Species selection



HT:- To summarise my point, if the fitness of a species is
to be understood in the way that I've interpreted it (more
descendant species means greater fitness), then a species
property that increases the likelihood of speciation is
still no more likely to be perpetuated than a property that
decreases it. It would be a mistake to assume that as the
number of species in a genus increases, the number of
individual organisms in that genus will also increase. One
does not entail the other. The number of species in a genus
may increase even as it marches toward extinction. Species
selection understood in these terms cannot therefore explain
any properties of species.

JE:- Absolutely correct. All group selection fails for the
same reason: when fitness of a species is the simple
addition of the individuals said to constitute
it. In this case, selection operates at the individual
level, FIRSTLY. The effect of this is that any group
selection between additive fitness groups must go with
and not against, selection at the lower, individual
level, i.e. the ONE Darwinian level of selection: the
level of the fertile form.

HT:- Note that although the number of species may increase
as the number of individuals decreases, this can't go on
forever. The number of species will ultimately be limited by
the number of individual organisms in the genus because you
have to have at least one (or two) individual per species.
The number of species will therefore cease to increase as
this limit is approached.

JE:- If you wish to make a more abstract view of this event,
the increase in species is only a relative fitness measure
but the increase or decrease of Darwinian individuals within
each species is an absolute fitness measure. Thus what you
are saying here is that it is pointless only increasing a
relative gain when it constitutes an absolute loss. You win
a battle but lose the war. This is why nature can only
select for an _absolute_ fitness increase. It does not
matter how many levels of fitness you suppose, only ONE can
constitute a valid measure of absolute fitness. Just this
one level, which empirically is the original Darwinian level
of selection: the fertile form level, constitutes an
absolute measure of fitness determining all selective
events. Only one absolute measure of fitness can logically,
exist. This measure is just an absolute assumption. Such
assumptions contest each other, such that all others refute
in favour of just one of them.

It is interesting to note that the reason why Hamilton's
rule fails is for the same reason. Hamilton et al only
measured a relative difference in fitness between rb and c,
i.e. they failed to include any measure for the absolute
fitness of the donor within his famous rule: rb>c. This
being the case, as the altruistic gene relatively spreads it
can suffer an absolute fitness loss sending both competing
genes into extinction along with that entire group of
individuals. Note that the Enron accountants destroyed Enron
for exactly the same reason. They allowed debits to become
credits. This was only possible because any measure of the
absolute wealth of Enron was allowed to be compromised.

At its most abstract level, the deletion of absolute
measures so that "everything is relative" is termed: Post
Modern Epistemology. It destroys everything it touches
because it is an utterly absurd view. This is because it is
100% non self consistent. The view that everything is
relative is an absolute assumption. This being the case it
100% contradicts itself. This ancient con has been known for
thousands of years. It is called Epimenides Paradox. Neo
Darwinian arguments are littered with such absurdities.

Respectfully,

John Edser Independent Researcher PO Box 266 Church Pt NSW
2105 Australia

[email protected]


[email protected] (Huck Turner) wrote in message
news:<[email protected]>...
> Tim Tyler <[email protected]> wrote in message
news:<[email protected]>...
> > Huck Turner <[email protected]> wrote or quoted:
> [snip]
> > > The only sense my feeble mind can make of the idea of
> > > opposing forces of selection is the case where there
> > > are two opposing forces at the same level. [...]
> >
> > Species-level selection should act to make populations
> > more likely to speciate.
> >
> > For example:
> >
> > * Members of a population should have genitalia that
> > quickly become incompatible with those of their host
> > population if they are isolated from it;
>
> So a species (or set of species) with more easily mutated
> genitalia will tend to give rise to a greater number of
> descendant species and we could say that this is what it
> means for a species property to have greater fitness.
> However, we have no reason to expect the total size of the
> population of individuals to change as a result of the
> speciation event unless the genital mutations also affect
> the chances of individuals reproducing. If it has an
> opposing negative effect on fitness at the level of the
> individual (the case that you take to be interesting), we
> should expect the total size of the population to decrease
> while the number of species in it increases until either a
> more stable species emerges with less mutable genitalia or
> the whole lot of them go extinct.
>
>
> >
> > * They should enjoy (and be well adapted to) living on
> > islands - since islands are where many new species are
> > born;
> >
> > * They should be prone to changes in ploidity;
> >
> > ...and so on.
> >
>
> The above reasoning could be applied to these examples
> too. As the number of species increases, the total number
> of individuals may decrease. This is a bit like dividing a
> cake into thirds, throwing away one third and then
> celebrating that now you have two pieces where before you
> had only one. Continue this process dividing each of the
> pieces into thirds and you'll have less and less cake, but
> more pieces. This is like the Cantor set
> (http://www.mathacademy.com/pr/prime/articles/cantset/).
>
>
> > These may not be the same traits that best serve
> > individual
reproduction.
> >
> > Indeed the traits the would best serve species
> > reproduction may
positively
> > /hinder/ individual level reproduction. Genitals that
> > are verging on mechanical incompatibility with other
> > members of your species might well make isolated
> > popualitons more likely to speciate - but they might
> > also make it more difficult to get a date.
>
> But if they're isolated already then it won't make matters
> worse if their genitalia are incompatible so there simply
> won't be a competing selection pressure for genital
> compatibility. It won't be present.
>
>
> >
> > Forces on the different levels can pull in different
> > directions - and favour different adaptations.
>
> Doesn't follow.
>
>
> H.
>
> ---
> Like-minds don't notice shared mistakes. Talk to
> someone else.
 
I'm hoping that someone, if not Tim, can respond to some of
the issues that I raised in my post (quoted below), which
has gone without a reply for the last couple of weeks.

To summarise my point, if the fitness of a species is to be
understood in the way that I've interpreted it (more
descendant species means greater fitness), then a species
property that increases the likelihood of speciation is
still no more likely to be perpetuated than a property that
decreases it. It would be a mistake to assume that as the
number of species in a genus increases, the number of
individual organisms in that genus will also increase. One
does not entail the other. The number of species in a genus
may increase even as it marches toward extinction. Species
selection understood in these terms cannot therefore explain
any properties of species.

