|
Author
|
Topic: Organisms using GAs vs. Organisms being built by GAs
|
Janitor@MIT
Member
Member # 125
|
posted 21. September 2002 13:38
Re: Dr. Taylor’s demand for an “ID” explanation.
Every design is an experiment; an experiment in the process of discovery/innovation. Design isn’t merely the rigorous and detailed planning and implementation of intents and purposes with well-defined explicit objectives and goals. Such a conception of design eliminates even the possibility of discovery. Design is highly contingent, extemporaneous, ad hoc, filled with unknowns and uncertainties, it is developmental and evolutionary, and obviously, this is where the possibility of discovery enters the whole process. Design is experimental in exactly the same way as the natural sciences are experimental, except often the experiments are considerably more elaborate than those in natural science. It is in the elaboration and complexity that discovery is made.
(This debate appears to be informed by some rigid preconceptions of both what design is and what evolution is (or must be). A broader conception of both may be necessary.)
One thing is clear, this peculiar network topology, discovered originally by linguists, then rediscovered by economists, then rediscovered again by sociologists, then rediscovered yet again by physicists, and then once more by the engineers, then finally by the biologists, can and does emerge by a process of design that does not have as its objective scale-invariance .*
Now, to my mind this represents quite a paradox, because, as we are always reminded, evolution is a “mindless” process, with no objectives, goals, or purposes, etc. And even though evolution may in its process and products resemble design, we usually discriminate between the two on this very basis: evolution has no objectives, etc., while designers do.
Yet here we have two systems (e.g., the Internet and the proteome), isomorphic in a definitive global property, but one basis for the usual distinction we make between the origins of these two systems has evaporated! A network designer can design and build a network with the identical generic properties but with no explicitly formulated plan or intention of doing so.
Now, there are some technical points of issue: Dr. Taylor’s “expectation” appears to me to ex post facto. Biologists have typically been modeling proteomic networks per Poisson for the last thirty years, following the tradition explored originally by Kauffman, if I recall correctly. (Now, if Dr. Taylor did anticipate the discovery I applaud her prescience.)
Preferential attachment is problematic on both design and evolutionary theoretic terms, as has been noted by many researchers, including Alberts and Barabasi themselves. They have attempted to give it a “fit-get-fitter” spin, but this is a simplism, and is to be rejected on both evolutionary and design theoretic terms. W/o explanation or PubMeds (!); I’ll simply refer the reader to the already unmanageably voluminous research on the subject. There are many ways to reproduce scale-invariance.
Just very generally, the arrival at and persistence of any global, time-independent state via the usual evolutionary dynamics is difficult to account for. These dynamics should significantly modify such a global state even over very brief evolutionary time scales. My “expectation” was that genomes shaped by chance and contingency should have resembled the antlers of the Irish elk, the elephant’s proboscis, the fish’s gills, and the cat’s pajamas. Was I wrong here?
As is even implicit in the PA rule, the impression is that genomes are largely shaped by purely intrinsic factors. This would explain the fine-tuning that effectively renders as a constant parameter what should be a stochastic variable and relation in a truly evolutionary model. But I don’t think that biologists generally are prepared for or anticipating a possible paradigm shift that this implies. Not meaning to be intentionally provocative, but I anticipate that resistance to the very idea will be shaped by more “metaphysical” considerations, than by the science. Tradition is a powerfully enervating force, even in science.
Dr. Dembski’s criticism is well taken, if I’ve understood him correctly. We have formed a bad habit of calling a description plus some traditional “formula” an explanation. But I am far more demanding than he is: “just so-just happened” narratives are to be tested experimentally. Until they are tested such hypothetical scenarios are just tales that we tell and not science.
*I forgot the ecologists and planetologists!
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 21. September 2002 14:00
I disagree. "Just so stories" are an intricate part of scientific progress. When confronted with an unknown situation science progresses through the formulation of hypotheses followed by attempts to support or more importantly disprove the hypotheses. Of course such ideas need to be tested. And yes they are explanations, perhaps bad ones but they are an attempt to incorporate our present knowledge with the facts. Which is why using elimination to ´prove´ design seems ultimately doomed to be a failure. This is why I have been asking for design to be given the same requirements as regularity-chance hypotheses.
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 21. September 2002 15:09
This is another topic, but I think that the term "just-so story" is much abused, and has shifted considerably from its original meaning. Gould & Lewontin's paper "The Spandrels of San Marco" (read the actual article online at AAAS), usually cited as the origin of the term, does not even contain it. What Gould & Lewontin were arguing, however, was that some features of organisms have dubious functions, and might therefor have "accidental" origins, being the by-products of other processes (human sacrifice and T. Rex limbs were two of his examples); and this was the logic behind the original point of labelling something a "just-so story".
This is all well and good, but the one thing this critique does not apply to is organismal features that *clearly* have an important function and are clearly highly refined and adapted to perform that function. This includes, by the way, all of the IC systems favored by Behe et al.
As Maynard Smith wrote:
quote:
Confronted with feathers, or eyes, or ribosomes, we cannot not ask what they are for. It would be no more plausible to suppose that they are accidental and non-selected byproducts of something else than it would be to suppose that the gyroscope in the V2 rocket was connected as it was because some German fitter made a mistake.
Now it may be that Gould (or others) expanded the term to include "weakly-supported origins scenarios" in general, but it is at least worth acknowledging that this is different from original usage.
If we do move to the broader usage, however, it surely should be noticed that "IDdidit" is a much more flagrant just-so-story than any of the evolutionary hypotheses for the origin of the immune system advanced here.
nic
[ 21 September 2002, 15:11: Message edited by: yersinia ]
IP: Logged
|
|
Deanne M. Taylor
Member
Member # 274
|
posted 21. September 2002 15:35
Janitor@MIT is quoted down the following:
quote:
Re: Dr. Taylor’s demand for an “ID” explanation.
Every design is an experiment; an experiment in the process of discovery/innovation.
So, here, you are defining "design" as a process of experimentation/innovation in stepwise fashion (process, you said above), which must be optimized as it was once imperfect, instead of created whole and perfect with aforethought and deliberate planning.
In that sense, then, the "irreducibly complex", intelligently designed systems posited by people like Mr. Dembski were once more imperfect and seperable than they are now? Are you saying the Designer needed to experiment to get his/her design to work?
Despite your objections, I believe that ID posits deliberate, intelligent design that produces complex irreducible systems that cannot be seperable or evolvable. That is, ID posits that the Designer was competently able to design a system from the start.
That is, the design was never haphazardly created to later evolve into a better form. It was not a "process". Conversely, evolution claims a process. I, personally, am arguing for a process of "design" by natural causes.
I believe that ID insinuates that the design was perfect from the get-go. Perhaps you'd like to speak to Mr. Dembski on this.
I am addressing this claim of "designed from the get-go" in the previous messages on this thread.
quote:
Design isn’t merely the rigorous and detailed planning and implementation of intents and purposes with well-defined explicit objectives and goals.
So, you are saying the Designer was goal-less, and didn't have an idea what he/she was trying to accomplish when he/she started the design? That the Designer set out to experiment from simple forms up through more complex forms? That would work if the Designer was evolutionary process.
quote:
Such a conception of design eliminates even the possibility of discovery.
Conception of design, if wide-reaching, negates any possibility of discovery, as you can posit the whole shebang was designed to be random and associative, and designed to evolve according to the mechanisms proposed by evolutionary theory. So, yes, you could say "design did it, but evolution is the result", which wouldn't do much for the current theories in Intelligent Design.
quote:
Design is highly contingent, extemporaneous, ad hoc, filled with unknowns and uncertainties, it is developmental and evolutionary, and obviously, this is where the possibility of discovery enters the whole process.
