Hi, I had this "insight" last night and and would like to read about
it, since I presume that it is not original.
The reason for redundancy in natural languages is that it is a
distinctive trait of a particular set of signals. Considering the first
encounter between two beings, the first problem that arises in
establishing a communication is the fact that advertising the intention
to communicate is an act of communication in itself and we get a kind
of hen and egg paradox. Now, redundancy is a way of sending a 'meta
signal' because redundancy is inefficient, and inefficiency is a
strident signal in a world where there is evolutionary pressure towards
efficiency. Therefore a set of redundant signals is more apt to be
noticed and thus stands out against the background noise.
Can anyone point to a reference or maybe share their own knowledge in
this respect?

"apc" <andres.paniagua@web.de> wrote in message 
news:1169465432.274445.165260@l53g2000cwa.googlegroups.com...
> Hi, I had this "insight" last night and and would like to read about
> it, since I presume that it is not original.
> The reason for redundancy in natural languages is that it is a
> distinctive trait of a particular set of signals. Considering the first
> encounter between two beings, the first problem that arises in
> establishing a communication is the fact that advertising the intention
> to communicate is an act of communication in itself and we get a kind
> of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> signal' because redundancy is inefficient, and inefficiency is a
> strident signal in a world where there is evolutionary pressure towards
> efficiency. Therefore a set of redundant signals is more apt to be
> noticed and thus stands out against the background noise.
> Can anyone point to a reference or maybe share their own knowledge in
> this respect?
>

I suspect it is more likely this:
language is very unstable, word pronuciation can vary to such a degree that 
after centuries or millenia, the language is no longer even recognizable;
natural language structure and semantics are essentially non-existant, and 
to what extent they do exist, they are often ad-hoc patterns and past 
convention.

I suspect language can't technically get much less verbose, because if it 
did, it would not be sufficiently flexible to represent all that much of 
anything (and can barely even do so as it is).


why does a typical english speaker not readily understand, say, german, 
spanish, or french? english has much similarity in terms of grammar and 
basic vocabulary with these languages.

yet your typical english speaker has no clue, even with such obvious words 
as: blut, haus, engel, und, hel, im, ...


synthetic example (sorry, not a true speaker):
und im mein Haus uber die Strasse ich sehen einen Mann und zu ihm "welcomm" 
ich sprache.

and, transcribed, this is roughly:
and in my house over the street I see a man and to him, "welcome" I say.

and note the similarity...


and I am still an american/native english speaker, never really bothering to 
actually learn any of these languages, yet even as such, these things I can 
recognize...

me remembering recently seeing part of a movie in dutch (I think), and 
noting that a good portion of the movie was basically about the same as what 
they would be saying in english (yes, the wording is different, and the 
pronunciation is different, yet why doesn't anyone seem to notice?...).

I guess it is inconvinient to read or listen to languages that are different 
than ones' own, but I suspect that many people manage to become exceedingly 
inflexible when it comes to these matters.

maybe if they would relax, and watch the movie with the subtitles off, they 
could start to understand.

much like anime (different in that for the most part, japanese is very 
different from english). watch enough for enough years, and one can turn off 
the subtitles and still have a basic idea what is going on, absent actually 
investing any real effort in learning the language (of course, the next step 
would be motivating oneself to give up the assistance of the subtitles).

of course, if they bother to learn, they can understand it that much better 
that much faster, but one finds that they are too lazy, and personally don't 
really care...


this is not to say the same rule necissarily applies to writing ability 
though. much like a programmer can understand a wide variety of programming 
languages, it is difficult to write in much more than the few they happen to 
know...

apc wrote:
> Hi, I had this "insight" last night and and would like to read about
> it, since I presume that it is not original.
> The reason for redundancy in natural languages is that it is a
> distinctive trait of a particular set of signals. Considering the first
> encounter between two beings, the first problem that arises in
> establishing a communication is the fact that advertising the intention
> to communicate is an act of communication in itself and we get a kind
> of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> signal' because redundancy is inefficient, and inefficiency is a
> strident signal in a world where there is evolutionary pressure towards
> efficiency. Therefore a set of redundant signals is more apt to be
> noticed and thus stands out against the background noise.
> Can anyone point to a reference or maybe share their own knowledge in
> this respect?
> 
It is hard for me to see what your conclusion related to redundancy in 
natural language has to do with the provided way of reasoning.
I suppose the reason for it in the way of understanding what does the 
word "redundancy" in context of natural language actually mean.
What in my eyes is worth to be called a deep insight is the discovered 
"chicken and egg" dilemma. I suggest that resolving this one (going as 
deep in the cause-effect chain as possible) will put also more light on 
subject of the redundancy, where you in your way of reasoning are maybe 
mixing redundancy of meaning (required for understanding) considering 
e.g. all words used to build a sentence and the redundancy of 
sound/text-font (required to pass a physical signal) in the 
spoken/written language.

Could it be, that your actual insight you would like to share here has a 
much more general nature (you haven't maybe got aware about yet) and the 
natural language came in only by chance as an example just because you 
considered it as you have got for the first time the enlightenment that

   # In every beginning of anything there must be a repetition #

?


Consider:
#) repetitions at the very beginning == obligatory precondition for 
knowledge, understanding, awareness (of an idea)
#) subsequent repetitions (considered as not obligatory necessary to 
convey the idea) == redundancy

REPETITION is in my eyes the magic intersection at which the basic ideas 
and concepts behind compression meet all of the other branches of 
science and knowledge (like for example linguistics).

Claudio Grondi

apc <andres.paniagua@web.de> wrote:

> Hi, I had this "insight" last night and and would like to read about
> it, since I presume that it is not original.
> The reason for redundancy in natural languages is that it is a
> distinctive trait of a particular set of signals.

Transmission of language suffers errors.  Redundancy allows these
errors to be corrected by the listener.

Here's some observations:

* Speakers who have an unusual (e.g., foreign) accent are easier
  to understand when they speak in sentences than when they pronounce
  individual words.

* People routinely mishear names (doubly so unfamiliar or unusual
  ones), because these are essentially random symbols that aren't
  limited by what makes sense in the context of the rest of a
  sentence or the general situational context of the conversation.

Listeners do not properly understand all phonemes and apply all
sorts of cross checks (valid grammar, known words, idiomatics, ...)
to extract the correct meaning from the jumbled sounds they hear.

-- 
Christian "naddy" Weisgerber                          naddy@mips.inka.de

cr88192 wrote:
> "apc" <andres.paniagua@web.de> wrote in message
> news:1169465432.274445.165260@l53g2000cwa.googlegroups.com...
> > Hi, I had this "insight" last night and and would like to read about
> > it, since I presume that it is not original.
> > The reason for redundancy in natural languages is that it is a
> > distinctive trait of a particular set of signals. Considering the first
> > encounter between two beings, the first problem that arises in
> > establishing a communication is the fact that advertising the intention
> > to communicate is an act of communication in itself and we get a kind
> > of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> > signal' because redundancy is inefficient, and inefficiency is a
> > strident signal in a world where there is evolutionary pressure towards
> > efficiency. Therefore a set of redundant signals is more apt to be
> > noticed and thus stands out against the background noise.
> > Can anyone point to a reference or maybe share their own knowledge in
> > this respect?
> >
>
> I suspect it is more likely this:
> language is very unstable, word pronuciation can vary to such a degree that
> after centuries or millenia, the language is no longer even recognizable;
> natural language structure and semantics are essentially non-existant, and
> to what extent they do exist, they are often ad-hoc patterns and past
> convention.
>
> I suspect language can't technically get much less verbose, because if it
> did, it would not be sufficiently flexible to represent all that much of
> anything (and can barely even do so as it is).
>
>
> why does a typical english speaker not readily understand, say, german,
> spanish, or french? english has much similarity in terms of grammar and
> basic vocabulary with these languages.
>
> yet your typical english speaker has no clue, even with such obvious words
> as: blut, haus, engel, und, hel, im, ...
>
>
> synthetic example (sorry, not a true speaker):
> und im mein Haus uber die Strasse ich sehen einen Mann und zu ihm "welcomm"
> ich sprache.
>
> and, transcribed, this is roughly:
> and in my house over the street I see a man and to him, "welcome" I say.
>
> and note the similarity...
>
>
> and I am still an american/native english speaker, never really bothering to
> actually learn any of these languages, yet even as such, these things I can
> recognize...
>
> me remembering recently seeing part of a movie in dutch (I think), and
> noting that a good portion of the movie was basically about the same as what
> they would be saying in english (yes, the wording is different, and the
> pronunciation is different, yet why doesn't anyone seem to notice?...).
>
> I guess it is inconvinient to read or listen to languages that are different
> than ones' own, but I suspect that many people manage to become exceedingly
> inflexible when it comes to these matters.
>
> maybe if they would relax, and watch the movie with the subtitles off, they
> could start to understand.
>
> much like anime (different in that for the most part, japanese is very
> different from english). watch enough for enough years, and one can turn off
> the subtitles and still have a basic idea what is going on, absent actually
> investing any real effort in learning the language (of course, the next step
> would be motivating oneself to give up the assistance of the subtitles).
>
> of course, if they bother to learn, they can understand it that much better
> that much faster, but one finds that they are too lazy, and personally don't
> really care...
>
>
> this is not to say the same rule necissarily applies to writing ability
> though. much like a programmer can understand a wide variety of programming
> languages, it is difficult to write in much more than the few they happen to
> know...
Hello, thank you very much for your reply. I think I was too concise in
my original article, therefore I will clarify a bit. When I spoke of
natural languages, I didn't mean only human spoken or written
languages, but all the languages that appear naturally among creatures
in nature; I should have stated that explicitly. I think your reply is
more relevant to the question of why human languages remain redundant
over time. Also I think that you are not considering all levels of
redundancy in natural language, at least you spoke mainly of  grammatic
redundancy and its immediate semantical consequences. Anyway, I am not
saying that the reason I give is the only reason for redundancy in
language, but one among many. Its special advantage being that of
helping to stress the intention of establishing communication. Let me
give you an example with the colour of flowers. The colour expresses
two things on two corresponding levels. The first and more basic one is
'pay attention this object is different from the background'. The
second, which is more in the nature of language is a message encoded by
the colour and intended for a specific recipient like 'I am your
favorite kind of flower'. The first kind of expression is what I was
thinking of as being accomplished by redundancy as a form of
inefficiency.

Claudio Grondi wrote:
> apc wrote:
> > Hi, I had this "insight" last night and and would like to read about
> > it, since I presume that it is not original.
> > The reason for redundancy in natural languages is that it is a
> > distinctive trait of a particular set of signals. Considering the first
> > encounter between two beings, the first problem that arises in
> > establishing a communication is the fact that advertising the intention
> > to communicate is an act of communication in itself and we get a kind
> > of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> > signal' because redundancy is inefficient, and inefficiency is a
> > strident signal in a world where there is evolutionary pressure towards
> > efficiency. Therefore a set of redundant signals is more apt to be
> > noticed and thus stands out against the background noise.
> > Can anyone point to a reference or maybe share their own knowledge in
> > this respect?
> >
> It is hard for me to see what your conclusion related to redundancy in
> natural language has to do with the provided way of reasoning.
> I suppose the reason for it in the way of understanding what does the
> word "redundancy" in context of natural language actually mean.
> What in my eyes is worth to be called a deep insight is the discovered
> "chicken and egg" dilemma. I suggest that resolving this one (going as
> deep in the cause-effect chain as possible) will put also more light on
> subject of the redundancy, where you in your way of reasoning are maybe
> mixing redundancy of meaning (required for understanding) considering
> e.g. all words used to build a sentence and the redundancy of
> sound/text-font (required to pass a physical signal) in the
> spoken/written language.
>
> Could it be, that your actual insight you would like to share here has a
> much more general nature (you haven't maybe got aware about yet) and the
> natural language came in only by chance as an example just because you
> considered it as you have got for the first time the enlightenment that

Hello there, thanks for your thoughts. What you are saying here, even
if it is the case (I must think about it yet), is not what I intended
to present as a newly gained insight. What I meant was the following:
Inefficiency works as a strong signal because it stands clearly out in
a world where there is high evolutional  pressure towards efficiency.
Redundancy is a form or a cause of inefficiency and is a characteristic
feature of natural language because it serves as a signal that stresses
the intention of communication.

>    # In every beginning of anything there must be a repetition #
>
> ?
>
>
> Consider:
> #) repetitions at the very beginning =3D=3D obligatory precondition for
> knowledge, understanding, awareness (of an idea)
> #) subsequent repetitions (considered as not obligatory necessary to
> convey the idea) =3D=3D redundancy
I am not very sure I agree with this. Once certain formalities have
been taken care of, repetitions are unnecessary even at the beginning.
Take for example mathematics. If you agree beforehand to the use of a
formal language, all you have to do is lay out the foundation in the
form of a list of axioms and from it develop everything else, without
ever having to recur to repetitions.

Andr=E9s
> REPETITION is in my eyes the magic intersection at which the basic ideas
> and concepts behind compression meet all of the other branches of
> science and knowledge (like for example linguistics).
>=20
> Claudio Grondi

Christian Weisgerber wrote:
> apc <andres.paniagua@web.de> wrote:
>
> > Hi, I had this "insight" last night and and would like to read about
> > it, since I presume that it is not original.
> > The reason for redundancy in natural languages is that it is a
> > distinctive trait of a particular set of signals.
>
> Transmission of language suffers errors.  Redundancy allows these
> errors to be corrected by the listener.
>
> Here's some observations:
>
> * Speakers who have an unusual (e.g., foreign) accent are easier
>   to understand when they speak in sentences than when they pronounce
>   individual words.
>
> * People routinely mishear names (doubly so unfamiliar or unusual
>   ones), because these are essentially random symbols that aren't
>   limited by what makes sense in the context of the rest of a
>   sentence or the general situational context of the conversation.
>
> Listeners do not properly understand all phonemes and apply all
> sorts of cross checks (valid grammar, known words, idiomatics, ...)
> to extract the correct meaning from the jumbled sounds they hear.
Hi, I already was aware of that fact. Robustness against transmission
errors is an advantage of redundant languages. However the question is
whether it is merely a consequence of natural languages being redundant
or a reason for their being redundant. Take the written form of human
languages for example. They are redundant partly because they are
directly derived from speech. But also they are redundant because of
the rules of orthography. This redundancy also derives in robustness
against transmission errors, but in this case this advantage is
definitely only consequential and not a determining cause. So, there
are certainly several causes for redundancy in natural language but the
question is, which of them were favored by evolution and which of them
only appeared as the consequences of other factors.
Andr=E9s
> --
> Christian "naddy" Weisgerber                          naddy@mips.inka.de

> Hi, I had this "insight" last night and and would like to read about
> it, since I presume that it is not original.
> The reason for redundancy in natural languages is that it is a
> distinctive trait of a particular set of signals. Considering the first
> encounter between two beings, the first problem that arises in
> establishing a communication is the fact that advertising the intention
> to communicate is an act of communication in itself and we get a kind
> of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> signal' because redundancy is inefficient, and inefficiency is a
> strident signal in a world where there is evolutionary pressure towards
> efficiency. Therefore a set of redundant signals is more apt to be
> noticed and thus stands out against the background noise.
> Can anyone point to a reference or maybe share their own knowledge in
> this respect?

