At 08:03 PM 11/14/98 -0500, you wrote:
>Having recently read Stephen Jay Gould's 1996 book "Full House", I
>thought it would be appropriate to re-formulate some of his thoughts and
>discuss what they may mean for the notion of meme and cultural
>evolution.
>
>Gould takes issue with the very idea of cultural evolution (he prefers
>the term cultural change) because this term implies a similar mechanism
>for biological evolution through natural selection and for cultural
>evolution. Needless to say, the notion of meme only emphasizes this
>supposed similarity.
>
>At the core of Gould's discontent is the notion of progress. The above
>mentioned book advances a tremendously sound argument for why biological
>evolution does not entail any progress or drive to increased complexity.
>I am not going to go into great detail here about this argument, but
>just say that it is based on the idea that given that life started with
>organisms almost as close as possible to a minimum threshold of
>complexity required for life, statistically we are to expect a strongly
>skewed distribution of complexity with an essentially constant mode
>close to low complexity and a few high complexity organisms. He points
>us to a lot of available evidence showing that such a distribution is
>expected from just random drift without recourse to explanations based
>on the supposed merits of higher complexity in traditional Darwinian
>discourse. This passive trend of biological evolution has also been
>defended by Dan McShea ("Mechanisms of large-scale evolutionary trends:,
>Evolution, Vol. 47, pp. 1747-1763).
You may recall a discussion between Don Mikuleky and I last summer on this
site about Gould's book. I expressed extreme discontent with Gould's
argument. I do not think it is "sound" at all, but perhaps I have trouble
proving my point. In re-thinking the issue now, I wonder if the following
might be a suitable counter-argument.
I accept the above description of Gould's argument. Well stated. The
passive process he thus concludes would produce the expectation that if
more simple states exist than complex ones, then evolution would fill more
simple states than complex ones. The first problem with this thinking goes
back to the idea of "pre-defined" niches - a debatable concept. Are the
"niche" possibilities for organisms pre-defined by the environment and
present complexity of the system? Or are they created as part of the
self-defining aspect of complex systems? In other words, did the exact
niche for an elephant, including all the "n-dimensional" characteristics of
that niche, which necessarily include behavior, learning ability, physical
size, etc., actually exist in the ecosystem before elephants evolved, or
did ecological complexity, and thus the niche, evolve along with the
elephant? Gould is thinking too mechanically, which is the Darwinian habit,
separating environment from organism. If we say that the niche did not
exist prior to the evolution of the elephant, than we are falsifying the
statistics. There are no empty states for Gould's passive process to fill.
On the other hand, if we accept the simplistic idea that niches of all
possible variety pre-exist in some "possibility space" (my term), then we
must say that we can have no idea of what those possibilities are until
they are filled. While conceptually this works (for Gould) it is
non-science, because it is not testable. It is like the ether before the
Mikelson-Morley experiment, except that there can be no experiment. Also,
what is the scientific basis for claiming there are more complex states
than simple ones? Mathematically one can claim this, but are we talking
about a mathematical possibility or something that can actually exist in
nature? Do we presume they are one in the same? If not, we do not know that
there are many possible complex states that are viable, nor what factors
determine their viability. So again, the statistics of possibilities would
depend on whether mathematical possibilities (which we can predict) are in
fact ecological possibilities (which perhaps we cannot predict).
The second argument I present is that, accepting for the moment Gould's
very simplistic idea of pre-existing niche possibility space, and that
there are more viable complex niches compared to a limited number of simple
ones, thus leading to the expectation that evolution will passively exhaust
the simple possibilities and begin filling the more complex ones, we can
thus examine the mechanism of Gould's passive progress. In particular, it
should be true that the probability of any given transition toward a more
complex organism is equal to the probability of a transition toward a more
simple organism, except for competition (i.e., the same thing as saying the
simple niches are "filled"). But modern ecology has greatly moved away from
"the dogmatic predictions of competition theory" (Simberlof). Competition
was treated in ecology almost like a religion - as is evident by Gould's
unquestioning acceptance of it in this case. Does he even mention that
competition is the mechanism for preventing retrograde evolution? What is
found is that there are many reasons why competition may not dominate in a
real ecosystem or community/assemblage. Micro-habitats afford locations
where otherwise "excluded" species in fact survive. Disturbance creates
micro-conditions for such disturbance. Complex organisms afford new
"niches" for simple organisms. In fact, there is a body of literature now
claiming that competition is an inadequate explanation for extinction,
except on a very local basis. There is also a body of literature claiming
that reduction of the gene pool due to low numbers of organisms (i.e., rare
or depleted species) is also an inadequate explanation for extinctions.
