|
Author
|
Topic: Eric Anderson: Irreducible Complexity Reduced: An Integrated Approach to the ...
|
Moderator
Administrator
Member # 1
|
posted 29. September 2004 13:51
Irreducible Complexity Reduced: An Integrated Approach to the Complexity Space
by Eric Anderson
Abstract: William Dembski’s and Michael Behe’s recent article, Irreducible Complexity Revisited , provides an important update on the irreducible complexity argument since the publication some 8 years ago of Behe’s book, Darwin’s Black Box . Their article, however, exhibits some confusion, or at least a lack of explicit clarification, regarding the interplay among specified complexity, cumulative complexity and irreducible complexity. In the present article, I analyze the relationship of these concepts and show that the argument from irreducible complexity cannot be divorced from the broader argument of specified complexity. While this has been previously acknowledged in a broad sense, I make explicit irreducible complexity’s dependence on specified complexity, including specified complexity as applied to cumulative complexity, and further demonstrate why this dependence causes the irreducible complexity argument to break down in the evolutionist’s mind. Indeed, this dependence is directly responsible for evolutionists’ ability to acknowledge the existence of irreducibly complex biological features while still rejecting the irreducible complexity argument. Finally, the present article demonstrates how the arguments from complexity can be better understood within the context of an integrated approach to the complexity space. With a better understanding of the complexity space, the concepts of irreducible complexity and specified complexity can be fortified and focused in order to bring the full weight of these arguments to bear on evolutionary claims.
To read the entire paper, click here.
[ 29. September 2004, 13:52: Message edited by: Moderator ]
IP: Logged
|
|
Eric Anderson
Member
Member # 1431
|
posted 08. October 2004 16:50
I thought I would begin the discussion with a brief explanation of the principal issues I discuss in my paper.
My principal goal with the paper is to step back from the irreducible complexity debate for a moment and take a broader view of the complexity space. This allows us to take a more integrated approach to the complexity space than what is often characterized as a simple irreducible v. non-irreducible dichotomy. I do retain the irreducible v. non-irreducible distinction, because I believe that drawing a reasonable distinction is central to our ability to even engage in the debate. However, in analyzing the complexity space in more detail, I argue that it is necessary to distinguish between different kinds of irreducible complexity, including General Irreducible Complexity (what ID proponents generally loosely think of when they discuss irreducible complexity), and two additional subcategories: Per Se Irreducible Complexity and Cumulative Irreducible Complexity.
I kick off the analysis by challenging certain examples in Dembski’s (and Behe’s) recent essay, Irreducible Complexity Revisited. I do this not to fundamentally dispute Dembski’s ultimate conclusions regarding irreducible complexity, but rather to underscore the challenge of appropriately defining irreducible complexity and to underscore the need for a more integrated view of the complexity space.
Although something of a disconnect in the irreducible complexity debate has been previously acknowledged, I attempt to bring together the various moving parts into a more integrated whole. I believe this approach not only correctly identifies and explains why evolutionary proponents and evolutionary critics often talk past each other, but, hopefully, this approach will also function as a useful framework for further evaluation of the argument from irreducible complexity.
Eric
IP: Logged
|
|
Scott
Member
Member # 1222
|
posted 08. October 2004 22:39
Eric,
I found your paper quite interesting and think it contains much of value. I can only hope that others will read it as well. Since you have decided to post here, I will re-read it and see if I can come up with any questions for you. Thank you for your effort in clarifying these important issues. Of course, I am very interested in any responses you may receive from either W. Dembski of M. Behe.
Have you read William Dembki's latest on the argument from Irreducible Complexity in the recent book Uncommon Dissent? (Amazingly, it is already available in an inexpensive paperback edition.)
Do you have any comment on that chapter?
regards
[ 08. October 2004, 22:40: Message edited by: Scott ]
IP: Logged
|
|
Eric Anderson
Member
Member # 1431
|
posted 14. October 2004 00:01
Thank you for your kind words, Scott. Although I have not yet had the opportunity to discuss my approach with them in detail, Bill Dembski and Mike Behe were the first to see my initial draft, and Bill is the one who encouraged me to keep developing my ideas and to submit my paper to ISCID. I have not yet read the chapter in Uncommon Dissent, but will try to check it out.
Regards,
IP: Logged
|
|
The Deuce
Member
Member # 992
|
posted 20. October 2004 11:11
Hey, Eric, I found your paper quite interesting and insightful, particularly the part where you point out that IC always occurs within a context.
Here's what I see the relation between IC and SC (based largely on having read your paper) being:
The argument from SC is that an evolutionary algorithm is no more likely to reach a complex specification than chance, unless the fitness function is loaded with the appropriate side information.
IC, on the other hand, additionally argues that the fitness function that is in effect currently (ie, the context in which the object is IC) isn't able to select subsets of the features comprising the whole IC object, and thus couldn't have been the fitness function used to put them together. Any thoughts on this?