Note that although the number of species may increase as the
number of individuals decreases, this can't go on forever.
The number of species will ultimately be limited by the
number of individual organisms in the genus because you have
to have at least one (or two) individual per species. The
number of species will therefore cease to increase as this
limit is approached.

[email protected] (Huck Turner) wrote in message
news:<[email protected]>...
> Tim Tyler <[email protected]> wrote in message
> news:<[email protected]>...
> > Huck Turner <[email protected]> wrote or quoted:
> [snip]
> > > The only sense my feeble mind can make of the idea of
> > > opposing forces of selection is the case where there
> > > are two opposing forces at the same level. [...]
> >
> > Species-level selection should act to make populations
> > more likely to speciate.
> >
> > For example:
> >
> > * Members of a population should have genitalia that
> > quickly become incompatible with those of their host
> > population if they are isolated from it;
>
> So a species (or set of species) with more easily mutated
> genitalia will tend to give rise to a greater number of
> descendant species and we could say that this is what it
> means for a species property to have greater fitness.
> However, we have no reason to expect the total size of the
> population of individuals to change as a result of the
> speciation event unless the genital mutations also affect
> the chances of individuals reproducing. If it has an
> opposing negative effect on fitness at the level of the
> individual (the case that you take to be interesting), we
> should expect the total size of the population to decrease
> while the number of species in it increases until either a
> more stable species emerges with less mutable genitalia or
> the whole lot of them go extinct.
>
>
> >
> > * They should enjoy (and be well adapted to) living on
> > islands - since islands are where many new species are
> > born;
> >
> > * They should be prone to changes in ploidity;
> >
> > ...and so on.
> >
>
> The above reasoning could be applied to these examples
> too. As the number of species increases, the total number
> of individuals may decrease. This is a bit like dividing a
> cake into thirds, throwing away one third and then
> celebrating that now you have two pieces where before you
> had only one. Continue this process dividing each of the
> pieces into thirds and you'll have less and less cake, but
> more pieces. This is like the Cantor set
> (http://www.mathacademy.com/pr/prime/articles/cantset/).
>
>
> > These may not be the same traits that best serve
> > individual reproduction.
> >
> > Indeed the traits the would best serve species
> > reproduction may positively /hinder/ individual level
> > reproduction. Genitals that are verging on mechanical
> > incompatibility with other members of your species
> > might well make isolated popualitons more likely to
> > speciate - but they might also make it more difficult
> > to get a date.
>
> But if they're isolated already then it won't make matters
> worse if their genitalia are incompatible so there simply
> won't be a competing selection pressure for genital
> compatibility. It won't be present.
>
>
> >
> > Forces on the different levels can pull in different
> > directions - and favour different adaptations.
>
> Doesn't follow.
>
>
> H.
>
> ---
> Like-minds don't notice shared mistakes. Talk to
> someone else.
 
[email protected] (Huck Turner) wrote in
news:[email protected]:

> I'm hoping that someone, if not Tim, can respond to
> some of the issues that I raised in my post (quoted
> below), which has gone without a reply for the last
> couple of weeks.
>
> To summarise my point, if the fitness of a species is to
> be understood in the way that I've interpreted it (more
> descendant species means greater fitness), then a species
> property that increases the likelihood of speciation is
> still no more likely to be perpetuated than a property
> that decreases it. It would be a mistake to assume that as
> the number of species in a genus increases, the number of
> individual organisms in that genus will also increase. One
> does not entail the other. The number of species in a
> genus may increase even as it marches toward extinction.
> Species selection understood in these terms cannot
> therefore explain any properties of species.

A species property that increases the likelihood of
speciation _is_ more likely to be perpetuated _unless_ it is
linked to a lesser chance of survival of the offspring
species - which may be the case, but needs to be explicitly
stated. The question then is whether simply blindly

relation to the success of those species. There is an
inverse relationship between the size of a population and
its likelihood of surviving , so an increase in number of
species without increasing the total population size of all
the species is unlikely to be successful.However there is a
possibility that isolation allows for drift or directed
selection or the Baldwin effect to move the newly isolated
species towards different niches. In this case it is
possible that increasing the number of species will not
result in a concomitant decrease in their survivability,
because the new species will move into new niches and
increase their numbers.

A downside of this type of speciation is that specialist
species are more susceptible to variable conditions than
generalist species, so any increase in overall numbers may
be short-lived. The ideal might be a species that remains a
generalist while spinning off specialist sub- species. This
might select for species with high developmental plasticity
(so a minor change in the genes that control development
could lead to ability to invade a new niche) combined with a
moderate resistance to

inbreeding, so the main population would remain generalist.

Yours,

Bill Morse
 
"John Edser" <[email protected]> wrote in message
news:<[email protected]>... [snip]
> HT:- Note that although the number of species may increase
> as the number of individuals decreases, this can't go on
> forever. The number of species will ultimately be limited
> by the number of individual organisms in the genus because
> you have to have at least one (or two) individual per
> species. The number of species will therefore cease to
> increase as this limit is approached.
>
> JE:- If you wish to make a more abstract view of this
> event, the increase in species is only a relative fitness
> measure but the increase or decrease of Darwinian
> individuals within each species is an absolute fitness
> measure. Thus what you are saying here is that it is
> pointless only increasing a relative gain when it
> constitutes an absolute loss. You win a battle but lose
> the war. This is why nature can only select for an
> _absolute_ fitness increase.

I'm not sure that absolute fitness is enough to explain how
selection occurs. Say there are two variants A and B, both
with high absolute fitness so that their counts are both
increasing in the population. If the fitness of the A
variants is higher relative to the B variants and they can
interbreed, the B variants will become progressively more
diluted in the population possibly disappearing altogether
despite having a high absolute fitness.

> It does not matter how many levels of fitness you suppose,
> only ONE can constitute a valid measure of absolute
> fitness. Just this one level, which empirically is the
> original Darwinian level of selection: the fertile form
> level, constitutes an absolute measure of fitness
> determining all selective events. Only one absolute
> measure of fitness can logically, exist.

Logically? Isn't 'absolute fitness' just whatever we
define it to be?

> This measure is just an absolute assumption. Such
> assumptions contest each other, such that all others
> refute in favour of just one of them.