 There is a discrepancy regarding de domains of 'natural language' and
the domain where you see redundancy. Within the natural language there
_does not exists_ redundancy in meaning. Everything means something
under the rules of the used language.
 In another place you take as example a flower and it's color.
 I'm going to take that as an example to explain the point by making
some (obviously non-real) assumptions:

 - there is only a set of three different flowers in nature
 - these flowers have all different colors red, green and blue
 - these flowers have all different shades near black, moderate and
very shiny

 as statistican you immediatly tend to say that this 'language'
(color+shade) contains redundancy.
 But that is wrong, as a a bio-scientist you immediatly understand that
the wide domain of language-possabilities leads automatically to
potential of evolutional advantages. You simply can't quit the color
and rely on the shade, if you have to expect that another fourth flower
can develop and compete with it in a way that proposes extinction or a
terrible price of adaption.
 Of course this is also true if a fourth type of flowers develops
'olor' in adition, but a mutation changing color is very much more
probable and very much faster achievable than smell.
 [What's going on with the flowers and their language is basically the
same that's going on in predictive compressors, they try to protect
themself from paying too high for a rapid change, but do not take care
so much of slow changes, because they may adapt accordingly.]

 So within the language there is _no redundancy_, because there is an
_all meaning_.

 This is the same for human language for example. Repetition for
example is not redundancy, it has a meaning; if you got good education
you got teached the rules of retoric, or learned pedagogic.
 In wichever way you express, the exact way you do express is
meaningfull. You may repeat because you want to hint importance, or
you're drunk, or you stutter, or you're unconcentrated, ...

 So within the language there is _no redundancy_, because there is an
_all meaning_. (I repeat and it has meaning and significance.)

 The reason we can compress languages (be it human or artificial) is
that compressors transcode them. They invent on the fly another
language that is capable of expressing the same (stationary) "things",
but shorter under the quality 'bits'.
 For lossless compressors there exists a 1:1 mapping of the rules in
the source language to the rules in the destination language, but it
introduces a complete different formalism. Imagine a human that would
be able to speak english-zip (which is not a fantastic impossability),
his output is allready formulated in a way of an 'optimal language'.
 But the cost to detect which of the tokens of his message are more
relevant than other is extremly high, and the english-zip message is
less error-prone, as well as not random-accessible. So compressors are
basically removing 'cost of features' from a language - compressors are
cost-optimizers, mostly with the cost-constrain being amount of bits;
most of the things/features that are important for human evolution are
unimportant for the compressor and vice versa.

 So compressors invent languages (formalisms) that transcendent the
_all meaning_ into another formalism, but generally without quiting
_all meaning_.
 Insofar I believe that the use(fullness) of the word 'redundancy' is
extremly limited, especially as synonym for 'quiting' something.
 
 Have a nice day
 	Niels

"apc" <andres.paniagua@web.de> wrote in message 
news:1169482539.298409.28510@v45g2000cwv.googlegroups.com...
> cr88192 wrote:
>> "apc" <andres.paniagua@web.de> wrote in message
>> news:1169465432.274445.165260@l53g2000cwa.googlegroups.com...
<snip>
>>
>> this is not to say the same rule necissarily applies to writing ability
>> though. much like a programmer can understand a wide variety of 
>> programming
>> languages, it is difficult to write in much more than the few they happen 
>> to
>> know...
> Hello, thank you very much for your reply. I think I was too concise in
> my original article, therefore I will clarify a bit. When I spoke of
> natural languages, I didn't mean only human spoken or written
> languages, but all the languages that appear naturally among creatures
> in nature; I should have stated that explicitly. I think your reply is
> more relevant to the question of why human languages remain redundant
> over time. Also I think that you are not considering all levels of
> redundancy in natural language, at least you spoke mainly of  grammatic
> redundancy and its immediate semantical consequences. Anyway, I am not
> saying that the reason I give is the only reason for redundancy in
> language, but one among many. Its special advantage being that of
> helping to stress the intention of establishing communication. Let me
> give you an example with the colour of flowers. The colour expresses
> two things on two corresponding levels. The first and more basic one is
> 'pay attention this object is different from the background'. The
> second, which is more in the nature of language is a message encoded by
> the colour and intended for a specific recipient like 'I am your
> favorite kind of flower'. The first kind of expression is what I was
> thinking of as being accomplished by redundancy as a form of
> inefficiency.
>

I was ambiguous I guess as well.

I guess my definition of redundancy is different.


I define redundancy more in terms of a long, drawn out, and repeating 
pattern (say, if whenever people greeted each other, they would go into a 
long drawn out, and largely pre-scripted set of monologues).

as a usual occurance, a short but obvious action is used (such as a hand 
wave, or making a noise). this is not redundant (unless a person, say, goes 
into a long repeating sequence of alternating tones).

as such, I say, the used used for getting attention is not redundant, 
rather, very much the inverse.


maybe a flower is 'redundant', if only because it does not change its color 
(or operate like an organic display device) to present information to other 
creatures.

however, I don't view it having a different color as a form of redundancy, 
in-fact, it containing a different color is a form of entropy.


however, note that even with said 'redundancy' in communication, there are 
many things in natural language that also suggest a form of compression, in 
particular:
notice how frequently used words or phrases are often truncated or otherwise 
deformed into a shorter form (I forget the term right now).

as such, one person will speak of decor, another of a prog lang, another of 
trogs, trolls, furries, and lollies, being emo, and so on...


under my definitions, word complexity and semantic meanings are what 
constitute redundancy.

even our communication system (speech) is fairly concise and information 
dense for an organic lifeform (we can, after all, transfer a range of ideas 
and information in only a matter of seconds, or maybe a few minutes).

I don't expect the redundancy can be reduced that much further absent a loss 
of flexibility.


for example, what if humans had instead started mapping numerical values to 
particular syllables, and when speaking would perform a number of 
calculations to represent a phrase or conversation in some minimum number of 
syllables, and where misinterpreting even a single syllable would render the 
whole conversation unintelligable, that or it would quickly descend into a 
state of chaos?...

or, more simply, what if we had some kind of complicated flagging system, 
where for example, single-sylable markers would be used which would define 
much of the grammar for a phrase, and where the various syllables had some 
number of overlapping namespaces and a 1 to 1 mapping to particular 
concepts.

for example, say we have 19 consonants and 7 distinctive vowel sounds 
(excluding schwa):
with both CV and CVC syllables, we have 2660 unique syllables.
nearly every concept in the language could be mappable to a single syllable 
(note the range would be expanded via a 'gramatic role', which could be, for 
example, defined by a single syllable at the start of the phrase).

such languages could encode an idea in far fewer syllables, but would be 
highly prone to error.


note that in true human languages, only even a small number of the possible 
syllables have any particular meaning, many not even occuring as a part of 
most words.

as such, we can readily import, and recognize, loanwords from other 
languages, which only rarely happen to sound like another word in the 
listeners native language (o-haio being a well-known example).


thus, I suspect, human languages do approach a sane minimum level of 
redundancy.

niels.froehling@seies.de wrote:
> > Hi, I had this "insight" last night and and would like to read about
> > it, since I presume that it is not original.
> > The reason for redundancy in natural languages is that it is a
> > distinctive trait of a particular set of signals. Considering the first
> > encounter between two beings, the first problem that arises in
> > establishing a communication is the fact that advertising the intention
> > to communicate is an act of communication in itself and we get a kind
> > of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> > signal' because redundancy is inefficient, and inefficiency is a
> > strident signal in a world where there is evolutionary pressure towards
> > efficiency. Therefore a set of redundant signals is more apt to be
> > noticed and thus stands out against the background noise.
> > Can anyone point to a reference or maybe share their own knowledge in
> > this respect?
>
>  There is a discrepancy regarding de domains of 'natural language' and
> the domain where you see redundancy. Within the natural language there
> _does not exists_ redundancy in meaning. Everything means something
> under the rules of the used language.
>  In another place you take as example a flower and it's color.
>  I'm going to take that as an example to explain the point by making
> some (obviously non-real) assumptions:
>
>  - there is only a set of three different flowers in nature
>  - these flowers have all different colors red, green and blue
>  - these flowers have all different shades near black, moderate and
> very shiny
>
>  as statistican you immediatly tend to say that this 'language'
> (color+shade) contains redundancy.
>  But that is wrong, as a a bio-scientist you immediatly understand that
> the wide domain of language-possabilities leads automatically to
> potential of evolutional advantages. You simply can't quit the color
> and rely on the shade, if you have to expect that another fourth flower
> can develop and compete with it in a way that proposes extinction or a
> terrible price of adaption.
>  Of course this is also true if a fourth type of flowers develops
> 'olor' in adition, but a mutation changing color is very much more
> probable and very much faster achievable than smell.
>  [What's going on with the flowers and their language is basically the
> same that's going on in predictive compressors, they try to protect
> themself from paying too high for a rapid change, but do not take care
> so much of slow changes, because they may adapt accordingly.]
>
>  So within the language there is _no redundancy_, because there is an
> _all meaning_.
>
>  This is the same for human language for example. Repetition for
> example is not redundancy, it has a meaning; if you got good education
> you got teached the rules of retoric, or learned pedagogic.
>  In wichever way you express, the exact way you do express is
> meaningfull. You may repeat because you want to hint importance, or
> you're drunk, or you stutter, or you're unconcentrated, ...
>
>  So within the language there is _no redundancy_, because there is an
> _all meaning_. (I repeat and it has meaning and significance.)
>
>  The reason we can compress languages (be it human or artificial) is
> that compressors transcode them. They invent on the fly another
> language that is capable of expressing the same (stationary) "things",
> but shorter under the quality 'bits'.
>  For lossless compressors there exists a 1:1 mapping of the rules in
> the source language to the rules in the destination language, but it
> introduces a complete different formalism. Imagine a human that would
> be able to speak english-zip (which is not a fantastic impossability),
> his output is allready formulated in a way of an 'optimal language'.
>  But the cost to detect which of the tokens of his message are more
> relevant than other is extremly high, and the english-zip message is
> less error-prone, as well as not random-accessible. So compressors are
> basically removing 'cost of features' from a language - compressors are
> cost-optimizers, mostly with the cost-constrain being amount of bits;
> most of the things/features that are important for human evolution are
> unimportant for the compressor and vice versa.
>
>  So compressors invent languages (formalisms) that transcendent the
> _all meaning_ into another formalism, but generally without quiting
> _all meaning_.
>  Insofar I believe that the use(fullness) of the word 'redundancy' is
> extremly limited, especially as synonym for 'quiting' something.

I don't know what "comp.compression" is, but it seems to be where the
original poster is coming from. From the point of view of human
language (the only proper topic for sci.lang), Niels seems not to
realize what is meant by "redundancy" in language. Here's an example:
In English, every voiceless stop consonant (p, t, k) is aspirated
(followed by a puff of air) almost every time -- as in pill, till,
kill. But after s, there is no aspiration after the consonant -- as in
spill, still, skill. Speakers of English are not even aware of the
aspiration or the lack of it. But the different quality of "h" noise
after each of the three consonants helps the ear detect and the brain
process which consonant is being said. And of course if someone puts or
doesn't put the aspiration in the wrong place, they have a "foreign
accent" -- and the English-speaker may be able to _imitate_ that accent
but almost certainly doesn't know how the imitation works. That's a
redundant phonetic feature that aids in comprehension and helps make up
for possible noise in the speech stream.

Another kind of redundancy in language is when German (for instance)
marks gender, case, and number on the article and all the adjectives
that go with a noun. You might not hear one of the endings, but another
one will clue you in. That's a redundant grammatical feature.

And so on. The leading American linguist Charles F. Hockett estimated
that "redundancy in human language hovers around 50 percent."

> Another kind of redundancy in language is when German (for instance)
> marks gender, case, and number on the article and all the adjectives
> that go with a noun. You might not hear one of the endings, but another
> one will clue you in. That's a redundant grammatical feature.

 I am german anyway but this doesn't restrict to german: the addition
of gender-postfixes and even gender-changes of prefixing articles
(which probably is unique for german) is not redundant per se.
Sometimes it is the only source of information about which previous
subject is referenced, especially in complicated sentences. So you
can't mark a subject by '<a_specific_article> boat' and be able to
determine which specific article is meant. There for sure is a
statistical distribution of how much neutral/female/male specific
articles occur in front of a male subject, but under the definition of
the german language I would not say that application of _the meaning_
of the article contains redundancies.
 [adjectives don't change gender in german, in fact they don't have
gender, and no quantity, there are only qualities applied to them:
rich, richer, richest]

 Maybe simply I can't express good enough why I would not apply de term
'redundancy' to the meanings within a language, to let other jump of
their math-statistics-grammar-syntax-thinking. I do not have whatsoever
problem to apply the term 'redundancy' to mathematical problems,
compression is one anyway.

 Ciao
	Niels

"Peter T. Daniels" <grammatim@verizon.net> wrote in message 
news:1169506570.500245.206060@a75g2000cwd.googlegroups.com...
>
<snip>

>
> I don't know what "comp.compression" is, but it seems to be where the
> original poster is coming from. From the point of view of human
> language (the only proper topic for sci.lang), Niels seems not to
> realize what is meant by "redundancy" in language.

comp.compression is a newgroup primarily talking about data compression (for 
example, huffman vs arithmetic coding, lz77 vs bwt, gzip, bzip2, zip, jpeg, 
png, and the like).

as such, the OP is OT for comp.compression as well. since I had figured the 
topic was human languages, that is mostly where I had directed my emphasis.

in my case, I had assumed redundancy in natural language to refer to the 
total number of syllables/pronouncable diphones/... needed to express an 
idea, so a language that requires more syllables would be rated as "more 
redundant", or at least more verbose, than one which requires less...

niels.froehling@seies.de wrote:
> > Another kind of redundancy in language is when German (for instance)
> > marks gender, case, and number on the article and all the adjectives
> > that go with a noun. You might not hear one of the endings, but another
> > one will clue you in. That's a redundant grammatical feature.
>
>  I am german anyway but this doesn't restrict to german: the addition
> of gender-postfixes and even gender-changes of prefixing articles
> (which probably is unique for german) is not redundant per se.

See what I mean? You use redundant expressions every day without being
aware of it -- even though redundancy in another realm is (apparently)
part of your daily work.

> Sometimes it is the only source of information about which previous
> subject is referenced, especially in complicated sentences. So you
> can't mark a subject by '<a_specific_article> boat' and be able to
> determine which specific article is meant. There for sure is a
> statistical distribution of how much neutral/female/male specific
> articles occur in front of a male subject, but under the definition of
> the german language I would not say that application of _the meaning_
> of the article contains redundancies.
>  [adjectives don't change gender in german, in fact they don't have
> gender, and no quantity, there are only qualities applied to them:
> rich, richer, richest]

(I think you'd better go back and study your Duden!)

In Die schoene blaue Donau, "feminine" and "singular" and "nominative"
are marked four times over, once on each word. If German had less
redundancy, it could use "Der schoen blau Donau," just like English
"the beautiful blue Danube."

English-speakers don't think we're missiing a thing by not marking
gender or case, or by not marking number anywhere but on the noun (and
the indefinite article, if you want to take "some" as the plural of
"a(n)").

>  Maybe simply I can't express good enough why I would not apply de term
> 'redundancy' to the meanings within a language, to let other jump of
> their math-statistics-grammar-syntax-thinking. I do not have whatsoever
> problem to apply the term 'redundancy' to mathematical problems,
> compression is one anyway.