Design done by optimization, perhaps.
Design done with acceptance of mistakes, perhaps.
I don't think "Intelligent Design" wants to see their "intelligent design" fall into the category of "intelligently designed to evolve exactly as the evolutionists are saying it all happens"
Design by random, natural causes are highly contingent, extemporaneous, ad hoc, filled with unknowns, etc. That's my point. I am saying that the "design" we see in biological networks can be explained by all of the above with the processes we have observed or have found proof of.
Do you have an alternative to the mechanisms I've named in previous posts to create a network with the same qualities we see in the protein-protein interaction network? Is design then to accomplish its objectives in the same way evolution does, or is it an event with a point of creation and no science involved?
Was the design of these irreducibly complex systems placed into being as a whole and total, irreducibly complex part, or did it evolve by often random process and sequential addition? You cannot defend intelligent design of irreducibly complex systems AND suggest that the irreducibly complex systems evolved through a process of optimization or discovery, &tc. That means that the system was once not irreducibly complex. Do you see the conundrum, there?
quote:
Design is experimental in exactly the same way as the natural sciences are experimental, except often the experiments are considerably more elaborate than those in natural science.
That's a whopper of an assumption. And completely contradictory to what you're saying. If design is ad-hoc in its experimentation, it's not planned. The "design experiments" as unplanned cannot be as elaborately planned, therefore, as most scientific experiments already are.
quote:
It is in the elaboration and complexity that discovery is made.
Discovery of what, the elaboration and complexity of random design? I still don't see how that supports an instantaneous creation of irreducibly complex systems by an intelligent designer.
Unless that intelligent designer was just like evolution is, without any of the science attached to back up proof of intelligent design. Then, I guess you could posit "intelligent design" except that you'd have to defend it using nothing but your gut instinct for what "Design" means. Not quite enough to justify teaching it in schools.
quote:
(This debate appears to be informed by some rigid preconceptions of both what design is and what evolution is (or must be). A broader conception of both may be necessary.)
You'd like to suggest a broader interpretation of Intelligent Design? Be my guest.
Evolution is already pretty broad. It posits everything you've been saying, except the "intelligent agent" are natural processes that follow certain rules of assembly, innovation and selection.
As for ID becoming more liberal in its definition, I'm not so sure that'd go over really well. Intelligent Design posits an intelligent Designer. Irreducibly complex systems are offered as PROOF for an Intelligent Designer. I would like to point out that what you're doing is removing the ability of the "IC" argument to defend ID. You're neutering it, essentially, and saying that an evolutionary process could be undertaken by an intelligent agent. Evolution is evolution. If you claim that irreducibly complex systems can evolve through experimentation, then that takes down Mr. Dembski's argument, and might also offend some people who might think that any powerful Designer was well-equipped to create perfection the first time around.
quote:
One thing is clear, this peculiar network topology, discovered originally by linguists, then rediscovered by economists, then rediscovered again by sociologists, then rediscovered yet again by physicists, and then once more by the engineers, then finally by the biologists, can and does emerge by a process of design that does not have as its objective scale-invariance .*
You know, removing the word "Design" up above describes evolution: "...can and does emerge by a process that does not have as its objective scale-invariance"
Of course, UNLESS you want to say that evolution is a process of design.
I have a feeling you are confusing the meaning behind "Intelligent Design". I have not read (yet, anyway) anyone purporting that Intelligent Design happened with random, imperfect processes to yield a final result! That's what evolution posits! Conflating the two removes the problem, but not, I think, for people like Mr. Dembski who may take exception.
I hope this does NOT evolve into a pedantic argument about what "design" means.
By the way, it was physicists who discovered scale invariance is pervasive through most networks that grow by random association and preferential attachment. Look up A.L. Barabasi at Notre Dame, and his papers over the past few years. If you know of earlier papers that he doesn't cite, please share them, so I can add the references to my collection.
quote:
Now, to my mind this represents quite a paradox, because, as we are always reminded, evolution is a “mindless” process, with no objectives, goals, or purposes, etc.
Ask yourself: Does consequence need purpose or goals?
If you make a decision in a completely random way you still have to take the consequences of your decision.
How are the consequences of a mindless decision any less important than the consequences of a mindful decision, if both decisions yield the same exact consequences?
quote:
And even though evolution may in its process and products resemble design, we usually discriminate between the two on this very basis: evolution has no objectives, etc., while designers do.
quote:
Yet here we have two systems (e.g., the Internet and the proteome), isomorphic in a definitive global property, but one basis for the usual distinction we make between the origins of these two systems has evaporated!
Yes, they both share a similar quality in that both were created by random processes. However, the Internet had a bit of designing here and there. The Internet, for the vast majority of it, is randomly "designed" and often at the mercy of selective processes. See Barabasi's papers on the WWW in the links I gave in earlier posts.
quote:
A network designer can design and build a network with the identical generic properties but with no explicitly formulated plan or intention of doing so.
The internet wasn't designed. But let's assume another network, say, in a University.
You're saying that a network designer is going to create a network that looks as if it was haphazardly designed on purpose. I seem to remember some papers on network design. They don't suggest scale-invariance as a good design principle. For instance, see Cisco.com's Campus network design white paper
In fact, I believe that a lot of "best practices" network design principles in IT these days seem to fall into the category of "bipartite graph", but someone out there who does that for a living should correct me if I'm wrong.
Anyway, good design principles of any network should NOT include hubs that can be hit and cause the network to break down completely. That's not a good network design, hence the bipartite graph structure, from what I remember.
As far as a designer designing a network to look as if it was randomly assembled, ID says that the Designer had a plan, and many of the components thus built must have been planned to be interlocking BECAUSE they are irreducibly complex. Yes, Irreducible Complexity says a designer must have planned these systems to be indivisible and perfect on design.
quote:
Now, there are some technical points of issue: Dr. Taylor’s “expectation” appears to me to ex post facto.
What, that I expect ID to explain why the concept of "irreducibly complex" design can explain the observed quality of scale invariance in biological networks?
Science provides proof and data on scale-invariance, gene duplication and retrotransposons which supports evolutionary theory. I'm waiting to see how this data supports ID while also supporting irreducibly complex systems.
quote:
Biologists have typically been modeling proteomic networks per Poisson for the last thirty years, following the tradition explored originally by Kauffman, if I recall correctly. (Now, if Dr. Taylor did anticipate the discovery I applaud her prescience.)
You recall incorrectly. Biologists were rarely modeling proteomic networks per Poisson for the last thirty years as there was precious little data to prove or disprove the models they generated.
I have a graph I made, about three months ago, of every "network" paper out there, on biological networks. Until the 1990s and the emergence of data from the Human Genome Project, there were very few papers, relative to papers in other subfields of biology. I have that graph, if you'd like it, and all my network/system biology references (all 800 of them...the graph is exponential!).
Additionally, I find it VERY interesting that you invoke Poisson models, which have been shown as inadequate to model biological networks. Poisson models are shown to be "old news", not sufficient, by several research groups, including Barabasi's papers, which I assumed you read, as you invoked him here.
Kauffman's work, which was indeed seminal, was established to study the distribution of connections, modeling gene networks as random boolean networks and was an early result in estimating their behavior.
HOWEVER, the scale-invariant property of protein-protein interaction networks does NOT fit Kauffman's model. The Kauffman model does not fit reality.
Scale invariance is an observed quality.