This is an interesting idea. I'd like to see the details of
this abduction. In particular, I would like to see why
selection at the gene level necessarily leads to a
contradiction in your view.

[snip]
> At its most abstract level, the deletion of absolute
> measures so that "everything is relative" is termed: Post
> Modern Epistemology. It destroys everything it touches
> because it is an utterly absurd view. This is because it
> is 100% non self consistent. The view that everything is
> relative is an absolute assumption. This being the case it
> 100% contradicts itself. This ancient con has been known
> for thousands of years. It is called Epimenides Paradox.
> Neo Darwinian arguments are littered with such
> absurdities.
>

I don't think the notion of 'relative fitness' has any more
to do with relativism (moral or epistemological) than
Einstein's theory of relativity.

Thanks for your interesting comments,

H.
 
Huck Turner <[email protected]> wrote or quoted:
> Tim Tyler <[email protected]> wrote in message
> news:<[email protected]>...

> > Species-level selection should act to make populations
> > more likely to speciate.
> >
> > For example:
> >
> > * Members of a population should have genitalia that
> > quickly become incompatible with those of their host
> > population if they are isolated from it;
>
> So a species (or set of species) with more easily mutated
> genitalia will tend to give rise to a greater number of
> descendant species and we could say that this is what it
> means for a species property to have greater fitness.

Counting species is not as good a measure of fitness as
counting individuals is. Species can vary much more in both
size and longevity then individuals of the same species do.

However # of species is still worth a fair bit.

> However, we have no reason to expect the total size of the
> population of individuals to change as a result of the
> speciation event unless the genital mutations also affect
> the chances of individuals reproducing. [...]

Which usually will happen as a *consequence* of speciation -
different species can better diversify into different niches
- and are less likely to be wiped out by a single pathogen,
predator, competitor or environmental fluctuation.

> > Forces on the different levels can pull in different
> > directions - and favour different adaptations.
>
> Doesn't follow.

It's true - species-level selection and individual level
selection can favour different adaptations.

ISTM that the examples I gave were reasonable ones of it
doing so - or were at least about as reasonable as any other
species-level selection examples would have been.
--
__________
|im |yler http://timtyler.org/ [email protected] Remove
lock to reply.
 
[email protected] (Huck Turner) wrote in message news:<[email protected]>...
> I'm hoping that someone, if not Tim, can respond to
> some of the issues that I raised in my post (quoted
> below), which has gone without a reply for the last
> couple of weeks.
>
> To summarise my point, if the fitness of a species is to
> be understood in the way that I've interpreted it (more
> descendant species means greater fitness), then a species
> property that increases the likelihood of speciation is
> still no more likely to be perpetuated than a property
> that decreases it. It would be a mistake to assume that as
> the number of species in a genus increases, the number of
> individual organisms in that genus will also increase. One
> does not entail the other. The number of species in a
> genus may increase even as it marches toward extinction.
> Species selection understood in these terms cannot
> therefore explain any properties of species.
>
> Note that although the number of species may increase as
> the number of individuals decreases, this can't go on
> forever. The number of species will ultimately be limited
> by the number of individual organisms in the genus because
> you have to have at least one (or two) individual per
> species. The number of species will therefore cease to
> increase as this limit is approached.
>
[Snip the post in question. The summary above raises enough
issues for one reply.]

I believe that you are right that a "successful" species,
measuring success by the number of descendent species, must
probably also increase the number of individuals. Hence, species-
level fitness has to have a big component of individual-
level fitness, thereby calling into question the reality of
species fitness as a concept distinct from individual
fitness. But, I don't think that your argument is decisive
and will forever banish consideration of species selection.
I am coming to realize that fitness is a very slippery
concept. Some random observations:
1. Fisher's notion of species fitness is that a fit
species minimizes its chance of extinction (in all
descendent branches). This notion of species fitness
avoids your argument.
2. A species of pollenating insects, say, that refuses to
spawn new species may well be less fit than one which is
more opportunistic - generating new descendent species
whenever its mutualist flowering plant generates new
descendents. An ability to spawn descendents may be
advantageous to a species even if total population
remains roughly constant over geological time. Even if
some of the descendent species become extinct, they have
cousins to carry on the genus.
3. For any unit of selection, it may be best to think of
fitness, not as a simple number w, but rather as a time
series w(t). For example,

may be tempted to roughly double its short term fitness w(1
generation). But its more foresighted sister looks at w(1
million generations) and

fitness in terms of w(1) applauds the first choice. One who
thinks of fitness in terms of w(1 million) will not be
surprised at the success of the second choice. Which aphid
(and which biologist) is right? They are both right! Both
get the correct answer, but they use different experimental
protocols to check the results. This applies to species
fitness as well as to gene fitness, and it applies to all
levels in between.
4. Sometimes, the argument over levels of selection is a
matter of taste rather than substance. You can
certainly apply a kind of reductionism to study
selection at the lowest possible level. Your low-level
models may be mathematically equivalent to a higher
level model. The question is which kind of math you
prefer - which set of equations provides the most
insight. In my limited experience, the best insights
are provided by looking at a phenomenon from multiple
viewpoints. Hamilton's rb>c can be derived at either
the gene level, the individual level, or a Price-style
group level. Which is best? Well, the gene-level
strikes me as most intuitive, but it isn't as good at
revealing the trade-off between the within-group and
the between-group effects as the Price approach. The
view from the individual level is the most thorny,
involving the wierd notion of "inclusive fitness". But,
IMHO, the effort one expends in puzzling out these
strange accounting rules is well worth the effort.
There is a subtraction of fitness that is received but
not "earned" and an adding back in of fitness earned,
but not received. Once you understand the logic, you
see applications of the trick to other problems.
5. Although it seems that high-level selection can always
be modeled at a low level (given a sufficiently complex
model), it does not seem to be true that low-level
selection can always be modeled successfully at a higher
level. So, the lower level view is, in this sense, more
fundamental. But not, by that, necessarily better than a
multi-level view.
 
"John Edser" <[email protected]> wrote in message
news:<[email protected]>... [snip]
> HT:- Note that although the number of species may increase
> as the number of individuals decreases, this can't go on
> forever. The number of species will ultimately be limited
> by the number of individual organisms in the genus because
> you have to have at least one (or two) individual per
> species. The number of species will therefore cease to
> increase as this limit is approached.

