Monday, December 07, 2009

The Searchers

AUTHORS: William A. Dembski and Robert J. Marks II

SOURCE: Proceedings of the 2009 IEEE International Conference on Systems, Man, and Cybernetics. San Antonio, TX, USA – October 2009, pp. 2647-2652

COMMENTARY: Allen MacNeill

First, congratulations to Drs. Dembski & Marks! Publication is the life blood of all career academics and the living heart of the intellectual process. It takes courage and hard work (and a little bit of luck) to get your original work published, and more of the same to weather the criticism that inevitably ensues. But, just as one cannot have a fencing match without an opponent, real progress in any intellectual endeavor cannot come from consensus, but only from the clash of ideas and evidence.

And so, to specifics:

I have no quibble with most of the mathematical analysis presented. Indeed, given the assumptions upon which the authors' Conservation of Information (COI) theory is based (with which I do not necessarily agree, but which are clearly presented in their paper), the analysis presented is apparently not completely outside the domain of No Free Lunch (NFL) theorems in general.

However, the same cannot be said for the application of these ideas to biological evolution. To be specific, consider the following quote [Dembski & Marks (2009) pg. 2651, lines 2-5]:
"From the perspective of COI, these limited number of endpoints on which evolution converges constitute intrinsic targets, crafted in part by initial conditions and the environment." [emphasis added]

This is indeed the crux of the issue vis-a-vis biological evolution. While it is clearly the case that Simon Conway-Morris asserts that there is an apparently limited number of biological "endpoints", it is neither the case that Morris' viewpoint represents the core of evolutionary theory, nor that his point is relevant to the analysis of COI presented in Dr. Dembski and Marks' paper.

To be specific, the highlighted qualifier from the quote above – crafted in part by initial conditions and the environment – is precisely the issue under debate between evolutionary biologists and supports of intelligent design (ID).

Taken at face value, this qualifying simply phrase means that, given specific starting conditions and a specific time-varying environmental context, the various mechanisms of evolution (e.g. mutation, natural selection, genetic drift, inbreeding, etc.) tend to converge on a relatively limited set of genotypic and phenotypic "endstates" (i.e. what could be loosely referred to as "evolutionary adaptations").

This is simply another way of defining evolutionary convergence, and in no way constitutes evidence for intrinsic evolutionary teleology. On the contrary, it simply provides support for the hypothesis that, given similar conditions, similar outcomes result.

Furthermore, it assumes that virtually all characteristics of living organisms are adaptations (that is, genotypic/phenotypic characteristics that fulfill some necessary function in the lives of organisms). However, this is manifestly not the case, nor is it an absolutely necessary component of current evolutionary theory. On the contrary, many (perhaps the majority) of the characteristics of living organisms are not adaptive. This is certainly the case at the level of the genome, as evidenced by the neutral and nearly neutraltheories of molecular evolution.

Finally, Morris' (and, by extension, Dembski and Marks') position completely omits any role for historical contingency, which both the fossil and genomic record indicate are of extraordinary importance in macroevolution. As Dembski and Marks state, the "endpoints" (perhaps it would be more precise to refer to them as "way stations") of macroevolution depend fundamentally on initial conditions and the environment. But this is not fundamentally different from Darwin's position in the Origin of Species:
"The complex and little known laws governing variation are the same, as far as we can see, with the laws which have governed the production of so-called specific forms. In both cases physical conditions seem to have produced but little direct effect; yet when varieties enter any zone, they occasionally assume some of the characters of the species proper to that zone." [Darwin, C. (1859) Origin of Species, pg. 472, emphasis added]

Moreover, Dembski and Marks' analysis completely ignores the appearance (or non-appearance) of new genotypic and phenotypic variations, and on the accidental disappearance of such characteristics (via extinction), without regard to the adaptive value of such characteristics, or the lack thereof.

In other words, Dembski and Marks' analysis, while interesting from the standpoint of what could be called "abstract" search algorithms, completely fails to address the central issues of evolutionary biology: the source of evolutionary novelty (i.e. the "engines of variation"), the effects of changing environmental conditions on the actual forms and functions of living organisms, and the fundamental importance of historical contingency in the ongoing evolution of genotypes and phenotypes.


As always, comments, criticisms, and suggestions are warmly welcomed!


Labels: , , , , , , , ,


At 12/07/2009 10:01:00 PM, Blogger Allen MacNeill said...

Furthermore, I believe that there is a fundamental and extremely important distinction that must be made between the concept of "target" and the concept of "outcome". Intelligent design (especially as promoted by Dembski) is all about the former, whereas evolution is all about the latter.

Ergo, some questions:

Are the concepts of "target" and "outcome" interchangeable?
If they are, how so, and if not, why not?
Is an "outcome" necessarily a "target"?
Is hitting (or missing) a "target" necessarily an "outcome"?
What part does intentionality play in hitting (or missing) a "target"? What part does intentionality play in causing an "outcome"?
Is it possible to distinguish between "targets" and "outcomes" using empirical observations and inductive inference?
If so, how, and if not, why not?

At 12/08/2009 09:07:00 AM, Blogger Joe G said...

Evolutionary biology doesn't address the central issues.

For example no one knows whether or not any amount of mutational accumulation can account for the transformations required.

Also you cannot search for something that doesn't exist.