Here at sci.lang, a number of posters would understand if you wrote in
German.

cr88192 wrote:
> "Peter T. Daniels" <grammatim@verizon.net> wrote in message
> news:1169506570.500245.206060@a75g2000cwd.googlegroups.com...
> >
> <snip>
>
> >
> > I don't know what "comp.compression" is, but it seems to be where the
> > original poster is coming from. From the point of view of human
> > language (the only proper topic for sci.lang), Niels seems not to
> > realize what is meant by "redundancy" in language.
>
> comp.compression is a newgroup primarily talking about data compression (for
> example, huffman vs arithmetic coding, lz77 vs bwt, gzip, bzip2, zip, jpeg,
> png, and the like).
>
> as such, the OP is OT for comp.compression as well. since I had figured the
> topic was human languages, that is mostly where I had directed my emphasis.
>
> in my case, I had assumed redundancy in natural language to refer to the
> total number of syllables/pronouncable diphones/... needed to express an
> idea, so a language that requires more syllables would be rated as "more
> redundant", or at least more verbose, than one which requires less...

I considered your initial reply perfectly adequate, so I didn't say
anything myself. Only when all other sorts of communication were
brought in did an alarm go off. They're not the subject matter of
linguistics -- they're studied in "semiotics" (or "semiology").

"Diphone," though, isn't a technical term in linguistics -- I don't
know what you're referring to.

The number of syllables in a word (not that "word" can be defined
universally for all languages!) depends in part on the number of
phonemes in the language, and on how complicated the language allows
syllables to be. In Japanese, the only consonant that can end a
syllable is n (and it doesn't even count as an entire consonant).
English can have very complicated syllables, like "strengths," so
English can have shorter words, all else being equal. Georgian (of the
Caucasus) can have something like six consonants in a row at the start
of a word (don't ask me for an example -- but if you happen to run into
someone from Tbilisi, ask them to say "orange peel").

Redundancy in language is more a matter of function than of form. In
languages of India, for instance, aspiration is as important as
anything else that distinguishes two consonants, and an English-speaker
has to make great efforts to learn to hear it and to pronounce it when
needed and not when not.

"Peter T. Daniels" <grammatim@verizon.net> wrote in message 
news:1169511023.822301.21640@q2g2000cwa.googlegroups.com...
>
> cr88192 wrote:
>> "Peter T. Daniels" <grammatim@verizon.net> wrote in message
>> news:1169506570.500245.206060@a75g2000cwd.googlegroups.com...
>> >
>> <snip>
>>
>> >
>> > I don't know what "comp.compression" is, but it seems to be where the
>> > original poster is coming from. From the point of view of human
>> > language (the only proper topic for sci.lang), Niels seems not to
>> > realize what is meant by "redundancy" in language.
>>
>> comp.compression is a newgroup primarily talking about data compression 
>> (for
>> example, huffman vs arithmetic coding, lz77 vs bwt, gzip, bzip2, zip, 
>> jpeg,
>> png, and the like).
>>
>> as such, the OP is OT for comp.compression as well. since I had figured 
>> the
>> topic was human languages, that is mostly where I had directed my 
>> emphasis.
>>
>> in my case, I had assumed redundancy in natural language to refer to the
>> total number of syllables/pronouncable diphones/... needed to express an
>> idea, so a language that requires more syllables would be rated as "more
>> redundant", or at least more verbose, than one which requires less...
>
> I considered your initial reply perfectly adequate, so I didn't say
> anything myself. Only when all other sorts of communication were
> brought in did an alarm go off. They're not the subject matter of
> linguistics -- they're studied in "semiotics" (or "semiology").
>

ok.


> "Diphone," though, isn't a technical term in linguistics -- I don't
> know what you're referring to.
>

http://en.wikipedia.org/wiki/Diphone


> The number of syllables in a word (not that "word" can be defined
> universally for all languages!) depends in part on the number of
> phonemes in the language, and on how complicated the language allows
> syllables to be. In Japanese, the only consonant that can end a
> syllable is n (and it doesn't even count as an entire consonant).
> English can have very complicated syllables, like "strengths," so
> English can have shorter words, all else being equal. Georgian (of the
> Caucasus) can have something like six consonants in a row at the start
> of a word (don't ask me for an example -- but if you happen to run into
> someone from Tbilisi, ask them to say "orange peel").
>

that is why I mentioned diphones...

I had assumed a diphone to be a more accurate measure than a syllable, as 
for syllables one has to define average syllable complexity, ... wheras a 
phone or diphone should be more at least a slightly more accurate measure.

> Redundancy in language is more a matter of function than of form. In
> languages of India, for instance, aspiration is as important as
> anything else that distinguishes two consonants, and an English-speaker
> has to make great efforts to learn to hear it and to pronounce it when
> needed and not when not.
>

yes.

Peter T. Daniels <grammatim@verizon.net> wrote in message
news:1169506570.500245.206060@a75g2000cwd.googlegroups.com...
> niels.froehling@seies.de wrote:
> > > Hi, I had this "insight" last night and and would like to read about
> > > it, since I presume that it is not original.
> > > The reason for redundancy in natural languages is that it is a
> > > distinctive trait of a particular set of signals. Considering the first
> > > encounter between two beings, the first problem that arises in
> > > establishing a communication is the fact that advertising the intention
> > > to communicate is an act of communication in itself and we get a kind
> > > of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> > > signal' because redundancy is inefficient, and inefficiency is a
> > > strident signal in a world where there is evolutionary pressure towards
> > > efficiency. Therefore a set of redundant signals is more apt to be
> > > noticed and thus stands out against the background noise.
> > > Can anyone point to a reference or maybe share their own knowledge in
> > > this respect?
> >
> >  There is a discrepancy regarding de domains of 'natural language' and
> > the domain where you see redundancy. Within the natural language there
> > _does not exists_ redundancy in meaning. Everything means something
> > under the rules of the used language.
> >  In another place you take as example a flower and it's color.
> >  I'm going to take that as an example to explain the point by making
> > some (obviously non-real) assumptions:
> >
> >  - there is only a set of three different flowers in nature
> >  - these flowers have all different colors red, green and blue
> >  - these flowers have all different shades near black, moderate and
> > very shiny
> >
> >  as statistican you immediatly tend to say that this 'language'
> > (color+shade) contains redundancy.
> >  But that is wrong, as a a bio-scientist you immediatly understand that
> > the wide domain of language-possabilities leads automatically to
> > potential of evolutional advantages. You simply can't quit the color
> > and rely on the shade, if you have to expect that another fourth flower
> > can develop and compete with it in a way that proposes extinction or a
> > terrible price of adaption.
> >  Of course this is also true if a fourth type of flowers develops
> > 'olor' in adition, but a mutation changing color is very much more
> > probable and very much faster achievable than smell.
> >  [What's going on with the flowers and their language is basically the
> > same that's going on in predictive compressors, they try to protect
> > themself from paying too high for a rapid change, but do not take care
> > so much of slow changes, because they may adapt accordingly.]
> >
> >  So within the language there is _no redundancy_, because there is an
> > _all meaning_.
> >
> >  This is the same for human language for example. Repetition for
> > example is not redundancy, it has a meaning; if you got good education
> > you got teached the rules of retoric, or learned pedagogic.
> >  In wichever way you express, the exact way you do express is
> > meaningfull. You may repeat because you want to hint importance, or
> > you're drunk, or you stutter, or you're unconcentrated, ...
> >
> >  So within the language there is _no redundancy_, because there is an
> > _all meaning_. (I repeat and it has meaning and significance.)
> >
> >  The reason we can compress languages (be it human or artificial) is
> > that compressors transcode them. They invent on the fly another
> > language that is capable of expressing the same (stationary) "things",
> > but shorter under the quality 'bits'.
> >  For lossless compressors there exists a 1:1 mapping of the rules in
> > the source language to the rules in the destination language, but it
> > introduces a complete different formalism. Imagine a human that would
> > be able to speak english-zip (which is not a fantastic impossability),
> > his output is allready formulated in a way of an 'optimal language'.
> >  But the cost to detect which of the tokens of his message are more
> > relevant than other is extremly high, and the english-zip message is
> > less error-prone, as well as not random-accessible. So compressors are
> > basically removing 'cost of features' from a language - compressors are
> > cost-optimizers, mostly with the cost-constrain being amount of bits;
> > most of the things/features that are important for human evolution are
> > unimportant for the compressor and vice versa.
> >
> >  So compressors invent languages (formalisms) that transcendent the
> > _all meaning_ into another formalism, but generally without quiting
> > _all meaning_.
> >  Insofar I believe that the use(fullness) of the word 'redundancy' is
> > extremly limited, especially as synonym for 'quiting' something.
>
> I don't know what "comp.compression" is, but it seems to be where the
> original poster is coming from. From the point of view of human
> language (the only proper topic for sci.lang), Niels seems not to
> realize what is meant by "redundancy" in language. Here's an example:
> In English, every voiceless stop consonant (p, t, k) is aspirated
> (followed by a puff of air) almost every time -- as in pill, till,
> kill. But after s, there is no aspiration after the consonant -- as in
> spill, still, skill. Speakers of English are not even aware of the
> aspiration or the lack of it. But the different quality of "h" noise
> after each of the three consonants helps the ear detect and the brain
> process which consonant is being said. And of course if someone puts or
> doesn't put the aspiration in the wrong place, they have a "foreign
> accent" -- and the English-speaker may be able to _imitate_ that accent
> but almost certainly doesn't know how the imitation works. That's a
> redundant phonetic feature that aids in comprehension and helps make up
> for possible noise in the speech stream.
>
> Another kind of redundancy in language is when German (for instance)
> marks gender, case, and number on the article and all the adjectives
> that go with a noun. You might not hear one of the endings, but another
> one will clue you in. That's a redundant grammatical feature.
>
> And so on. The leading American linguist Charles F. Hockett estimated
> that "redundancy in human language hovers around 50 percent."

Would you know if he based his estimate on some (other
people's) calculations? Would his redundancy have some
relationship with language entropy as calculated by some
(theory of information) people?

pjk

Peter T. Daniels wrote:
> cr88192 wrote:
> > "Peter T. Daniels" <grammatim@verizon.net> wrote in message
> > news:1169506570.500245.206060@a75g2000cwd.googlegroups.com...
> > >
> > <snip>
> >
> > >
> > > I don't know what "comp.compression" is, but it seems to be where the
> > > original poster is coming from. From the point of view of human
> > > language (the only proper topic for sci.lang), Niels seems not to
> > > realize what is meant by "redundancy" in language.
> >
> > comp.compression is a newgroup primarily talking about data compression=
 (for
> > example, huffman vs arithmetic coding, lz77 vs bwt, gzip, bzip2, zip, j=
peg,
> > png, and the like).
> >
> > as such, the OP is OT for comp.compression as well. since I had figured=
 the
> > topic was human languages, that is mostly where I had directed my empha=
sis.
> >
> > in my case, I had assumed redundancy in natural language to refer to the
> > total number of syllables/pronouncable diphones/... needed to express an
> > idea, so a language that requires more syllables would be rated as "more
> > redundant", or at least more verbose, than one which requires less...
>
> I considered your initial reply perfectly adequate, so I didn't say
> anything myself. Only when all other sorts of communication were
> brought in did an alarm go off. They're not the subject matter of
> linguistics -- they're studied in "semiotics" (or "semiology").

I guess the objection that my article is out of topic is partially true
for both groups. But I wouldn't know where else to ask for references.
Participants of both groups are familiar with the phenomenon of
redundancy in language and I think it is a legitimate expectation that
they should have come across the question of how or why natural
language (and natural communications) acquired their redundancy. And I
think the answer to that question would give the areas of study of both
groups valuable insights.
Andr=E9s

> 
> (I think you'd better go back and study your Duden!)
> 
> In Die schoene blaue Donau, "feminine" and "singular" and "nominative"
> are marked four times over, once on each word. If German had less
> redundancy, it could use "Der schoen blau Donau," just like English
> "the beautiful blue Danube."
> 
> English-speakers don't think we're missiing a thing by not marking
> gender or case, or by not marking number anywhere but on the noun (and
> the indefinite article, if you want to take "some" as the plural of
> "a(n)").

At the risk of making myself look ignorant, this English speaker thinks 
we *are* missing something — the clarity you get with a morphology that 
clearly illustrates the relationship between parts of speech by marking 
case, gender, and number. Of course, that clarity can (must?) be made 
up for in other ways (e.g. more rigid word order). Speaking as one who 
hasn't deeply studied the issue, I suppose that in language, 
"redundancy" works out to the same thing as "clarity" and 
"flexibility", and thinking about it that way takes the focus off the 
(possibly) pejorative connotation of the former term.

A brief example, and possibly not the best one but the first I thought of:
la chère maison (dear house)
la maison chère (expensive house)

Speakers of article-less languages might say "la" is redundant, but it 
tells me a noun's coming up — quite useful indeed, especially if it's 
the first time I've heard that word (not likely here, but still…). And 
since "chère" agrees with the gender of "maison", that tells me it's an 
adjective instead of a noun. While English would indicate the same 
adjective-noun relationship via strict word order, French is free to 
implement both word orders and have the word assume two meanings.

Paul

Paul D wrote:

> At the risk of making myself look ignorant, this English speaker thinks
> we *are* missing something - the clarity you get with a morphology that
> clearly illustrates the relationship between parts of speech by marking
> case, gender, and number.

Give us an example of something you can say in French that you can't
say in English.

> Of course, that clarity can (must?) be made
> up for in other ways (e.g. more rigid word order). Speaking as one who
> hasn't deeply studied the issue, I suppose that in language,
> "redundancy" works out to the same thing as "clarity" and
> "flexibility", and thinking about it that way takes the focus off the
> (possibly) pejorative connotation of the former term.
>
> A brief example, and possibly not the best one but the first I thought of:
> la ch=E8re maison (dear house)
> la maison ch=E8re (expensive house)
>
> Speakers of article-less languages might say "la" is redundant, but it

No, no one suggests that marking definiteness is "redundant." But
marking "feminine" twice is "redundant" at least once over. Marking
definiteness is not, however, _necessary_ for communication; neither
Latin nor Russian does so.

> tells me a noun's coming up - quite useful indeed, especially if it's
> the first time I've heard that word (not likely here, but still...). And
> since "ch=E8re" agrees with the gender of "maison", that tells me it's an
> adjective instead of a noun.

No, it doesn't; but the fact that it occurs immediately after a noun,
or, unusually, between the article and the noun, _does_ tell you it's
an adjective.

French nouns can be marked for sex where appropriate (though sometimes
in French, gender and sex do not agree).

> While English would indicate the same
> adjective-noun relationship via strict word order, French is free to
> implement both word orders and have the word assume two meanings.

Not exactly a major feature of French adjectives!

