Ecological Inheritance

[Jochen Fromm]

>
> The new point they have made is that the Niche and the Gene are two
> things that can be inherited, for which the rules of inheritence
> differ, and for which they can be formalized in some cases, so you
> have committed to specific dynamical models, as opposed to just being
> descriptive.  [..]  It seems like the place
> Feldman et al have gained a pawn on the problem is expanding this to
> include models of inherited environments within a similarly specific
> framework, to see how that alters the evolution of the distributions.

How can there be an inheritance in environments ?
What exactly is inherited in "ecological inheritance" ?
There is no ecological replicator. You can only construct
an ecological "inheritance tree" which is based on artificial
classifications.

There seems to be a kind of natural duality between
species and ecological niches, firms and economic niches,
job holders and jobs, and between populations and habitats.
What is the special approach from Feldman et al ? They
say niches can be described by inheritance trees, because
species are subject to inheritance, and species are adapted
to niches ? All species together form a big tree of life (see
http://www.sciencemag.org/feature/data/tol/ ), and
each species lives in a certain niche, so we have a
tree of niches, too ?

[Roger Critchlow-2]

Jochen Fromm wrote:

>>The new point they have made is that the Niche and the Gene are two
>>things that can be inherited, for which the rules of inheritence
>>differ, and for which they can be formalized in some cases, so you
>>have committed to specific dynamical models, as opposed to just being
>>descriptive.  [..]  It seems like the place
>>Feldman et al have gained a pawn on the problem is expanding this to
>>include models of inherited environments within a similarly specific
>>framework, to see how that alters the evolution of the distributions.
>
>
> How can there be an inheritance in environments ?
> What exactly is inherited in "ecological inheritance" ?
> There is no ecological replicator. You can only construct
> an ecological "inheritance tree" which is based on artificial
> classifications.

I definitely need to read this book.

Well, a tree is barely a reasonable structure for representing organic
inheritance.  The inheritance relations of sexual species require a
lattice.  Exactly how we're going to sort out the lateral transfer
relations among microbes is an open question, and opening still further
now that we're getting the environmental sequencing experiments that are
showing how few of the organisms present have ever been censused and how
many lateral transfer paths there are.

The niche inheritance relation will be further complicated by the action
of causes at different time scales.

> There seems to be a kind of natural duality between
> species and ecological niches, firms and economic niches,
> job holders and jobs, and between populations and habitats.
> What is the special approach from Feldman et al ? They
> say niches can be described by inheritance trees, because
> species are subject to inheritance, and species are adapted
> to niches ? All species together form a big tree of life (see
> http://www.sciencemag.org/feature/data/tol/ ), and
> each species lives in a certain niche, so we have a
> tree of niches, too ?

One of the most striking examples of niche construction that I found
while googling the term is one which was studied by Rene Dubos many
years ago: the mammalian gut and the bacterial flora that live in it.
The morphology of the intestinal wall which is medically recognized as
"normal" is only achieved when a "normal" ecosystem of gut bacteria is
in residence.  The gut wall and the gut ecosystem co-develop into
different steady states depending on the initial culture of flora
transferred to the infant and the dietary practices of the culture.  A
"normal" morphology in one culture might appear as a pathology in
another culture.

-- rec --

[David Eric Smith]

Hi Jochen,

> > The new point they have made is that the Niche and the Gene are two
> > things that can be inherited, for which the rules of inheritence
> > differ, and for which they can be formalized in some cases, so you
> > have committed to specific dynamical models, as opposed to just being
> > descriptive.  [..]  It seems like the place
> > Feldman et al have gained a pawn on the problem is expanding this to
> > include models of inherited environments within a similarly specific
> > framework, to see how that alters the evolution of the distributions.
>
> How can there be an inheritance in environments ?
> What exactly is inherited in "ecological inheritance" ?
> There is no ecological replicator. You can only construct
> an ecological "inheritance tree" which is based on artificial
> classifications.

Yes, over the last couple of days I have been thinking about how I
would have wanted to lay this out more carefully.  Inheritence as I
was using the term need not mean duplication or replication, with the
replicate instantiating whatever is inherited.  Biological inheritence
through those mechanisms is one instance of inheritence, in the sense
that a pattern is preserved through time as other details of
population structure change.  But mere physical preservation is also
inheritence in the looser sense I intended.

The way I probably could have said it better was this: Why _not_ just
talk about the "extended phenotype"?  In a standard population-genetic
model, selection acts by determining whose phenotype will be
reproduced before the individual dies.  No matter what, the
instantiation of the phenotype by each individual dies in each
generation.  The interesting properties of that phenotype survive only
if represented in the offspring, so here the phenotype is passed
through time only via reproduction, which means that the number of its
instances is exactly the same as the number of individuals who are,
themselves, then subject to selection.  There can also be variation at
the reproduction stage, but I suspect the more important constraint is
that the redundancy of the instantiation of the phenotype is closely
tied to the redundancy of the units of selection, and that is an
extremely specific model of boundaries for properties within the
population.

But a niche that some individual actively participates in shaping
doesn't vanish at the death of the individual, getting reproduced with
variation in the body of the individual's offspring.  So it need not
satisfy the same constraints and channels of passage through time that
the population-genetic models take as starting structure.  

I think that the mathematical thing the Feldman etal book want to
capture is those aspects of the niche that are "as manufactured" as
the rest of the phenotype, but that pass to the next generations
without the death/reproduction interface, and thus not necessarily
with a redundancy in proportion to the number of dead/reproduced
individuals.  (For Niches, suddently it isn't even obviously important
that there is a measure of "redundancy" that is especially important.
Perhaps volume and physical inertia instead.)  

> There seems to be a kind of natural duality between
> species and ecological niches, firms and economic niches,
> job holders and jobs, and between populations and habitats.

Yes, I think this is a deep statement.  My worry in the Feldman etal
development (I work on metabolism) is that the very act of being alive
is the act of impacting the environment, so it is very hard not to tie
every living act to something inherited outside the individual.  In
some sense, Niche Construction seems to have a Goldilocks point where
you don't try to do too much with it: look at the metabolically
neutral and relatively decoupled innovations by species, and ask where
a few details modify the strict phenotype-only models for persistence
through time, and decouple the constraints on persistence from those
on the unit of selection.

> What is the special approach from Feldman et al ? They
> say niches can be described by inheritance trees, because
> species are subject to inheritance, and species are adapted
> to niches ? All species together form a big tree of life (see
> http://www.sciencemag.org/feature/data/tol/ ), and
> each species lives in a certain niche, so we have a
> tree of niches, too ?

Hmm.  

Eric