
extensional stance which I think is pretty much what we all mean by
science and its technology.
There's only one world Patty (like it or not, as John Martyn said ;-)
--
David Longley
http://www.longley.demon.co.uk/Frag.htm
patty
<http://search.barnesandnoble.com/booksearch/isbnInquiry.asp?isbn=0385315147>
Sure, there is only one world, David, and i take that on faith.
However, you do not know that world, nor do i. To hold forth from that
world is to talk from the God's-eye-view. In fact anything you say of
that world, claiming it is of that world, is a bold face lie. It is
more intellectually honest not to talk of such things; for nothing from
there can ever become a variable in any equation.
patty
Neil W Rickert
>In fact, Turing (and Godel's) work on undecidability can be seen as
>the rigorous explanation for some of the famous philosophical
>arguments of Wittgenstein and Quine.
I cannot find any basis for that.
David Longley
You appear to be going down the "possible worlds" of modal logic route.
If so, as a prophylactic, I suggest you look into what Quine has said
about modal logic. As good a place as any is his response to Hintikka's
contribution "Quine on Who's Who" in "The Philosophy of W.V.O Quine" in
the Schlipp series. Another place to look is in "Worlds Away" in his
book "Theories and Things".
At times, you appear to be suggesting that I don't know what the
implications of intensional opacity are ;-) You'll have me playing
computer games next!
--
David Longley
Eray Ozkural exa
I wonder what Stern thinks about it.
JXStern
wrote:
>In fact, Turing (and Godel's) work on undecidability can be seen as
>the rigorous explanation for some of the famous philosophical
>arguments of Wittgenstein and Quine.
That's the conventional view, but I'm talking about Turing's work
being an extension of the themes of instrumentalism that appear
variously in Wittgenstein, Brouwer, Hilbert, and Tarski, to name a
few. The undecidability is really not the issue, it's a matter of how
decidability works, when it works. Constructivism.
> On the other hand, I think
>Wittgenstein did make some sound metaphysical observations which we
>might be only beginning to discover mathematically. [By this I mean
>that there is a semi-Wittgensteinean interpretation of recent results
>in metamathematics, perhaps at a different level than he imagined:
>language game/algorithmic independence. no rules/randomness of
>mathematical knowledge. Did you notice?]
I'm not familiar with this, but it sounds like the same direction in
which I'm going.
J.
patty
After all my tirades against formal logic i fail to see how you could
accuse me of going down any logic route. No, what appears to be going
on is that you have trigger words, and unfortunately i tripped your
trigger that leads to the response: Quine was against modal logic.
> At times, you appear to be suggesting that I don't know what the
> implications of intensional opacity are ;-) You'll have me playing
> computer games next!
>
Well i don't see where you got that from what i wrote above. But in any
case, i would not accuse you of not knowing *some* implications of
intensional opacites. But I would accuse you of not liking them on
those grounds. Rather like avoiding a race of people based on the color
of their skin.
patty
David Longley
I wonder....perhaps you'd like to ponder on this?
"As early as 1939, Quine posed the question to which he returns over and
over again (Ways of Paradox, 1966, p.68): "How economical an ontology
can we achieve and still have a language adequate to all purposes of
science?"
From Hao Wang's (1986) "Quine's Logical Ideas in Historical
Perspective", p. 624 ch. 23 in "The Philosophy of W V Quine" in the P.
Schlipp series.
I reckon you need to find out what the answer to that is. It's stated in
the next few sentences of Wang's chapter (it's pretty clear to me that
you don't know much about Quine's work, and with your interests, you
should). I suggest that you (and Stern) read Quine's reply to Wang's
chapter as well, (Ullian's chapter may be worth a look too).
Perhaps the answer might help you to write/talk more sensibly about
"behaviourism"!
--
David Longley
JXStern
<Da...@longley.demon.co.uk> wrote:
>>> >In fact, Turing (and Godel's) work on undecidability can be seen as
>>> >the rigorous explanation for some of the famous philosophical
>>> >arguments of Wittgenstein and Quine.
>>>
>>> I cannot find any basis for that.
>>
>>I wonder what Stern thinks about it.
Aha, I saw my name, I'll respond below. Well, I responded already in
the main thread (6/7 7:25AM)(Neil's message starts a new thread on my
reader, I think he removed an extraneous space in the subject line).
>I wonder....perhaps you'd like to ponder on this?
