Re: Minimum Cell?

On 8 Dec, 00:12, Tim Tyler <seemy...@xxxxxxxxxxxxxx> wrote:
r norman wrote:
Tim Tyler <seemy...@xxxxxxxxxxxxxx> wrote:
r norman wrote:
On Fri, 07 Dec 2007 10:14:04 GMT, Tim Tyler <seemy...@xxxxxxxxxxxxxx> wrote:
Any physical system has a state. Throw a bunch of nails on the
floor and the resulting configuration has a state.

[...]





The continuous state variable issue is a straw man because I specified
two stable states.
The "continuous state variable" issue is a figment of your
imagination - as far as I can see. I never mentioned any
such thing.

I was talking about copying. To "remember" things for
very long in biology, you *have* to make copies. The system
can't just lie there "remembering" its state and expect to
attain immortality - that's not a viable long-term strategy.

The error correction described seems to be
referring to something like homeostasis or negative feedback
regulation, the ability to continue relatively unchanged
in spite of environmental disturbances.
Error correction is, um, error correction: the ability to
correct errors. In this context, it is not "homeostasis" or
"negative feedback". It is the ability to find and fix
copying errors. To quote from my page on the subject:

``There is really only one known example of an error correction
mechanism which occurs in nature outside of biology.

That is the error correction process that is associated with
crystal growth.''

-http://originoflife.net/error_correction/

There's a section there on reversibility. [...]

We seem to be on totally different wavelengths, here.

As to states: you said "Throw a bunch of nails on the floor and the
resulting configuration has a state" Now tell me that the
configuration of nails on the floor is not a continuously variable
condition.

Right. The point of that example was that "having a state"
is a trait shared by any physical system.

True, but there is loads of room for confusion. Philosophers
distinguish between defining and nondefining characteristics of a
*type* of thing. When they do this, they are kind of making the point
that natural language should grow up and know how to choose a primary
key.

So you say "having a state", but you could just as well have said
"having attributes" - loads of people don't understand the difference
(although I know rnorman does). A thing can be the same sort of thing
but come in the colour red instead of the usual blue. That isn't the
same as state. A thing can be red and heavy OR red and light - they
aren't mutually exclusive. We reserve the word "state" for attributes
which are mutually exclusive (and therefore time dependent).

Now I can
imagine how you interpreted it as raising the issue of
discrete systems vs continuous ones - which wasn't the
intended point of it at all.

The /intended/ point was: "having a state" is easy -
it's *copying* it that can get tricky.

As to copying, the cellular automata I mentioned indeed produce
copies.

Yes - but flames don't - or at least they don't do so naturally.

The hypothetical flame-inheritance system under discussion
is not a known natural system. You might be able to make a flame
inheritance system out of multiple gas rings, with a lab assistant
putting them next to each other - but that is a system based on
elements from modern biology, with no known prebiotically-plausible
counterpart. If you think otherwise, then show me a natural
flame-based inheritance system.

By contrast, crystals are a *real* physical system that *actually*
produces high-fidelity copies of information during natural crystal
growth processes.

I suppose they do. Is it fair to call it "seeded self organisation"?
I don't think it is fair to call it even the loosest metaphor for
abiogenesis.

Such hi-fidelity copying is one of the key requirements of the
earliest living system.

As to why error correction should occur best in near equilibrium
systems, I am at a total loss to understand the motivation for such a
statement.

That was *not* my statement. "Reversibility" != "near equilibrium".

Like you say, you seem to be on a pretty different wavelength.

The link between near-reversibility and natural error correction
is /fairly/ well known:

Natural error correction typically requires that incorrect units
can easily become detached and replaced with correct ones.

If the operation attaching "incorrect" units is too strongly
energetically favoured (over the operation that detaches
them again), mistakes are liable to go uncorrected. Similar logic
applies to "correct" units - else the addition of a "correct"
unit would lock an "incorrect" unit in place - leading to another
type of error. For the best chance of success, "correct" and
"incorrect" attachment and detachment operations need to be close
to reversible - allowing a "three steps forward, two steps back"
construction regime - that offers multiple chances to correct any
errors.

http://originoflife.net/error_correction/offers more details.
--
__________
|im |yler http://timtyler.org/ t...@xxxxxxxxxxx Remove lock to reply.

.

Relevant Pages