Thus, claiming that a simple niche is "filled" is inadequate to explain the
absence of "reverse-evolved" organisms. Are there examples of algae that
evolved from elephants? Or have we just not waited long enough? A more
simple form of elephant evolved from a more complex one would probably
survive long enough to be in the geologic record. We should find conpelling
evidence that nearly as many species evolve toward simple states at evolve
toward more complex states. Only their relative abundances would be
different, but they should still be in evidence. Hence, Gould's conclusion
would apply to the differences in relative abundance and thus the overall
affect on the total gene pool, but a study of the probabilities of
transitions to complex or simple states should show that his mechanism
doesn't work. Where are the examples of reverse-engineered organisms that
this passive process would imply? The ones Gould supplies are very limited
examples. A shark existing for a long time as an example of a stable simple
form -- but shark research is showing that they are quite complex
creatures, not just the "eating machines" we are used to think of them as.
The evidence for evolution of simple forms from complex ones is very weak.
Especially if Gould is correct that there is better survival probability in
the simpler states, one would expect all organisms to be vying to evolve
more simple forms and competing for these highly prized simple states. We
should see thousands of examples of these attempts that became
out-competed. In fact all of the cases of simplification seem tracable to
very unique environmental conditions. The "simplification" is really
adaptation to stable conditions over a long time and is of a very limited
nature.
Let's borrow some of Rosen's "relational biology" theory, which states that
a unique property of all organisms is that they are "closed under efficient
cause," i.e., they are self-defining, and that they are "anticipatory
systems" because their make-up involves the internal generation of a model
of external reality in relation to self. Organisms thus do not respond
directly to the environment, according to Rosen, but to their internal
model of it, and presumably the internal model is subject to improvements
much the way a scientist improves his/her scientific models, by testing.
The difference (I argue) is that the organism is both scientist and
lab-rat, whereas we prefer to subject other organisms to the testing
process (perhaps because we have less confidence in our theories!). In this
view, a "niche" is truly a dynamically created thing. It is the result of
the modeling relation between an organism's model-based perception and the
true environment. This gets at what "causes" or produces the niche.
Gould doesn't say where niches come from, he just uses his assumption about
them in a statistical arguement. The traditional view involves nothing more
than arm waving when it comes to explaining what a niche is. No ecologist
has ever been able to completely define a niche or describe one except in
general characteristics, without assuming there are more subtle
characteristics always at work. You can't study the niche space because it
does not exist objectively. It is subjectively defined by the interaction
between organism and environment. The niche defines the niche. The theory
on which statistics is based requires that there be pre-existing countable
states to which the statistics are applied. Thus the mathematics itself is
inapplicable to the case Gould is discussing.
Now, when such self-generating situations occur, one can assign
probabilities observationally and then base statistics on that. The
statistics apply then to what we observe, not what might be causing it.
This is done in quantum physics. The probability density function refers to
known possible states that have been observed. But at such a fundamental
level of physical reality, the possible physical states (after quantum
collapse) are observable and repeatable. They are elements of the
macroscopic physical universe, which appears to be extremely stable in its
possibilities. In other words, we do not observe a high rate of newly
evolved physical states (it may actually be null). There is positive spin
and negative spin. We don't expect or observe a new kind of spin evolving.
In biology, however, we're not dealing with such fundamental physical
states, but a much greater and unpredictable variety of states that can
neither be repeatably observed (evolution will likely not repeat the exact
same organism if re-done), nor are they predictable from prior observation.
We have to accept that the possibilities are simply unknown and unknowable.
