Re: Evolution without either drift or selection?


William Morse wdmorse@xxxxxxxxxxxx wrote:-

You could avoid drift by using artificial selection.
E.g. sequence all the genomes involved, and then use a
deterministic algorithm to produce the genomes of the
next generation from the genomes of the previous one.

JE:-
You cannot EMPIRICALLY stop ANY random process. Science is only
concerned with the empirical facts. Hypothetical events can be used
or misused.

[moderator's puzzlement: Can you unempirically stop any random
process? In other words, whachu talkin bout, John? - JAH]

JE:-
Josh,
I'm "talkin bout" heuristic models. Random events can be halted within
them by definition.
Heuristic events are not empirical events. If they were, the word
"heuristic" would become redundant.

If I understand your explanation,it is that what you meant to say was
that you cannot stop any empirically random process.

JE:-
Hello William,

Randomness in nature, i.e. _empirical_ randomness cannot be halted via a
simple definition (or using any other way) as indeed it can be within a
heuristic/hypothetical model. As an example, no way exists of removing
genetic drift or mutation entirely within a natural population because both
are defined to be empirical random processes. Only NON random events can
become entirely halted within nature, e.g. selection. I have outlined an
experiment which can halt all selection within an empirical population, i.e.
not just a heuristic or hypothetical population such as a computer
simulation.


I disagree with this
statement as well. ISTM that I can stop an empirically random process
by, for instance, stopping throwing dice or flipping a coin.

JE:-
Picking up a dice or a coin need not be a random event. What produces the
randomness you are describing here is not caring what you do with the
coin/dice after you have picked it up. Whenever you care you can select say,
the heads side of the coin or the six on the dice. If this doesn't matter
then it is just waste of time and resources making either of these
selections. The point is: non random processes produce random patterns only
because they cannot be bothered carefully controlling every possible event.
OTOH random processes can only produce random patterns because they are
barred entirely from controlling anything by _definition_. If they do
control something then the non random pattern this must produce means that
the processes cannot any longer be defined as random. Random processes are
epistemologically unique because they can be refuted but they cannot be
verified (the reverse of what most people consider to be "normal": can be
verified but not refuted).

You could
also interfere with otherwise random processes by for instance putting a
magnet under a roulette wheel. Tim appears to be talking about just this
kind of deliberate interference - and I fail to see how your comment
applies.

JE:-
If the empirical process is just a defined random process then all you can
do is reduce the level of randomness, you cannot eliminate it. For example,
drift via the founding effect can be reduced by increasing the mean size of
migrating populations. Random mutation can be reduced by shielding. Neither
can be empirically eliminated but both can be removed entirely just using
heuristics.

Regards,

John Edser
Independent Researcher

edser@xxxxxxxxxxxxxx




.