> JE:- If you wish to make a more abstract view of this
> event, the increase in species is only a relative fitness
> measure but the increase or decrease of Darwinian
> individuals within each species is an absolute fitness
> measure. Thus what you are saying here is that it is
> pointless only increasing a relative gain when it
> constitutes an absolute loss. You win a battle but lose
> the war. This is why nature can only select for an
> _absolute_ fitness increase.

HT:- I'm not sure that absolute fitness is enough to explain
how selection occurs. Say there are two variants A and B,
both with high absolute fitness so that their counts are
both increasing in the population. If the fitness of the A
variants is higher relative to the B variants and they can
interbreed, the B variants will become progressively more
diluted in the population possibly disappearing altogether
despite having a high absolute fitness.

JE:- Relative fitness must and is, also be included. This
the default comparison to each other of all parental fitness
totals within the same population.

Each parents total fitness count within one population
(parental absolute fitness) must be compared with every
other before a selective event is completed. Comparing these
counts at any two points in time such that some totals are
completed but others are not will not produce as accurate
account as only comparing completed totals.

> JE:- It does not matter how many levels of fitness you
> suppose, only ONE can constitute a valid measure of
> absolute fitness. Just this one level, which empirically
> is the original Darwinian level of selection: the fertile
> form level, constitutes an absolute measure of fitness
> determining all selective events. Only one absolute
> measure of fitness can logically, exist.

HT:- Logically? Isn't 'absolute fitness' just whatever we
define it to be?

JE:- Yes, but everything "is whatever we define it to be".
What is required by the sciences are _different_ definitions
embedded within _contestable_ theories such that one theory
must refute in favour of another.

> JE:- This measure is just an absolute assumption. Such
> assumptions contest each other, such that all others
> refute in favour of just one of them.

HT:- This is an interesting idea. I'd like to see the
details of this abduction. In particular, I would like to
see why selection at the gene level necessarily leads to a
contradiction in your view.

JE:- Only ONE total count of something can represent the
absolute fitness of whatever you have defined as a selectee.
This one total count is the absolute fitness assumption of
the proposition. If two absolute counts exist then the
proposition can eternally evade refutation. As one refutes,
just go to the other! Multi levels of selection are much
more popular than single levels because of this fact. You
have to be brave to propose only a single level of selection
because now, you _can_ be refuted. Multi levels of selection
can never be refuted because always, as one level fails
another can be substituted. Darwin only implicitly allowed a
single level of selection. I _explicitly_ only allow one
level for the reasons given. In most Neo Darwinian dialog
you will find that gene centric Neo Darwinists who argue the
gene level is the level of selection operating in nature
quietly switch levels as and when it suits them. For
example, Hamilton's b is an organism _group_ level. In one
swoop Hamilton switches from the gene level used to argue
organism fitness altruism (OFA) to classical group
selection, and then back gain to his gene level. This is
just intellectual slight of hand. Dawkins suggests that the
Hamiltonian view is classically Darwinistic when OFA is
entirely gene centric. No classically Darwinistic argument
allows OFA! That is why selection at the group level and
then Hamilton's hypothetical selection at the gene level,
were invented. Imagine running a company that had TWO profit
totals at the end of the financial year, one making a
massive loss and the other a massive profit! Clearly such a
Mad Hatter result reduces accounting to Lunacy. Yet, we are
seriously asked to allow such massive contradictions within
gene centric Neo Darwinistic fitness accounting.

> JE:-
[snip]
> At its most abstract level, the deletion of absolute
> measures so that "everything is relative" is termed: Post
> Modern Epistemology. It destroys everything it touches
> because it is an utterly absurd view. This is because it
> is 100% non self consistent. The view that everything is
> relative is an absolute assumption. This being the case it
> 100% contradicts itself. This ancient con has been known
> for thousands of years. It is called Epimenides Paradox.
> Neo Darwinian arguments are littered with such
> absurdities.

HT:- I don't think the notion of 'relative fitness' has any
more to do with relativism (moral or epistemological) than
Einstein's theory of relativity.

JE:- "Einstein's theory of relativity" was _not_ just
"relative". The testable focus in the Special Theory Of
Relativity was entirely based on the absolute assumption of
the maximum speed of light in a vacuum. This was not
relative to anything so it became a new, unexpected absolute
assumption of physics. "Everything" that we got used to
being constant suddenly became relative to just c. Hence the
title of Einstein's view. Suddenly, mass and time were
reduced to only being variables when in Newton's mechanics
they were constants. Most people have never gotten over the
shock of it. Newton's absolute assumption of the constancy
of time and mass was refuted.

Respectfully,

John Edser Independent Researcher

PO Box 266 Church Pt NSW 2105 Australia

[email protected]
 
[email protected] (Jim Menegay) wrote in message news:<[email protected]>...
> [email protected] (Huck Turner) wrote in message
> news:<[email protected]>...
[snip]
> > To summarise my point, if the fitness of a species is to
> > be understood in the way that I've interpreted it (more
> > descendant species means greater fitness), then a
> > species property that increases the likelihood of
> > speciation is still no more likely to be perpetuated
> > than a property that decreases it. It would be a mistake
> > to assume that as the number of species in a genus
> > increases, the number of individual organisms in that
> > genus will also increase. One does not entail the other.
> > The number of species in a genus may increase even as it
> > marches toward extinction. Species selection understood
> > in these terms cannot therefore explain any properties
> > of species.
> >
> > Note that although the number of species may increase as
> > the number of individuals decreases, this can't go on
> > forever. The number of species will ultimately be
> > limited by the number of individual organisms in the
> > genus because you have to have at least one (or two)
> > individual per species. The number of species will
> > therefore cease to increase as this limit is approached.
> >
> [Snip the post in question. The summary above raises
> enough issues for one reply.]
>
> I believe that you are right that a "successful" species,
> measuring success by the number of descendent species,
> must probably also increase the number of individuals.
> Hence, species-level fitness has to have a big component
> of individual-level fitness, thereby calling into question
> the reality of species fitness as a concept distinct from
> individual fitness. But, I don't think that your argument
> is decisive and will forever banish consideration of
> species selection. I am coming to realize that fitness is
> a very slippery concept. Some random observations:
> 1. Fisher's notion of species fitness is that a fit
> species minimizes its chance of extinction (in all
> descendent branches). This notion of species fitness
> avoids your argument.