I quote an excellent short explanation of this for you, from this link
If you read Barabasi's papers you'll notice that he compares the Poisson distribution to the actual observed distribution, and it was clearly not a Poisson distribution. It is a power law.
See various papers on Poisson modeling of biological networks. For instance, here's an abstract at Cornell that says,
"Using three different measures, namely types of attractors, fraction of elements that are active, and length of period, we show that finite, scale- free networks are more ordered than either the Poisson or delta function networks below the critical point. Thus the topology of scale-free biochemical networks, characterized by a wide distribution in the number of inputs per element, may provide a source of order in living cells."
Other work has shown that Poisson models are not good approximations, first off. They've failed for modeling the internet which is scale invariant. See for instance this paper at citeseer.
Poisson models are not good enough. That's because the network is already scale invariant, and Kauffman's work was not based on the kind of data that Barabasi's group had access to.
quote:
Preferential attachment is problematic on both design and evolutionary theoretic terms, as has been noted by many researchers, including Alberts and Barabasi themselves.
I don't recall them ever saying it was problematical to the point of being impossible, which is what you're implying. Can you give me a direct reference and the context in which they're saying it, and then scan later papers to see what they say, later on, as more data came to light? Citing an earlier paper when they say "we haven't figured it out yet" is not acceptable as "problematical". It's "lack of data".
But before you bother, consider this. Preferential attachment in biological terms is perfectly reasonable because the hubs which are preferred are already, by definition, more ripe for gene duplications than their less-connected peers. A hub protein that has the most interaction partners will have a greater chance of receiving a gene duplication as a new partner since duplication events are generally random, and these hub proteins have more protein interaction parterns TO duplicate. That's not problematical. Were you reading my posts?
quote:
They have attempted to give it a “fit-get-fitter” spin, but this is a simplism, and is to be rejected on both evolutionary and design theoretic terms.
Um, by who, and by what? What's your measure? Simple disbelief is not allowed, here. You're invoking ignorance to explain away something you can't understand.
quote:
W/o explanation or PubMeds (!); I’ll simply refer the reader to the already unmanageably voluminous research on the subject.
You don't have to cite PubMed. Give us a name and a quote, at least, that says that biological networks can't evolve by gene duplication and attachment. I'll supply my own in kind that says they can. You'll have to start by negating Eichler's work (Case Western) or Gerstein's work (Yale). No, actually, you'll have to begin by negating gene duplication in the first place, which means negating the data coming out of the human genome project, as it provides the proof.
Start with Eichler's work. Maybe you can explain why gene duplication and rapid Darwinian selection cannot explain gene duplication and innovation. It's been observed, so why not begin by criticizing the experimental data.
quote:
There are many ways to reproduce scale-invariance.
Key word there: Reproduce. So?
There are many ways to reproduce a truly random number, too, and they involve writing a computer program and hoping your algorithim is sufficient (I remember the saying: rand() is NOT.)
quote:
Just very generally, the arrival at and persistence of any global, time-independent state via the usual evolutionary dynamics is difficult to account for.
Again, what do you base that conjecture on?
There's plenty of evidence that a complex network can have "attractors" that create a stable state out of apparent nonlinear behavior. Are you saying attractors cannot exist?
Biological networks, however, are feedback control networks, for all their complexity. Are you saying feedback control cannot stabilize a steady-state solution? Argue with the engineers.
If the state is to time-independent as a requirement for survival, why wouldn't regulatory control systems establish as neccesary dynamic-range tolerance over physiological states, and steady-state solutions of functional parameters?
In terms of systems, that's what evolution claims happens, essentially: survival of a lineage by a system's adaptation to changes that perturb the system from stasis. Sometimes that adaptation causes phenotypic change that reproductively isolates the descendents from the ancestral lineage.
Do you have any evidence that a feedback control network, created under selective pressure, can not become a more complex network by all the mechanisms I've been naming, and maintain a steady-state solution by selection?
There are papers on evolution of complex networks, including a model of the evolution by gene duplication, one out of the Santa Fe Institute:
http://www.santafe.edu/sfi/publications/wpabstract/200202008
quote:
These dynamics should significantly modify such a global state even over very brief evolutionary time scales.]
So what you're saying is that you believe that an evolutionary network cannot evolve as steady-state and therefore any impetus in this non-steady-state solution will modify the global state over brief evolutionary time?
Besides the fact that you're basing YOUR argument on "it can't happen because I said so", this is (almost) exactly what I said in a previous post.
If a system is perturbed with a gene duplication that becomes necessary and that evolves away from its duplicate, it would be expected the system will undergo a pendulum swing of evolutionary change to accomodate the duplication's drift and change.
And that's what we see.
See human chorionic gonadatropin as one example. If you want me to get more specific on this, just ask. It was a recent duplication in the primate lineage and there's a rapid (and traceable) positive Darwinian evolutionary change in the receptor-substrate systems up through the primates.
These changes are not expected to be found if a system is in evolutionary stasis.
Perturb the stasis and you see change.
Let things continue on in happy continuity and any impetus for change does not exist. Genes that could get fixed into that scale-invariant network do not need to get fixed. That's the general idea behind punctuated equilibrium: evolution happens not by gradualism but by fast spurts.
I might mention that drug companies use this principle -- that essential proteins are evolutionarily static -- and therefore do not evolve quickly in relation to their lineage. Those essential proteins are then studied in that context as targets (or dangerous targets). I know this, because I do that regularly at work. It's homology modeling. In fact, UCSC's genome browser has a whole track of human/mouse similarity for that very purpose...find the proteins that are conserved through evolution...those will likely be the hub proteins.
And to make matters even nicer, Barabasi's gone and done a study on yeast to show that the hub proteins are indeed important for cellular function. See the links in previous posts, under "cellular networks"...there's a very fine knockout paper in there.
quote:
My “expectation” was that genomes shaped by chance and contingency should have resembled the antlers of the Irish elk, the elephant’s proboscis, the fish’s gills, and the cat’s pajamas. Was I wrong here?
Um, they do. That's why we talk about PROTEIN FAMILIES. That's basic undergraduate protein structure, there. Go look up protein families.
quote:
As is even implicit in the PA rule, the impression is that genomes are largely shaped by purely intrinsic factors. This would explain the fine-tuning that effectively renders as a constant parameter what should be a stochastic variable and relation in a truly evolutionary model.
What parameter? What variable? What fine-tuning? Genomes are full of crap, generally. The human genome even more so. Which supports, by the way, swaths of segmental duplication. But we'll leave that alone for now.
Please explain the above paragraph, it's very...fuzzy.
quote:
But I don’t think that biologists generally are prepared for or anticipating a possible paradigm shift that this implies.
I could figure out what you were implying if you'd come out, directly, and say what "this" is that is being implied.
quote:
Not meaning to be intentionally provocative, but I anticipate that resistance to the very idea will be shaped by more “metaphysical” considerations, than by the science. Tradition is a powerfully enervating force, even in science.
I'm still looking to see what the "very idea" is that biologists would resist. Please try to be more explicit for people like me, who cannot read minds.
quote:
Dr. Dembski’s criticism is well taken, if I’ve understood him correctly. We have formed a bad habit of calling a description plus some traditional “formula” an explanation. But I am far more demanding than he is: “just so-just happened” narratives are to be tested experimentally. Until they are tested such hypothetical scenarios are just tales that we tell and not science.