And of course an adjective can follow its noun in English for special
effect: "I sing the body electric."

cr88192 wrote:
> "Peter T. Daniels" <grammatim@verizon.net> wrote in message
> news:1169511023.822301.21640@q2g2000cwa.googlegroups.com...
> >
> > cr88192 wrote:
> >> "Peter T. Daniels" <grammatim@verizon.net> wrote in message
> >> news:1169506570.500245.206060@a75g2000cwd.googlegroups.com...
> >> >
> >> <snip>
> >>
> >> >
> >> > I don't know what "comp.compression" is, but it seems to be where the
> >> > original poster is coming from. From the point of view of human
> >> > language (the only proper topic for sci.lang), Niels seems not to
> >> > realize what is meant by "redundancy" in language.
> >>
> >> comp.compression is a newgroup primarily talking about data compression
> >> (for
> >> example, huffman vs arithmetic coding, lz77 vs bwt, gzip, bzip2, zip,
> >> jpeg,
> >> png, and the like).
> >>
> >> as such, the OP is OT for comp.compression as well. since I had figured
> >> the
> >> topic was human languages, that is mostly where I had directed my
> >> emphasis.
> >>
> >> in my case, I had assumed redundancy in natural language to refer to the
> >> total number of syllables/pronouncable diphones/... needed to express an
> >> idea, so a language that requires more syllables would be rated as "more
> >> redundant", or at least more verbose, than one which requires less...
> >
> > I considered your initial reply perfectly adequate, so I didn't say
> > anything myself. Only when all other sorts of communication were
> > brought in did an alarm go off. They're not the subject matter of
> > linguistics -- they're studied in "semiotics" (or "semiology").
> >
>
> ok.
>
>
> > "Diphone," though, isn't a technical term in linguistics -- I don't
> > know what you're referring to.
> >
>
> http://en.wikipedia.org/wiki/Diphone

Ah, yes. A gimmick used in speech synthesis that corresponds to no
linguistic reality at all.

> > The number of syllables in a word (not that "word" can be defined
> > universally for all languages!) depends in part on the number of
> > phonemes in the language, and on how complicated the language allows
> > syllables to be. In Japanese, the only consonant that can end a
> > syllable is n (and it doesn't even count as an entire consonant).
> > English can have very complicated syllables, like "strengths," so
> > English can have shorter words, all else being equal. Georgian (of the
> > Caucasus) can have something like six consonants in a row at the start
> > of a word (don't ask me for an example -- but if you happen to run into
> > someone from Tbilisi, ask them to say "orange peel").
> >
>
> that is why I mentioned diphones...
>
> I had assumed a diphone to be a more accurate measure than a syllable, as
> for syllables one has to define average syllable complexity, ... wheras a
> phone or diphone should be more at least a slightly more accurate measure.

"Measure" of what? What does it (as the jargon was in the mid-70s) "buy
you" to introduce a new level of analysis?

> > Redundancy in language is more a matter of function than of form. In
> > languages of India, for instance, aspiration is as important as
> > anything else that distinguishes two consonants, and an English-speaker
> > has to make great efforts to learn to hear it and to pronounce it when
> > needed and not when not.
> >
> 
> yes.

Paul J Kriha wrote:
> Peter T. Daniels <grammatim@verizon.net> wrote in message

> > And so on. The leading American linguist Charles F. Hockett estimated
> > that "redundancy in human language hovers around 50 percent."
>
> Would you know if he based his estimate on some (other
> people's) calculations? Would his redundancy have some
> relationship with language entropy as calculated by some
> (theory of information) people?

I _suspect_ that it was a gur feeling based on working with dozens of
diverse languages over a lifetime (remember in those days we subscribed
to Martin Joos's formulation, "Languages can differ from each other
without limit and in upredictable ways"). However, he did write the
monograph on Mathematical Linguistics in Current Trends in Linguistics
3 (1966), and his response to Chomsky, State of the Art (1968), meets
him on his own formalistic ground.

On 2007-01-23 22:23:46 +0900, "Peter T. Daniels" <grammatim@verizon.net> said:

> 
> Paul D wrote:
> 
>> At the risk of making myself look ignorant, this English speaker thinks
>> we *are* missing something - the clarity you get with a morphology that
>> clearly illustrates the relationship between parts of speech by marking
>> case, gender, and number.
> 
> Give us an example of something you can say in French that you can't
> say in English.

I wouldn't claim there's any such thing; just that French has slightly 
more elegant morphology than English. (Subjective, I know.)

>> 
>> A brief example, and possibly not the best one but the first I thought of:
>> la chère maison (dear house)
>> la maison chère (expensive house)
>> 
>> Speakers of article-less languages might say "la" is redundant, but it
> 
> No, no one suggests that marking definiteness is "redundant." But
> marking "feminine" twice is "redundant" at least once over. Marking
> definiteness is not, however, _necessary_ for communication; neither
> Latin nor Russian does so.

Wouldn't a Latin speaker or Russian then be tempted to declare articles 
to be redundant? Or at least superfluous?

> 
>> tells me a noun's coming up - quite useful indeed, especially if it's
>> the first time I've heard that word (not likely here, but still...). And
>> since "chère" agrees with the gender of "maison", that tells me it's an
>> adjective instead of a noun.
> 
> No, it doesn't; but the fact that it occurs immediately after a noun,
> or, unusually, between the article and the noun, _does_ tell you it's
> an adjective.

Well, just to make my point clear, I'm suggesting that if the words 
"chère maison" were otherwise ambiguous, the e on the "chère" would 
mean that that word must be the adjective describing "maison", and not 
the other way around. The gender declension is communicating 
information about the word's part of speech. By what standard is 
clarifying a word's part-of-speech a redundancy?

> 
> French nouns can be marked for sex where appropriate (though sometimes
> in French, gender and sex do not agree).
> 
>> While English would indicate the same
>> adjective-noun relationship via strict word order, French is free to
>> implement both word orders and have the word assume two meanings.
> 
> Not exactly a major feature of French adjectives!

No, but frequent enough.

> 
> And of course an adjective can follow its noun in English for special
> effect: "I sing the body electric."

Only if the context makes it quite clear, though. With many English 
words doing double duty as nouns and adjectives, the situation could 
get confusing and you'd have to rely on rigid word order again.

Paul

On 23 Jan 2007 05:23:46 -0800, "Peter T. Daniels"
<grammatim@verizon.net> wrote in
<news:1169558626.820276.19230@m58g2000cwm.googlegroups.com>
in comp.compression,sci.lang:

[...]

> And of course an adjective can follow its noun in English for special
> effect: "I sing the body electric."

Or for historical reasons: court(s) martial.

Brian

On Tue, 23 Jan 2007 23:34:50 +0900, Paul D
<paul@hiddenfortress.ten> wrote in
<news:2007012323345075249-paul@hiddenfortressten> in
comp.compression,sci.lang:

> On 2007-01-23 22:23:46 +0900, "Peter T. Daniels" <grammatim@verizon.net> said:

>> Paul D wrote:

[...]

>>> A brief example, and possibly not the best one but the first I thought of:
>>> la ch�re maison (dear house)
>>> la maison ch�re (expensive house)
 
>>> Speakers of article-less languages might say "la" is redundant, but it

>> No, no one suggests that marking definiteness is "redundant." But
>> marking "feminine" twice is "redundant" at least once over. Marking
>> definiteness is not, however, _necessary_ for communication; neither
>> Latin nor Russian does so.

> Wouldn't a Latin speaker or Russian then be tempted to declare articles 
> to be redundant? Or at least superfluous?

The latter, but hardly the former.  Note that the two are
quite different.

>>> tells me a noun's coming up - quite useful indeed, especially if it's
>>> the first time I've heard that word (not likely here, but still...). And
>>> since "ch�re" agrees with the gender of "maison", that tells me it's an
>>> adjective instead of a noun.

>> No, it doesn't; but the fact that it occurs immediately after a noun,
>> or, unusually, between the article and the noun, _does_ tell you it's
>> an adjective.

> Well, just to make my point clear, I'm suggesting that if the words 
> "ch�re maison" were otherwise ambiguous, the e on the "ch�re" would 
> mean that that word must be the adjective describing "maison", and not 
> the other way around. 

Doesn't follow, since there are masculine nouns that end in
<-e> (e.g., <murmure>).

> The gender declension is communicating  information about
> the word's part of speech. 

'Declension' refers to the inflection of verbs.

> By what standard is clarifying a word's part-of-speech a
> redundancy?

First, that isn't happening here.  Secondly, no one
suggested that marking a word's part of speech was a
redundancy; marking it more than once, however, is.

[...]

Brian

Peter T. Daniels wrote:
> Marking definiteness is not, however, _necessary_ for communication;

If the definite article were removed from English, how would one
distinctly communicate these two?
1) I am Carpenter.
2) I am the carpenter.

> neither Latin nor Russian does so.

Brian M. Scott wrote:


>
> 'Declension' refers to the inflection of verbs.
> 

Are you sure about that?

John

"bulkington63" <john_66044@yahoo.com> wrote in message 
news:1169576965.885689.150240@j27g2000cwj.googlegroups.com...
>
> Brian M. Scott wrote:
>
>
>>
>> 'Declension' refers to the inflection of verbs.
>>
>
> Are you sure about that?
>
> John

Don't know if he's sure about that, but I'm sure he's wrong.
Nouns are declined.
Verbs are conjugated. 

On 23 Jan 2007 10:29:26 -0800, bulkington63
<john_66044@yahoo.com> wrote in
<news:1169576965.885689.150240@j27g2000cwj.googlegroups.com>
in comp.compression,sci.lang:

> Brian M. Scott wrote:

>> 'Declension' refers to the inflection of verbs.

> Are you sure about that?

Jesus.  Did I write that?!

Brian

On 23 Jan 2007 10:10:25 -0800, "ranjit_mathews@yahoo.com"
<ranjit_mathews@yahoo.com> wrote in
<news:1169575825.597235.151870@m58g2000cwm.googlegroups.com>
in comp.compression,sci.lang:

> Peter T. Daniels wrote:

>> Marking definiteness is not, however, _necessary_ for communication;

> If the definite article were removed from English, how would one
> distinctly communicate these two?
> 1) I am Carpenter.
> 2) I am the carpenter.

My name is 'Carpenter'.  I do carpentry / My job is
carpentry / My position here is carpenter / ... .

>> neither Latin nor Russian does so.

Brian

> >  I am german anyway but this doesn't restrict to german: the addition
> > of gender-postfixes and even gender-changes of prefixing articles
> > (which probably is unique for german) is not redundant per se.
> See what I mean? You use redundant expressions every day without being
> aware of it -- even though redundancy in another realm is (apparently)
> part of your daily work.

 Under what condition do we use redundant expressions?

 Is it still redundant if you quit half the message?
 Is it still redundant if you analyse the speak-style of somebody?
 Is it still redundant if you check the english-exercises of your
students?
 Is it still redundant if you translate from english to german?
 Is it still redundant if you make an analysis of word-distribution in
books?

> > Sometimes it is the only source of information about which previous
> > subject is referenced, especially in complicated sentences. So you
> > can't mark a subject by '<a_specific_article> boat' and be able to
> > determine which specific article is meant. There for sure is a
> > statistical distribution of how much neutral/female/male specific
> > articles occur in front of a male subject, but under the definition of
> > the german language I would not say that application of _the meaning_
> > of the article contains redundancies.
> >  [adjectives don't change gender in german, in fact they don't have
> > gender, and no quantity, there are only qualities applied to them:
> > rich, richer, richest]
>
> (I think you'd better go back and study your Duden!)

 I'm sorry, I was not thinking long enough about it. I had only these
fragments in mind:

 Der Mann ist reich.
 Diese Frau dort ist reich.
 Das Kind ist noch nicht reich an Erfahrung.

> In Die schoene blaue Donau, "feminine" and "singular" and "nominative"
> are marked four times over, once on each word. If German had less
> redundancy, it could use "Der schoen blau Donau," just like English
> "the beautiful blue Danube."

 Der schoenen blauen Donau		 , geht es nicht gut.
 Der schoenen blauen Donau		 wegen, bin ich hierher gezogen.
 Die schoene blaue Donau		 , ist verschmutzt.
 Die schoen-blau gefaerbte Donau , ist der laengste Fluss Europas (oder
die Wolga?).
 Die schoene blau gefaerbte Donau, macht viele Boegen.
 ...

 I think it's not very reasonable to define a strict relationship of
what's happened synchron to 'schoen' and to 'blau', and it doesn't
directly inherit the gender of 'Donau', only indirect through the rules
of accusativ/dativ/nominativ/genetiv.
 These loose/indirect relationships within grammar isn't my point.
 It's about meaning, you can't preserve meaning (in it's absolute
sense, not in a specific self- and well-defined environment) while on
the other side you remove meaning.
 You can say a language is redundant in compairison with a different
language expressing the exact same message, but shorter message size.
 You can't say that a language is redundant in itself, because every
formulation of the same message with different tokens: is another
message! It may contain the same information (under a very specific
point of view), but it has another identity.

> English-speakers don't think we're missiing a thing by not marking
> gender or case, or by not marking number anywhere but on the noun (and
> the indefinite article, if you want to take "some" as the plural of
> "a(n)").

 I can sing whale-songs if I'm a whale and don't miss a bit what those
humans have.

> Here at sci.lang, a number of posters would understand if you wrote in
> German.

 Okay, so I did a bit. :-)
 
 Ciao
 	Niels

On 23 Jan 2007 11:12:54 -0800, <niels.froehling@seies.de>
wrote in
<news:1169579574.025793.49320@m58g2000cwm.googlegroups.com>
in comp.compression,sci.lang:

>>>  I am german anyway but this doesn't restrict to german: the addition
>>> of gender-postfixes and even gender-changes of prefixing articles
>>> (which probably is unique for german) is not redundant per se.

>> See what I mean? You use redundant expressions every day without being
>> aware of it -- even though redundancy in another realm is (apparently)
>> part of your daily work.

>  Under what condition do we use redundant expressions?

Virtually every time we say something.

[...]

>>> Sometimes it is the only source of information about which previous
>>> subject is referenced, especially in complicated sentences. So you
>>> can't mark a subject by '<a_specific_article> boat' and be able to
>>> determine which specific article is meant. There for sure is a
>>> statistical distribution of how much neutral/female/male specific
>>> articles occur in front of a male subject, but under the definition of
>>> the german language I would not say that application of _the meaning_
>>> of the article contains redundancies.
>>>  [adjectives don't change gender in german, in fact they don't have
>>> gender, and no quantity, there are only qualities applied to them:
>>> rich, richer, richest]

>> (I think you'd better go back and study your Duden!)

>  I'm sorry, I was not thinking long enough about it. I had only these
> fragments in mind:

>  Der Mann ist reich.
>  Diese Frau dort ist reich.
>  Das Kind ist noch nicht reich an Erfahrung.

>> In Die schoene blaue Donau, "feminine" and "singular" and "nominative"
>> are marked four times over, once on each word. If German had less
>> redundancy, it could use "Der schoen blau Donau," just like English
>> "the beautiful blue Danube."

>  Der schoenen blauen Donau		 , geht es nicht gut.
>  Der schoenen blauen Donau		 wegen, bin ich hierher gezogen.
>  Die schoene blaue Donau		 , ist verschmutzt.
>  Die schoen-blau gefaerbte Donau , ist der laengste Fluss Europas (oder
> die Wolga?).
>  Die schoene blau gefaerbte Donau, macht viele Boegen.
>  ...

Gender, number, and case are not *uniquely* specified in
'der sch�nen blauen Donau', but they are *redundantly*
specified.

[...]

> These loose/indirect relationships within grammar isn't my
> point.

Ignoring such redundancy won't make it go away.

[...]

Brian

> Gender, number, and case are not *uniquely* specified in
> 'der sch=F6nen blauen Donau', but they are *redundantly*
> specified.

 'Meaning' is uniquely defined.