Yes, the argument is irrelevant if species fitness is
defined that way.

> 2. A species of pollenating insects, say, that refuses to
> spawn new species may well be less fit than one which
> is more opportunistic - generating new descendent
> species whenever its mutualist flowering plant
> generates new descendents. An ability to spawn
> descendents may be advantageous to a species even if
> total population remains roughly constant over
> geological time. Even if some of the descendent
> species become extinct, they have cousins to carry on
> the genus.

Both Bill Morse and Tim Tyler made similar arguments in
their responses to me and I had a few things to say about
this in my response to Bill. I argued that this seems to be
an issue about the ability to adapt rather than of the
ability to speciate per se and suggested a context in which
that ability might be selected against, but the issues still
aren't clear to me.

> 3. For any unit of selection, it may be best to think of
> fitness, not as a simple number w, but rather as a
> time series w(t). For example,

> may be tempted to roughly double its short term fitness
> w(1 generation). But its more foresighted sister looks at
> w(1 million generations) and

> fitness in terms of w(1) applauds the first choice. One
> who thinks of fitness in terms of w(1 million) will not be
> surprised at the success of the second choice. Which aphid
> (and which biologist) is right? They are both right! Both
> get the correct answer, but they use different
> experimental protocols to check the results. This applies
> to species fitness as well as to gene fitness, and it
> applies to all levels in between.

I think that measuring fitness as increase in numbers over
time is a good methodological step, but I think there are
some problems with the way you have applied it. It hardly
needs to be pointed out that aphids and evolution don't have
foresight so can't literally 'choose' to pursue a short or
long term survival strategy.

It is also worth pointing out that you can measure the
fitness of a single variant sampling at different time
intervals. You might for instance observe a fitness decrease
if you observe numbers from summer to winter, but an
increase if you observe from winter to summer or from one
summer to the next and so on. Fitness might decrease in the
short term, but increase in the longer term or vice versa,
or it might exhibit cyclic behaviour associated with the
seasons, the reproductive cycles of prey, larger scale
climate changes, and so on.

> 4. Sometimes, the argument over levels of selection is a
> matter of taste rather than substance. You can
> certainly apply a kind of reductionism to study
> selection at the lowest possible level. Your low-level
> models may be mathematically equivalent to a higher
> level model. The question is which kind of math you
> prefer - which set of equations provides the most
> insight. In my limited experience, the best insights
> are provided by looking at a phenomenon from multiple
> viewpoints. Hamilton's rb>c can be derived at either
> the gene level, the individual level, or a Price-style
> group level. Which is best? Well, the gene-level
> strikes me as most intuitive, but it isn't as good at
> revealing the trade-off between the within-group and
> the between-group effects as the Price approach. The
> view from the individual level is the most thorny,
> involving the wierd notion of "inclusive fitness".
> But, IMHO, the effort one expends in puzzling out
> these strange accounting rules is well worth the
> effort. There is a subtraction of fitness that is
> received but not "earned" and an adding back in of
> fitness earned, but not received. Once you understand
> the logic, you see applications of the trick to other
> problems.

I see no problem with talking about selection at different
levels if it is just a methodological convenience. It is the
more substantial claims that I'm not so convinced by, the
claims that group, species and individual level theories
account for data that cannot in principle be explained by
gene level theories.

> 5. Although it seems that high-level selection can always
> be modeled at a low level (given a sufficiently
> complex model), it does not seem to be true that low-
> level selection can always be modeled successfully at
> a higher level. So, the lower level view is, in this
> sense, more fundamental. But not, by that, necessarily
> better than a multi-level view.

I agree.

H.
 
William Morse <[email protected]> wrote in message news:<[email protected]>...
> [email protected] (Huck Turner) wrote in
> news:[email protected]:
[snip]
> > To summarise my point, if the fitness of a species is to
> > be understood in the way that I've interpreted it (more
> > descendant species means greater fitness), then a
> > species property that increases the likelihood of
> > speciation is still no more likely to be perpetuated
> > than a property that decreases it. It would be a mistake
> > to assume that as the number of species in a genus
> > increases, the number of individual organisms in that
> > genus will also increase. One does not entail the other.
> > The number of species in a genus may increase even as it
> > marches toward extinction. Species selection understood
> > in these terms cannot therefore explain any properties
> > of species.
>
> A species property that increases the likelihood of
> speciation _is_ more likely to be perpetuated _unless_ it
> is linked to a lesser chance of survival of the offspring
> species - which may be the case, but needs to be
> explicitly stated.

And a property that increases the likelihood of speciation
is _less_ likely to be perpetuated unless it is linked to a
_greater_ chance of survival of the offspring species. You
can phrase it either way, and I'm not sure that one
perspective is more natural than the other.

> The question then is whether simply blindly

> relation to the success of those species.

I think this is the central question. It is not obvious that
properties facilitating speciation should have greater
longevity. This would have to be demonstrated empirically
because it doesn't follow logically. If true, we could then
ask the separate question of whether it is necessary to view
this phenomenon at the species level rather than at the
individual or gene level.

> There is an inverse relationship between the size of a
> population and its likelihood of surviving , so an
> increase in number of species without increasing the total
> population size of all the species is unlikely to be
> successful.However there is a possibility that isolation
> allows for drift or directed selection or the Baldwin
> effect to move the newly isolated species towards
> different niches. In this case it is possible that
> increasing the number of species will not result in a
> concomitant decrease in their survivability, because the
> new species will move into new niches and increase their
> numbers.

I think the issue you're getting at is a bit more general
than speciation. I think it is an issue of evolvability
which may or may not be associated with speciation.

I can imagine a scenario in which a single species inhabits
an environment with two food sources, one of which is
untapped. The population size is limited by the availability
of food so a speciation event that produced a second species
that could take advantage of the other food source would
allow the population to grow. But in this case, whether
speciation occurs or not won't really be important. What is
important is that some mutation arises that gives
individuals the capacity to exploit the other food source
and this may or may not be associated with a speciation
event. The species that were limited by the availability of
the original food source might simply change as a result of
the greater reproductive success of the individuals that
could exploit both sources.