Exactly my point. I've pointed out evidence for all of the things I am talking about. ID and IC does not have any data that provides evidence that supports intelligent design and excludes evolution (can you name some that I'm not aware of?). However, evolution has plenty of data that is evidence to exclude intelligent design. When you suggest that science provide proof, perhaps you'd like to demand it of evolution AND intelligent design, simultaneously. Evolution will present you with thousands of tests of evolutionary theory. The person who can prove that evolutionary theory is wrong, using ordinary scientific means, will get famous and also rich. That nobody has been able to do that thus far should say something to you.
You make some grand pronouncements on what is or what is not possible, but you do not present explicit examples or name specific scientific findings. If you feel the need to debate with me again on widely-published results, please name your sources, and I will (again, as I have been doing), reply with mine.
Deanne
[ 21 September 2002, 16:53: Message edited by: Deanne M. Taylor ]
IP: Logged
|
|
William A. Dembski
Member
Member # 7
|
posted 22. September 2002 08:23
I'm directing these remarks to Deanne Taylor's most recent posts. I had expected to have more time this weekend, but I don't, so I'm going to have to make it short.
It seems that Deanne works with a very limited conception of design and how it would play itself in natural history. For her, design creates things from scratch. Otherwise, it's evolution, and naturalistic evolution at that. Thus she will claim that the Internet was not designed. True, there was no micromanaged design in the sense that some one designer or central committee planned it out down to the last details. But the broad contours were designed and then subsequently intelligent agents acted within those contours. So, to claim that the Internet was designed is misleading. The Internet is chock-full of design.
With regard to biology, ID is certainly compatible with a nonnaturalistic form of evolution. The model I'm working with these days is that of technological evolution as advanced by TRIZ (see http://www.ideationtriz.com/). This nonnaturalistic evolution works by naturalistic processes solving "routine problems" and design solving "inventive problems" (this is a standard distinction within the TRIZ literature -- the one is solvable by trial and error, the other requires a conceptual breakthrough). This form of evolution therefore will have many of the features that Deanne cites (like scale invariance) but also exhibit clear marks of intelligence. At any rate, the design possibilities for how life emerged and developed are much richer than Deanne indicates or seems to grasp (she might have a look at chapter 6 of my book No Free Lunch).
One final remark: when I said that scale invariance "disappears" in Deanne's account of how IC systems arise when she lays out bridge principles to connect her scale invariance to actual biological systems, I didn't mean that scale invariance ceases to exist. I was simply saying that scale invariance did not figure in substantively or even appear in her account of those bridge principles. Indeed, she merely reverted to the standard coevolution/cooption explanation. I therefore continue to maintain that scale invariance has very little empirical content but merely serves as a convenient framework for organizing certain facts.
IP: Logged
|
|
Deanne M. Taylor
Member
Member # 274
|
posted 22. September 2002 10:02
Mr. Dembski writes: quote:
It seems that Deanne works with a very limited conception of design and how it would play itself in natural history. For her, design creates things from scratch. Otherwise, it's evolution, and natural evolution at that.
Please keep my statements in context. I was discussing the general definition of "design" vs. "naturalistic evolution" and drawing a line between the two in that one supposes a deliberate design, while the other supposes that natural process over extended time produces the patterns as seen in nature.
If you call my conversations on "design", as supposing a designer, as NOT the idea behind Intelligent Design, then I was perhaps being too narrow in my assumptions. Are you telling me that "Intelligent Design" does not need to suppose an intelligence behind the design?
If not, then how do you seperate a design with no intelligence behind it, from natural unintelligent process?
Mr. Dembski writes:
quote:
Thus she will claim that the Internet was not designed. True, there was no micromanaged design in the sense that some one designer or central committee planned it out down to the last details.
I don't have to "claim" that. It's not even in question. Why are you setting up this strawman?
Mr. Dembski writes:
quote:
But the broad contours were designed and then subsequently intelligent agents acted within those contours. So, to claim that the Internet was designed is misleading. The Internet is chock-full of design.
To claim the contrary is actually what is misleading. If you check your history of the Internet, it was NOT created as a large superstructure that was later filled in. It was a simple network connecting Universities and government labs, only 30 of them, from my memory, not the thousands it does, now. It was ARPNet, I believe, that connected Universities who already had the infrastructure in place internally and the internal support hook in. Thus, it was preferentially created at first to link together nodes that were ready. This was not "design", this was "let's hook up whomever is ready and we can try this out".
And it just got more haphazard from there. Do you remember the early days of the internet when one router going down could wipe out half the country? I do. I was using the Internet during its tender years.
I remember we used to curse Sprint for effectively knocking out half the entire US network with their crappy routers and service. That's because the early internet was poor in design -- it wasn't designed at all. It was slapped together in stages of adaptation and cooptation. That was my point. Not designed as a complete whole with an objective in mind.
One router, one hub, would take down the entire network. That's not intelligently designed Internet, there. That's haphazardly slapped together early infrastructure, and that's the way it was.
If you are saying, now, that ID accepts that things can be slapdashed together as design principle, then as I'll point out in several places in this post ![[Smile]](smile.gif) that you cannot use IC as an argument for intelligent design, since intelligent design would never create IC from the get-go. IC would have once had to have been reducible complexity.
I expect you set up the strawman to knock it over with the fact that the Internet had to have people to fill in the connections in between the loose "design" of the Internet.
That's a weak argument. The Internet structure sprang up out of necessity and underwent a lot of pruning and ...well, pardon the expression, evolution. The backbones followed where the Universities had already established data-sharing strategies.
The ironic thing behind your statement is that the basic layout of the Internet was based on necessity, on a layout of what Universities and hubs were already established and had been added to through ARPAnet.
In other words, this "basic design" of the first Internet was based on cooptation of early architectures that already existed and were subsequently added to through necessity and opportunity.
See, for example, Walt Howe's "History of the Internet".
http://www.walthowe.com/navnet/history.html
or the PBS timeline on the Internet:
http://www.pbs.org/internet/timeline/
So, if in the light of that, if you want to claim that the Internet was "designed", by cooptation and constant adaptation and re-deployment, I certainly do not object. You are saying design requires constant adaptation and addition from simple systems to more complex, "irreducibly complex" in some cases, systems.
Therefore, are you admitting that at a certain points in these "adapted designs" with or without intelligent agents, "irreducibly complex" systems were not once NOT irreducibly complex? That the intelligent agents BUILT UP the irreducibly complex systems from simpler systems?
This kind of architecture-building requires that once simple systems can become complex. Even in systems like the Internet. Simply saying "See, the Internet IS designed" does not support the claim that "irreducibly complex" systems prove designers, if the designers themselves had to build up these "irreducibly complex" systems from simpler systems. This means you allow that simpler systems yield more complex systems, that fully "irreducibly complex systems" are able to be built out of smaller components.
I imagine you can see the contradiction, here. You cannot claim intelligent design on the basis of irreducible complexity if you say that design must admit an evolutionary change, since you therefore must acknowkedge that irreducible complexity is incompatible with this alternative idea of gradualist design!
Unless, of course, you claim that design is full of slow gradualist evolution intespersed with sudden appearence of IC design as soon as the designer thinks the system is "ready" for it, like modular plug-ins on occasion.
In that case, how can you tell which systems are IC and which are not, as there seems to be a nice blend of "IC" to "not IC" along the power-law spectrum that scale-invariance observes. See the Barabasi paper on yeast knockouts, I believe they cover that in there.
Life seems very willing and able to adapt to perturbation, as knock-out studies show.
How would you seperate out such an approach from punc eek, which posits the same exact thing,except punctuated equilibrium holds that natural processes, such as bottlenecking of populations, causes these rapid and sudden changes in system architectures?