> Brian

 Ciao
    Niels

On 23 Jan 2007 11:43:48 -0800, <niels.froehling@seies.de>
wrote in
<news:1169581428.859058.269060@q2g2000cwa.googlegroups.com>
in comp.compression,sci.lang:

>> Gender, number, and case are not *uniquely* specified in
>> 'der sch�nen blauen Donau', but they are *redundantly*
>> specified.

>  'Meaning' is uniquely defined.

Not in general, no.  In general meaning is heavily dependent
on context.

Brian

[I wrote the following before anyone else responded to Paul, but google
groups decreed that I had exceeded their chimerical "posting limit," so
here it unfortunately belatedly is.]

Paul D wrote:
> On 2007-01-23 22:23:46 +0900, "Peter T. Daniels" <grammatim@verizon.net> =
said:
>
> >
> > Paul D wrote:
> >
> >> At the risk of making myself look ignorant, this English speaker thinks
> >> we *are* missing something - the clarity you get with a morphology that
> >> clearly illustrates the relationship between parts of speech by marking
> >> case, gender, and number.
> >
> > Give us an example of something you can say in French that you can't
> > say in English.
>
> I wouldn't claim there's any such thing; just that French has slightly
> more elegant morphology than English. (Subjective, I know.)

Linguists' descriptions of languages may be categorized as more or less
elegant. The actual grammars of languages -- what goes on in people's
heads -- can't.

I suspect that what you're favoring is _explicitness_, not "elegance"
or "redundancy." You'd be very pleased with "agglutinative" languages
like Turkish, where each separate little bit of grammatical information
is given by a separate little ending on the word, all strung in a
specific order.

> >> A brief example, and possibly not the best one but the first I thought=
 of:
> >> la ch=E8re maison (dear house)
> >> la maison ch=E8re (expensive house)
> >>
> >> Speakers of article-less languages might say "la" is redundant, but it
> >
> > No, no one suggests that marking definiteness is "redundant." But
> > marking "feminine" twice is "redundant" at least once over. Marking
> > definiteness is not, however, _necessary_ for communication; neither
> > Latin nor Russian does so.
>
> Wouldn't a Latin speaker or Russian then be tempted to declare articles
> to be redundant? Or at least superfluous?

Superfluous, yes. Redundant, no.

There's a regular poster to sci,lang (he hasn't been around for a few
weeks) who loves to complain about having to remember to reveal the
_number_ of what he's talking about when he's speaking English, or the
_gender_ of what he's talking about when he's speaking German --
neither of which is needed in his native Chinese. Contrariwise, the
English- or German-speaker has to remember to specify whether the
brother being discussed in Chinese is an older or a younger brother.

> >> tells me a noun's coming up - quite useful indeed, especially if it's
> >> the first time I've heard that word (not likely here, but still...). A=
nd
> >> since "ch=E8re" agrees with the gender of "maison", that tells me it's=
 an
> >> adjective instead of a noun.
> >
> > No, it doesn't; but the fact that it occurs immediately after a noun,
> > or, unusually, between the article and the noun, _does_ tell you it's
> > an adjective.
>
> Well, just to make my point clear, I'm suggesting that if the words
> "ch=E8re maison" were otherwise ambiguous,

That's a pretty big if. Can you construct a French sentence in which
that could happen? (Don't restrict yourself to expensive houses -- try
it with any of the small number of movable adjectives.)

> the e on the "ch=E8re" would
> mean that that word must be the adjective describing "maison", and not
> the other way around. The gender declension is communicating
> information about the word's part of speech. By what standard is
> clarifying a word's part-of-speech a redundancy?

If any item of information (I won't say "bit," because you're computer
guys and you could go in some other direction) is communicated more
than once, then there is redundancy.

Note that grammatical gender itself is not "information" (i.e.,
semantic content) per se; it's a traffic-control device used precisely
for coordinating things that go together.

> > French nouns can be marked for sex where appropriate (though sometimes
> > in French, gender and sex do not agree).
> >
> >> While English would indicate the same
> >> adjective-noun relationship via strict word order, French is free to
> >> implement both word orders and have the word assume two meanings.
> >
> > Not exactly a major feature of French adjectives!
>
> No, but frequent enough.
>
> >
> > And of course an adjective can follow its noun in English for special
> > effect: "I sing the body electric."
>
> Only if the context makes it quite clear, though. With many English
> words doing double duty as nouns and adjectives, the situation could
> get confusing and you'd have to rely on rigid word order again.

What do you mean by "have to" and "again"? That's exactly how English
does it! It's very rare that you don't know immediately what part of
speech a word is used as in any particular instance, even though we
have almost no inflectional morphology (grammatical endings) at all.
(Cf. Shakespeare's "But me no buts.")

ranjit_mathews@yahoo.com wrote:
> Peter T. Daniels wrote:
> > Marking definiteness is not, however, _necessary_ for communication;
>
> If the definite article were removed from English, how would one
> distinctly communicate these two?
> 1) I am Carpenter.
> 2) I am the carpenter.

Using whatever device the language had evolved that made the definite
article redundant, so that it could be lost.

niels.froehling@seies.de wrote:
> > >  I am german anyway but this doesn't restrict to german: the addition
> > > of gender-postfixes and even gender-changes of prefixing articles
> > > (which probably is unique for german) is not redundant per se.
> > See what I mean? You use redundant expressions every day without being
> > aware of it -- even though redundancy in another realm is (apparently)
> > part of your daily work.
>
>  Under what condition do we use redundant expressions?
>
>  Is it still redundant if you quit half the message?
>  Is it still redundant if you analyse the speak-style of somebody?
>  Is it still redundant if you check the english-exercises of your
> students?
>  Is it still redundant if you translate from english to german?
>  Is it still redundant if you make an analysis of word-distribution in
> books?

I really don't know what you're getting at -- redundancy has nothing to
do with meaning. It has to do with the form, not the function, of
utterances.

In article <1169465432.274445.165260@l53g2000cwa.googlegroups.com>,
apc <andres.paniagua@web.de> wrote:
>Hi, I had this "insight" last night and and would like to read about
>it, since I presume that it is not original.
>The reason for redundancy in natural languages is that it is a
>distinctive trait of a particular set of signals. Considering the first
>encounter between two beings, the first problem that arises in
>establishing a communication is the fact that advertising the intention
>to communicate is an act of communication in itself and we get a kind
>of hen and egg paradox. Now, redundancy is a way of sending a 'meta
>signal' because redundancy is inefficient, and inefficiency is a
>strident signal in a world where there is evolutionary pressure towards
>efficiency. Therefore a set of redundant signals is more apt to be
>noticed and thus stands out against the background noise.
>Can anyone point to a reference or maybe share their own knowledge in
>this respect?

How would you avoid getting redundancy in a natural
language?  It would have to be highly deliberate, and
even then it would be difficult.  

Lack of redundancy means that any collection of 
signals can have at most one meaning, and that
additional signals cannot change that meaning.





-- 
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
hrubin@stat.purdue.edu         Phone: (765)494-6054   FAX: (765)494-0558

"Peter T. Daniels" <grammatim@verizon.net> wrote in message 
news:1169558884.503456.110680@q2g2000cwa.googlegroups.com...
>
> cr88192 wrote:
>> "Peter T. Daniels" <grammatim@verizon.net> wrote in message
>> news:1169511023.822301.21640@q2g2000cwa.googlegroups.com...
>> >

<snip>

>>
>> ok.
>>
>>
>> > "Diphone," though, isn't a technical term in linguistics -- I don't
>> > know what you're referring to.
>> >
>>
>> http://en.wikipedia.org/wiki/Diphone
>
> Ah, yes. A gimmick used in speech synthesis that corresponds to no
> linguistic reality at all.
>

I would have figured that since they exist, they correspond to something?...


>> > The number of syllables in a word (not that "word" can be defined
>> > universally for all languages!) depends in part on the number of
>> > phonemes in the language, and on how complicated the language allows
>> > syllables to be. In Japanese, the only consonant that can end a
>> > syllable is n (and it doesn't even count as an entire consonant).
>> > English can have very complicated syllables, like "strengths," so
>> > English can have shorter words, all else being equal. Georgian (of the
>> > Caucasus) can have something like six consonants in a row at the start
>> > of a word (don't ask me for an example -- but if you happen to run into
>> > someone from Tbilisi, ask them to say "orange peel").
>> >
>>
>> that is why I mentioned diphones...
>>
>> I had assumed a diphone to be a more accurate measure than a syllable, as
>> for syllables one has to define average syllable complexity, ... wheras a
>> phone or diphone should be more at least a slightly more accurate 
>> measure.
>
> "Measure" of what? What does it (as the jargon was in the mid-70s) "buy
> you" to introduce a new level of analysis?
>

well, it can be assumed to be a more accurate measure of pronounced 
complexity than is the syllable.

then again, I am more just a programmer, not really a linguist in any real 
sense.


sorry, mood is too poor right now to right a better response.
technically everything is going well enough right now, and I have managed to 
be helpful to others, but right now I still feel rather terrible, and I 
still feel alone...

Am 23 Jan 2007 12:54:37 -0800 schrieb Peter T. Daniels:

> 
> There's a regular poster to sci,lang (he hasn't been around for a few
> weeks) who loves to complain about having to remember to reveal the
> _number_ of what he's talking about when he's speaking English, or the
> _gender_ of what he's talking about when he's speaking German --
> neither of which is needed in his native Chinese. Contrariwise, the
> English- or German-speaker has to remember to specify whether the
> brother being discussed in Chinese is an older or a younger brother.
> 

In fact, they need not _remember_ to specify these things - the
languages just don't offer the possibility not to specify them.

Joachim

Am 23 Jan 2007 12:59:30 -0800 schrieb Peter T. Daniels:

> niels.froehling@seies.de wrote:
>>> >  I am german anyway but this doesn't restrict to german: the addition
>>> > of gender-postfixes and even gender-changes of prefixing articles
>>> > (which probably is unique for german) is not redundant per se.
>>> See what I mean? You use redundant expressions every day without being
>>> aware of it -- even though redundancy in another realm is (apparently)
>>> part of your daily work.
>>
>>  Under what condition do we use redundant expressions?
>>
>>  Is it still redundant if you quit half the message?
>>  Is it still redundant if you analyse the speak-style of somebody?
>>  Is it still redundant if you check the english-exercises of your
>> students?
>>  Is it still redundant if you translate from english to german?
>>  Is it still redundant if you make an analysis of word-distribution in
>> books?
> 
> I really don't know what you're getting at -- redundancy has nothing to
> do with meaning. It has to do with the form, not the function, of
> utterances.

This is a newsgroup posting I wrote answering another newsgroup
posting by a newsgroup poster. 

I see the redundancies in my previous sentence (plus the fact that it
is redundent itself) purely in meaning. 

This holds even in a comp.compression context, I feel: aren't the
lossy compression algorithms (like in mp3) based on considerations of
function?

Joachim

Am 23 Jan 2007 16:30:12 -0500 schrieb Herman Rubin:

> 
> How would you avoid getting redundancy in a natural
> language?  It would have to be highly deliberate, and
> even then it would be difficult.  
> 
> Lack of redundancy means that any collection of 
> signals can have at most one meaning, and that
> additional signals cannot change that meaning.

I would say, additional signals would necessarily change that meaning;
if they sidn't, they would be redundant, hence not grammatical in your
redundancy-free language.

Joachim

On Jan 22, 7:31 pm, "cr88192" <cr88...@NOSPAM.hotmail.com> wrote:
> "apc" <andres.pania...@web.de> wrote in messagenews:1169482539.298409.285=
10@v45g2000cwv.googlegroups.com...
>
>
>
> > cr88192 wrote:
> >> "apc" <andres.pania...@web.de> wrote in message
> >>news:1169465432.274445.165260@l53g2000cwa.googlegroups.com...
> <snip>
>
> >> this is not to say the same rule necissarily applies to writing ability
> >> though. much like a programmer can understand a wide variety of
> >> programming
> >> languages, it is difficult to write in much more than the few they hap=
pen
> >> to
> >> know...
> > Hello, thank you very much for your reply. I think I was too concise in
> > my original article, therefore I will clarify a bit. When I spoke of
> > natural languages, I didn't mean only human spoken or written
> > languages, but all the languages that appear naturally among creatures
> > in nature; I should have stated that explicitly. I think your reply is
> > more relevant to the question of why human languages remain redundant
> > over time. Also I think that you are not considering all levels of
> > redundancy in natural language, at least you spoke mainly of  grammatic
> > redundancy and its immediate semantical consequences. Anyway, I am not
> > saying that the reason I give is the only reason for redundancy in
> > language, but one among many. Its special advantage being that of
> > helping to stress the intention of establishing communication. Let me
> > give you an example with the colour of flowers. The colour expresses
> > two things on two corresponding levels. The first and more basic one is
> > 'pay attention this object is different from the background'. The
> > second, which is more in the nature of language is a message encoded by
> > the colour and intended for a specific recipient like 'I am your
> > favorite kind of flower'. The first kind of expression is what I was
> > thinking of as being accomplished by redundancy as a form of
> > inefficiency.I was ambiguous I guess as well.
>
> I guess my definition of redundancy is different.
>
> I define redundancy more in terms of a long, drawn out, and repeating
> pattern (say, if whenever people greeted each other, they would go into a
> long drawn out, and largely pre-scripted set of monologues).
>
> as a usual occurance, a short but obvious action is used (such as a hand
> wave, or making a noise). this is not redundant (unless a person, say, go=
es
> into a long repeating sequence of alternating tones).
>
> as such, I say, the used used for getting attention is not redundant,
> rather, very much the inverse.
>
> maybe a flower is 'redundant', if only because it does not change its col=
or
> (or operate like an organic display device) to present information to oth=
er
> creatures.
>
> however, I don't view it having a different color as a form of redundancy,
> in-fact, it containing a different color is a form of entropy.
Hi, I didn't mean to say that the color of the flower was a form of
redundancy. I only used that example in order to illustrate what I
meant about a possible function of redundancy in language; To stress
and mark the difference of intentional signals so that a potential
recipient may discriminate them from background noise.

> however, note that even with said 'redundancy' in communication, there are
> many things in natural language that also suggest a form of compression, =
in
> particular:
> notice how frequently used words or phrases are often truncated or otherw=
ise
> deformed into a shorter form (I forget the term right now).
>
> as such, one person will speak of decor, another of a prog lang, another =
of
> trogs, trolls, furries, and lollies, being emo, and so on...
>
> under my definitions, word complexity and semantic meanings are what
> constitute redundancy.
>
> even our communication system (speech) is fairly concise and information
> dense for an organic lifeform (we can, after all, transfer a range of ide=
as
> and information in only a matter of seconds, or maybe a few minutes).
>
> I don't expect the redundancy can be reduced that much further absent a l=
oss
> of flexibility.
>
> for example, what if humans had instead started mapping numerical values =
to
> particular syllables, and when speaking would perform a number of
> calculations to represent a phrase or conversation in some minimum number=
 of
> syllables, and where misinterpreting even a single syllable would render =
the
> whole conversation unintelligable, that or it would quickly descend into a
> state of chaos?...
>
> or, more simply, what if we had some kind of complicated flagging system,
> where for example, single-sylable markers would be used which would define
> much of the grammar for a phrase, and where the various syllables had some
> number of overlapping namespaces and a 1 to 1 mapping to particular
> concepts.
>
> for example, say we have 19 consonants and 7 distinctive vowel sounds
> (excluding schwa):
> with both CV and CVC syllables, we have 2660 unique syllables.
> nearly every concept in the language could be mappable to a single syllab=
le
> (note the range would be expanded via a 'gramatic role', which could be, =
for
> example, defined by a single syllable at the start of the phrase).
>
> such languages could encode an idea in far fewer syllables, but would be
> highly prone to error.
>
> note that in true human languages, only even a small number of the possib=
le
> syllables have any particular meaning, many not even occuring as a part of
> most words.
>
> as such, we can readily import, and recognize, loanwords from other
> languages, which only rarely happen to sound like another word in the
> listeners native language (o-haio being a well-known example).
>
> thus, I suspect, human languages do approach a sane minimum level of
> redundancy.