So, in view of these arguments, to me, Gould's argument isn't so much
wrong, as it is completely irrelevant. One can apply a statistical picture
to any observed pattern (say the distribution of rocks in your driveway)
after the fact, but that picture is explanatory only if there can be some
underlying cause associated between outcome and statistical model. I
believe I have shown above that there is at best only a weak demonstration
of passive mechanism in evolution as Gould discusses it, and that the
statistical argument has no basis. However, the statistical description of
the pattern can be accurate -- it was designed to fit the observations so
it must be accurate. The problem is it doesn't explain anything. It doesn't
lead to any further insight into biology or evolution. It is a circular bit
of mathematical trivia, that tells us something about the current state of
the world, but nothing about what produced it. It is also circular in the
respect that the very concept of complexity is relative to simplicity and
depends on it. By definition (I suspect most or all 31 definitions of
complexity agree in this principle) you can't have complexity without
simplicity. But, conceptually, you can have simplicity without complexity.
Hence, it is our concept that Gould is mapping out statistically and
demonstrating this necessity in our definitions. So, again, the argument
tells us nothing about nature.
Taking this a bit farther, Gould actually DOES seem to imply a claim about
evolutionary preferences, even though he also strongly denies it. He claims
that the simple states are preferred by nature over the more complex ones,
and that organisms are only filling the complex states as a matter of
chance. Hence there are far more organisms (numbers and species) in the
better simple states. But again, the argument is based on the observed
statistical pattern without any reference to something explanatory -- i.e.,
why simple is better. In essence he is saying there is no explanation, the
pattern is simply the way it is and organisms conform to this non-causal
model by randomly filling the pre-existing possibilities, the more
prevalent ones in greater abundance and the less prevelant ones in lesser
abundance, and the numbers of organisms in each being proportional to the
relative value accorded these states by their pre-existing capacity and
distribution. In other words, there is nothing dynamic in the process of
evolution, which is characterized by a conceptually stable background
"potential environment," and thus niche space, which gets filled. It is a
paradigm of a completely dead system, and again, a denial not just of the
idea that evolution may have directions or that complexity may confer
advantage, but the more serious allegation that living systems involve
something besides mechanical materiality. The view, demonstrably
un-scientific, is a defense of the traditional view of nature which must
separate organism and environment, and I strongly suspect that is its real
purpose.
Let's look at this in yet another way. The claim that natural selection
favors simple states over more complex ones (because they are more
prevalent) is supported by the greater numbers and biomass of simple
organisms. This presumes that numbers of individuals, or biomass, is
naturally important - the right measure to use. Ecology has not been able
to establish many fundamental value measures and continues to struggle in
this regard. All possible measures are questionable in their general
importance - number of organisms, biomass, size of organism, within-species
genetic diversity, species diversity, etc. Each seems to have relative
importance in specific circumstances. Perhaps it is better in some
circumstances to live as a colony, in others as an individual. In some
circumstances high numbers are beneficial, in others heardiness and low
numbers, and so forth. Why is number of individuals, or their mass
important to Gould? Because it produces a skewed statistical distribution
toward the "simple" organism end of Gould's scale. There are other measures
that would turn out to be very low for simple organisms and high for more
complex ones -- skewed in the opposite way. Why aren't those the important
ones? The assumption is that more = better = nature's preference. How else
can we measure "survival"? Fitness has to ultimately be defined in terms of
numbers. In other words, Gould is using the same argument he is attacking.
He is saying the really important thing for nature is numbers and mass, and
that conforms to a very vague concept of what is simple vs. what is
complex. Hence there is no real preference for complex. Well, doesn't a
simple measure, which we assume, determine the outcome of the statistics?
Of course it does! If we used a complex measure, we would see nature
favoring complexity. We're still attributing some kind of "preference" to
nature either way, based on our measure. Simple measures and passive
processes are just a more acceptable preference for mechanists because it
has mechanical analogies that we are used to.
