niche construction

Kent · September 13, 2012, 12:43am

Hi to CSGNET and PCT modelers,

I was teaching a seminar course with a biologist last spring, so I've been reading a lot of biology lately, and a new development in evolutionary theory has caught my attentionï¿½something called "niche construction," which seems closely related to PCT.

This new theory is an extension of standard evolutionary theory, which is all about how natural selection happens, how factors in the environment of a population of organisms give a reproductive advantage to some combinations of genes over other genes, so that organisms with those genes have more descendants, and thus the gene pool of that species will change over time to include a larger proportion of the advantageous genes, so that evolution has taken place. In standard evolutionary theory, the causal arrow is all one way, from changes in the environment to changes in the genetic makeup of organisms, or E->O.

The biologists and evolutionary anthropologists interested in the new theory (the two biggest names seem to be a couple of Brits: F. John Odling-Smee from Cambridge University and Kevin N. Laland from the University of St. Andrews) have observed that the behavior of organisms can have a big impact on their own environments, either by compensating for random changes in their own environments (which sounds a lot like control to me!) or moving to a different environment, when the environment they're in is not to their liking (more control). The theorists then hypothesize that these changes that organisms make in their own environments, which they call niche construction, can have an impact on the course of evolution. In other words, they're interested in the causal arrows from the organism to the environment (O->E), as well as the feedback effects in this causal loop, E->O->E.

These biological theorists cite work by a group of ecologists, Jones, Lawton, and Shanuck, who talk about "organisms as ecosystem engineers" and focus on such animals as beavers, whose dam-building activities can transform local ecosystems. Odling-Smee and Laland give a variety of other examples, as well, of organisms whose niche-constructing activities can have big impacts on their local environments: leaf-cutter ants that build enormous nests and practice a kind of agriculture, raising a crop of mold that they eat; bacteria that produce bacteriocins, chemicals lethal to other bacteria in the neighborhood but not to themselves; earthworms that have a breathing apparatus inherited from their aquatic worm ancestors but can live in underground because of their own tunneling, urine, and dragging of leaf litter below ground, which changes the soil dramatically enough that they can survive and thrive in it; and dozens of other examples. The changes that these organisms make in their own environments, according to Odling-Smee and Laland, have to have a big impact on the evolution of subsequent generations of their species, as well as on the evolution of other species in the immediate ecosystem. Thus, evolutionary theorists should be paying attention to these feedback effects.

This theory of niche construction really caught my interest, because it seems to dovetail nicely with my own recent work on "stabilization of the environment." As I've been arguing in the essays I've written about stabilization, we in the PCT community might do well to emphasize that control of perceptions has some real and important impacts on the environments in which humans are doing their controlling, and that for most people most of the time, the purpose of their control efforts is to transform their own environments to match their own reference conditions and then maintain those preferred conditions as best they can.

Unfortunately, from my point of view, Odling-Smee and Laland themselves seem unaware of PCT and its relevance to their work, and I don't get the impression that they've really thought through the idea of behavior as control. They do say that "natural selection has furnished organisms with onboard guidance systems that allow them to learn about their environment and, within some constraints, to adjust their behavior accordingly during their lives", and they compare the behavioral process to the on-board guidance systems of smart missiles (Odling-Smee et al. 2003, p. 256). But then they go on to describe those onboard guidance systems as working by the "law of effect", which "states that actions that are followed by a positive outcome are likely to be repeated, while those followed by a negative outcome will be eliminated" (p. 256) (sounds a lot like behaviorism to me).

In any case, their recognition that organisms have purposes, that they affect their own environments, and that the causal arrows aren't just one way from environment to organism, all seem like steps in the right direction to me. Gary Cziko's books have already done a nice job of relating PCT to evolutionary theory, but this new development seems to tie it even closer.

Here are some relevant citations:

Odling-Smee, F. John, Kevin N. Laland and Marcus W. Feldman. 2003. Niche Construction: The Neglected Process in Evolution. Princeton, NJ: Princeton University Press.

Odling-Smee, F. John and Kevin N. Laland. 2009. ï¿½Cultural Niche Construction: Evolutionï¿½s Cradle of Language.ï¿½ Pp. 99-121 in The Prehistory of Language, edited by Rudolf Botha and Chris Knight. Oxford: Oxford University Press.