I agree with much of what you say. Note however that redundant has a
widely accepted meaning which is 'exceeding what is necessary of
normal', to quote it from Merriam-Websters dictionary. What do you
think about my hypothesis that redundancy   is a characteristic of
language which at least originally served the purpose of making
messages or signals distinct?
Andr=E9s

On Jan 23, 10:30 pm, hru...@odds.stat.purdue.edu (Herman Rubin) wrote:
> In article <1169465432.274445.165...@l53g2000cwa.googlegroups.com>,
>
> apc <andres.pania...@web.de> wrote:
> >Hi, I had this "insight" last night and and would like to read about
> >it, since I presume that it is not original.
> >The reason for redundancy in natural languages is that it is a
> >distinctive trait of a particular set of signals. Considering the first
> >encounter between two beings, the first problem that arises in
> >establishing a communication is the fact that advertising the intention
> >to communicate is an act of communication in itself and we get a kind
> >of hen and egg paradox. Now, redundancy is a way of sending a 'meta
> >signal' because redundancy is inefficient, and inefficiency is a
> >strident signal in a world where there is evolutionary pressure towards
> >efficiency. Therefore a set of redundant signals is more apt to be
> >noticed and thus stands out against the background noise.
> >Can anyone point to a reference or maybe share their own knowledge in
> >this respect?

> How would you avoid getting redundancy in a natural
> language?  It would have to be highly deliberate, and
> even then it would be difficult.
>
> Lack of redundancy means that any collection of
> signals can have at most one meaning, and that
> additional signals cannot change that meaning.

That is a good question too. Could natural language have evolved
without redundancy? I suppose so, if using language were so expensive
that evolutionary pressure towards efficiency played a decisive role in
its development. But my question is just the other way around. Is
redundancy a feature of language because it stresses the difference
between intentional signals and background noise; and is this so
because redundancy is a form of inefficiency and thus stands out in a
world where there is a strong evolutionary pressure towards efficiency?
What do you think?
Andr=E9s


> --
> 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

Am 23 Jan 2007 22:32:43 -0800 schrieb apc:

> 
> That is a good question too. Could natural language have evolved
> without redundancy? I suppose so, if using language were so expensive
> that evolutionary pressure towards efficiency played a decisive role in
> its development. But my question is just the other way around. Is
> redundancy a feature of language because it stresses the difference
> between intentional signals and background noise; and is this so
> because redundancy is a form of inefficiency and thus stands out in a
> world where there is a strong evolutionary pressure towards efficiency?
> What do you think?
> Andr�s
> 

Redundancy is not a form of inefficiency, as you have to have some
means to cope with transmission errors. Using a redundancy-free
language, the impact of these errors would be more expensive than
redundancy.

Joachim

On Jan 24, 9:18 am, Joachim Pense <s...@pense-mainz.eu> wrote:
> Am 23 Jan 2007 22:32:43 -0800 schrieb apc:
>
>
>
> > That is a good question too. Could natural language have evolved
> > without redundancy? I suppose so, if using language were so expensive
> > that evolutionary pressure towards efficiency played a decisive role in
> > its development. But my question is just the other way around. Is
> > redundancy a feature of language because it stresses the difference
> > between intentional signals and background noise; and is this so
> > because redundancy is a form of inefficiency and thus stands out in a
> > world where there is a strong evolutionary pressure towards efficiency?
> > What do you think?
> > Andr=E9s
> Redundancy is not a form of inefficiency, as you have to have some
> means to cope with transmission errors. Using a redundancy-free
> language, the impact of these errors would be more expensive than
> redundancy.
I cannot debate this at this point, because I lack the knowledge
necessary to do so. But let me turn the question on you. Is there any
evidence that the impact of transmission errors on natural language
would be more expensive than the redundancy that it presently has? Is
there evidence that this was also the case at the initial stages when
natural language first developed? Or is there enough room for the
hypothesis that I presented? I would be very grateful if you could
provide some references.
Andr=E9s
> Joachim

In message <1169575825.597235.151870@m58g2000cwm.googlegroups.com>, 
"ranjit_mathews@yahoo.com" <ranjit_mathews@yahoo.com> writes
>Peter T. Daniels wrote:
>> Marking definiteness is not, however, _necessary_ for communication;
>
>If the definite article were removed from English, how would one
>distinctly communicate these two?
>1) I am Carpenter.
>2) I am the carpenter.

Not a valid comparison: (2) makes no sense as an isolated utterance, so 
there must have been some previous statement defining the context within 
which "the" carpenter is unique.

Once you've filled in that detail, why assume that a hypothetical 
"article-free English" would use the same verb in both sentences? Or 
that the spoken language wouldn't have a way of marking the capital 
letter? Or indeed any number of other devices.

>
>> neither Latin nor Russian does so.
>

-- 
Richard Herring

On Jan 24, 3:25 pm, "Peter T. Daniels" <gramma...@verizon.net> wrote:
> On Jan 24, 4:12 am, "apc" <andres.pania...@web.de> wrote:
>
>
>
> > On Jan 24, 9:18 am, Joachim Pense <s...@pense-mainz.eu> wrote:
>
> > > Am 23 Jan 2007 22:32:43 -0800 schrieb apc:
>
> > > > That is a good question too. Could natural language have evolved
> > > > without redundancy? I suppose so, if using language were so expensi=
ve
> > > > that evolutionary pressure towards efficiency played a decisive rol=
e in
> > > > its development. But my question is just the other way around. Is
> > > > redundancy a feature of language because it stresses the difference
> > > > between intentional signals and background noise; and is this so
> > > > because redundancy is a form of inefficiency and thus stands out in=
 a
> > > > world where there is a strong evolutionary pressure towards efficie=
ncy?
> > > > What do you think?
> > > > Andr=E9s
> > > Redundancy is not a form of inefficiency, as you have to have some
> > > means to cope with transmission errors. Using a redundancy-free
> > > language, the impact of these errors would be more expensive than
> > > redundancy.I cannot debate this at this point, because I lack the kno=
wledge
> > necessary to do so. But let me turn the question on you. Is there any
> > evidence that the impact of transmission errors on natural language
> > would be more expensive than the redundancy that it presently has? Is
> > there evidence that this was also the case at the initial stages when
> > natural language first developed? Or is there enough room for the
> > hypothesis that I presented? I would be very grateful if you could
> > provide some references.
> It seems obvious that, evolutionarily speaking, successful
> communication is far more important than whatever "cost" you seem to
> think is involved in redundancy!
It seems obvious, but is it true? I guess it is so let us just take it
for granted for now.
> To take a silly example, what if the
> pre-language "word" for 'dangerous saber-tooth tiger' differed only
> minimally from the "word" for 'delicious bunny-rabbit' (as would
> necessarily be the case in a pre-language that hadn't yet evolved
> "duality of patterning," where a small group of meaningless elements --
> "phonemes," think of them as "letters" -- combine into an indefinitely
> large group of meaningful elements -- "morphemes," think of them as
> "words")?
I think we are a lot closer in our thoughts than you seem to believe.
Above, you stress the importance of the ability to distinguish among
signals precisely. Certainly the addition of redundancy to language is
one solution. But, do we have evidence for this? Can you provide some
references for me to read up on the subject? In any case, one could
think of other mechanisms available to evolution, like the development
of improvements in the sensory organs in order to increase that
ability. What I am talking about, is the ability to distinguish signals
from noise. How can a consensus  on which signals carry information be
arrived at if there is no means of communication available? That is the
hen and egg problem I was talking about somewhere else. And this is
where I believe that the introduction of inefficiency in the form of
redundancy may have played a role. Because acting inefficiently is
acting against the norm, and acting against the norm is more prone to
attract attention.
>
> Isn't the whole point of "compression" in your sense the _addition_ of
> redundancy to help guard against error, with check-digits and all that?
I don't exactly understand what you mean by my sense, but the point of
compression is the removal of redundancy in order to transmit a signal
or to store a message at the least possible expense.

> >  Under what condition do we use redundant expressions?
>
> >  Is it still redundant if you quit half the message?
> >  Is it still redundant if you analyse the speak-style of somebody?
> >  Is it still redundant if you check the english-exercises of your students?
> >  Is it still redundant if you translate from english to german?
> >  Is it still redundant if you make an analysis of word-distribution in books?

 I'm sorry, because I really would like to be understood.
 Look, for us IT guys it's really simple:

  redundancy_in_bits =
 	length_of_expression_of(message1,in_language1) -
 	length_of_expression_of(message1,in_language2);

 this means:

  message2 =
 	expression_of(message1,in_language2)
  message3 =
 	expression_of(message2,in_language1)

  message3 === message1 (in all and every aspect, syntactical and
semantical)

 The interpretation of IT-redundancy is:

> > > > > The reason we can compress languages (be it human or artificial) is
> > > > > that compressors transcode them. They invent on the fly another
> > > > > language that is capable of expressing the same (stationary) "things",
> > > > > but shorter under the quality 'bits'.

 Thus is removing IT-redundancy.

> > > > >  For lossless compressors there exists a 1:1 mapping of the rules in
> > > > > the source language to the rules in the destination language, but it
> > > > > introduces a complete different formalism.

 This is bijectivity.

 So IT-redundancy does only exists, because we have a stationary source
(we're not able to determine the properties of non-stationary sources)
and because we are free to develop our own language with our own tokens
and gramar, syntax, without touching the semantics.
 If we (IT-guys) were not allowed to use another language - like this:

  message2 =
 	another_expression_of(message1,in_language1)
  message3 =
 	another_expression_of(message2,in_language1)

 we receive:

  message3 !== message1

 Bijectivity is destroyed. You can't map multiple messages to the same
message, and extract the exact same multiple messages on the final step
of the process.
 This are very basic logical/combinatorical rules.
 So based on this, I only say, that in no whatsoever language you can
map from multiple meaning to one, and back to multiple. How do you
define redundancy in this aspect within a language?

> I really don't know what you're getting at -- redundancy has nothing to
> do with meaning. It has to do with the form, not the function, of
> utterances.

 Redundancy has everything to do with form, and nothing with function,
as you don't know it's function, only it's form.
 Redundancy has everything to do with meaning, if you accept that every
form-ulation 'means' something, contains information, that can't be
thrown away without throwing away meaning.
 I guess we're just faceing the ambiguity of the term 'redundancy' in
language science and in information theory.
 Redundancy in IT is hard, mathematical, it's the result of a formula.

 If you quit information/meaning from a sentence but still pretend you
speak the same language, it's not redundancy. If one writes an
algorithm that quits eyes from face-pictures (because it's a fact that
humans have eyes, the information 'human' would be enough to know
'there are eyes') it's not redundancy-removal, it's controlled
information-destruction.
 Computers are easy, way easier than humans. In compression we also
have speaker (compressors) and interpreter (de-compressors), but the
machine is strict and able to stay strictly in bijectivity - human seem
not, so translations of german to english to german back again isn't
bijective, you loose information, destroy information, you're not
removing redundancy in the first step and adding redundancy in the
second step.

 If you call the controlled information-destruction within a language:
'redundancy'; I would resist.

 Ciao
 	Niels

On Jan 24, 6:40 pm, niels.froehl...@seies.de wrote:
> > >  Under what condition do we use redundant expressions?
>
> > >  Is it still redundant if you quit half the message?
> > >  Is it still redundant if you analyse the speak-style of somebody?
> > >  Is it still redundant if you check the english-exercises of your students?
> > >  Is it still redundant if you translate from english to german?
> > >  Is it still redundant if you make an analysis of word-distribution in books? I'm sorry, because I really would like to be understood.
>  Look, for us IT guys it's really simple:
>
>   redundancy_in_bits =
>         length_of_expression_of(message1,in_language1) -
>         length_of_expression_of(message1,in_language2);
>
>  this means:
>
>   message2 =
>         expression_of(message1,in_language2)
>   message3 =
>         expression_of(message2,in_language1)
>
>   message3 === message1 (in all and every aspect, syntactical and
> semantical)
Your explanation is not really very clear, but it would seem that it is
also wrong. According to Information Theory redundancy is the
predictability of the next symbol in a stream or in a string.

>  The interpretation of IT-redundancy is:
>
> > > > > > The reason we can compress languages (be it human or artificial) is
> > > > > > that compressors transcode them. They invent on the fly another
> > > > > > language that is capable of expressing the same (stationary) "things",
> > > > > > but shorter under the quality 'bits'.
> Thus is removing IT-redundancy.
This is not true, compression is possible only if a certain property of
the source is known and the messages of the source display redundancy
under that property. There is nothing stationary about natural
languages or transmissions thereof.

>
> > > > > >  For lossless compressors there exists a 1:1 mapping of the rules in
> > > > > > the source language to the rules in the destination language, but it
> > > > > > introduces a complete different formalism.
> This is bijectivity.
Bijectivity is a 1 to 1 mapping between two sets. It has nothing to do
with rules of languages.


>  So IT-redundancy does only exists, because we have a stationary source
> (we're not able to determine the properties of non-stationary sources)
> and because we are free to develop our own language with our own tokens
> and gramar, syntax, without touching the semantics.
>  If we (IT-guys) were not allowed to use another language - like this:
>
>   message2 =
>         another_expression_of(message1,in_language1)
>   message3 =
>         another_expression_of(message2,in_language1)
>
>  we receive:
>
>   message3 !== message1
>
>  Bijectivity is destroyed. You can't map multiple messages to the same
> message, and extract the exact same multiple messages on the final step
> of the process.
>  This are very basic logical/combinatorical rules.
>  So based on this, I only say, that in no whatsoever language you can
> map from multiple meaning to one, and back to multiple. How do you
> define redundancy in this aspect within a language?
You appear to be trying to say that meanings are distinctly unique. But
that is quite obvious and is also true for expressions. Two expressions
that are different from each other, are in fact different from each
other. But what you are overseeing is that this is not true for
significations, that is, the mapping of meanings to expressions. A
single expression can have several meanings, depending on context. In
the same manner a single meaning can be signified by several different
expressions. And also note that this is not the only form of redundancy
in languages.
> > I really don't know what you're getting at -- redundancy has nothing to
> > do with meaning. It has to do with the form, not the function, of
> > utterances. Redundancy has everything to do with form, and nothing with function,
> as you don't know it's function, only it's form.
>  Redundancy has everything to do with meaning, if you accept that every
> form-ulation 'means' something, contains information, that can't be
> thrown away without throwing away meaning.
>  I guess we're just faceing the ambiguity of the term 'redundancy' in
> language science and in information theory.
>  Redundancy in IT is hard, mathematical, it's the result of a formula.
But you don't seem to be going by that formula, otherwise you would not
be stating that languages are not redundant.