So, design must hold that:
- Design springs forth in the instantaneous, complete and limited sense (the "limited view" you were accusing me of), or
- Design proceeds by processes that go from simpler to more complex as we observe, and "irreducible complexity" cannot prove design in and of itself, thus removing IC from your stable of "proof" for ID, or
- you (colloq.) assume that design proceeds by gradual adaptation and "filling in" (evolution) of networks in early life forms to gradually build up a more complex life form but you assume it does not do so by random natural process. You assume that occasionally in that gradual process an "intelligent agent" inserts something that is so amazingly unadaptable that you view it as proof for intelligent design.
The only problem with the last point, of these sudden insertions of IC, is that if they are sudden insertions into a network, they must be seperable by measure from a hub or IC system that was evolved by natural processes. For instance, do you see any hubs that lay in illogical places according to the random model of gene insertion, drift, innovation? Thus far, that has not been observed.
In biological systems, such definition of "design" by incremental change is what is posited by evolutionary theory in the first place, only evolution requires on natural causes, not specific design.
In the case of evolutionary theory, we have evidence for what those processes originate from.
We have evidence or have observed the mechanisms in action, as I have posted previously.
In the case of that new idea of 'design', you have no evidence that biological networks have an intelligent agent behind them, nor that "design" excludes evolution from explaining the same processes. Both theories on that level ONLY are indistinguishable.
Mr. Dembski writes:
quote:
With regard to biology, ID is certainly compatible with a nonnaturalistic form of evolution.
Then you are saying that "irreducible complexity" can be generated from reducible complexity and ID is okay with that?
Mr. Dembski writes:
quote:
The model I'm working with these days is that of technological evolution as advanced by TRIZ (see http://www.ideationtriz.com/). This nonnaturalistic evolution works by naturalistic processes solving "routine problems" and design solving "inventive problems" (this is a standard distinction within the TRIZ literature -- the one is solvable by trial and error, the other requires a conceptual breakthrough). This form of evolution therefore will have many of the features that Deanne cites (like scale invariance) but also exhibit clear marks of intelligence.
You will have to establish how "inventive problem solving" through naturalistic process will be discernable from "inventive problem solving" by an intelligent agency. Do you have any details on how you'll go about trying to tell the difference between the two cases as to exclude one over the other with a scientific method?
Mr. Dembski writes:
quote:
At any rate, the design possibilities for how life emerged and developed are much richer than Deanne indicates or seems to grasp (she might have a look at chapter 6 of my book No Free Lunch).
Please be attentive to the fact that here, you are thinly veiling an insult established by assuming that I can indicate the entire sum of my knowledge and expectations in ISCID posts on one subject. I was attending, in earlier posts, to the fact that IC, in support of ID, does not tolerate the idea of gradual addition. That was your entire thesis in the first place. Are you saying you are now not requiring systems to be always irreducibly complex in their entire histories? If you are requiring that, then your accusations of my limited view are groundless, since that was directly what I was addressing. I would expect you know that.
As to what we were discussing in other posts, I asked you about mathematical properties of observed networks, and the mechanisms that have been studied and have generated data in terms of how this observed data fits into ID and IC systems.
You have, thus far, given me a thinly veiled insult, a wave of a hand back to your book, and a reference to a theory of problem-solving. None of them really address how ID and IC fit into the observed characteristics of the natural world that I have cited, including gene duplication, scale invariance, retrotransposons, segmental duplications, and the like
I'm willing to wait for ID to catch up to recent events in genomics. I might mention, aside, that evolutionary theory hasn't worked very hard to catch up...the observations were unexpected but not unpleasantly surprising.
I'm willing to wait to hear an explanation for all these events on solid mathematical terms from the ID side, how any ID theory can be supported by mathematics intimately tied with observation of the natural world.
Mr. Dembski writes:
quote:
One final remark: when I said that scale invariance "disappears" in Deanne's account of how IC systems arise when she lays out bridge principles to connect her scale invariance to actual biological systems, I didn't mean that scale invariance ceases to exist. I was simply saying that scale invariance did not figure in substantively or even appear in her account of those bridge principles. Indeed, she merely reverted to the standard coevolution/cooption explanation.
You didn't read carefully enough, then, or you missed the point entirely. I linked my explanation TO an example of how a HUB protein, an "irreducibly complex" system, can change and indeed confer a new pathway into a system without ever needing to change in a discontinuous way, and how the mechanism of gradual change can add to the qualities of a scale-free network. I used observed natural processes to explain this.
ISCID cites the Santa Fe Institute on its website when it talks about complex systems. Perhaps you'd like to read the working paper on gene duplication creating complexity in biological systems at that very same site...
http://www.santafe.edu/sfi/publications/wpabstract/200202008
Abstract of the Santa Fe Working Paper,
"Evolving Protein Interaction Networks through Gene Duplication"
quote:
The topology of the proteome map revealed by recent large-scale hybridization methods has shown that the distribution of protein-protein interactions is highly heterogeneous, with many proteins having few links while a few of them are heavily connected. This particular topology is shared by other cellular networks, such as metabolic pathways, and it has been suggested to be responsible for the high mutational homeostasis displayed by the genome of some organisms. In this paper we explore a recent model of proteome evolution that has been shown to reproduce many of the features displayed by its real counterparts. The model is based on gene duplication plus re-wiring of the newly created genes. The statistical features displayed by the proteome of well-known organisms are reproduced, suggesting that the overall topology of the protein maps naturally emerges from the two leading mechanisms considered by the model.
That is, a mutation in the system will NOT kill the system, especially if the mutation is necessary. You will notice I termed everything in the context of irreducibly complex systems.
As an aside, it's nice to see Kauffmann will be talking at ISCID's chat room on the 25th of November. The kind of adjacency he discusses in his new book supports the gradual adaptation of complex networks through adjacent states...something you dismiss as "the usual" up above.
Mr Dembski writes:
quote:
I therefore continue to maintain that scale invariance has very little empirical content but merely serves as a convenient framework for organizing certain facts.
Am I correct in assuming you think that protein-protein interaction networks, real physical quantities, operate in a framework that is mere organizational convenience? That's like claiming that the US highway system is mere organizational convenience, instead of being an entity unto itself.
Again, I'm a patient person. I can wait until ID publishes something that can provide mathematical, discerning analysis that will be able to seperate out the conditions for the emergence of these scale-invariant networks between Intelligent Design and evolutionary theory.
Several people I know, however, call that patient part of me an "idealist". Being the cynic I am, I take exception to that.
[ 22 September 2002, 11:53: Message edited by: Deanne M. Taylor ]
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 22. September 2002 10:31
Dr Dembski,
I find it interesting to hear you propose a more detailed idea of ID. With TRIZ you seem to be suggesting that naturalistic processes will solve routine problems and that design solves "inventive problems". Does this include the possibility that design can involve naturalistic processes as well? But more interestingly you seem to be proposing a design in which the designer(s) intervened and various intervals.
I would like to hear more about the detailed pathways to support such a design hypothesis that surely must exist for it to be more than a 'just-so' story.
I do agree with your statement that design possibilities may be ´rich´ in the sense that design can explain anything and thus nothing as long as it relies on elimination rather than on a positive approach.
Perhaps we can learn from nature how it ´designed´ and apply this in useful manners to improve our own capabilities of design. Looking through the TRIZ website I am not too surprised to see that creative techniques include
2. Randomization
6. Evolutionary directed techniques
An example of an innovative-creative step found in nature would for instance be gene duplication which now expands the dimensions of the original search space. Or HOX genes which seem to have been instrumental in the Cambrian explosion.
Another creative step involves the reuse of available materials in the jawbone of reptiles to form the inner ear bones of the mammal.
