This is Lucien's latest reply.
>From: Lucien Carroll [mailto:email@example.com]
>Sent: Sunday, May 12, 2002 12:16 AM
>To: Glenn Morton
>Subject: Re: Emergence of information out of nothing?
>I forgot to change the reply field last time, so neither my posting nor
>yours made it to the list. I don't have a copy of what I wrote. Feel
>free to forward the conversation to the list, but since this wont make
>sense to the list without the previous posts, i'm just sending it
>to you directly.
>Glenn Morton wrote:
>> > The third thing, conventions, is likewise an
>> >objective feature of the environment.
>> I think I only partly agree with this. Conventions are
>objective, but they
>> can't be recognized a priori and thus are also part of the subjective.
>> Obviously for us, the language you and I are communicating in has an
>> objective standard (dictionary) for what things mean. But the
>> my professor who taught me Wittgenstein said, is a great
>tautology. It is a
>> game in which all words are defined by reference to other words.
>> this, the structure of a language simply can't be objectivized
>in the way in
>> which Shannon entropy can be. Shannon entropy is defined such
>that a given
>> sequence will have a given informational content.
>Doesn't Shannon entropy also depend on the system that a state is drawn
>from? The microstate 153832 has one information content if it is a
>decimal string and another if it is a hexadecimal string, because of the
>partition function. If you take a protein, can you determine the
>information content without knowing that the amino acids represented are
>the only ones possible?
>> But when we speak of
>> semantic information, we have moved into several different games, as
>> Wittgenstein called it, in which tautological definitions rule
>and there is
>> no objective standard. Chinese zi dan's [I think that is the word for
>> dictionary] define words in reference to chinese zi's[characters].
>usually cidian in pinyin
>> I think you miss the problem.Don't limit yourself to known
>> on earth. Let's say we receive a string of 1's and 0's from
>> on 1630 mHz. People record it and if you bunch it in bytes of 5
>bits it fits
>> Zipf's law. For those who don't know, Zipf's law is a
>> which fits languages. Anyway, if we find this, does it mean that we have
>> found those proverbial little green men? I don't think so. In
>fact I am not
>> sure we would understand that we have a language. How do we
>know that the
>> sequence is to be 5 bit bytes? Zipf's law fits many other
>things, like the
>> population of cities and no one 'designed' that relationship.
>Like the cyphered string, we would likely not be able recognize the
>meaning of this string (assuming it did have meaning). A string that we
>can find no meaning in, may or may not have meaning, but one in which we
>can find meaning does have it. We can certainly find good reason to say
>a particular string might or might not have meaning, such as the Zipf's
>law thing perhaps, but we are left in doubt.
>> Wo you ee ge wen tea. Ni hui shou Zhong guo hua, ma?
>Wo de Zhongwen hai hen cha, danshi wo hui shuo. :)
>> >Perhaps Shannon information would be easier to explain to people if
>> >semantic information were better understood.
>> Probably, but as Shannon said, his form of 'information' has
>nothing what so
>> ever to do with semantic information. The problem is the logical
>> equivocation upon the word 'information' and there is where the confusion
>> comes from.
>I still don't think they are totally unrelated. I'm getting out of my
>depth here, (or maybe i have been all along) but i think this is a valid
>example: There is a limit to how much you can compress a megapixel 16
>million color image. There are all kinds of ways of making it fill more
>space, but the image (and its compression algorithm, say, to avoid
>cheating) must occupy more than some minimum amount of space. I suspect
>that if that minimum could be achieved, the shannon information content
>of that space would be at its maximum.
>Semantic shift and such in language (such as the word "information", for
>example) makes it clear to me that the semantic kind of information does
>arise naturally in this kind of system. (Unlike the case of
>"information", the vast majority of language change is not conscious or
>by design). Denying that it exists or denying that it is a valid field
>of scientific study won't make it go away.
>Lucien S Carroll firstname.lastname@example.org
>"All mankind is stupid, devoid of knowledge."
This archive was generated by hypermail 2b29 : Sun May 12 2002 - 16:58:07 EDT