What if we really ask why ecology hasn't found any fundamental value
measure that would directly map into survival? Let's examine the
fundamental value of survival for a moment. Survival is presumed as a
fundamental value because it is synonymous with existence, and existence is
all we can talk about. So in essence we're really placing value on our
talking, not survival -- but put that aside for the moment. Let's accept
that existence is a fundamental value. There are still no ecological
measures that universally map into existence or survival. Not numbers,
size, mass, complexity, style, or anything. Why has ecology failed to find
such a fundamental measure? The default answer is the "it doesn't exist."
But there is another possibility. That we are looking in the wrong place
for it. We're looking for material measures and we've shown convincingly
that there aren't any. What about non-material measures? What about ideas?
Is there some basis for saying that thought maps into existence (Descart)
and that good ideas map into survival? OK, we still have the circular
problem that "good" is itself defined in terms of existence and survival.
But, an "idea" has the property that it can encompass the entire organism,
both psychologically and physically (non-materiality and materiality). It
is not much different than saying that the organism itself is the
fundamental measure of value, which I subscribe to, but by thinking of it
as an idea rather than a physical thing, we can include the important
properties that characterize living systems.
For example, organisms are self-defining. In Rosen's terms, they are closed
to efficient cause. They provide their own semantics. They don't need
statistics to provide their semantics. They create niches then occupy them,
then modify them and evolve. To accept this thesis we have to accept the
notion that there can be something "new" as opposed to only new
combinations of pre-existing states. Our traditional scientific model of
the universe does not allow this because it is based on rigid determinism -
the idea that everything is ultimately derivable from pre-existing
conditions. The "strategy" or "idea" that an organism represents is
irrelevant because it is merely a result of the organization. But I and
many others are beginning to strongly suspect it is the other way around
with living things (and perhaps all things). Or let me put it this way: I
strongly suspect that THINKING of it the other way around will be by far
more productive from here on. That mode of thought has a future, something
to investigate, more to learn. The other mode is the past -- a very
necessary background, but we're about done with it and ready to move on to
more advanced levels of thought. All due respect for age, wisdom, and past
work, but Gould's idea is part of the past and offers little avenue for
future exploration. That's my way of saying it is "wrong," but putting it
this way is the more epistemologically correct way to say it
(philosophically why no view is exactly "wrong"). So I respect Gould for
the "right" ideas he has produced, but reject his attempts to limit the
future to his own accomplishments.
>In contrast, cultural evolution (or change), he defends, show an active,
>rapid, change and directionality (or progress). He points out two main
>differences between cultural and biological evolution (in the following,
>text in between quotes is from Gould's book above pages 221 and 222):
>
>1) Topology. Biological evolution through natural selection is a
>"process of constant separation and distinction". Species interact
>ecologically, "but they cannot physically join into a single
>reproductive unit".
Counter-example: Mitochondria were probably originally separate species
that "joined" another species to produce a new kind of life form that is
reproductive as a unit. There are reasons to think that strong symbiosis
can in fact lead to a unification between two species.
>Species do not amalgamate, join, or crossover
>systemic information: "once a species becomes separated from an
>ancestral line, it remains distinct forever."
Again, an observed tendency, due to physical limitations, but not a
fundamental law. Why does Gould tell species what they can or can't do?
They determine that.
>Cultural evolution (or
>change), on the other hand, is defined by amalgamation and crossover of
>traditions, ideas, information, etc.
Counter example: Island cultures have become isolated by their geography
and evolved without "cross-over." Other examples can be found where
isolation was by preference rather than by environment. Nevertheless, the
culture evolved.
>This cultural proliferation is of
>an active, explosive nature.
Only in some cases, in others it is immeasurably slow.
>The impact of this process "powers cultural
>change by a mechanisms unknown in the slower world of Darwinian
>evolution". Examples of this abound, such as the introduction of fire
>arms into Japan by portuguese sailors, which caused a long lasting
>period of multi-partite civil war to change into a more unified
>shogunite rule.
>
>2) Mechanism of Inheritance. Natural selection provides local adaptation
>to an environment by elimination of most variants and preservation of
>those individuals fortuitously better adapted (not necessarily more or
>less complex) to changing local environments.