IP: Logged
|
|
Krauze
Member
Member # 1119
|
posted 21. October 2004 06:55
It's a privilege to have the author of a paper ready to discuss and clarify his or her work. Now that Deuce has offered his words of encouragement, I feel more comfortable advancing my (hopefully positive) criticism.
As I read this paper, my impression was, "Where's the biology?" Talking about stools and cities is all well and good, but unless one can go back and apply one's points to real biological systems, I see little point in the discussion. One can blame people like Thornhill, Ussery, or Gross for introducing analogies like chairs and cities, or one can blame Behe for using a non-biological system to explain his definition of IC in the first place. But Behe then spent most of the rest of his book applying his definition to various biological systems, showing its relevance. A similar engagement with biological detail is almost completely lacking in this paper.
Take the authors' suggestion that the context must be considered when deciding if something is IC. He makes some interesting points about how a stool needs to be constructed to fit into different context; the author's bar, by a camp fire, etc. But the ID theorist is left in the dark as to what insights these concepts will bring in an investigation of the biotic world.
To be fair, the author does make one application of his views to biology, noting that "if the bacterial flagellum were used simply to tickle other bacteria, then perhaps any number of similar-looking whip-like filaments might do the trick [thereby illustrating the need to consider the context in defining an ID system]" (p. 6). But what is the relevance of this? Has any ID-critic seriously proposed that a simpler proto-flagellum served to "tickle other bacteria"?
In short, I am doubtful that the concepts suggested in this paper will aid the ID community in its exploration of teleology. However, let me clarify that I speak from a position of ignorance in trying to apply the author's concepts to biology, and that my critique shouldn't be taken as a downright dismissal. If I were to give Eric some advice, it would be to "show the way" for skeptical ID theorists like myself, by taking some questions raised in the discussion between ID-proponents and ID-critics over specific biological features, and using his concepts to throw light on them. I'm sure PCID would like to publish a follow-up. ![[Smile]](smile.gif)
IP: Logged
|
|
Eric Anderson
Member
Member # 1431
|
posted 25. October 2004 16:36
Thank you, Deuce, for your encouraging comments. Perhaps I can offer a clarification or two.
The Deuce wrote:
quote:
The argument from SC is that an evolutionary algorithm is no more likely to reach a complex specification than chance, unless the fitness function is loaded with the appropriate side information.
I think it is correct that with evolutionary algorithms the relevant information is generally loaded up front. If I might focus on a nuance for a moment, however, I am not sure that SC is making any statement about evolutionary algorithms as such. The argument from SC simply says that with the probabilistic resources at hand, certain instances of SC cannot be generated without the input of an intelligent agent (i.e., chance and natural law alone are insufficient). (Incidentally, Dembski’s design filter operates as a negative filter. In other words, eliminate chance and natural law, and design is left. It is not a logical requirement that the filter be constructed in this manner – we could have design as the first cone of the filter. I believe Dembski constructed his filter as he did, however, as a pragmatic accommodation of two factors (i) the desire to eliminate the two most common explanations in the natural world, and (ii) the desire to minimize false positives.)
Evolutionary algorithms, do not try to avoid chance. Rather, they try to use chance to derive SC by means of a designer substitute -- in the case of Darwin’s theory, natural selection. The problem with evolutionary algorithms, from my perspective, is not so much that they violate SC, but that they are generally built upon circular reasoning. They tend to incorporate in the assumptions (in the case of programs, the initial programming parameters), the very intelligence they claim to be doing without. This is true of algorithms from Dawkins’ “methinks it is a weasel” to more sophisticated efforts, like Avida. Thus, with respect to IC, I think evolutionary algorithms generally fail as a matter of logic, independent of any discussion of SC. But this is really a topic for another time . . .
The Deuce wrote:
quote:
IC, on the other hand, additionally argues that the fitness function that is in effect currently (ie, the context in which the object is IC) isn’t able to select subsets of the features comprising the whole IC object, and thus couldn’t have been the fitness function used to put them together. Any thoughts on this?
In a cumulative scenario, the current function exhibited by an IC core must be different than the previous function, otherwise we are but refining our definition of the IC core. In that sense, then, the current function could be subject to selection after it has arisen in the current environment, but not before. As a result, you are right that the current function could not have been the fitness function used to cobble together the IC core historically, because by definition the IC core did not exist prior to the cobbling together. I believe this follows as a matter of logic. (Again, when we appropriately parse out the definitions, the oft-used phrase “improved function” really means new function for purposes of identifying an IC core.)
What I am calling attention to, however, is that when evolutionary critics note that the current function of an IC core could not have been the function selected in assembling the IC core in the first place, this does precious little to challenge the traditional evolutionary story.
Indeed, folks like Kenneth Miller do not propose that the current function was responsible for the current IC core. Rather, they imagine a long chain of unknown functions in ages past that slowly change over time as new parts happen to come on the scene.