So if we take evolvability in general (understood in terms
of some perhaps nontrivial relationship to mutation rates)
as the important property, we could argue that a species
with greater evolvability would be more inclined to survive
than another when the environment changes, but during a
period of environmental stability, higher mutation rates
would constantly lead individuals away from locally optimal
designs thus presumably favouring a species (or individuals)
with lower mutation rates (evolvability). Would this count
as species/group selection? I don't know.

H.
 
Huck Turner <[email protected]> wrote or quoted:
> William Morse <[email protected]> wrote in message
> news:<[email protected]>...
> > [email protected] (Huck Turner) wrote in

> > The question then is whether simply blindly

> > relation to the success of those species.
>
> I think this is the central question. It is not obvious
> that properties facilitating speciation should have
> greater longevity. This would have to be demonstrated
> empirically because it doesn't follow logically. [...]

It doesn't adress the question of whether you can have too
much speciation - but I should thing a casual look at nature
should tell you that not speciating is either bad or
difficult - since most of the things you see in the world
have numerous close relative species.

As has been mentioned, speciation allows diversification and
subsequent specialisation. This isn't a one-size-fits-all
world - and if you can't do diversification, it's a good bet
that one of your competitors will manage it the trick at no
extra cost.
--
__________
|im |yler http://timtyler.org/ [email protected] Remove
lock to reply.
 
[email protected] (Huck Turner) wrote in message news:<[email protected]>...
> [email protected] (Jim Menegay) wrote in message
> news:<[email protected]>...

[snip areas of agreement, leaving this]

> > 3. For any unit of selection, it may be best to think
> > of fitness, not as a simple number w, but rather as
> > a time series w(t). For example,

> > may be tempted to roughly double its short term fitness
> > w(1 generation). But its more foresighted sister looks
> > at w(1 million generations) and

> > fitness in terms of w(1) applauds the first choice. One
> > who thinks of fitness in terms of w(1 million) will not
> > be surprised at the success of the second choice. Which
> > aphid (and which biologist) is right? They are both
> > right! Both get the correct answer, but they use
> > different experimental protocols to check the results.
> > This applies to species fitness as well as to gene
> > fitness, and it applies to all levels in between.
>
> I think that measuring fitness as increase in numbers over
> time is a good methodological step, but I think there are
> some problems with the way you have applied it. It hardly
> needs to be pointed out that aphids and evolution don't
> have foresight so can't literally 'choose' to pursue a
> short or long term survival strategy.

I choose to be a deliberate heretic here. I claim that
evolution DOES have metaphorical foresight! Our ancestors
had metaphorical foresight! They chose not to be tempted by
short-term advantage, and instead chose (perhaps by sheer
luck) the best long term strategy. (Of course, they had to
be careful not to be so oblivious of the short term as to
drive their line to extinction.)

The often repeated dogma in evolutionary theory - that
"foresight" is impossible - is simply a mistake, IMHO. It is
as as silly as the anti-dogma that evolution cannot be
"creative". Certainly, it is not the case that today's
organisms, on average, have either foresight or creativity.
But it most certainly IS the case, after the fact, that we
can look back and say that of yesteryear's organisms, the
tiny fraction that have left surviving descendents exhibited
both creativity and foresight.

The process of natural selection is not purely a short term
process! It appears to be a short term process if you take
short term measurements. But it appears to be a long term
process if you take long term measurements.

As a thought experiment, ask yourself just what is the
proper "term" for understanding selection. Is it a single
generation? Ten generations, say? A year? A century? What
criterion do you use in making this judgement? Clearly, for
a Monarch butterfly, a year is a better choice than a
generation. But for a cicada, a generation is better than a
year. Are there species in which fitness is best measured
over a time scale of a few ice-ages? What advantage does a
short time scale have over a long one?
 
Jim Menegay <[email protected]> wrote or quoted:
> [email protected] (Huck Turner) wrote in message
> news:<[email protected]>...
> > [email protected] (Jim Menegay) wrote in message
> > news:<[email protected]>...

> > > For example, an aphid considering switching over to an

> > > double its short term fitness w(1 generation). But its
> > > more foresighted sister looks at w(1 million
> > > generations) and

> > > thinks of fitness in terms of w(1) applauds the first
> > > choice. One who thinks of fitness in terms of w(1
> > > million) will not be surprised at the success of the
> > > second choice. Which aphid (and which biologist) is
> > > right? They are both right! Both get the correct
> > > answer, but they use different experimental protocols
> > > to check the results. This applies to species fitness
> > > as well as to gene fitness, and it applies to all
> > > levels in betwen.
> >
> > I think that measuring fitness as increase in numbers
> > over time is a good methodological step, but I think
> > there are some problems with the way you have applied
> > it. It hardly needs to be pointed out that aphids and
> > evolution don't have foresight so can't literally
> > 'choose' to pursue a short or long term survival
> > strategy.
>
> I choose to be a deliberate heretic here. I claim that
> evolution DOES have metaphorical foresight! Our ancestors
> had metaphorical foresight! They chose not to be tempted
> by short-term advantage, and instead chose (perhaps by
> sheer luck) the best long term strategy. (Of course, they
> had to be careful not to be so oblivious of the short term
> as to drive their line to extinction.)
>
> The often repeated dogma in evolutionary theory - that
> "foresight" is impossible - is simply a mistake, IMHO. It
> is as as silly as the anti-dogma that evolution cannot be
> "creative". Certainly, it is not the case that today's
> organisms, on average, have either foresight or
> creativity. But it most certainly IS the case, after the
> fact, that we can look back and say that of yesteryear's
> organisms, the tiny fraction that have left surviving
> descendents exhibited both creativity and foresight.

The evolutionary process has more than just metaphorical
foresight - now that is has created intelligent life. It has
completely literal foresight - and can make predictions.

Aphids don't have foresight. However they /may/ have
"metaphorical foresight" - assuming that means that they
behave as though they can predict the future.