We may find evidence of creative or innovative solutions but are such evidence supportive of design and not evolution? Or do we need detailed pathways for either to assign probabilities allowing us to determine which of these hypotheses is most likely?
IP: Logged
|
|
William A. Dembski
Member
Member # 7
|
posted 22. September 2002 20:52
Briefly Deane,
**You write explicitly "The internet wasn't designed." Granted, you probably meant that there was no grand top-down design architecture, and that is true. But with the internet there are elements of real design through and through (not random design, natural design, or some other reductionist conception of design). For instance, the transfer protocols had to be designed, promulgated, and rendered conventional. The point at issue is whether such real design pervades biology.
**As for your bridge principles, I looked at them again, and scale invariance in your account ends up dispensable as an analytic tool in accounting for the origin of IC systems. You are telling a standard coevolution-cooption story. You then try to make it still more plausible by pointing to scale invariance, as though this will further enhance natural selection's ability to suitably channel the evolution of biological systems. But that added plausibility depends on having accepted the standard evolutionary story in the first place. I see no causal power in scale invariance. I still see only a framework for rendering plausible what you already believe according to general evolutionary theory, and not a means of independent verification. Your point #7 is unanchored speculation.
**I looked at the Santa Fe Institute article you cited in the last post. I don't think the mileage you are trying to get from it is warranted. Protein-protein interaction maps are fine and well -- they describe a biological reality. But then, in standard Santa Fe fashion, we go to a mathematical model, and suddenly everything we want to be true about evolution is confirmed (with considerable fine-tuning of the mathematical model, I might add; though you, no doubt, will justify such fine-tuning by a programmer as mirroring fine-tuning due to natural selection). This is a toy model, as the authors themselves admit: "The model is a simple approximation to the evolution of the real proteome map, and no functionality is considered (i.e. no dynamics is explicitly introduced)." Note well: "No functionality is considered."
**As for TRIZ, I take it that IC provides the analytic cut between routine and inventive problems in biology. This requires justification and I'm working on it.
IP: Logged
|
|
Janitor@MIT
Member
Member # 125
|
posted 23. September 2002 11:33
No, Dr. Taylor, I’ll decline your invitation to debate. You’re quite a formidable debater!
As you’ve indicated, a “war of authorities” would appear to be pointless. And some of the participants have indicated that it’s tiresome. I will however add to the ISCID Bibliography the references that support the statements of fact that I made, and I will include comments as indicators.
Just a few observations: Obviously, the IDers conception of design is far broader than mine. My ideas about design, such as they are, are informed by own limited and mundane experience. But they are also informed by what little I know of biology. It seems to me that biologists’ understanding is similarly informed, as biologists seem to be quite familiar with engineering design methods, concepts, and theories: codes, programs, networks, communications, control, regulation, strategies, optimization, robustness, etc. I find it ironic that biologists’ understanding of life has been so profoundly shaped by design, even since antiquity, and never more than now, but they feel compelled to deny the obvious: Life is design.
It’s not difficult to trace the source of this denial, because it wasn’t always true that biologists were so insistent that life is not designed. But then when one refers to the source one finds something very interesting: Charles Darwin cannot be the source of the “anti-design” notion, because he makes of his theory an argument from design : “I have called this principle, by which each slight variation, if useful, is preserved, by the term of Natural Selection, in order to mark its relation to man's power of selection.” (He even closes his opus with a “god-of-the-gaps” argument!)
Here, Darwin has had the perfectly naïve insight, which he can’t make explicit, because it never appears to rise to the level of consciousness, that the intentions, designs, actions, and productions of mundane intelligent agents are ultimately explicable in terms of a principle of variation and selection. (I had a college instructor who made the very point: All mathematics, e.g., is “combinatorics.” I had another instructor concur when he informed us that all engineering is “optimization.”)
Darwin’s motives, as revealed in his Autobiography, are quite plain, but also quite confused, even on his own admission. He’s plainly identified traditional, “revealed” religion with the natural theology that was so influential of his scientific contemporaries. But Darwin was a little myopic. His target was revealed religion, he took aim at natural theology, but his dart struck at the very heart of biological science . Darwin’s poorly conceived and executed attack upon religion has misfired disastrously. The IDers are quite correct to point this out. Among other things, what has suffered is the science, because the biologists cannot fully explore and exploit their own design paradigm as long as they are influenced by Darwin’s “muddled” theology.
To an outsider such as myself, this confused and contradictory etiology of ideas is quite evident and paradoxical: Biologists often invoke design and occasionally argue it with some sophistication and in detail, all the while denying that it is, indeed, what they are talking about! In my own round about way, what I’ve suggested to the IDers is that the biologists might be led to a less “muddled” conception of the design of life via baby steps; via the very design “metaphors” they routinely employ—beginning with natural selection, e.g.
Now, none of this addresses Dr. Dembski’s topical question. But no, GA’s do not explain the origin of anything. If we say that they do, we beg the next embarrassing question that he no doubt already has in mind. If science must assume an agnostic position wrt to origins, well fine. Let’s not compound embarrassment with absurdity. Scientifically, we can only treat life as a found object and as found, it is an intelligent design indeed. Darwin’s emphasis upon the chance, contingency, and necessity as eliminative of design is nothing more than an argument from ignorance. And I have repeatedly and rather pointedly criticized it: the most persistent argument from ignorance made in this “debate,” Dr. Taylor, is an argument from ignorance of design.
IP: Logged
|
|
Deanne M. Taylor
Member
Member # 274
|
posted 29. September 2002 09:16
I was away at a conference all week, so I'm back to continue.
Mr. Dembski writes:
quote:
**You write explicitly "The internet wasn't designed." Granted, you probably meant that there was no grand top-down design architecture, and that is true. But with the internet there are elements of real design through and through (not random design, natural design, or some other reductionist conception of design). For instance, the transfer protocols had to be designed, promulgated, and rendered conventional. The point at issue is whether such real design pervades biology.
I wasn't talking about the specifics of network component protocols in earlier posts (in the context of specific interactions between network components). I was talking about the entire network's observed architecture, and the need to develop theories based on what is observed.
However, if you say transfer protocols had to be designed (and might I add, optimized!) during the maturation of the current state of the Internet, that is very true. But moving into this new analogy, let's examine the parallels in evolutionary biology to TCP/IP and other protocols. Transfer protocols in protein-protein interaction networks need no such special creation design: "communication protocols" in biological networks are simple physical interactions: negative charges are attracted by positive charges, hydrophobic surfaces attract through Van der Waals forces, proteins interact by simple physical interactions. Ab initio, the "protocols" already exist in nature in the very descriptors of physical interactions and behave in the same way that interactions between inorganic molecules behave. That is, the physical laws apply to proteins as well as chemicals as well as, well, rocks.
Using the protocol analogy, if you want to argue that the entire Universe was designed before it sprang into being to specifically allow proteins to interact through intermolecular forces that act as "interaction protocols", that's a design argument that can never enter the realm of science, as it can never be isolated from the naturalistic origin theories of the Universe nor independently studied and compared to "another" Universe.
Mr. Dembski writes:
quote:
**As for your bridge principles, I looked at them again, and scale invariance in your account ends up dispensable as an analytic tool in accounting for the origin of IC systems.
Except that scale invariance in networks has been mathematically shown to occur in those systems that are randomly assembled with preferential attachment. Translating that into evolutionary theory, that means gene duplication on well-populated nodes and then mutational drift, to follow the example you were talking about.