At 12/08/2009 09:29:00 AM, Blogger Allen MacNeill said...

On the contrary, evolutionary biology directly addresses the most central issue of all: which processes are legitimately "intentional" and which are not.

It is certainly not the case that all processes that converge upon similar outcomes are necessarily teleological. For example, all dropped rocks fall to the ground. That is, they "converge" on the lowest point in the local topography, because they are drawn by the force of gravity toward the center of mass of the planet. Does this mean that falling rocks are necessarily teleological? Or, if you prefer a semantic argument, does it make sense to say that dropped rocks fall in order to reach the ground?

Similarly, all of the sand in the top lobe of an hourglass "converges" on the hole in the constriction of the hourglass before falling through it to the lower lobe. Does this mean that the sand pouring through an hourglass is necessarily teleological? Yes, the hourglass itself is a designed (i.e. teleological) object; it exists in order to precisely demarcate a period of elapsed time. However, the movement of the sand per se does not happen in order to precisely demarcate a period of elapsed time. On the contrary, it simply converges on the constriction of the hourglass because of its shape and the force of gravity.

The hourglass exhibits a property that could be referred to as "designed teleology"; that is, it has been designed in order to accomplish some pre-existing goal (i.e. precisely demarcating a period of elapsed time). The same is not the case for the movement of the sand, however. By itself (i.e. not considered as part of the hourglass), the movement of the sand is purely teleomatic[1]; it reaches an endpoint as the result of external causes and conditions (i.e. the force of gravity, which does not exist in order to cause the sand to converge on and fall through the constriction in the hourglass).

In the case of the sand in the hourglass, the movement of the sand exhibits a property that could be referred to as "instrumental teleology"; that is, it has been co-opted by an intentional agent in order to accomplish some pre-existing goal (i.e. precisely demarcating a period of elapsed time). The same thing would be the case for a pebble collected from a streambed and used as a paperweight. The shape, size, and composition of the pebble are entirely the result of purely natural (i.e. non-teleological) processes, whereas its use as a paperweight is the result of instrumental teleology.


[1] Mayr, E. (1974) Teleological and teleonomic: A new analysis. In Boston Studies in the Philosophy of Science, XIV, pp. 91 to 117 (available here: nomi.rtf).

At 12/08/2009 09:29:00 AM, Blogger Allen MacNeill said...

A commentator at another website wrote:

"I sense that a profound misunderstanding of teleology pervades the scientific world."

Not just the scientific world; we are all very confused about the differences between what Ernst Mayr called teleomatic processes and teleonomic processes (the former are caused by purely natural/physical causes, whereas the latter are caused by intentional causes). I believe that Mayr was on the right track in sorting out the differences, and am currently extending Mayr's analysis in a monograph which I intend to call On Purpose: The Evolution of Design by Means of Natural Selection, or the Proliferation of Intentional Agents in the Struggle for Existence (all puns and similarities intentional, of course).

Outcomes are simply what happens at the end of a sequence of events, which may or may not be intentional. If the outcome is the result of purely natural causes (i.e. it is teleomatic, using Mayr's terminology), it does not qualify as a "target". Extending the analogy used in my previous comment, the ground upon which a dropped rock lands is not a target.

However, if a person throws a stone at something (e.g. a window), then that "something" is a target for the intentional agent throwing the stone, regardless of whether or not the stone actually hits the window.

In the difference between these two cases lies the entire source of the disagreement between ID and evolutionary biology...indeed, it is the central argument in virtually all of western philosophy. The Ionians (Democritus of Abdera, et al) took the position later taken by Newton, Mendeleev, Darwin, and virtually all natural scientists since the 18th century: that there is a class of phenomena (i.e. teleomatic processes, using Mayr's terminology) in which intentional teleology plays no part. An example of such teleomatic processes is rocks falling as the result of the force of gravity.

There is also a class of phenomena (i.e. teleonomic processes, again using Mayr's terminology) in which the processes are guided by a pre-existing program. Biological development is an example of such teleonomy, as it is guided by the genome of the developing organism, which pre-exists it.

According to the empirical natural science of evolutionary biology, evolution itself is a teleomatic process (i.e. the result of mutation, natural selection, genetic drift, inbreeding, etc.), whereas the products of evolution (i.e. living organisms) are the result of a teleonomic process (i.e. development guided by the genome as it interacts with the environment of the developing organism).

At 12/08/2009 09:33:00 AM, Blogger Allen MacNeill said...

And I completely agree that it is impossible to search for something that doesn't exist. This is precisely why evolution does not qualify as some kind of "search". The outcomes of any evolutionary process are not "searched for" in any way. Rather, they are the outcomes of teleomatic processes which, as Mayr pointed out in 1974, have no intentional outcome at all, but rather are the result of purely teleomatic processes.

At 12/25/2009 01:31:00 PM, Blogger RBH said...

Allen wrote

Furthermore, I believe that there is a fundamental and extremely important distinction that must be made between the concept of "target" and the concept of "outcome".

"Target" presupposes intentionality, so using it rather than "outcome" begs the question of teleomaticity vs. teoleonomy.

Once again, conceiving of biological evolution as "search," particularly search for some identified outcome (the eye, the flagellum) is a snare and a deception.


Post a Comment

Links to this post:

Create a Link

<< Home