Odling-Smee, John. 2009. ï¿½Niche Construction in Evolution, Ecosystems and Developmental Biology.ï¿½ Pp. 69-91 in Mapping the Future of Biology: Evolving Concepts and Theories, edited by Anouk Barbarousse, Michel Morange and Thomas Pradeau. New York: Springer.

Jones, Clive. G., Lawton, John. H. and Moshe Shachak. 1994. "Organisms as Ecosystem Engineers." Oikos 69:373-386.

Jones, Clive. G., Lawton, John. H. and Moshe Shachak. 1997. "Positive and Negative Effects of Organisms as Physical Ecosystem Engineers." Ecology 78(7):1946ï¿½1957.

Kent

Martin_Lewitt · September 15, 2012, 3:42am

This theory doesn't seem that new. Dawkin's "Extended Phenotype"
discusses genes for things outside the organism, such as nests, bowers,
burrows, etc. Another concept that appears related is "niche reduction",
e.g., evolution has constructed an environment of parental care, which
reduces the need for adaptations to the intermediate environmental niches
between infancy and adulthood.

-- Martin L

···

On 9/12/12 6:43 PM, "McClelland, Kent" <MCCLEL@GRINNELL.EDU> wrote:

Hi to CSGNET and PCT modelers,

I was teaching a seminar course with a biologist last spring, so I've
been reading a lot of biology lately, and a new development in
evolutionary theory has caught my attentionï¿½something called "niche
construction," which seems closely related to PCT.

This new theory is an extension of standard evolutionary theory, which is
all about how natural selection happens, how factors in the environment
of a population of organisms give a reproductive advantage to some
combinations of genes over other genes, so that organisms with those
genes have more descendants, and thus the gene pool of that species will
change over time to include a larger proportion of the advantageous
genes, so that evolution has taken place. In standard evolutionary
theory, the causal arrow is all one way, from changes in the environment
to changes in the genetic makeup of organisms, or E->O.

The biologists and evolutionary anthropologists interested in the new
theory (the two biggest names seem to be a couple of Brits: F. John
Odling-Smee from Cambridge University and Kevin N. Laland from the
University of St. Andrews) have observed that the behavior of organisms
can have a big impact on their own environments, either by compensating
for random changes in their own environments (which sounds a lot like
control to me!) or moving to a different environment, when the
environment they're in is not to their liking (more control). The
theorists then hypothesize that these changes that organisms make in
their own environments, which they call niche construction, can have an
impact on the course of evolution. In other words, they're interested in
the causal arrows from the organism to the environment (O->E), as well as
the feedback effects in this causal loop, E->O->E.

These biological theorists cite work by a group of ecologists, Jones,
Lawton, and Shanuck, who talk about "organisms as ecosystem engineers"
and focus on such animals as beavers, whose dam-building activities can
transform local ecosystems. Odling-Smee and Laland give a variety of
other examples, as well, of organisms whose niche-constructing activities
can have big impacts on their local environments: leaf-cutter ants that
build enormous nests and practice a kind of agriculture, raising a crop
of mold that they eat; bacteria that produce bacteriocins, chemicals
lethal to other bacteria in the neighborhood but not to themselves;
earthworms that have a breathing apparatus inherited from their aquatic
worm ancestors but can live in underground because of their own
tunneling, urine, and dragging of leaf litter below ground, which changes
the soil dramatically enough that they can survive and thrive in it; and
dozens of other examples. The changes that these organisms make in their
own environments, according to Odling-Smee and Laland, have to have a big
impact on the evolution of subsequent generations of their species, as
well as on the evolution of other species in the immediate ecosystem.
Thus, evolutionary theorists should be paying attention to these feedback
effects.

This theory of niche construction really caught my interest, because it
seems to dovetail nicely with my own recent work on "stabilization of the
environment." As I've been arguing in the essays I've written about
stabilization, we in the PCT community might do well to emphasize that
control of perceptions has some real and important impacts on the
environments in which humans are doing their controlling, and that for
most people most of the time, the purpose of their control efforts is to
transform their own environments to match their own reference conditions
and then maintain those preferred conditions as best they can.