Mr. Dembski writes:
quote:
You are telling a standard coevolution-cooption story. You then try to make it still more plausible by pointing to scale invariance, as though this will further enhance natural selection's ability to suitably channel the evolution of biological systems. But that added plausibility depends on having accepted the standard evolutionary story in the first place.
Can you come up with an alternative explanation that would fit the observed data? Can you give another line of plausible reasoning that can link together "genetic fossils" which are evidence of gene duplication, junk DNA, scale-invariant networks that assemble by preferential attachment, and mutational drift, also specific conservations of proteins in special hub-positions of that interaction network?
Mr. Dembski writes:
quote:
I see no causal power in scale invariance. I still see only a framework for rendering plausible what you already believe according to general evolutionary theory, and not a means of independent verification. Your point #7 is unanchored speculation.
Scale invariance in biological networks has an immediate causal power in the requirement that it arise from random processes with preferential attachment, which is anchored quite firmly to the mathematics. And gene duplication is anchored quite firmly in mathematical analysis of the human genome. And those are just two examples, not to say that everything else I've discussed isn't also experimentally verified. I just don't feel like repeating myself, see previous posts for references.
If you want to label the linking together of several observed phenomena with a system that has an observed architecture that is best FIT by those observed phenomena as "unanchored speculation", then I would ask what you call "Irreducible complexity" in that light? It's certainly much weaker in position by its lack of supporting evidence in the first place, especially since the same characteristics of "irreducible complexity" arise from a natural model that includes protein-protein network scale invariance.
Mr. Dembski writes:
quote:
**I looked at the Santa Fe Institute article you cited in the last post. I don't think the mileage you are trying to get from it is warranted. Protein-protein interaction maps are fine and well -- they describe a biological reality. But then, in standard Santa Fe fashion, we go to a mathematical model, and suddenly everything we want to be true about evolution is confirmed (with considerable fine-tuning of the mathematical model, I might add; though you, no doubt, will justify such fine-tuning by a programmer as mirroring fine-tuning due to natural selection).
To say that the researchers had to do "considerable fine-tuning of the mathematical model" is disingenuous.
For those who want the reference, here's the paper and abstract again:
http://www.santafe.edu/sfi/publications/wpabstract/200202008
Let's examine their assumptions in detail:
That correlation of gene duplication events to re-wiring events have to do with the assumption of evolutionary significance and an examination of how uncorrelated rewiring has no support in observed network structure:
From the paper:
quote:
"The reason to consider correlations has to do with the assumption that the evolutionary significance of gene duplication lies in the fact that changes in the newly created genes can lead to the emergence of novelty (Patthy, 1999). After gene duplication, one of the two copies becomes redundant and either one of them becomes non-functional (i.e. a pseudogene) or accumulates molecular changes that provide a new function. The new function might be very different. An example is provided by mouse lysozyme genes. One of them has a digestive function in the intestine and the second has a bactericide action in myeolid tissues. Strong divergences from the original function displayed by the ancestor can develop. Moreover, from a numerical point of view, the analysis of the models in which creation or deletion of links is uncorrelated yield results which are in disagreement with the experimental observations in real proteome maps."
So their first assumption is based on observation: that correlated addition and loss of links is supported by the observation, the alternative is NOT supported.
The article also says (page 9):
quote:
Two essential components define the model: growth by single gene duplication plus correlated re-wiring. Unequal cross-over is actually known to be the dominant contribution to genome growth and dynamics (Ohno, 1970). The second rule is inspired in the assumption that
novelties derived from changes in regulation patterns will be constrained by the functional properties present in already established interacting networks or subnetworks. Such constraints are likely to be relaxed when new genes are created through duplication.
The authors assume that the rate of gene duplication and re-wiring is not important -- it is time independent. (ie, you can model the current state of the protein-protein interaction network by sequential events with no assumed rate.) This "collapses time". Time is not important to this model.
The authors assume:
- Gene duplication occurs among the members of a protein-protein interaction network.
- Areas of gene duplications in a network are re-wired, preferentially, over areas that were not duplicated because of the implied freedom that area of the network now has to move away from its initial function. That is, the correlation exists between adding a duplicate and then re-wiring the local network, versus adding a duplicate and re-wiring another part of the network. This is also borne up by observation.
As a programmer, I do not argue here that fine-tuning a model is akin to evolution fine-tuning a model. Besides, the researchers here did not fine tune as much as compare their model to the data.
They then go on to say they derived the rate constants in this model from real data:
From the article:
quote:
We derived the rate equations for the evolution of the degree distribution nk(t) and its stationary states under some constraints imposed by available data from the analysis of yeast proteome.
I find it interesting that you suggest they "fine-tuned" the model, when their only assumptions were gene duplication (preferential attachment) and mutational drift and comparison then to the real data.
The "fine tuning" you might be referring to is their test of the assumption that there's a correlation between attachment and reordering, which isn't a fine-tuning, it's a test of the model against the real data. They assumed several models, ran them, and compared them to the data.
They tested their very simple "toy model" against reality and got a good match using known mechanisms of gene duplication and innovation.
There was no big black box with a dozen dials and arbitrary tweaking to get the results they were looking for. These are good scientists, and they're not going to waste time. As any physicist is taught early on, "Five arbitrary parameters can fit an elephant." But that's not the case, here, and they explicitly address that fact.
Mr. Dembski writes:
quote:
This is a toy model, as the authors themselves admit: "The model is a simple approximation to the evolution of the real proteome map, and no functionality is considered (i.e. no dynamics is explicitly introduced)." Note well: "No functionality is considered."
My goodness, you truncated that quote and insinuated a different meaning than what the authors intended. Here's the real version:
From the paper (page 9):
quote:
"The model is a simple approximation to the evolution of the real proteome map, and no functionality is considered (i.e. no dynamics is explicitly introduced). "
Read further, and they explain WHAT they mean by dynamics:
From the article:
quote:
Further developments of this model should consider different components of proteome structure and the underlying dynamics of protein-protein interactions.
To say they're somehow limited in that they're not examining the underlying dynamics of protein-protein interactions or time-dependent behavior of the gene duplication networks says nothing about their success in modeling the observed architecture from a simple model that went very far in its simulation based on two tightly constrained assumptions.
In other words, no functionality OF the network is presented. That little "i.e." at the end gives what they really meant. They're saying they're not modeling the dynamics, just the method of generating the scale invariance through gene duplication events. If they considered dynamics, it would strengthen their argument, not weaken it.
And it goes both ways. That a simple approximation can work so well with a minimum number of parameters is startling. That functionality didn't have to be considered in developing this model should tell you that the architecture is easily generated from a minimum of assumptions. That makes the model more robust than if they had to assume a dozen parameters and fine-tuned each of them, as I said above.
Mr. Dembski writes:
quote:
**As for TRIZ, I take it that IC provides the analytic cut between routine and inventive problems in biology. This requires justification and I'm working on it.
I would be careful in pointing fingers at scientists who make assumptions to generate their data, when you are, out loud, musing on doing exactly the same thing with much less supporting evidence...arbitrary assumptions that IC is true in the first place and can provide any kind of "analytic cut".
[ 29. September 2002, 10:21: Message edited by: Deanne M. Taylor ]
IP: Logged
|
|
Deanne M. Taylor
Member
Member # 274
|
posted 29. September 2002 11:11
Janitor@MIT writes:
quote:
Just a few observations: Obviously, the IDers conception of design is far broader than mine. My ideas about design, such as they are, are informed by own limited and mundane experience. But they are also informed by what little I know of biology. It seems to me that biologists’ understanding is similarly informed, as biologists seem to be quite familiar with engineering design methods, concepts, and theories: codes, programs, networks, communications, control, regulation, strategies, optimization, robustness, etc. I find it ironic that biologists’ understanding of life has been so profoundly shaped by design, even since antiquity, and never more than now, but they feel compelled to deny the obvious: Life is design.