>  If you quit information/meaning from a sentence but still pretend you
> speak the same language, it's not redundancy. If one writes an
> algorithm that quits eyes from face-pictures (because it's a fact that
> humans have eyes, the information 'human' would be enough to know
> 'there are eyes') it's not redundancy-removal, it's controlled
> information-destruction.
>  Computers are easy, way easier than humans. In compression we also
> have speaker (compressors) and interpreter (de-compressors), but the
> machine is strict and able to stay strictly in bijectivity - human seem
> not, so translations of german to english to german back again isn't
> bijective, you loose information, destroy information, you're not
> removing redundancy in the first step and adding redundancy in the
> second step.
>
>  If you call the controlled information-destruction within a language:
> 'redundancy'; I would resist.
> 
>  Ciao
>         Niels

> Your explanation is not really very clear, but it would seem that it is
> also wrong. According to Information Theory redundancy is the
> predictability of the next symbol in a stream or in a string.

 May you be so kind and edit this article:

http://en.wikipedia.org/wiki/Redundancy_%28information_theory%29

 to correct this miss-information (which was not my source of
information, but maybe for a lot others). It's possible with anonymous
access rightaway.

> This is not true, compression is possible only if a certain property of
> the source is known and the messages of the source display redundancy
> under that property. There is nothing stationary about natural
> languages or transmissions thereof.

 Compression is a transform, mapping symbols in a source domain to
symbols in the destination domain. Predictability is not a requirement
for compression.

> > > > > > >  For lossless compressors there exists a 1:1 mapping of the rules in
> > > > > > > the source language to the rules in the destination language, but it
> > > > > > > introduces a complete different formalism.
> > This is bijectivity.
> Bijectivity is a 1 to 1 mapping between two sets. It has nothing to do
> with rules of languages.

 I used the term 'rules' to seperate from 'meaning', two messages in
english and english-zip, have the exact same 'meaning', but are built
under different 'rules' or syntax and gramar.

> You appear to be trying to say that meanings are distinctly unique. But
> that is quite obvious and is also true for expressions. Two expressions
> that are different from each other, are in fact different from each
> other.

 You are right, I mean that.
 So why do you say that this:

 english-input => redundancy-remover => english-output

 and

 english-input == english-output

> But what you are overseeing is that this is not true for
> significations, that is, the mapping of meanings to expressions. A
> single expression can have several meanings, depending on context. In
> the same manner a single meaning can be signified by several different
> expressions. And also note that this is not the only form of redundancy
> in languages.

 It's not possible to define significance in itself, significance is
given in the relationsship between message and observer, it's not the
producer who's giving significance. He may have intentions, but no
control over the use of his message.

 I understood that you were connecting significance and redundancy in
languages, that's why I pointed, or asked under which significance a
message is redundant. Seems like there is a number of distinct
redundancies equal to the number of observers for a given message. Can
you, under that condition, safely throw away a part of a message?

> >  Redundancy in IT is hard, mathematical, it's the result of a formula.
> But you don't seem to be going by that formula, otherwise you would not
> be stating that languages are not redundant.

 Preserve all information, using the same language and having a smaller
message is oxymoronic. You are trowing away information, or using
another language, or end up with the same message! But not all three.
 If I use my own developed redundant-free english, and start talking to
you, you are not going to understand it.

 Ciao
    Niels

On Jan 24, 9:58 am, "Andres" <andres.pania...@web.de> wrote:
> On Jan 24, 3:25 pm, "Peter T. Daniels" <gramma...@verizon.net> wrote:
>
> > On Jan 24, 4:12 am, "apc" <andres.pania...@web.de> wrote:
>
> > > Is there any
> > > evidence that the impact of transmission errors on natural language
> > > would be more expensive than the redundancy that it presently has? Is
> > > there evidence that this was also the case at the initial stages when
> > > natural language first developed? Or is there enough room for the
> > > hypothesis that I presented? I would be very grateful if you could
> > > provide some references.

> > It seems obvious that, evolutionarily speaking, successful
> > communication is far more important than whatever "cost" you seem to
> > think is involved in redundancy!

> It seems obvious, but is it true? I guess it is so let us just take it
> for granted for now.

Why would you suppose it might not be true?

> > To take a silly example, what if the
> > pre-language "word" for 'dangerous saber-tooth tiger' differed only
> > minimally from the "word" for 'delicious bunny-rabbit' (as would
> > necessarily be the case in a pre-language that hadn't yet evolved
> > "duality of patterning," where a small group of meaningless elements --
> > "phonemes," think of them as "letters" -- combine into an indefinitely
> > large group of meaningful elements -- "morphemes," think of them as
> > "words")?

> I think we are a lot closer in our thoughts than you seem to believe.
> Above, you stress the importance of the ability to distinguish among
> signals precisely.

Not really -- I'm not concerned with the _signals_, but with the
_message_. The signals are a means to an end (the end being
communication).

> Certainly the addition of redundancy to language is
> one solution. But, do we have evidence for this?

Every human language exhibits considerable redundancy -- which one
knowledgeable authority estimated at 50%. Surely that means
_something_.

> Can you provide some
> references for me to read up on the subject?

I don't know that it's a subject linguists have bothered to explore,
because it's simply so obvious everywhere.

> In any case, one could
> think of other mechanisms available to evolution, like the development
> of improvements in the sensory organs in order to increase that
> ability. What I am talking about, is the ability to distinguish signals
> from noise. How can a consensus  on which signals carry information be
> arrived at if there is no means of communication available?

Vision and touch (hearing is a specialized kind of touch) and
taste/smell existed from the very beginning, so the channels existed
long, long before there were consciousnesses or proto-consciousnesses,
so this is a non-problem.

> That is the
> hen and egg problem I was talking about somewhere else. And this is
> where I believe that the introduction of inefficiency in the form of
> redundancy may have played a role. Because acting inefficiently is
> acting against the norm, and acting against the norm is more prone to
> attract attention.

I think you're playing with words there! (And why should "efficiency"
be the "norm"? Evolution certainly doesn't trend toward "efficiency"
overall -- look at the mechanism for the clotting of blood!)

> > Isn't the whole point of "compression" in your sense the _addition_ of
> > redundancy to help guard against error, with check-digits and all that?I don't exactly understand what you mean by my sense, but the point of
> compression is the removal of redundancy in order to transmit a signal
> or to store a message at the least possible expense

Ah. So you guys aren't concerned with accuracy, just efficiency. (Don't
forget about those random cosmic rays!)

> > compression is the removal of redundancy in order to transmit a signal
> > or to store a message at the least possible expense
> Ah. So you guys aren't concerned with accuracy, just efficiency. (Don't
> forget about those random cosmic rays!)

 We are concerned about accuracy, very much. But for languages we don't
have programs able to fetch meaning, and to apply an inaccurate (lossy)
transform to the message, so we're only permiting full accuracy while
trying to be as efficient as possible.

 Ciao
    Niels

 P.S.: the new Google-Interface has bugs, no crosspost-replys and
quotation breaks on every change in indention.

"apc" <andres.paniagua@web.de> wrote in message 
news:1169619210.662803.18280@k78g2000cwa.googlegroups.com...


On Jan 22, 7:31 pm, "cr88192" <cr88...@NOSPAM.hotmail.com> wrote:
> "apc" <andres.pania...@web.de> wrote in 
> messagenews:1169482539.298409.28510@v45g2000cwv.googlegroups.com...
>
>
>
> > cr88192 wrote:
> >> "apc" <andres.pania...@web.de> wrote in message
> >>news:1169465432.274445.165260@l53g2000cwa.googlegroups.com...
> <snip>
>
<snip>

>
> maybe a flower is 'redundant', if only because it does not change its 
> color
> (or operate like an organic display device) to present information to 
> other
> creatures.
>
> however, I don't view it having a different color as a form of redundancy,
> in-fact, it containing a different color is a form of entropy.

<
Hi, I didn't mean to say that the color of the flower was a form of
redundancy. I only used that example in order to illustrate what I
meant about a possible function of redundancy in language; To stress
and mark the difference of intentional signals so that a potential
recipient may discriminate them from background noise.
>

I don't think this is it exactly. if so, I would not call this redundancy, 
personally.

<snip>

>
> as such, we can readily import, and recognize, loanwords from other
> languages, which only rarely happen to sound like another word in the
> listeners native language (o-haio being a well-known example).
>
> thus, I suspect, human languages do approach a sane minimum level of
> redundancy.

<
I agree with much of what you say. Note however that redundant has a
widely accepted meaning which is 'exceeding what is necessary of
normal', to quote it from Merriam-Websters dictionary. What do you
think about my hypothesis that redundancy   is a characteristic of
language which at least originally served the purpose of making
messages or signals distinct?
Andr�s
>

yes, in a weak sense. however, this amounts to little more than jerryrigging 
and redefinition, mainly because this particular word has a much more 
specific meaning than this.

http://en.wikipedia.org/wiki/Redundancy

not the 'or duplication part'.

it means that we may be exceeding the normal, or because something is 
duplicated, but most often duplication is a strongly implied component. to 
use the term in a way inconsistent with this notion (not referring either to 
duplication, or by extension, potentially unnecessary parts by some means of 
either duplication or some easily derived process), then the usage is 
suspect.


note here:
http://en.wikipedia.org/wiki/Redundancy_%28language%29

the term is used very differently than the way in which you have presented.


more so, I suspect even your basic premise is wrong, in that most background 
noise is, itself, redundant (either a constant signal or a repeating 
pattern).

so, note how what the people at SETI are looking for, is not a redundant 
signal, as these are everywhere (spectrum given off by stars, pulsars, ...). 
what they are looking for is, very much, a signal not exhibiting this 
property (a signal containing some level of informational entropy).

so, what they are looking for does in-fact 'go beyond what is normal', 
however, by a means directly opposed to that of repetition, and thus the 
term redundancy is inappropriate in this case.


and so, if you are going to use a word, I strongly suggest sticking to how 
it is used, and not trying to bend some existing definition to what you want 
it to mean.

On Jan 24, 10:11 pm, "cr88192" <cr88...@hotmail.com> wrote:
> "apc" <andres.pania...@web.de> wrote in messagenews:1169619210.662803.182=
80@k78g2000cwa.googlegroups.com...
>
> On Jan 22, 7:31 pm, "cr88192" <cr88...@NOSPAM.hotmail.com> wrote:> "apc" =
<andres.pania...@web.de> wrote in
> > messagenews:1169482539.298409.28510@v45g2000cwv.googlegroups.com...
>
> > > cr88192 wrote:
> > >> "apc" <andres.pania...@web.de> wrote in message
> > >>news:1169465432.274445.165260@l53g2000cwa.googlegroups.com...
> > <snip><snip>
>
>
>
> > maybe a flower is 'redundant', if only because it does not change its
> > color
> > (or operate like an organic display device) to present information to
> > other
> > creatures.
>
> > however, I don't view it having a different color as a form of redundan=
cy,
> > in-fact, it containing a different color is a form of entropy.<
> Hi, I didn't mean to say that the color of the flower was a form of
> redundancy. I only used that example in order to illustrate what I
> meant about a possible function of redundancy in language; To stress
> and mark the difference of intentional signals so that a potential
> recipient may discriminate them from background noise.
>
> I don't think this is it exactly. if so, I would not call this redundancy,
> personally.
I didn't call it redundancy. Read carefully, I explained that the color
was not an example of redundancy but of the role redundancy could play
in communication.

> <snip>
>
>
>
> > as such, we can readily import, and recognize, loanwords from other
> > languages, which only rarely happen to sound like another word in the
> > listeners native language (o-haio being a well-known example).
>
> > thus, I suspect, human languages do approach a sane minimum level of
> > redundancy.<
> I agree with much of what you say. Note however that redundant has a
> widely accepted meaning which is 'exceeding what is necessary of
> normal', to quote it from Merriam-Websters dictionary. What do you
> think about my hypothesis that redundancy   is a characteristic of
> language which at least originally served the purpose of making
> messages or signals distinct?
> Andr=E9s
>
> yes, in a weak sense. however, this amounts to little more than jerryrigg=
ing
> and redefinition, mainly because this particular word has a much more
> specific meaning than this.
> http://en.wikipedia.org/wiki/Redundancy
I didn't redefine redundancy. I use the term in the sense you point to,
which is by the way almost exactly what I quoted from Merrian-Webster.
I also use it in the Information theoretical sense, which is a
specialization of the former.

>
> not the 'or duplication part'.
>
> it means that we may be exceeding the normal, or because something is
> duplicated, but most often duplication is a strongly implied component. to
> use the term in a way inconsistent with this notion (not referring either=
 to
> duplication, or by extension, potentially unnecessary parts by some means=
 of
> either duplication or some easily derived process), then the usage is
> suspect.
>
> note here:http://en.wikipedia.org/wiki/Redundancy_%28language%29
>
> the term is used very differently than the way in which you have presente=
d=2E

I guess you are totally misreading me. I am not explaining what I
believe that redundancy is; I am not trying to give the term a new
meaning. I am talking about an hypothesis concerning a reason for the
redundancy of languages.

> more so, I suspect even your basic premise is wrong, in that most backgro=
und
> noise is, itself, redundant (either a constant signal or a repeating
> pattern).
Noise is a matter of the model you are working with. Within a certain
model noise is random. And in that context that means unpredictable and
thus not redundant.

> so, note how what the people at SETI are looking for, is not a redundant
> signal, as these are everywhere (spectrum given off by stars, pulsars, ..=
..).
> what they are looking for is, very much, a signal not exhibiting this
> property (a signal containing some level of informational entropy).
And that is the trouble with noise. It looks like information and we as
humans have no information about what signals might have been intended
as legitimate messages by a possible alien transmitter.


> so, what they are looking for does in-fact 'go beyond what is normal',
> however, by a means directly opposed to that of repetition, and thus the
> term redundancy is inappropriate in this case.

I don't know anything about SETI but I would suppose that they are
making some assumptions to use as clues as to which signals might be
actually intentional messages. But they do whatever they do as the
result of a rational decision. That tells us nothing about whether my
hypothesis is true or false.

> and so, if you are going to use a word, I strongly suggest sticking to how
> it is used, and not trying to bend some existing definition to what you w=
ant
> it to mean.
As said above, you must have heavily misread me. I use the word
according to its common acceptions.

On Jan 24, 8:24 pm, niels.froehl...@seies.de wrote:
> > Your explanation is not really very clear, but it would seem that it is
> > also wrong. According to Information Theory redundancy is the
> > predictability of the next symbol in a stream or in a string. May you be so kind and edit this article:
>
> http://en.wikipedia.org/wiki/Redundancy_%28information_theory%29
>
>  to correct this miss-information (which was not my source of
> information, but maybe for a lot others). It's possible with anonymous
> access rightaway.

I will not correct that page for several reasons.
First: I do not consider myself qualified enough to publish what others
might consider reliable information on the subject.
Second: I do not use Wikipedia as a source but on rare occasions, as I
consider it a very unreliable source.
Third: I think the contents of the article are correct.
The definition I gave above is not formally stated. However, if you
understand the article you pointed to  you can see that the definition
I gave is not wrong either.