>
>"As early as 1939, Quine posed the question to which he returns over and
>over again (Ways of Paradox, 1966, p.68): "How economical an ontology
>can we achieve and still have a language adequate to all purposes of
>science?"
I like the quotation, but it's not on page 68 of my (1976) copy, where
p 68 is the first page of "Necesary Truth", which is discussing
necessity. In fact, it doesn't sound like Quine to me, and seems to
mix categories of ontology and language - which I am all in favor of,
but am not sure that Quine ever did! I see this is Wang's
(mis)citation, but if anyone can relocate it for me, I'd appreciate
it.
> From Hao Wang's (1986) "Quine's Logical Ideas in Historical
>Perspective", p. 624 ch. 23 in "The Philosophy of W V Quine" in the P.
>Schlipp series.
Yes, that citation works for me.
>I reckon you need to find out what the answer to that is. It's stated in
>the next few sentences of Wang's chapter (it's pretty clear to me that
>you don't know much about Quine's work, and with your interests, you
>should). I suggest that you (and Stern) read Quine's reply to Wang's
>chapter as well, (Ullian's chapter may be worth a look too).
Wang's chapter is all about Quine and formal logic, and neither
chapter nor response has any bearing on any issue discussed here, that
I can see.
--
Back to Turing/Wittgenstein/Quine and undecidability, I gather that
what Eray had in mind was some variety of undecidability equating to
skepticism for Wittgenstein, insofar as that's a concise way of
explaining why his linguistic turn makes no ontological commitments,
and the somewhat similar verificationism of Quine (*not* behaviorism,
folks, if there is a difference, which I do sometimes doubt).
But I don't see undecidability that way, and even if I did would not
find it relevant to an instrumentalist theory of computation, which
for me is about just those things that are decidable.
J.
ot - lots of noise outside my window, apparently President Reagan's
body is being transported from Santa Monica to Simi Valley by way of
the freeway instead of up the coast and over the hills, which would be
shorter and more scenic, and enough news helicopters to pacify
Afghanistan (well, I count five, at least) are fighting low clouds and
fog to shadow the procession a couple of blocks past my house.
David Longley
<JXSternC...@gte.net> writes
>On Mon, 7 Jun 2004 16:11:43 +0100, David Longley
><Da...@longley.demon.co.uk> wrote:
>>>> >In fact, Turing (and Godel's) work on undecidability can be seen as
>>>> >the rigorous explanation for some of the famous philosophical
>>>> >arguments of Wittgenstein and Quine.
>>>>
>>>> I cannot find any basis for that.
>>>
>>>I wonder what Stern thinks about it.
>
>Aha, I saw my name, I'll respond below. Well, I responded already in
>the main thread (6/7 7:25AM)(Neil's message starts a new thread on my
>reader, I think he removed an extraneous space in the subject line).
>
>>I wonder....perhaps you'd like to ponder on this?
>>
>>"As early as 1939, Quine posed the question to which he returns over and
>>over again (Ways of Paradox, 1966, p.68): "How economical an ontology
>>can we achieve and still have a language adequate to all purposes of
>>science?"
>
>I like the quotation, but it's not on page 68 of my (1976) copy, where
>p 68 is the first page of "Necesary Truth", which is discussing
>necessity. In fact, it doesn't sound like Quine to me, and seems to
>mix categories of ontology and language - which I am all in favor of,
>but am not sure that Quine ever did! I see this is Wang's
>(mis)citation, but if anyone can relocate it for me, I'd appreciate
>it.
>
It's from "A Logistical Approach to the Ontological Problem" 1939, and
in the revised and enlarged edition of 1976 it's on page 201 end of the
second paragraph. It is pretty standard Quine. What you say above tells
me you don't know as much Quine as you claim. Still, that's no big deal,
we can all learn here.
>> From Hao Wang's (1986) "Quine's Logical Ideas in Historical
>>Perspective", p. 624 ch. 23 in "The Philosophy of W V Quine" in the P.
>>Schlipp series.
>
>Yes, that citation works for me.
>
>>I reckon you need to find out what the answer to that is. It's stated in
>>the next few sentences of Wang's chapter (it's pretty clear to me that
>>you don't know much about Quine's work, and with your interests, you
>>should). I suggest that you (and Stern) read Quine's reply to Wang's
>>chapter as well, (Ullian's chapter may be worth a look too).
>
>Wang's chapter is all about Quine and formal logic, and neither
>chapter nor response has any bearing on any issue discussed here, that
>I can see.