Further, I argue that a cumulative pathway to IC is a live possibility. As a result, the mere existence of a system that exhibits Current Irreducible Complexity tells us nothing about whether it is an example of Cumulative Irreducible Complexity or Per Se Irreducible Complexity. Dembski and Behe (and most ID proponents) tend to start from the position that it must be Per Se Irreducible Complexity. However, I argue that such a position is logically problematic, is internally inconsistent, and leads to a great deal of disconnect in the overall IC debate.
The key is to focus on SC, because at the end of the day if a system exhibits Current Irreducible Complexity (as a specific example of SC), it is irrelevant to the design inference, whether the system is an example of Per Se Irreducible Complexity or Cumulative Irreducible Complexity.
Eric
IP: Logged
|
|
Eric Anderson
Member
Member # 1431
|
posted 25. October 2004 17:41
Thank you, Krauze, for your critique.
You are certainly right that my paper does not detail any new biological insights, as that was not the intent. Scott Minnich and others are heavily involved in researching the biological and biochemical foundations of the formation of the bacterial flagellum, for example, and can bring more to bear than could I in this regard. Thus, you are correct to note that I am not proposing new research techniques or insights into “investigation of the biotic world.” What I hope I have been able to do, however, is outline the logical framework within which the biological results are debated and interpreted.
I was concerned that my discussion of Dembski’s stool and city may have been too lengthy, which concern may be validated by your comments. However, I felt it important to follow these examples to their logical conclusion, as it is my position that these examples underlie a more significant problem in the overall IC debate. Specifically, IC as generally discussed by Behe and Dembski (and other ID proponents) is a non-cumulative process, and yet, I show that it is logically problematic for ID proponents to take this position. This is true both because it results in a miscommunication with evolutionary proponents, and, more importantly, because it unintentionally undercuts the fundamental basis of ID, which is SC. The fact that Dembski has continued to use his city example over the years shows that he has not thought through these examples to their logical conclusion and has not recognized how this failure undercuts his broader argument and leads to a disconnect in the IC debate. It is in helping to clarify and focus the key ID arguments that I hope to make a contribution.
With regard, to my comments about the bacterial flagellum, I may be fairly accused of a hint of artistic license, but I am not suggesting that any biologist actually proposed that a proto-flagellum served to “tickle other bacteria.” Rather, I was making an ad absurdum argument to highlight that our observation of the existence of other similar systems in nature is irrelevant to the question of whether our current system exhibits irreducible complexity. Yet it is the existence of similar looking systems that evolutionary proponents often latch onto as “proof” of an evolutionary relationship. Thus, my comment goes to the heart of the co-option hypothesis championed by Miller and others. Further, my comment is relevant to the broader concepts of homology and convergence, though I do not discuss that aspect in my paper.
Finally, my point is instructive to ID proponents in their efforts to move forward their position, because it is very tempting to get into a debate about whether such and such a structure evolved from another. While it is intellectually interesting and publicly important to counter evolutionary proponents who assert an evolutionary relationship (think Minnich’s recent paper challenging Miller’s notions on the bacterial flagellum), this is secondary (potentially even irrelevant) to the question of whether the systems in question were designed.
IP: Logged
|
|
Krauze
Member
Member # 1119
|
posted 31. October 2004 14:31
Hi Eric, and thanks for your reply.
"Thus, you are correct to note that I am not proposing new research techniques or insights into "investigation of the biotic world.""
Actually, I wasn't talking about "research techniques", but simply applications. For example, you applied your concepts to stools and cities, but I doubt you did any research in those areas.
"What I hope I have been able to do, however, is outline the logical framework within which the biological results are debated and interpreted."
As they say, "The proof of the pudding is in the eating." I see a number of problems in trying to apply IC to things like stools and cities. (When noting that "if the stool is used at a bar, it will need to be of a particular height", does that mean that "height" is a part? In the case of the city, is the entire road-net one single part, despite having been constructed by adding streets and extensions of these over time? etc.) I guess we could discuss these problems and maybe even come to an agreement as to how IC should be applied to stools and cities. But I', afraid that it'll leave us with a feeling of "okay, now what?" Knowing that a stool is IC (or not, if that's the case) will tell us what about, say, the flagellum? So instead of spending time discussing the application of IC to human-made systems, then discussing how to export our insights to biology, what about if we saved ourselves a lot of time and just started discussing the biology?
There's plenty of biological systems that needs in investigation based on a IC-related methodology. I can provide a list, if anyone's interested.
"Thus, my comment goes to the heart of the co-option hypothesis championed by Miller and others."
Yes, this was one aspect where I thought your apporach could bring some light, and I was surprised that you didn't bring it up in your paper.
PS. If you are interested in discussing some telic thoughts, you might want to check your e-mail.
IP: Logged
|
|
|
Printer-friendly view of this topic