> > This is not true, compression is possible only if a certain property of
> > the source is known and the messages of the source display redundancy
> > under that property. There is nothing stationary about natural
> > languages or transmissions thereof. Compression is a transform, mapping symbols in a source domain to
> symbols in the destination domain. Predictability is not a requirement
> for compression.
>
> > > > > > > >  For lossless compressors there exists a 1:1 mapping of the rules in
> > > > > > > > the source language to the rules in the destination language, but it
> > > > > > > > introduces a complete different formalism.
> > > This is bijectivity.
> > Bijectivity is a 1 to 1 mapping between two sets. It has nothing to do
> > with rules of languages.

> I used the term 'rules' to seperate from 'meaning', two messages in
> english and english-zip, have the exact same 'meaning', but are built
> under different 'rules' or syntax and gramar.
>
> > You appear to be trying to say that meanings are distinctly unique. But
> > that is quite obvious and is also true for expressions. Two expressions
> > that are different from each other, are in fact different from each
> > other.
> You are right, I mean that.
>  So why do you say that this:
>
>  english-input => redundancy-remover => english-output
>
>  and
>
>  english-input == english-output
>
> > But what you are overseeing is that this is not true for
> > significations, that is, the mapping of meanings to expressions. A
> > single expression can have several meanings, depending on context. In
> > the same manner a single meaning can be signified by several different
> > expressions. And also note that this is not the only form of redundancy
> > in languages.

> It's not possible to define significance in itself, significance is
> given in the relationsship between message and observer, it's not the
> producer who's giving significance. He may have intentions, but no
> control over the use of his message.
>
>  I understood that you were connecting significance and redundancy in
> languages, that's why I pointed, or asked under which significance a
> message is redundant. Seems like there is a number of distinct
> redundancies equal to the number of observers for a given message. Can
> you, under that condition, safely throw away a part of a message?

The meaning assigned to symbols or signals is a matter of consensus.
Language works because 'everyone' agrees as to the meaning of signals.


> > >  Redundancy in IT is hard, mathematical, it's the result of a formula.
> > But you don't seem to be going by that formula, otherwise you would not
> > be stating that languages are not redundant.
> Preserve all information, using the same language and having a smaller
> message is oxymoronic. You are trowing away information, or using
> another language, or end up with the same message!
Can't you see how this is obviously wrong?
'Mary drives the car' has the same meaning as 'Mary drives the
Automobile' but is in fact 6 letters shorter.

> But not all three.
>  If I use my own developed redundant-free english, and start talking to
> you, you are not going to understand it


> 
>  Ciao
>     Niels

On Jan 24, 9:06 pm, "Peter T. Daniels" <gramma...@verizon.net> wrote:
> On Jan 24, 9:58 am, "Andres" <andres.pania...@web.de> wrote:
>
>
>
> > On Jan 24, 3:25 pm, "Peter T. Daniels" <gramma...@verizon.net> wrote:
>
> > > On Jan 24, 4:12 am, "apc" <andres.pania...@web.de> wrote:
>
> > > > Is there any
> > > > evidence that the impact of transmission errors on natural language
> > > > would be more expensive than the redundancy that it presently has? =
Is
> > > > there evidence that this was also the case at the initial stages wh=
en
> > > > natural language first developed? Or is there enough room for the
> > > > hypothesis that I presented? I would be very grateful if you could
> > > > provide some references.
> > > It seems obvious that, evolutionarily speaking, successful
> > > communication is far more important than whatever "cost" you seem to
> > > think is involved in redundancy!
> > It seems obvious, but is it true? I guess it is so let us just take it
> > for granted for now.
> Why would you suppose it might not be true?
That is not the point. The point is that it is not enough to rely on
what is obvious or what seems plausible enough. If you want to know
whether a hypothesis is true you have to test it.
>
> > > To take a silly example, what if the
> > > pre-language "word" for 'dangerous saber-tooth tiger' differed only
> > > minimally from the "word" for 'delicious bunny-rabbit' (as would
> > > necessarily be the case in a pre-language that hadn't yet evolved
> > > "duality of patterning," where a small group of meaningless elements =
--
> > > "phonemes," think of them as "letters" -- combine into an indefinitely
> > > large group of meaningful elements -- "morphemes," think of them as
> > > "words")?
> > I think we are a lot closer in our thoughts than you seem to believe.
> > Above, you stress the importance of the ability to distinguish among
> > signals precisely.
> Not really -- I'm not concerned with the _signals_, but with the
> _message_. The signals are a means to an end (the end being
> communication).
> > Certainly the addition of redundancy to language is
> > one solution. But, do we have evidence for this?
> Every human language exhibits considerable redundancy -- which one
> knowledgeable authority estimated at 50%. Surely that means
> _something_.
Exactly, and my question was whether part of that _something_ could be
that redundancy helped to stress the intention to communicate at the
early development stages of language. What do you think?

> > Can you provide some
> > references for me to read up on the subject?
> I don't know that it's a subject linguists have bothered to explore,
> because it's simply so obvious everywhere.
It is also obvious that the sun circles around the earth and that time
flows equally everywhere. It turns out that modern science considers
these obvious facts to be wrong. It is often rewarding to look a second
time.
>
> > In any case, one could
> > think of other mechanisms available to evolution, like the development
> > of improvements in the sensory organs in order to increase that
> > ability. What I am talking about, is the ability to distinguish signals
> > from noise. How can a consensus  on which signals carry information be
> > arrived at if there is no means of communication available?
> Vision and touch (hearing is a specialized kind of touch) and
> taste/smell existed from the very beginning, so the channels existed
> long, long before there were consciousnesses or proto-consciousnesses,
> so this is a non-problem.
I did not mean the channels or the physical means. I meant language.
Imagine you meet an extraterrestrial being. How would you initiate
communication with it? How would you make it understand that you are
trying to communicate? Couldn't it be that whatever you do will be
considered by it to be your usual behavior and totally unrelated to it
and its presence?
> > That is the
> > hen and egg problem I was talking about somewhere else. And this is
> > where I believe that the introduction of inefficiency in the form of
> > redundancy may have played a role. Because acting inefficiently is
> > acting against the norm, and acting against the norm is more prone to
> > attract attention.

> I think you're playing with words there! (And why should "efficiency"
> be the "norm"? Evolution certainly doesn't trend toward "efficiency"
> overall -- look at the mechanism for the clotting of blood!)
Of course not overall, but in general. And I am just asking, could it
be? I am not saying this is certainly so.

> > > Isn't the whole point of "compression" in your sense the _addition_ of
> > > redundancy to help guard against error, with check-digits and all tha=
t?
>> I don't exactly understand what you mean by my sense, but the point of
> > compression is the removal of redundancy in order to transmit a signal
> > or to store a message at the least possible expense

> Ah. So you guys aren't concerned with accuracy, just efficiency. (Don't
> forget about those random cosmic rays!)
Who do you mean by "you guys"? And who says compression is not
concerned with accuracy? In fact lossless compression requires 100%
accuracy. But that doesn't make what you said right. The whole point of
"compression" is NOT the_addition_of redundancy.
Andr=E9s

"apc" <andres.paniagua@web.de> wrote in message 
news:1169707490.065306.7720@s48g2000cws.googlegroups.com...


<big snip>

<
As said above, you must have heavily misread me. I use the word
according to its common acceptions.
>

this, here, is what I am debating.

you are using the word in a way, theoretically applicable to this 
definition, but in a way completely different than how it is typically used 
within most contexts (the connotations and semantics are not quite right).


it is almost like, you had little or no conception of the word as commonly 
used, picked it because it seemed "about right", and used it as is.

I am saying, however, that this is not adequate, and has IMO lead to 
inappropriate usage of this word.


others may judge differently, but this is how it would seem to me.

note that I am a native english speaker, and I have read more than enough to 
understand the ways in which this way is used.

some may debate this due to my lack of proper capitalization, ..., but I am 
lazy, and don't like having to hit shift for every sentence, and have always 
done things this way, among other things...

basically, this is my casual style, and I typically switch to using proper 
capitalization, paragraphs, ... when forced to write documentation or 
papers.

On Jan 25, 9:53 am, "cr88192" <cr88...@NOSPAM.hotmail.com> wrote:
> "apc" <andres.pania...@web.de> wrote in messagenews:1169707490.065306.772=
0@s48g2000cws.googlegroups.com...
>
> <big snip>
>
> <
> As said above, you must have heavily misread me. I use the word
> according to its common acceptions.
>
> this, here, is what I am debating.
>
> you are using the word in a way, theoretically applicable to this
> definition, but in a way completely different than how it is typically us=
ed
> within most contexts (the connotations and semantics are not quite right).
>
> it is almost like, you had little or no conception of the word as commonly
> used, picked it because it seemed "about right", and used it as is.
But then please point out an instance of this inadequate usage of the
term on my part.
Andr=E9s
> I am saying, however, that this is not adequate, and has IMO lead to
> inappropriate usage of this word.
>
> others may judge differently, but this is how it would seem to me.
>
> note that I am a native english speaker, and I have read more than enough=
 to
> understand the ways in which this way is used.
>
> some may debate this due to my lack of proper capitalization, ..., but I =
am
> lazy, and don't like having to hit shift for every sentence, and have alw=
ays
> done things this way, among other things...
>
> basically, this is my casual style, and I typically switch to using proper
> capitalization, paragraphs, ... when forced to write documentation or
> papers.

 Hy Andre;

 Tut mir Leid, dass Du genervt bist. Lass uns das Thema am besten
auslaufen.

>>> Your explanation is not really very clear, but it would seem that it is
>>> also wrong. According to Information Theory redundancy is the
>>> predictability of the next symbol in a stream or in a string.
>> May you be so kind and edit this article:
>> http://en.wikipedia.org/wiki/Redundancy_%28information_theory%29
>>
>>  to correct this miss-information (which was not my source of
>> information, but maybe for a lot others). It's possible with anonymous
>> access rightaway.
>
> I will not correct that page for several reasons.
> First: I do not consider myself qualified enough to publish what others
> might consider reliable information on the subject.

 Du machst aber doch auch Aussagen im Usenet, die korrigiert werden,
oder bestaetigt. Nichts anderes ist Wikipedia.

> Second: I do not use Wikipedia as a source but on rare occasions, as I
> consider it a very unreliable source.

 Es ist eine evolutionaere Quelle, so lange niemand Misstaende
beseitigt, weil es ja den Misstand beinhaltet, wird die Qualitaet von
Wikipedia auf niedrigem Niveau bleiben.

> Third: I think the contents of the article are correct.
> The definition I gave above is not formally stated. However, if you
> understand the article you pointed to  you can see that the definition
> I gave is not wrong either.

 Ich war nicht darauf aus Deine Aussage zu bestreiten, Du hast meine
bestritten:

.... but it would seem that it is also wrong.

 Falls Vorhersagbarkeit Teil der Definition von Redundanz ist/waere
solltest Du dies' (weil es ja kein anderer macht) in Wikipedia
klarstellen/korrigieren/hinzufuegen.
 Ich kann das nicht machen, weil ich nicht wuesste das Vorhersagbarkeit
Teil der Redundanz-Definition ist.

>>>>>>>>>  For lossless compressors there exists a 1:1 mapping of the rules in
>>>>>>>>> the source language to the rules in the destination language, but it
>>>>>>>>> introduces a complete different formalism.

 Hier waere es besser zu verstehen gewesen, wenn ich:

 ... 1:1 mapping of a message under the rules of the source language to
a message under the rules of the destination language

 geschrieben haette.

> The meaning assigned to symbols or signals is a matter of consensus.
> Language works because 'everyone' agrees as to the meaning of signals.

 Dieses 'everyone' schliesst alle Wesen aus, die keine Ahnung von
'meaning' haben. Also Kinder, Aliens, Computer, Affen und Wale, und
viele viele mehr.
 Es gibt die Moeglichkeit 'meaning' zu lernen, allerdings nur weil
Sprachen 'verbose' sind. Ein Kind haette enorme Schwierigkeiten eine
Sprache zu erlernen, die optimal Kurz ist. Wir Erwachsenen vergessen
und erlernen Sprache jeden Tag aufs neue. Wir haben keinen
nicht-fluechtigen statischen Speicher im Kopf. Alles dort ist dynamisch
und in Bewegung und wird staendig neu manifestiert. Das was als
Redundanz in einer Sprache beschrieben wird, ist genau das was
Bedeutung stiftet und erhaelt.

>> Preserve all information, using the same language and having a smaller
>> message is oxymoronic. You are trowing away information, or using
>> another language, or end up with the same message!
> Can't you see how this is obviously wrong?
> 'Mary drives the car' has the same meaning as 'Mary drives the
> Automobile' but is in fact 6 letters shorter.

 And in fact has exchanged information in a destructive way.

 Wie steht diese Aussage on Dir dieser:

>>> You appear to be trying to say that meanings are distinctly unique. But
>>> that is quite obvious and is also true for expressions. Two expressions
>>> that are different from each other, are in fact different from each
>>> other.

 gegenueber?

 Ciao & suggesting EOT
	Niels

 P.S.: Sorry for the german, I got the feeling that my german is
understood better.

apc wrote:
> Peter T. Daniels wrote:
>> cr88192 wrote:
>>> "Peter T. Daniels" <grammatim@verizon.net> wrote in message
>>> news:1169506570.500245.206060@a75g2000cwd.googlegroups.com...
>>> <snip>
>>>
>>>> I don't know what "comp.compression" is, but it seems to be where the
>>>> original poster is coming from. From the point of view of human
>>>> language (the only proper topic for sci.lang), Niels seems not to
>>>> realize what is meant by "redundancy" in language.
>>> comp.compression is a newgroup primarily talking about data compression (for
>>> example, huffman vs arithmetic coding, lz77 vs bwt, gzip, bzip2, zip, jpeg,
>>> png, and the like).
>>>
>>> as such, the OP is OT for comp.compression as well. since I had figured the
>>> topic was human languages, that is mostly where I had directed my emphasis.
>>>
>>> in my case, I had assumed redundancy in natural language to refer to the
>>> total number of syllables/pronouncable diphones/... needed to express an
>>> idea, so a language that requires more syllables would be rated as "more
>>> redundant", or at least more verbose, than one which requires less...
>> I considered your initial reply perfectly adequate, so I didn't say
>> anything myself. Only when all other sorts of communication were
>> brought in did an alarm go off. They're not the subject matter of
>> linguistics -- they're studied in "semiotics" (or "semiology").
> 
> I guess the objection that my article is out of topic is partially true
> for both groups. But I wouldn't know where else to ask for references.
> Participants of both groups are familiar with the phenomenon of
> redundancy in language and I think it is a legitimate expectation that
> they should have come across the question of how or why natural
> language (and natural communications) acquired their redundancy. And I
> think the answer to that question would give the areas of study of both
> groups valuable insights.
> Andr�s
> 
But as the many responses have indicated, people are interpreting 
'redundancy' in many different ways. While there's clearly a common core 
(content in the information stream which could be eliminated without 
compromising the information conveyed) it is not obvious that any 
narrower meaning for 'redundancy' is useful or even possible in 
different fields.

My take on your original question is that the vast majority of 
linguistic communication is between individuals who do share the 
language, and therefore it cannot be a principal purpose of redundancy 
(whatever it means) to facilitate communication between those who don't!

Colin

ranjit_mathews@yahoo.com wrote:

> If the definite article were removed from English, how would one
> distinctly communicate these two?
> 1) I am Carpenter.
> 2) I am the carpenter.

Taking ambiguity in language to the extreme:

http://en.wikipedia.org/wiki/Lion-Eating_Poet_in_the_Stone_Den

-t