There seems to be quite a lot that's been discussed, or referenced here,
that's either passed you by (for one reason or another) or which you
don't see as having any bearing on what's discussed. That's why I said
what I said above (and don't take offence at my saying so - just quote
the first paragraph of Quine's reply to Wang. Why do you think I cited
this?).
The thread title, incidentally, is about whether we should "avoid
attributing mental states to AI mechanisms" is it not?
Here and elsewhere, Ozkural and yourself have made quite "eccentric"
remarks about Quine, Wittgenstein (and in other places Skinner and
"behaviourism"). Furthermore, I was tacitly suggesting to Ozkural
(knowing of his obsession with Kolmogorov Complexity etc) that he might
be wise to actually look into what an austere behaviourist is prepared
to accept into his ontology and why (not to mention his contribution to
programming). Ozkural's simple minded conception of "behaviourism" is
not because "behaviourism" is "simply minded", it is just an instance of
Ozkural being simple minded! What does Quine accept in his 1939 paper,
and what did he abandon later? Why did he (and Roger Gibson) describe
himself as being a behaviourist (as much as anyone in their right mind
could be)?
<snip rest of post which really doesn't make any sense>
--
David Longley
Eray Ozkural exa
> Curt Welch wrote:
> > patty <pat...@SPAMicyberspace.net> wrote:
> >
>
> >>I have no problem with saying it that way. Note your usage of the word
> >>"idea". A circle is an ideal. If you want to disallow ideals in your
> >>ontology, then that is fine with me, but i think i can function with
> >>those in my ontology and not get confused.
> >
> >
> > I don't know what an ontology is so I can't use that idea to discuss this
> > with you.
> >
>
> Formal ontologies can be very useful; they can eliminate most of the
> kinds of misunderstandings that have been happening in this train.
> Check out the one that the IEEE is working on:
> <http://ontology.teknowledge.com/>
Or they can eliminate most of the kinds of meaning that are essential
to how we communicate.
> Well there is some precedence for that. Pat Hayes
> <http://www.ihmc.us/users/user.php?UserID=42> who is a respected
> authority in this area once said something to the effect that if you
> don't want Unicorns to exist, then don't talk about them. Whenever we
> use language we invent worlds (formal semantics calls these
> interpretations) and we talk of whether a thing exists in a particular
> world or not.
Yes, we have to construct a model. But how we construct the model is
knowledge, it is not existence.
> > I describe existence based on what we can sense.
[snip]
> >
> > We can sense circular objects, but we could not sense a perfect circle.
> >
> > We learn to first describe those things which we can sense, and then use
> > that lanague to describe things which we have never sensed, and in some
> > cases, like the perfect circle, can never sense.
> >
>
> Fine.
So you agree that mathematical realism is bankrupt. Fine indeed, but
the above is probably not the complete explanation. Most definitely, I
perceive an "ideal circle" when I look at a high-resolution circle on
a computer screen. It depends on what you mean by a perfect circle and
perception. If you mean the continuous one, well, it probably does not
exist, because it requires the continuum to exist in the first place.
If you mean the discrete perfect circle, it probably does exist, and
it probably can be sensed JUST AS WELL AS ANY OTHER ABSTRACT OBJECT.
[snip]
> Whether "wants" and other mental states exist in the natural world is,
> as you now doubt know, a hotly debated topic on this forum. There is no
> doubt in my mind that these terms can be mapped to objects in some
> interpretation. As such they are useful tokens in a language game. Is
> that interpretation (world) scientific and extensional. I think not.
> Should we avoid these terms when trying to understand human psychology?
> Luckily i do not need to answer that question because it is not my
> field of study. Should we avoid attributing these states to our AI
> mechanisms? I don't know Curt, what do you think?
The answer is easy. We should not avoid attributing mental states to
our AI mechanisms any more than we attribute mental states to people.
