Re: OOL X - The origin of the RNA world.

> From: Tim Tyler <tim@xxxxxxxxxxx>
> Evolving systems don't necessarily have to have clearly defined
> generations

If you are referring to generations in a human sense, that is a set of
people born about the same time are considered one generation, then
their children (by sexual reproduction) are a second generation, etc.,
with never a mating between people of different generations hence
always an unambiguous decision which generation to assign each child to
(the next generation after the child's parents), then you are correct,
you can't partition people into distinct generations like that in a
fully consistent way.

But for asexually reproducing life, you *can* define generations from
any single individual starting point. It doesn't matter that people of
a "later" generation may actually be born before people of an "earlier"
generation due to faster reproduction along one branch than along
another. Defining generations in this way is useful for the
mathematics. For example, with bacterial fission, you can state from
one individual you get 2 of next generation, then 4 of next generation
after that, 8, 16, etc. assuming no deaths. So long as you are just
using this in a logical way, and you realize the specific "generations"
might overlap chronologically, there's no problem.

For independent replicators, reproduction of the whole set might not
make sense, but for each single replicator which is a simple catalytic
loop, you can define generations per the operation of that loop. If
catalyst A begats catalyst B which begats catalyst C which begats
another copy of catalyst A, then the A which comes out from that is one
generation after the A which went in. I use this definition of
"generation" when calculating fecundities of such catalytic loops.

> nor do they need to replicate anything.

I disagree! Without replication, there's no way to amplify any adaption
that may have occurred, no way to get progressive adaption which is
what we normally call evolution. If you start with 5 copies of
something, and there's no replication, then eventually one decays and
you have 4, then 3, then 2, then 1, then 0 and it's over. Even if there
was variation among the five, and some survived longer than others
because they were more stable (or more "fit"), in the end there's
nothing left. It's only with replication that losses of less-fit
individuals are replaced by more copies of more-fit, so that you can
maintain the numbers of the best-fit and have anything remaining after
a long time. I don't consider an "evolving system" to be one which just
dies out.

> What they do need to do is preserve information across reasonable
> spans of time.

If they completely die out, then it's impossible for them to have
preserved information.

> > So for just the definitin of "genetic", a single bit turned on is
> > sufficient, but for something that can evolve via group selection of
> > the many replicators sharing a "cell", you need several replicators not
> > just one.

> The issue isn't simple. For one thing, you can't assume identifiable
> replicators exist in the system.

In the case I was referring to, one (of many theoretically possible)
replicator, which is a simple catalytic loop (a hyper-cycle of chemical
loops), I've presumed a clear distinction this particular species of
catalytic loop and any other species of catalytic loop.

Alternately, I considered catalytic-type loop, where a generic class of
catalysts all catalysts the same generic kind of chemical reaction,
such that a class or two of abiotically produced reactants are
converted to a different class or two of products, one of which is a
second class of catalysts which do the same sort of thing, etc., until
at the end of a chain we have the original class of catalysts being
produced, which then loop around. Perhaps we can't define precisely
what these classes of chemicals are, to where we can define a single
bit per class with no overlap, is that what you are complaining about?
If so, maybe I should change the bitmap to represent all the chemical
species which sustain the catalytic loop, such that any one by itself
could re-start the loop if all the others disappeared, so then a
catalytic-type loop would have not just a single bit turned on but a
group of self-reenforcing bits?

> If you can only go on information in the system which is preserved
> over time - and if you fail to specify some minimum threshold of
> such information - then you wind up with rivers storing their paths
> "genetically" and the earth remembering which way it is spinning
> "genetically".

The key is to consider the "basins" in which the genome is usually
locked. With just one single replicator, a catalytic loop or a
catalytic-type loop, the genome is locked into that single replicator
for a long time, possibly varying the specific proportions of the
various catalysts among the catalytic type, surely varying the
proportions of activity around the loop depending on variation of
"food" available hence speed-up of one step of loop and slow-down of
another step of loop. But at all times the genome remains in that basin
of these catalysts in some mix and large amount of all but none of any
other in any significant quantity. If and when a second capable
replicator spontaneously forms and grows exponentially to become
"successful" too, then that new replicator is in a different basis from
the first. There's a clear gap between the two basins. No ambiguity
whether the bit is "on" or "off" for each (except during early stages
of exponential growth when we might debate the threshold for "success",
and during final stages toward extinction when we can't measure whether
it's really extinct yet).

But with river channels, there's no clearcut basin in this sense. Start
two parallel simulations with a single river in two different locations
on otherwise identical continents, and they might by chance miagrate
toward the same final state, thereby contradiction any assumption that
they were in different evolutionary basins in the first place. Start
with multiple copies of all possible starting positions for a river on
otherwise identical continents, running all the simulations in
parallel, and whenever a position P "evolves" to a position Q then
equate P and Q, and apply transitivity, and viola *all* river positions
are equal, that is between any Pa and Pz there is a chain of Pa -> Qa
<- Pa'=Pb -> Qb <- Pb'=Pc -> ... <- Pz.

In conclusion, my usage of "genome" or "genetic" doesn't include the
short-term or sloppy-medium-term memory that river paths have, so I
regard such an example as a strawman.

Now if a river must cross a ridge, which has a very hard obstacle in
the middle, but two relatively easier passes on either side, and the
river must breach either one or the other, and once breaching one it
wears it down making it impossible to ever rise enough to ever breach
the other, then there would be a permanent bit of information "etched"
in the landscape. But rivers don't reproduce such decisions, so they
don't qualify for evolution by natural selection which requires
reproduction to amplify the decisions so that more-fit genomes don't
die out. So in this case the river is a strawman for a different
reason.

.

Relevant Pages

  • Re: Precision is not likely in the OOL
    ... the first successful replicator would come into existance more quickly, ... chemicals that are catalysts but don't fit into that initial ... combination of primary high-energy chemicals and secondary low-energy ...
    (sci.bio.evolution)
  • Re: Precision is not likely in the OOL
    ... The first successful replication was probably very sloppy. ... >> provides a model of a successful sloppy replicator. ... > - Etc. other catalysts, each rather general, but given only one ... inherent specificities are modest, ...
    (sci.bio.evolution)
  • Re: I thought this NG was about Origins?
    ... would suddenly start manufacturing whatever large molecule it was ... I use the word "replicator" to refer to the entire auto-catalytic set ... of chemicals, i.e. it's the system that synthesizes more copies of all ... reaction products are likewise generic catalysts, ...
    (talk.origins)
  • Re: I thought this NG was about Origins?
    ... would suddenly start manufacturing whatever large molecule it was ... auto-catalytic set, which I don't see happening until after replicators ... I use the word "replicator" to refer to the entire auto-catalytic set ... reaction products are likewise generic catalysts, ...
    (talk.origins)
  • Re: OOL X - The origin of the RNA world.
    ... I agree the enzymes in the ocean wouldn't ... and the secondary reactions from when heavy chemicals ... ocean that could be consumed by the first replicator to provide atoms ... But as soon as these simple catalysts (each part of its own catalytic ...
    (sci.bio.evolution)