The aphid's future predictions will typically be that the
future will be much the same as how their ancestors
experienced the past. Most of the time, that's a pretty good
bet about how things are likely to turn out.
--
__________
|im |yler http://timtyler.org/ [email protected] Remove
lock to reply.
 
Tim Tyler <[email protected]> wrote in message news:<[email protected]>...
> Huck Turner <[email protected]> wrote or quoted:
> > William Morse <[email protected]> wrote in message
> > news:<[email protected]>...
> > > [email protected] (Huck Turner) wrote in
>
> > > The question then is whether simply blindly

> > > relation to the success of those species.
> >
> > I think this is the central question. It is not obvious
> > that properties facilitating speciation should have
> > greater longevity. This would have to be demonstrated
> > empirically because it doesn't follow logically. [...]
>
> It doesn't adress the question of whether you can have too
> much speciation - but I should thing a casual look at
> nature should tell you that not speciating is either bad
> or difficult - since most of the things you see in the
> world have numerous close relative species.

This sounds suspiciously like Dr. Pangloss's "as all things
have been created for some end, they must necessarily be
created for the best end".

Perhaps speciation is just something that happens because
species get physically isolated from each other from time to
time. We don't need an adaptive explanation for it. It may
just be a group-level spandrel.

H.
 
[email protected] (Jim Menegay) wrote in message news:<[email protected]>...
> [email protected] (Huck Turner) wrote in message
> news:<[email protected]>...
> > [email protected] (Jim Menegay) wrote in message
> > news:<[email protected]>...
>
> [snip areas of agreement, leaving this]
>
> > > 3. For any unit of selection, it may be best to think
> > > of fitness, not as a simple number w, but rather
> > > as a time series w(t). For example,

> > > may be tempted to roughly double its short term
> > > fitness w(1 generation). But its more foresighted
> > > sister looks at w(1 million generations) and

> > > fitness in terms of w(1) applauds the first choice.
> > > One who thinks of fitness in terms of w(1 million)
> > > will not be surprised at the success of the second
> > > choice. Which aphid (and which biologist) is right?
> > > They are both right! Both get the correct answer, but
> > > they use different experimental protocols to check the
> > > results. This applies to species fitness as well as to
> > > gene fitness, and it applies to all levels in between.
> >
> > I think that measuring fitness as increase in numbers
> > over time is a good methodological step, but I think
> > there are some problems with the way you have applied
> > it. It hardly needs to be pointed out that aphids and
> > evolution don't have foresight so can't literally
> > 'choose' to pursue a short or long term survival
> > strategy.
>
> I choose to be a deliberate heretic here. I claim that
> evolution DOES have metaphorical foresight! Our ancestors
> had metaphorical foresight! They chose not to be tempted
> by short-term advantage, and instead chose (perhaps by
> sheer luck) the best long term strategy. (Of course, they
> had to be careful not to be so oblivious of the short term
> as to drive their line to extinction.)
>
> The often repeated dogma in evolutionary theory - that
> "foresight" is impossible - is simply a mistake, IMHO. It
> is as as silly as the anti-dogma that evolution cannot be
> "creative". Certainly, it is not the case that today's
> organisms, on average, have either foresight or
> creativity. But it most certainly IS the case, after the
> fact, that we can look back and say that of yesteryear's
> organisms, the tiny fraction that have left surviving
> descendents exhibited both creativity and foresight.

What you are saying is similar to what Dennett says about
"as-if intentionality" in biological traits and to Dawkins's
re-evaluation of the argument from the appearance of design
as evidence of cumulative selection, but these authors still
argue that evolution is just a hill climbing process at the
whim of short-term pressures.

>
> The process of natural selection is not purely a short
> term process! It appears to be a short term process if you
> take short term measurements. But it appears to be a long
> term process if you take long term measurements.

Not necessarily. Some design 'choices' look pretty stupid a
few million years down the track. The usual example is the
vertebrate eye which is committed to a apparently
suboptimal design with photoreceptors pointing away from
the light. There might be some undiscovered functional
reason for it being this way, but this is doubtful because
the cephalopod eye (which evolved independently but which
is otherwise extremely similar) has the receptors round the
way you'd expect.

But then again perhaps if we wait another 100 million years
the vertebrate eye will be an optimal solution to some yet-to-be-
realised lighting conditions. Even a stopped clock is right
twice a day!

>
> As a thought experiment, ask yourself just what is the
> proper "term" for understanding selection. Is it a single
> generation? Ten generations, say? A year? A century? What
> criterion do you use in making this judgement? Clearly,
> for a Monarch butterfly, a year is a better choice than a
> generation. But for a cicada, a generation is better than
> a year. Are there species in which fitness is best
> measured over a time scale of a few ice-ages? What
> advantage does a short time scale have over a long one?

I think you are confusing your metaphor with the real
thing. No species ever literally makes a decision about how
far ahead to consider. A design 'choice' might _look_
clever after some long delay, but you can't predict how
long that delay will be or that a given 'choice' will ever
look clever.

H.
 
Huck Turner <[email protected]> wrote or quoted:
> [email protected] (Jim Menegay) wrote in message
> news:<[email protected]>...

> > The often repeated dogma in evolutionary theory - that
> > "foresight" is impossible - is simply a mistake, IMHO.
> > It is as as silly as the anti-dogma that evolution
> > cannot be "creative". Certainly, it is not the case that
> > today's organisms, on average, have either foresight or
> > creativity. But it most certainly IS the case, after the
> > fact, that we can look back and say that of yesteryear's
> > organisms, the tiny fraction that have left surviving
> > descendents exhibited both creativity and foresight.
>
> What you are saying is similar to what Dennett says about
> "as-if intentionality" in biological traits and to
> Dawkins's re-evaluation of the argument from the
> appearance of design as evidence of cumulative selection,
> but these authors still argue that evolution is just a
> hill climbing process at the whim of short-term pressures.

These days evolution is a hillclimbing exercise on a fitness
landscape *with* directed mutation.

Directed mutation represents a substantial addition.

Even *before* directed mutations were readily available, it
was /still/ possible for foresight to influence the course
of evolution.

Basically organisms with large brains could deform the
fitness landscape in a manner that facilitated access to
the unlikely, foresight-requiring solution (which would
then be reached via effects such as mate choice or the
Baldwin effect).
--
__________
|im |yler http://timtyler.org/ [email protected] Remove
lock to reply.
 