David Longley
>David Longley wrote:
>> In article <%fMwc.14898$Sw.14753@attbi_s51>, patty
>><pat...@SPAMicyberspace.net> writes
>>
>>> David Longley wrote:
>>>
>>>> In article <bmLwc.10171$HG.7181@attbi_s53>, patty
>>>><pat...@SPAMicyberspace.net> writes
>>>>
>>>>> David Longley wrote:
>>>>>
>>>>>> In article <MiKwc.10102$HG.3436@attbi_s53>, patty
>>>>>><pat...@SPAMicyberspace.net> writes
>>>>>>
>>>>>>>> In article <k976c0h95t8dgvp58...@4ax.com>,
>>>>>>>>JXStern <JXSternC...@gte.net> writes
>>>>>>>>
>>>>>>>>> I like patty's formulation better. I know there is literature in
>>>>>>>>> various fields about fictional worlds we invent and use, but by far
>>>>>>>>> the majority of semantic and semiotic materials, like most logicist
>>>>>>>>> and philosophical ones, worry primarily about how we relate to The
>>>>>>>>> Real World, and consider hypothetical ones difficult at best,
>>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> We have a couple of definitions of this alleged "Real World"
>>>>>>>running around now and perhaps we should distinguish between
>>>>>>>them. I say start with what we know - our own "experiential
>>>>>>>bubble". There really is nothing outside of that which we can
>>>>>>>talk authoritative about - and that includes our understandings
>>>>>>>which we get from reading and writing language. But there is
>>>>>>>another world which is populated with those points (events)
>>>>>>>which have matches (however approximate) with other experiential
>>>>>>>bubbles. Until a better term comes along, let us call that the
You're now telling me you think its OK to be completely irrational as
well as kidding me that you're an Eskimo! It's amazing what holes people
will dig themselves into rather than admit they're made a mistake. We
all seem a bit too prone to that, but I thought you were one of the
smarter ones (actually, I *know* you're one of the smarter ones - so
take this as a friendly slap - I'm caring that way you know! <g>)
>
>> At times, you appear to be suggesting that I don't know what the
>>implications of intensional opacity are ;-) You'll have me playing
>>computer games next!
>>
>
>Well i don't see where you got that from what i wrote above. But in
>any case, i would not accuse you of not knowing *some* implications of
>intensional opacites. But I would accuse you of not liking them on
>those grounds. Rather like avoiding a race of people based on the
>color of their skin.
>
>patty
It has nothing to do with "liking", but it's got everything to do with
being able to make anything *reliable* or do anything reliably.
Surely that matters in science and engineering (as with everything
else)?
--
David Longley
David Longley
>our AI mechanisms any more than we attribute mental states to people.
>
Do you like it when Glen and I attribute "mental states" to you?
--
David Longley
Gary Forbis
Oh come on. You're just jealous.
dan michaels
And if you want to see what Pat has been up to of late, google
> would be your best bet; he is all over the place.
>
> patty
While testing the validity of your statement above, I ran across this
site, might be something useful:
Eray Ozkural exa
LOL
Now, we are attributing "mental states" to Longley. Maybe jealousy
wasn't quite accurate (although I suspect he could envy anyone), but
this one will work: "Longley believes that behaviorism is the correct
philosophical foundation for psychology and AI". That's also
attributing mental states, namely belief, and we all know that a state
of belief exists and how it feels like.
In my opinion, a machine could believe, too. It's hard to imagine
intelligence without belief, for no entity can ever be certain of the
world. Naturally, we'd prefer machines intelligent enough to
understand behaviorism's limitations and look at other philosophical
theories.
David Longley
Ozkural exa <er...@bilkent.edu.tr> writes
>forbi...@msn.com (Gary Forbis) wrote in message
>news:<5a1238fe.04060...@posting.google.com>...
>> David Longley <Da...@longley.demon.co.uk> wrote in message
>>news:<gnIzGlSj...@longley.demon.co.uk>...
>> > In article <fa69ae35.04060...@posting.google.com>, Eray
>> > Ozkural exa <er...@bilkent.edu.tr> writes
>> > >
>> > >The answer is easy. We should not avoid attributing mental states to
>> > >our AI mechanisms any more than we attribute mental states to people.
>> > >
>> >
>> > Do you like it when Glen and I attribute "mental states" to you?
>>
>> Oh come on. You're just jealous.
>
>LOL
>
>Now, we are attributing "mental states" to Longley. Maybe jealousy
>wasn't quite accurate (although I suspect he could envy anyone), but
>this one will work: "Longley believes that behaviorism is the correct
>philosophical foundation for psychology and AI". That's also
>attributing mental states, namely belief, and we all know that a state
>of belief exists and how it feels like.
>
>In my opinion, a machine could believe, too. It's hard to imagine
>intelligence without belief, for no entity can ever be certain of the
>world. Naturally, we'd prefer machines intelligent enough to
>understand behaviorism's limitations and look at other philosophical
>theories.