Look, as I said in the previous post, I was talking about "design" from the ID perspective, design was a catch-all I was using to indicate design by intelligent agency. NOT design by naturalistic process. Life is design in the general term, yes, if you want to go OUTSIDE my clearly-stated definition of the same in the earlier post.
So, here, in this post, you conflate my stated definition of "design" (which was CLEARLY STATED as being limited to Intelligent design). Can I make that very clear to you?
In Previous Posts I Indicated that "DESIGN" meant "DESIGN BY SPECIAL AGENCY". Let's re-read that.
In previous posts I indicated that design meant "design by special agency", de novo design, springing forth outside the agency of evolutionary process.
Now that we have that clear, let me make this next point clear:
You're arguing that "Life is Design". Nobody is questioning that. You're basing some wide pronouncements on a section of the argument that I had never addressed. I never said that life did not show an architecture of some kind. That architecture, I argued, showed RANDOM process "designing" it by some kind of self-organizational principle (as Kauffmann might say, I think) rather than some intelligent agency appearing and saying "voila".
That I'm making such a clear point on all of this is because, reading through your letter, I can see that I have to explain this once again.
EVOLUTIONISTS DO NOT argue that life does not have *A* design in terms of "an architecture". Of course life has an architecture of some kind, besides the architectures we've explicitly observed, including Davidson's work on regulatory networks in sea urchin, how else could life function and behave in all the ways we observe?
But by re-stating that "life shows design" you're implicitly arguing semantics, here.
I am asking if life has a special creation kind of design. Was it DESIGNED by an external agency outside of naturalistic process? That's the question. That's the point of "intelligent design". If you ask an IDer if "design" was naturally evolved or specially created, they'd say life has a creation point, a designer of some kind. An intelligent agent in the sense that there's an intelligence behind the design.
To say that a piece of pure crystalline rock salt shows "a design" would get you agreement from crystallographers who can point out the regularities of spacing in the crystal among the atoms. Would they say that the crystalline solid shows "design" in the sense that it infers intelligence? Would you say that an invisible intelligent agent had to design that crystal to crystallize, or would you say that natural processes caused the crystal to assume its natural shape? That's the kind of questions we're talking about, here.
As I indicated previously, I do NOT want this to become some kind of debate on the definition of "design" and I clearly stated what I meant by design. But I've felt it is necessary to exhibit that the two meanings of "design" meaning some kind of ordered architecture vs. "design" in the idea of an intelligent agent creating that ordered architecture are conflated in your description, and that is a misleading conflation.
quote:
It’s not difficult to trace the source of this denial, because it wasn’t always true that biologists were so insistent that life is not designed.
You're switching definitions again.
Are you saying that life has some kind of A DESIGN, or are you saying LIFE IS DESIGNED? I don't think this is getting through to you, here, and it's worrisome.
Life can show a specific "design" without an intelligent entity designing it. Life can show "a design" from naturalistic processes, like a crystal can. It does not have to be DESIGNED by an outside entity, outside of natural processes.
There is no contradiction to evolutionary theory to say that life shows "design by natural processes". NOT by intelligent agents. Most biologists (note "most", there) would disagree with your assumptions.
And it wasn't so long ago that we thought the earth was flat, that the sun went around the earth, and big sky-turtles held up the world.
Doesn't mean we should go back to those erstwhile time-honored traditions, no matter how insistent the learned heads of Europe were on those issues at the time.
quote:
But then when one refers to the source one finds something very interesting: Charles Darwin cannot be the source of the “anti-design” notion, because he makes of his theory an argument from design : “I have called this principle, by which each slight variation, if useful, is preserved, by the term of Natural Selection, in order to mark its relation to man's power of selection.” (He even closes his opus with a “god-of-the-gaps” argument!)
Yes, and you must get out of your self-referential illogic, here. First, evolutionary biology has come quite a long way since Darwin. Gould's book "The Structure of Evolutionary Theory" does a nice job in showing how it's changed.
Second, neo-Darwinian evolutionary theory does posit the idea of general design arising from natural selection. The argument is where that "biological design" originates from. Darwin cited design by Natural Selection.
Evolution does not posit a design that springs forth outside of the agency of evolution or natural selection. Evolution does not accept a designer who creates biological systems entire by special agency. That is outside the realm of observation and of science.
You are conflating the issue of "observed design" which is a completely different issue than the AGENCY of design.
Design is either brought about by intelligent agency outside of the natural process of evolution and natural selection, or it's a natural process called Darwinian evolution.
quote:
Here, Darwin has had the perfectly naïve insight, which he can’t make explicit, because it never appears to rise to the level of consciousness, that the intentions, designs, actions, and productions of mundane intelligent agents are ultimately explicable in terms of a principle of variation and selection.
In which way are you using "intelligent agents", here?
quote:
(I had a college instructor who made the very point: All mathematics, e.g., is “combinatorics.” I had another instructor concur when he informed us that all engineering is “optimization.”)
And all evolved life forms show "design" by th unintelligent agents of random process and selection as suggested by the data.
quote:
Darwin’s motives, as revealed in his Autobiography, are quite plain, but also quite confused, even on his own admission. He’s plainly identified traditional, “revealed” religion with the natural theology that was so influential of his scientific contemporaries. But Darwin was a little myopic. His target was revealed religion, he took aim at natural theology, but his dart struck at the very heart of biological science . Darwin’s poorly conceived and executed attack upon religion has misfired disastrously. The IDers are quite correct to point this out. Among other things, what has suffered is the science, because the biologists cannot fully explore and exploit their own design paradigm as long as they are influenced by Darwin’s “muddled” theology.
Besides the fact that that is a muddled bit of rhetoric, above, making more assumptions than I think the gentle reader will care that I dissect, I'd like to politely ask that you cite your reference where ANY biologist is constrained by the simple premise that life evolves. In fact, I'd like you to show me some data (and feel free to use the ID'ers data) to somehow "prove" that biologists are constrained in interpreting the data based on evolutionary theory, because I all I hear from ID proponents is how widely accomodating evolutionary theory IS, in that it seems to fit data so well. You can't have it both ways. Either evolutionary theory is constraining and prevents biologists from interpreting data or it's so widely adaptable that it can fit any data.
Or are you suggesting, in the alternative, that evolutionary theory is constraining in that it is so adapatable to observation, that evolutionary theory is constraining to the point where we can't accept "Intelligent Design" where we're not looking? Find us some actual data on "intelligent design" and someone will pay money to study it and publish it in peer-reviewed journals. Until then, you're pointing at nothing.
What you're saying, is comparable in my mind to saying that Gravitational Theory is constraining to the point where we can't accept frogs fly when we're not looking. Sure, we might be missing the fact that frogs fly. Want to pay some scientist a research budget to spend all their time monitoring frogs for flight? There's no EVIDENCE for flying frogs. Find us evidence for flying frogs and we'll pay to study them. Until then, stop warning us that we might be missing the flying frogs. You're pointing at nothing.
quote:
To an outsider such as myself, this confused and contradictory etiology of ideas is quite evident and paradoxical: Biologists often invoke design and occasionally argue it with some sophistication and in detail, all the while denying that it is, indeed, what they are talking about!
Sigh. Again, you're conflating "design arising from natural process" to "design because someone put down the blueprints and manufactured it"
| | |
Topic Closed