Does he mean this to imply intention of Nature? Obviously not, given his
other arguments. Thus he means it again only as an observed pattern, and a
questionable one at that. Locally adapted populations also preserve genetic
diversity - a counter-example to his claim. If species become too narrow in
the conditions to which they are adapted, they become more vulnerable to
change. So the degree of disturbance affects the optimal balance between
adaptation and diversity. There may be, and likely are, many other factors
that also affect this balance; which is a dynamic that is best thought of
in terms of ideas, not forms. Again, the argument reduces to "Nature
prefers what survives, hence has no particular preference as a whole, and
survival is ultimately unpredictable -- so, it is what it is!" This is true
but not very informative. Keep this in the materialist tradition and it is
simply a circular arguement that can go nowhere except to insist that
nature is "nothing but" what it is, and since we can describe anything that
is with statistics, nature is nothing but statistics. I find that
incredibly stupid. I believe living nature is unpredictable because it is
partly the result of innovation of ideas. Innovation also involves
intention. Intention can exist within an organism (and demonstrably does).
Organismic intentions are ontologically derivable from universal intention,
which we can identify as the will to exist (which we interpret as a will to
survive because of our heightened awareness of, and attachment to, time). I
see the will to exist as being reflected in ecological strategy, which then
is the true driver of physical evolution, which is what converts existence
into survival. This reverses the common assumption that the physical
environment strictly and ontologically determines which forms survive, and
that strategies are somehow derived from the form. In reality, I believe
both are true -- like all living and complex systems, causation is not
linear, it is mutual, hence living systems are in part self-defining and in
part constrained to physical laws. I know this language is offensive to
many of us raised in the strong materialist traditions, with a very
specific linear concept of cause and effect, but my point is that it is a
more useful paradigm for the things we need to study next.
"Local improvement rises
>upon the hecatomb of countless deaths; we get to a 'better' place by
>removing the ill-adapted, not by actively constructing an improved
>version."
There is no evidence to support the second part of this statement, and in
contrast many counter-examples. Certainly there are human examples of
"constructing" improved life styles, which ultimately affect selective
forces and lead to selection for those adapted to the improved styles. This
is termed "cultural evolution" primarily to safely isolate the human case
from our consideration of the natural world. But humans are animals which
have evolved just like everything else. Rather than supposing some
fundamentally unique new factor introduced to humans by Divine providence,
complexity, or whatever other untestable explanation, it makes far more
sense to treat humans as evidentiary of processes that likely exist in more
primative states in other organisms. That assumption should be made until
proven false. Are chimps not capable of constructing new and improved
chimps through their "cultural" evolution? Some religioiusly say they are
not, because their religion is based on yesterday's materialistic science,
and a strong disrespect for chimps.
>In contrast, cultural evolution potentially follows a more
>active mechanisms of Lamarckian evolution: the inheritance of acquired
>characteristics. "Any cultural knowledge acquired in one generation can
>be passed directly to the next by what we call, in a most noble word,
>education. If I invent the first wheel, my brainchild is not condemned
>to oblivion by hereditary impassability (as any purely bodily
>improvement would be). I just teach my children, my apprentices, my
>social group, how to make more wheels."
Yes. This is Gould's concession to the obvious. But don't miss the fact
that his "cultural evolution" is neatly separated from the rest of Nature.
Why doesn't this also work for gorillas, chimps, oranguthans, dogs, cats,
other animals, plants, microbes? Nobody can cleanly draw the line, because
there isn't one in Nature.
>
>Framing this issue more on semiotics, one can say that both biological
>and cultural evolution are established by some kind of selection on some
>self-organizing basis (with material, environmental, and cultural
>constraints), what I have referred to elsewhere as semiotic selected
>self-organization. The transmission of variants and possibilities relies
>on the existence of some symbolic elements (genes and language) which
>decode with repercussions on some organism/environment coupling
>(pragmatics). However, the selection process is remarkably different:
>biological evolution relies on an implicit selection of locally fit
>organisms (with encoded phenotypes in genotypes), while cultural
>evolution relies on a very explicit Lamarckian selection of ideas and
>artifacts (encoded in symbols: language is a retrovirus!).
