Re: Lame philosphy

In article <dugb0u$jia$1@xxxxxxxxxxxxxxxxxxxx>,
John Wilkins <john@xxxxxxxxxxxxx> wrote:
Friar Broccoli wrote:
Hi John;

Just want to check on something else.

I am just now reading:

http://www.talkorigins.org/origins/postmonth/oct00.html

Since you (with Mark Isaak) nominated it, I assume that you agree
that DNA is unequivocally "Shannon information" so the only thing you
are arguing is that it is not the more constrained "meaningful"
information?

Any ordered sequence can be given a Shannon entropy; inversely any Shannon
entropy can be expressed as a sequence (or, I guess, the set of all sequences
of equivalent Shannon entropy).

Oh dear. I'm about to disagree with a philosopher about a word use ...

Shannon entropy doesn't even need ordering, and I think that's crucial
to some following points. H is defined by -sum(p_i* log(p_i)) where
i is an index to the state referenced. There's nothing here necessarily
about ordering or sequencing. These probabilities are computed by
finding global frequency of occurrance of states. (Global w.r.t.
system being encoded that is.)

We could take, for instance, a strand of DNA, count up the A, T, C, and
Gs and see what the probability of each was, then execute the sum. Only
4 states involved. This H would be invariant with respect to swapping
base pairs around. But the resultant biological entity would, quite
probably, be different. So as far as information meaning 'build an
organism' goes, this elementary sort of Shannon information is fairly
useless.

We could get slightly more elaborate and pay attention to the sequencing
base pair to base pair. The coding problem is then the transition set
A->A, A->T, ... So 16 possible elements for the index i to range over.
Again we could compute the H (larger, I expect). But, again, we could
swap around chunks of base pairs without modifying H at all -- these
probabilities are global statements -- and, again, the resulting
organism would be changed. The biologically meaningful information isn't
in the Shannon sense of base pairs, nor even in the base pair by base pair
sequencing. (Likewise for the triplets.)


But is DNA a Shannon sequence? Well, no, not
really. When we assign it an entropy we do so by specifying what changes of
state affect the entropy. So the replacement of a Guanine by a Cytosine
molecule in a DNA strand affects the Shannon entropy (henceforth H) of the
DNA. Replacing a Guanine molecule with a Cytosine in the surrounding nclear
medium may not affect the H of the nuclear medium much, because it is not
ordered. By identifying the strand as the "message", we decide that it has an H.

It is a fact that of all ordered systems that happen to match the
preconditions for having an H, DNA sequences are instances. So they have H.
But *are* they H-bearers? I think not. They are molecular structures and
systems. H resides in the way we describe DNA because it is epistemically
useful to us to do so.

Agreed, per above. We could choose many states to represent, and then
apply a Shannon encoding to. But it is _we_ who are choosing the states
to represent. The H that results is largely a statement about how we
happen to be choosing states, not about the thing being represented.

[kept for completeness, no further comment]

Yes, I deny that DNA is "meaningful", the so-called teleosemantic notion of
information, but even here there are caveats. Usually this is a claim that DNA
sequences are "about" some aspect of the organism, the endophenotype or the
exophenotype. In that respect you can say that AGG is "about" Arginine (except
in mitochondria, where it is a stop codon), because that is how it is
transcribed. But the description here is better expressed as the sum of the
vectors of all the physical properties - of hydrogen bonds, strong bonds, van
der Waals forces, and the valencies of the molecules, and how they act as the
catalysts for mRNA and tRNA, etc. And when you have done this, there is
nothing left to explain. So you'd be better off saying that AGG acts as a
catalytic substrate for Arginine in nuclear eukarutoic chemistry of the nucleus.

That DNA is "about" blue eyes or sexual propensity is just stretching it too
far. It certainly contributes to the causal processes. Sometimes it acts as a
flag that a certain causal process is going on (like when we use the COI gene
to indicate that speciation has taken place, even when COI is not directly
involved in speciation, which is laways so far as I know). But DNA is not
information except in these very restricted causal or diagnostic ways.
--
John S. Wilkins, Postdoctoral Research Fellow, Biohumanities Project
University of Queensland - Blog: evolvethought.blogspot.com
Servum tui ero, ipse vespera



--
Robert Grumbine http://www.radix.net/~bobg/ Science faqs and amateur activities notes and links.
Sagredo (Galileo Galilei) "You present these recondite matters with too much
evidence and ease; this great facility makes them less appreciated than they
would be had they been presented in a more abstruse manner." Two New Sciences

.

Relevant Pages

  • A 28,000 Years Old Cro-Magnon mtDNA Sequence Differs from All Potentially Contaminating Modern S
    ... Differs from All Potentially Contaminating Modern ... DNA sequences from ancient speciments may in fact result ... radically differs from those of the almost contemporary Neandertals, ...
    (sci.archaeology)
  • Re: Evolution Theory Has Scientificly Collapsed
    ... plants and simple dna type organisms.) ... What non-dna type organisms are you aware of? ... You know, anything above species level is macro, below species is ... Different species have slightly different sequences, ...
    (talk.origins)
  • Re: Humans, chimps, wheat and frogs
    ... homologs present in both species. ... conserved) sequences are a tiny minority of the genome, ... an assay of junk DNA. ... of the numbers you have quoted have anything to do with counting indels. ...
    (talk.origins)
  • Re: Humans, chimps, wheat and frogs
    ... Depends on what part of the genome you're looking at. ... conserved) sequences are a tiny minority of the genome, ... These stretches of DNA ... As the species specific miRNAs described here are ...
    (talk.origins)
  • Re: DNA carries information
    ... Shannon like systems to semantic systems ... Thus your claim that 'DNA is NOT ... information' is completely empty of meaning, ... How much more explicit could I be? ...
    (talk.origins)