Unfortunately, from my point of view, Odling-Smee and Laland themselves
seem unaware of PCT and its relevance to their work, and I don't get the
impression that they've really thought through the idea of behavior as
control. They do say that "natural selection has furnished organisms with
onboard guidance systems that allow them to learn about their environment
and, within some constraints, to adjust their behavior accordingly during
their lives", and they compare the behavioral process to the on-board
guidance systems of smart missiles (Odling-Smee et al. 2003, p. 256). But
then they go on to describe those onboard guidance systems as working by
the "law of effect", which "states that actions that are followed by a
positive outcome are likely to be repeated, while those followed by a
negative outcome will be eliminated" (p. 256) (sounds a lot like
behaviorism to me).

In any case, their recognition that organisms have purposes, that they
affect their own environments, and that the causal arrows aren't just one
way from environment to organism, all seem like steps in the right
direction to me. Gary Cziko's books have already done a nice job of
relating PCT to evolutionary theory, but this new development seems to
tie it even closer.

Here are some relevant citations:

Odling-Smee, F. John, Kevin N. Laland and Marcus W. Feldman. 2003. Niche
Construction: The Neglected Process in Evolution. Princeton, NJ:
Princeton University Press.

Odling-Smee, F. John and Kevin N. Laland. 2009. ï¿½Cultural Niche
Construction: Evolutionï¿½s Cradle of Language.ï¿½ Pp. 99-121 in The
Prehistory of Language, edited by Rudolf Botha and Chris Knight. Oxford:
Oxford University Press.

Odling-Smee, John. 2009. ï¿½Niche Construction in Evolution, Ecosystems and
Developmental Biology.ï¿½ Pp. 69-91 in Mapping the Future of Biology:
Evolving Concepts and Theories, edited by Anouk Barbarousse, Michel
Morange and Thomas Pradeau. New York: Springer.

Jones, Clive. G., Lawton, John. H. and Moshe Shachak. 1994. "Organisms as
Ecosystem Engineers." Oikos 69:373-386.

Jones, Clive. G., Lawton, John. H. and Moshe Shachak. 1997. "Positive and
Negative Effects of Organisms as Physical Ecosystem Engineers." Ecology
78(7):1946ï¿½1957.

Kent

Bill_Powers1 · September 15, 2012, 2:28pm

[From Bill Powers (2012.0-9.15.0742 MDT)]

This theory doesn’t seem that
new. Dawkin’s “Extended Phenotype”

discusses genes for things outside the organism, such as nests,
bowers,

burrows, etc.

Doesn’t this concept skip over something important? The genes can’t, of
course, specify anything outside an organism. Genes specify patterns,
proteins, organizations, inside organisms. What is skipped over is the
mechanism that can make it appear, superficially, that the gene
itself systematically affects something external. This makes the
“extended phenotype” idea into a metaphor, not a fact. When we
try to deal with “is” rather than “as if it is”, we
have to focus on the actual physical system and try to understand how it
works.
There’s a great deal of fuzzy science that works in terms of metaphors, I
would guess because nobody has any idea of how these superficial
appearances are generated. But we shouldn’t allow that limitation to fool
us into thinking that this sort of image actually explains anything. It’s
a poetic description. Its main striking feature is that it implies
something that we know is pretty certainly not true. That’s what calls
our attention to it.

Another example is “retroactive inhibition,” a descriptive term
that makes it seem as if one stimulus can reach back through time to
alter the effect of a previous one on behavior. The only virtue of that
term is that it signals us, with clumsy subtlety, that the appearance is
NOT what is actually happening, because we know that backward causation
through time does not happen. The term should really be expanded to
“What seems to look like retroactive inhibition but is not.”
Similarly for “What seems to work like an extended phenotype but
clearly can’t be that.”

Another concept that
appears related is “niche reduction”,

e.g., evolution has constructed an environment of parental care,
which

reduces the need for adaptations to the intermediate environmental
niches

between infancy and adulthood.

Again, evolution has done no such thing, and can’t do that sort of thing.
Evolution consists of alterations in organisms, not environments. It is
the evolved organisms that alter environments according to the laws of
perceptual control theory and reorganization. And despite the purposive
language, evolution does not have a goal of reducing some objective need
for adaptations, even though an organism may change its organization so
that it can accomplish certain goals without having to develop the
ability to adapt to an environment (see LCS3, Demo5-1).

The metaphorical language allows us to give the appearance of
understanding something without actually doing so. It encourages sloppy
thinking and misleading communication. I’m against it in general, though
as for all stylistic rules, communication is sometimes improved by
breaking the rule. Sometimes.

