It might be helpful, then, to state up front that information a la
information theory isn't the thing being referred to, but instead
something of the 'kind' variety--a goal-post-moving kind of construct
in which it is possible to shift contexts at will to rule out any
sort of counterargument, is being used.
[now quoting from other post:]
> It would seem that the technical definition of information is
> somewhat outmoded. For a clearer picture of the difference between
> real information and the statistical level (which
> is the subject of Claude Shannon's useful, but inadequate theory), I
> suggest you read the book by Werner Gitt, who is a specialist in
Bwahaha. If information theory doesn't make sense when applied to
biology, then it doesn't make sense. That doesn't make it inadequate.
So far, there has been a struggle to fruitfully apply the idea, but
(in my opinion) it is too early to rule out any such application.
[...]
> the Wall Street Journal (WSJ). Now I start to print the pages. This is
> like the example of autopolyploidy, making multiples copies of the same
> information. By the time I have 2 copies, do I now have double the
> information? When I have run 100,000 copies, do I have 100,000 times the
> amount? Of course not. Now what about allopolyploidy? It is like I take
> the WSJ and the New York Times (NYT) and tape them to each other. In the
> most naive sense, I have 'more information', but not in any sense which is
> useful for evolutionary theory. Let me explain. The theory posits that
How unprofound. Since all the information there is is in the positions
and velocities of fundamental particles, then rearranging them cannot
possibly result in any gain or loss in information.
Where do I line up to get my Nobel Prize? :-)
Applying information theory to DNA is an interesting puzzle, because
there are lots of dependencies which are unknown, and it is therefore
hard to describe the context in which the code is being decoded, and
so a straightforward application of information theory is impossible.
This shouldn't be news to anyone who has worked in the field, but I'd
expect a bit more appreciation for the actual underpinnings of
information theory. Calling it 'inadequate' is hardly a good sign,
unless one wishes to merely argue that television didn't have as
much to do with DNA as we thought.
[creation of information 'for' lungs and feathers]
> content in the only sense that counts in this debate - i.e. the creation of
> new information. Dawkins' neo-Darwinism insists that random mutations
> filtered by selection have raised the information in the biosphere (and in
> lines of phyletic descent). As Spetner points out, to have a net increase
> in information, one needs to demonstrate that the processes one is relying
> upon to generate this are capable of individually increasing information.
This is fairly incredible. One of the central dogmas of information theory
is that information isn't *for* anything. While it may mean a lot to
a person to receive the message 'you have just won $10,000,000,000' there
is no special significance to that message in information theory.
Actually, this was one of the factors which made people think there was
possible biological application: mutations appear to be random, meaning
that they aren't Lamarckian-ly directed at anything in particular.
It is not unusual for people to misunderstand the basics of information
theory, but it is disappointing that someone who has done research in
the field can bungle it so.
-Greg