[email protected] (Huck Turner) wrote in message news:<[email protected]>...
> A design 'choice' might _look_ clever after some long
> delay, but you can't predict how long that delay will be
> or that a given 'choice' will ever look clever.

I'm not trying to predict anything. But, to paraphrase
someone who shall go nameless, "That which is mostly
observed looks pretty damn clever in retrospect". I am
offering a partial explanation of what IS, not a prediction
of what WILL BE.
 
> As a thought experiment, ask yourself just what is the
> proper "term" for understanding selection. Is it a single
> generation? Ten generations,
say?
> A year? A century? What criterion do you use in making
> this judgement? Clearly, for a Monarch butterfly, a year
> is a better choice than a >>
generation.
> But for a cicada, a generation is better than a year. Are
> there species
in
> which fitness is best measured over a time scale of a few
> ice-ages? What advantage does a short time scale have over
> a long one?

HT:- I think you are confusing your metaphor with the real
thing. No species ever literally makes a decision about how
far ahead to consider. A design 'choice' might _look_
clever after some long delay, but you can't predict how
long that delay will be or that a given 'choice' will ever
look clever.

JE:- The absolute Darwinian fitness of any parent is the
total number of fertile forms that parent reproduces into
one population. Thus the time unit is _exact_, within
Darwinism. It is the time required for a parent to complete
a reproductive total for the number of fertile_ forms it has
reproduced into one population. Depending on the species
concerned this time unit can be quite large, e.g. seeds may
remain dormant for centuries. Also, because of migration, a
parent can reproduce into separate populations providing a
different absolute fitness for that parent in each
population.

Respectfully,

John Edser Independent Researcher

PO Box 266 Church Pt NSW 2105 Australia

[email protected]
 
[email protected] (Huck Turner) wrote in
news:[email protected]:

> [email protected] (Jim Menegay) wrote in message
> news:<[email protected]>...

>> 4. Sometimes, the argument over levels of selection is a
>> matter of taste rather than substance. You can
>> certainly apply a kind of reductionism to study
>> selection at the lowest possible level. Your low-
>> level models may be mathematically equivalent to a
>> higher level model. The question is which kind of
>> math you prefer - which set of equations provides the
>> most insight. In my limited experience, the best
>> insights are provided by looking at a phenomenon from
>> multiple viewpoints. Hamilton's rb>c can be derived
>> at either the gene level, the individual level, or a
>> Price-style group level. Which is best? Well, the gene-
>> level strikes me as most intuitive, but it isn't as
>> good at revealing the trade-off between the within-
>> group and the between-group effects as the Price
>> approach. The view from the individual level is the
>> most thorny, involving the wierd notion of "inclusive
>> fitness". But, IMHO, the effort one expends in
>> puzzling out these strange accounting rules is well
>> worth the effort. There is a subtraction of fitness
>> that is received but not "earned" and an adding back
>> in of fitness earned, but not received. Once you
>> understand the logic, you see applications of the
>> trick to other problems.

> I see no problem with talking about selection at different
> levels if it is just a methodological convenience. It is
> the more substantial claims that I'm not so convinced by,
> the claims that group, species and individual level
> theories account for data that cannot in principle be
> explained by gene level theories.

Sorry for the delay in responding - I have been sidetracked
by other discussions.

While I am willing to entertain the idea of multiple levels
of selection, I also would like to see more examples of
higher level selection.With this in mine, I would like to
note a mention by Wilson in "Sociobiology" of something I
had previously noted - a lot of animals don't spend all that
much time working hard at surviving. If you watch the animal
specials, you note that many animals spend a lot of time
doing nothing.

Wilson's explanation is that this represents a form of group
selection: if the animals were working hard to survive
during normal times, they would starve when times got bad.
But there are some problems with this. One would think that
an animal that hustled all the time during good times and
bad would still outcompete the one that only hustled during
bad times. Wilson falls back on the argument that
populations with boom and bust cycles are more likely to go
extinct than stable populations - basically species level
selection. But some species have survived for many millions
of years - why don't they continually tend to revert to boom
and bust behavior if that is favored by individual
selection?

Now part of the explanation may lie in the "worst year in
ten" effect. That is, there is sufficient variation in
conditions that "bad" years will occur several times during
an animal's typical life span. In that case I would think
you would see a higher level of "hustle" in animals with a
shorter life span, and that may in fact be true - I am not
enough of a naturalist to know (note that this is different
from the higher metabolic rate seen in small animals due to
scaling effects).

Yours,

Bill Morse

Note Wilson's explanation of a point I had previously raised
- most animals do not spend all their time working hard at
surviving. Why not? If due to worst year in ten, how does
this get passed on to short term species.
 
William Morse <[email protected]> wrote or quoted:

> Sorry for the delay in responding - I have been
> sidetracked by other discussions.
>
> While I am willing to entertain the idea of multiple
> levels of selection, I also would like to see more
> examples of higher level selection.With this in mine, I
> would like to note a mention by Wilson in "Sociobiology"
> of something I had previously noted - a lot of animals
> don't spend all that much time working hard at surviving.
> If you watch the animal specials, you note that many
> animals spend a lot of time doing nothing.
>
> Wilson's explanation is that this represents a form of
> group selection: if the animals were working hard to
> survive during normal times, they would starve when times
> got bad. But there are some problems with this. One would
> think that an animal that hustled all the time during good
> times and bad would still outcompete the one that only
> hustled during bad times. Wilson falls back on the
> argument that populations with boom and bust cycles are
> more likely to go extinct than stable populations -
> basically species level selection. But some species have
> survived for many millions of years - why don't they
> continually tend to revert to boom and bust behavior if
> that is favored by individual selection?

I believe spiders spend a fair bit of their time doing
nothing.

AFAICS, they are conserving their energy for the next
time they have to do battle, seek out a mate or repair
their traps.

Between them, limited energy resources and the need for
recuperation and digestion probably explain much of the
observed levels of physical inactivity.
--
__________
|im |yler http://timtyler.org/ [email protected] Remove
lock to reply.
 
You must log in or register to reply here.
Top Bottom Back