I would like to see a list of examples of non-biological "cultures."
Although a matter of interpretation, it is also difficult to construct a
list of non-cultural organisms. If either of these poses some
difficulties, why are we separating biological evolution from cultural
evolution at such a fundamental level? There is only one evolution. The
point is to understand it. By "biological" we mean a Darwinian, passive,
material mechanism based on randomly introduced innovation solely at the
genetic level. By "cultural" we mean innovation can be introduced
creatively (from an unknown source) at the phenotypic level. My point is
that what is called "cultural" can causally affect the "biological"
evolution of that species (and vice-versa), and this is true for all
organisms, not just humans. This captured in what I have called
"autevolution."
By separating out "cultural" evolution, many people feel safe because they
can restrict discussion to the human case and computers that we want to
make human-like. The more troublesome aspects of combining biological and
cultural phenomena into a more robust understanding of evolution is avoided
by a religioius taboo that scientists seem to pay their dues to far too
heavily.
>
>Now, if these two types of semiotic selected self-organization are so
>distinct -- I re-emphasize that one leads to a passive form of evolution
>while the other to a complexifying, potentially progressive one -- why
>do we insist on blurring the differences? Why refer to symbols,
>language, and ideas as memes, when the cultural mechanism of
>transmission of the first is so very different from that of genes?
>Furthermore, from a semiotic perspective, we can see genes as symbols
>and genotypes as descriptions of organisms, but what about memes? Is the
>meme of the firearm the human language symbols for this concept, or is
>it the firearm itself? I cannot find in the notion of meme any
>distinction between signifier and signified (memotype and ????). Aren't
>we better off explaining cultural evolution with the more traditional
>semiotic categories of syntax, semantics, and pragmatics?
I basically agree that the term "meme" seems pretty redundant with other
concepts that may have better scientific definition, but perhaps it helps
people latch onto the concept. I think the idea of memes was invented
specifically to draw some parallelism with genes and introduce the concept
that there is something else going on.
Unfortunately, most people do not seem to have gotten the main point of
this -- that there is something else going on in EVOLUTION (considered as
one thing, not separate varieties of evolution) than is supposed by the
Modern Synthesis (neo-Darwinism). Instead, it was taken literally and
SEPARATED from biological evolution. Why? Because we really don't want
other organisms to have minds. These troubling studies conducted by pesky
women in the jungles of Africa, or sign-language labs in Berkley, or other
animal mind and behavior studies, are "interesting" and mostly annoying,
but so far we can cleaverly avoid having them change the way we think about
nature. I think we're too cleaver in this regard. We're out-smarting
ourselves. Nature is thus dead, passive, and nothing more than statistics -
right? We'll explain the exceptions later when we figure out chaotic
mechanics or complexity theory. Good idea. I prefer to work with the
evidence and a rational ontology based on our knowledge of where organisms
come from and what they can do.
So if the use of "memes" or "cultural evolution" will continue to have the
effect of separating our concepts of evolution and falsely supposing that
they apply to different things, then the words do more harm than good. It
is always reasonable to make distinctions for working purposes, but this
goes beyond that convenience. The distinction is treated as a fundamental
matter. Living organisms evolve by both means simultaneously, with each
having causal affect on the other. We have to study the system too, not
just separated causation lines. We have to modify our concept of
evolutionary mechanism to include the uncertainties introduced by
independent psychological phenomena.
>
>Cheers,
>Luis
>--
>______________________________________________________
>
>Luis Mateus Rocha (Postdoctoral Associate)
>Los Alamos National Laboratory
>Computer Research and Applications Group (CIC-3)
>Los Alamos, NM 87545, USA
>e-mail: rocha@lanl.gov or rocha@santafe.edu
>www: http://www.c3.lanl.gov/~rocha
>
>
-----------------------------------------------
John J. Kineman, Physical Scientist/Ecologist
National Geophysical Data Center
325 Broadway E/GC1 (3100 Marine St. Rm: A-152)
Boulder, Colorado 80303 USA
(303) 497-6900 (phone)
(303) 497-6513 (fax)
jjk@ngdc.noaa.gov (email)