Best,

Bill P.

···

At 09:42 PM 9/14/2012 -0600, Martin Lewitt wrote:

Martin_Lewitt · September 15, 2012, 6:11pm

[Martin Lewitt 2012 Sep 15 1155am]

It is not that big a mystery. Genes can specify things outside an organism, by specifying the behaviors that result it them. The selective advantage of the genes for those behaviors is that they result in a nest. A lot of what Dawkin’s does, is explain why these metaphors work. When we understand why metaphors work, they become productive conceptual conveniences, equivalent to the underlying explanation.

Humans are tropical animals that have extended our range by creating or finding tropical microenvironments for ourselves. In our case it is general intelligence and propensities that the genes specify.

– Martin L

···

On 9/15/12 8:28 AM, “Bill Powers” powers_w@FRONTIER.NET wrote:

[From Bill Powers (2012.0-9.15.0742 MDT)]

At 09:42 PM 9/14/2012 -0600, Martin Lewitt wrote:

This theory doesn’t seem that
new. Dawkin’s “Extended Phenotype”

discusses genes for things outside the organism, such as nests,
bowers,

burrows, etc.

Doesn’t this concept skip over something important? The genes can’t, of
course, specify anything outside an organism. Genes specify patterns,
proteins, organizations, inside organisms. What is skipped over is the
mechanism that can make it appear, superficially, that the gene
itself systematically affects something external. This makes the
“extended phenotype” idea into a metaphor, not a fact. When we
try to deal with “is” rather than “as if it is”, we
have to focus on the actual physical system and try to understand how it
works.
There’s a great deal of fuzzy science that works in terms of metaphors, I
would guess because nobody has any idea of how these superficial
appearances are generated. But we shouldn’t allow that limitation to fool
us into thinking that this sort of image actually explains anything. It’s
a poetic description. Its main striking feature is that it implies
something that we know is pretty certainly not true. That’s what calls
our attention to it.

Another example is “retroactive inhibition,” a descriptive term
that makes it seem as if one stimulus can reach back through time to
alter the effect of a previous one on behavior. The only virtue of that
term is that it signals us, with clumsy subtlety, that the appearance is
NOT what is actually happening, because we know that backward causation
through time does not happen. The term should really be expanded to
“What seems to look like retroactive inhibition but is not.”
Similarly for “What seems to work like an extended phenotype but
clearly can’t be that.”

Another concept that
appears related is “niche reduction”,

e.g., evolution has constructed an environment of parental care,
which

reduces the need for adaptations to the intermediate environmental
niches

between infancy and adulthood.

Again, evolution has done no such thing, and can’t do that sort of thing.
Evolution consists of alterations in organisms, not environments. It is
the evolved organisms that alter environments according to the laws of
perceptual control theory and reorganization. And despite the purposive
language, evolution does not have a goal of reducing some objective need
for adaptations, even though an organism may change its organization so
that it can accomplish certain goals without having to develop the
ability to adapt to an environment (see LCS3, Demo5-1).

The metaphorical language allows us to give the appearance of
understanding something without actually doing so. It encourages sloppy
thinking and misleading communication. I’m against it in general, though
as for all stylistic rules, communication is sometimes improved by
breaking the rule. Sometimes.

Best,

Bill P.

_Boris_Hartman · September 15, 2012, 8:39pm

Hi, Bill

I have some troubles understanding this.

BP : Genes specify patterns, proteins, organizations, inside organisms.

BP : Again, evolution has done no such thing, and can’t do that sort of thing. Evolution consists of alterations in organisms, not environments. It is the evolved organisms that alter environments according to the laws of perceptual control theory and reorganization.

BH : Does what you are saying mean that autonomy of organisms organization is geneticaly specified, and if this genetic mechanisms are able to maintain “homeostasis” in organisms under “the disturbances” of environment, organisms survive ? Genetic mechanisms and organisms as a whole are working according to laws of PCT and reorganization ?

If genetically specified organisms survive (manage to maintain “homeostasis” with counteracting) and are able to reproduce, than they can “transfer” characteristics on offsprings. Is this the mechanism of evolution ?