>
Read this carefully:
I have written that you have misrepresented radical behaviourism. I have
told you that you would be wise to correct this. It is irrelevant (in
this respect) whether or not I "believe that" radical or evidential
behaviourism is the correct foundation for psychology and AI. Most of
what I have said has been said to correct falsehoods and to point out
the propensity of people to both articulate, promulgate and reinforce
falsehoods and what that tells us about behaviour.
In other words, what I have been doing is considerably more subtle than
(most of) you have grasped.
--
David Longley
Victor Nazarov
> patty <pat...@SPAMicyberspace.net> wrote:
>
>>Curt Welch wrote:
> The words exist. But like "perfect circle" the thing they imply exists
> doesn't exist. What they are used for however is to understand and describe
> behavior, which does exist.
>
> When I tie a weight to the end of a rope and spin it in a circle, I'm using
> the idea of a "perfect circle" to help undertand the path of the weight at
> the end of the rope. The weight does not move in a perfect circle, but
> yet, it's LIKE a perfect circle - meaning it shares features in common with
> this thing called a perfect circle which doesn't actually exist - only the
> description of it exists. The idea of a perfect circle is useful as a tool
> for understanding other real things in the world. If we study the
> properties of a perfect circle, we can gain understanding about the real
> things in the world which are LIKE a perfect circle.
>
> The idea of "wants" is like the idea of a perfect circle. Wants are not
> real, but the the idea is useful to help us understand real things - our
> behavior.
I think you can't name any thing that really exist. You can touch the
will and the touch will exist at that moment not later or sooner. If we
are talking about the human mind then no space or time exist. Future has
never existed and will not exist in the future. All we have is our
fillings. The problem of words is that words refer to nothing. They are
closed in your mind and can't escape. The one word in one language can
be assosiated with different thoughts fillings among different people.
The words of even one language don't map exactly to each other betwean
heads of two different peaple. The only magic of words is that we can
use it and they work! You can't change this situation using words in
this conference.
--
vir
Eray Ozkural exa
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1> >er...@bilkent.edu.tr (Eray Ozkural exa) wrote in message news:<fa69ae35.04060...@posting.google.com>...
> >> Greetings,
>
> >> I'm surprised that some people take "abstract" as a vague common-sense
> >> concept. To me, it has a precise technical meaning: lossy compression.
>
> >> A program is abstract, because it *loses* the architectural details of
> >> a computation, and it is concise. A blueprint of a house is abstract
> >> because it *loses* the architectural and material properties of an
> >> actual house, and it is concise.
>
> The trouble is that you have confused the technical term "abstract"
> from object oriented programming, with the ordinary word "abstract".
No, I haven't. The abstraction property of classes discussed in OOP
literature is hardly abstract.
> >Maybe, I should state my question in a more mathematical way.
>
> >Suppose that we have a meta-mathematical theory of theorem proving. In
> >this system, there are entities like proofs, which are derivations in
> >a formal axiomatic system. Now, suppose that some of the proofs we
> >would like to observe are practically impossible to understand due to
> >their excessive length. [+] We would like to construct a complementary
> >system that will summarize these proofs in another language [*],
> >giving us proof sketches, which will be comprehensible to
> >mathematicians. What is the best way to formalize this process of
> >abstraction?
>
> It's a process of summarization, not of abstraction. And you probably
> cannot formalize it.
Yes, my argument was that summarization is identical to abstraction.
If we cannot formalize it, then how can minds use abstraction? By
means other than mechanical?
eagleso...@yahoo.com
> Greetings,
>
> I'm surprised that some people take "abstract" as a vague common-sense
> concept. To me, it has a precise technical meaning: lossy compression.
>
> A program is abstract, because it *loses* the architectural details of
> a computation, and it is concise. A blueprint of a house is abstract
> because it *loses* the architectural and material properties of an
> actual house, and it is concise.
>
> counter-factuals in its meaning. An "abstract" entity represents
> another entity in a purposeful way, it is a sign that points to
> another object or sign.
>
> Maybe we should all study semiotics instead of Platonist computer
> science and mathematics! Maybe that is how one truly becomes a
> hard-core materialist!
>
> Regards,
>
> PS: Dan, have I gotten materialist enough now?
The term abstract is indeed an exactly defined one. It means that
which is generalized. And it is a form of thought always.
An exact abstract form is called a relation. And these are either
subjective or objective abstractions. And here the meaning of
scienctific abstract versus subjective abstract becomes another topic.