Best,

Boris

···

----- Original Message -----

From:
Bill Powers

To: CSGNET@LISTSERV.ILLINOIS.EDU

Sent: Saturday, September 15, 2012 4:28 PM

Subject: Re: niche construction

[From Bill Powers (2012.0-9.15.0742 MDT)]

At 09:42 PM 9/14/2012 -0600, Martin Lewitt wrote:
This theory doesn't seem that new.  Dawkin's "Extended Phenotype"
discusses genes for things outside the organism, such as nests, bowers,
burrows, etc.
Doesn’t this concept skip over something important? The genes can’t, of course, specify anything outside an organism. Genes specify patterns, proteins, organizations, inside organisms. What is skipped over is the mechanism that can make it appear, superficially, that the gene itself systematically affects something external. This makes the “extended phenotype” idea into a metaphor, not a fact. When we try to deal with “is” rather than “as if it is”, we have to focus on the actual physical system and try to understand how it works.
There’s a great deal of fuzzy science that works in terms of metaphors, I would guess because nobody has any idea of how these superficial appearances are generated. But we shouldn’t allow that limitation to fool us into thinking that this sort of image actually explains anything. It’s a poetic description. Its main striking feature is that it implies something that we know is pretty certainly not true. That’s what calls our attention to it.

Another example is “retroactive inhibition,” a descriptive term that makes it seem as if one stimulus can reach back through time to alter the effect of a previous one on behavior. The only virtue of that term is that it signals us, with clumsy subtlety, that the appearance is NOT what is actually happening, because we know that backward causation through time does not happen. The term should really be expanded to “What seems to look like retroactive inhibition but is not.” Similarly for “What seems to work like an extended phenotype but clearly can’t be that.”
Another concept that appears related is "niche reduction",
e.g., evolution has constructed an environment of parental care, which
reduces the need for adaptations to the intermediate environmental niches
between infancy and adulthood.
Again, evolution has done no such thing, and can’t do that sort of thing. Evolution consists of alterations in organisms, not environments. It is the evolved organisms that alter environments according to the laws of perceptual control theory and reorganization. And despite the purposive language, evolution does not have a goal of reducing some objective need for adaptations, even though an organism may change its organization so that it can accomplish certain goals without having to develop the ability to adapt to an environment (see LCS3, Demo5-1).

The metaphorical language allows us to give the appearance of understanding something without actually doing so. It encourages sloppy thinking and misleading communication. I’m against it in general, though as for all stylistic rules, communication is sometimes improved by breaking the rule. Sometimes.

Best,

Bill P.

Bill_Powers1 · September 15, 2012, 10:34pm

[From Bill Powers ()2012.09.15.1603 MST)]

[Martin Lewitt 2012 Sep 15
1155am]

It is not that big a mystery. Genes can specify things outside an
organism, by specifying the behaviors that result in them. The
selective advantage of the genes for those behaviors is that they result
in a nest.

But if there’s anything we PCTers can be sure of, it’s that no result can
be made to repeat unless something can sense the result and adjust its
action to repeat the result despite disturbances and other changes in
environmental characteristics, which are always present in the real
world. In order for a gene to cause behaviors that result in building a
nest (like the bower bird’s fancy love-nest), it would have to cause
exactly the right variations in behavior that are needed to keep
the same nest pattern occurring (not repeat the same behaviors). We know
that genes can’t do that – they have no eyes, ears, noses, tactile
organs, or other senses. The required variations must be an effect of the
kind of organization that genes do generate: control systems with
exteroceptive senses, comparators, and so on. There can be no direct
connection between a gene and a controlled result of action.
Of course the superficial appearance is as if such a direct connection
does exist. But that is only a metaphor, and as soon as we ask how that
can possibly work, we can see that the metaphor has to be false to fact.
Genes cannot in fact specify anything but the structures and
relationships inside the organism. In fact, you have provided a clear
example of how the metaphor misleads us. It entices you into an
explanation that can’t possibly work. Repeatable results are not achieved
by, as you put it, specifying the behaviors that result in them. The
result can be repeated only by changing the behavior in exactly the ways
necessary to counteract the effects of external disturbances, which in
general are different each time the same result is generated. That can be
done only by a negative feedback control system capable of sensing the
result, comparing it with a reference condition, and modifying its
actions so as to reduce the error. When you ask how the
observed appearance can be created, the answer exposes the metaphor as a
falsehood.

Best,

Bill P.

···

At 12:11 PM 9/15/2012 -0600, Martin Lewitt wrote: