Uncommon Descent Serving The Intelligent Design Community

My Wistar Retrospective Talk

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

For those following this thread at the Panda’s Thumb, I’m providing here (as a pdf) the slides from my talk at the Wistar Retrospective meeting, held this past June in Woburn, Massachusetts.

Pay attention to the puzzle described in slides 14-21. Here’s a brief outline of the problem:

1. To establish cellular differentiation in a metazoan (i.e., an animal), instructions must be provided to the starting cell.

2. Natural selection is one possible process by which this occurred, when the metazoan in question first appeared.

3. A necessary condition for natural selection is reproductive capability.

4. But reproductive capability (in an animal) requires cellular differentiation.

5. Thus, a necessary condition for natural selection lies causally downstream from the phenotypic outcome — namely, cellular differentiation — natural selection was supposed to explain.

Those who have thought hard about this problem (e.g., Wolpert 1994, Schlichting 2003) have appealed to the Baldwin effect to generate the instructions at step (1).

Wolpert writes:

The key to all development is the generation of differences between the cells, that is, making them non-equivalent (Lewis and Wolpert, 1976). Only if the cells are different can the organism be patterned so that there are organized changes in shape, and cells at specific sites differentiate into different cell types. How could this have evolved? …

Whatever the advantage, an environmental signal brings about a localized change in the organism which becomes elaborated with time. It could even lead to suppression of growth of adjacent cells and so the restriction of reproduction to cells at the opposite end of the embryo. An embryonic axis could be specified. Evolution of patterning has occurred….

Invoking again, the Baldwin effect, we can see how this process could be constitutive, that is the cells at the contact site could be genetically specified. This illustrates the advantages and economy for evolution of development that the Baldwin effect provides. In its absence it would be necessary to first genetically specify one group of cells as being different but without there being any selective advantage. Only then could there be changes in these cells similar to those described above. This sequence of events is unlikely in the extreme. The reason for invoking the Baldwin effect is that an environmental signal provides the initial basis for a developmental alteration, which could have a selective advantage.

(L. Wolpert, “The evolutionary origin of development: cycles, patterning, privilege and continuity,” in M. Akam et al., eds., The Evolution of Developmental Mechanisms [Cambridge, Company of Biologists Ltd., 1994], pp. 79-84; pp. 80-81, emphasis added.)

See also Carl D. Schlichting, “Origins of differentiation via phenotypic plasticity,” Evolution & Development 5 (2003):98-105.

Although he does not appeal to the Baldwin effect, Dan Brooks appears to agree that natural selection does not explain the origin of metazoan ontogenetic networks:

Nelson also claimed that natural selection is not capable of creating nematodes, which is true but irrelevant. In Darwinism, natural selection is not called upon to create, but to select. And yes, that does beg the question of where the entities acted upon by selection come from. Darwin suggested that the major mechanism was the then-unknown laws of inheritance, coupled with what he called the mysterious laws of growth, the mysterious law of the correlation of parts, and the cohesion of homologies. Today, those phenomena are less mysterious, but still fascinating and incompletely understood.

I look forward to further clarification from Dan.

Comments
"But a change in initial conditions — even a big change — does not necessarily change the attractor." This is true. However, if you have a system that is complex enough to require a class 4 system to implement, then to implement it would require actually making use of the chaotic areas of the class 4 system, which means that at least one if not multiple would indeed have the chaotic effects if you tried to evolve it. In all cases I know about, when a system is evolved using a Class 4 base system (AVIDA and a few others) the systems they evolve do not depend on the sorts of instructions which introduce chaos on their level of instruction. For instance, if you can find an AVIDA-evolved program which the end program _required_ a loop to be implemented (like say for computing the factorial function), I will be rather surprised.johnnyb
February 9, 2008
February
02
Feb
9
09
2008
02:22 AM
2
02
22
AM
PDT
johnnyb, I think your vagueness about the "drastic consequences" of changing initial conditions may be a source of confusion. Sensitivity to initial conditions means that system trajectories diverge exponentially when the initial conditions are changed ever so slightly. But a change in initial conditions -- even a big change -- does not necessarily change the attractor. In a natural and important sense, the system "does the same thing" when a change in initial conditions does not change the attractor. A reason that chaos is essential to homeostasis is that a chaotic system tends to return to its attractor after perturbation. Again, perturbation has a radical impact on the precise system trajectory, but as long as the system returns to the attractor, it has "recovered," and there are no "drastic consequences" of perturbation. Off the top of my (sleepy) head, I'd say it's more likely the attractor than the trajectory that contributes to fitness in a living organism.Cloud of Unknowing
February 9, 2008
February
02
Feb
9
09
2008
01:59 AM
1
01
59
AM
PDT
Cloud of Unknowing: I am full aware of all of the facts you state in #15 but yet do not see how they affect the argument. The behavior of Class 4 systems do not become less complex simply because resources are limited. And it still takes a class 4 system to create any sort of interesting program. This again is irrelevant to the fact that it has limited memory. In #17: "You seem to be reading Wolfram selectively." I think you are misunderstanding both Wolfram and my arguments about Class 4 systems. "He gives huge emphasis to the simplicity of systems that may be thought of as control units of universal computers. These systems have specified complexity far below the threshold of 400 bits Dembski uses in design inference. That they may arise by chance, even under a uniform probability model, is not a good result for ID." You are confusing the control systems with the software that runs on them. For example, Rule 110 is a very simple system which has been proven by Matthew Cooke to be Universal (the original conjecture was by Wolfram). This is a trivial system. What is not trivial, however, is creating the initial state (i.e. the "program") to get it to do anything interesting. My point is that systems like Rule 110 behave chaotically, and what Wolfram's results indicate is that without the chaotic behavior of systems like Rule 110, interesting programs would not be possible. However, because systems like Rule 110 are chaotic, we cannot expect natural selection to be able to build the program, because even slight changes to initial conditions will have drastic consequences, and in fact _must_ have drastic consequences somewhere along the line in order to develop a program that required a class 4 system to begin with.johnnyb
February 8, 2008
February
02
Feb
8
08
2008
10:10 PM
10
10
10
PM
PDT
johnnyb, Regarding #15, I inadvertently gave the base-2 logarithm of the estimated maximum number of quantum states available to the known universe. You seem to be reading Wolfram selectively. He gives huge emphasis to the simplicity of systems that may be thought of as control units of universal computers. These systems have specified complexity far below the threshold of 400 bits Dembski uses in design inference. That they may arise by chance, even under a uniform probability model, is not a good result for ID. By the way, uniform priors on physical events in a finite quantum universe are often inappropriate. It is common to see most of the probability mass assigned to events that require more bits to describe than the known universe registers. At least some of our perception of order in nature is due to the fact that we easily allude to disorderly events that are impossible in a universe as small as the one we know. Part of my "cloud of unknowing" is that I cannot tease apart imputed order and created order. Regarding chaos, there's nothing in chaotic individuals that throws selection for a loop. I once obtained apparent chaos in artificial neural nets with a small programming error. I meant to update neuronal outputs in parallel, but I instead updated them sequentially, and the rippling of activations through the system gave complex dynamics. Computational evolution of a population of the mutant nets yielded a substantial improvement on the best published results in prediction of famous chaotic time series derived from observations of nature. No human has an inkling of physical understanding of the natural system. In short, a very small "mutation" can engender complex dynamics, and an evolutionary system can obtain nonlinear models of a natural chaotic system in the complete absence of human understanding of the system and the dynamics of the models themselves.Cloud of Unknowing
February 8, 2008
February
02
Feb
8
08
2008
07:14 PM
7
07
14
PM
PDT
Bob O'H, I would be interested in just what the gene pool of cichlids is and what types of new elements are now present. For all I know no estimate of a gene pool for cichlids exists or likely to in the near future. I assume this is a complicated and massive effort since there are so many varieties. My gut instinct is that the so called species of cichlids are just refinements of an original gene pool and that no real novelty was created in the processes that led to so many species/varieties. Occasional mutations are surely the cause of some of the species but I bet all are trivial. Now I am not a biologist or have any technical knowledge in all of this but my understanding of human behavior leads me to my conclusions. If there were good information that novelty was being created in the origin of all these cichlid species, we would have never heard the end of it. It would be the smoking gun for Darwinian processes. My conclusions could be modified if any definitive empirical evidence exists. I have no ideological basis for my conclusions and accept Darwinian processes for providing much of the variety in nature and in the formation of all the cichlid varieties but I consider this trivial in the overall evolutionary picture. I have never seen Darwinian processes working to produce novelty. I have been searching for 9 years and have not found any evidence yet or anyone who could provide some. What I see is that people are sucked into the belief in Darwinian processes by the formation of all these species which is fairly obvious to me. Because this process produces the great variety of life we see we then assume that it works for the really hard stuff without any empirical evidence. Instead evolutionary biologists resort to arguments of deep time and fanciful story telling.jerry
February 8, 2008
February
02
Feb
8
08
2008
02:46 PM
2
02
46
PM
PDT
johnnyb, There is no empirical basis for claiming that a universal computer exists. The closest physical approximation is a finite-state system that follows instructions specified in a Turing-complete language. Universal computers are by definition infinite-state systems. When people repeat Wolfram's statements without repeating his provisos, they engender ignorance. It's awfully important to keep in mind the simplifying assumptions of models. Wolfram treats the "memory" of the known universe as effectively infinite. Whether that is reasonable depends heavily on what is being modeled. It has been estimated that there are at most 10^120 quantum states available to the known universe. Believe it or not, that number poses a very stringent limit on some everyday problems in computing. Arguments hinging on the finitude of the known universe are central to ID theory. No ID advocate should suggest that there are natural processes with infinite resources.Cloud of Unknowing
February 8, 2008
February
02
Feb
8
08
2008
02:18 PM
2
02
18
PM
PDT
jerry - the obvious example of an expanded gene pool is cichlids, which have (apparently) diversified massively. I haven't followed that literature closely, but I think there's evidence that loci with faster mutation rates are more diverse, which is what would be expected if the gene pool has been increasing. I would have to check out the details though, but not tonight... BobBob O'H
February 8, 2008
February
02
Feb
8
08
2008
12:20 PM
12
12
20
PM
PDT
Bob O'H, That would be great. It would have nothing to do with my proposition though unless you could show that all/most gene pools are now essentially larger with novelty and complexity in the body plans. The gene pool of multi-celled animals does get larger through mutations but from what I understand these are all trivial. If you have some contradictory information where sub species are more sophisticated in their body plans than their parent species, I and probably many others would be interested in the examples. If you have information that the gene pool of sub species are larger than their ancestors in any meaningful way, it would be interesting to know about it. From what I understand the concept of a gene pool so far has little concrete empirical evidence but as more individual genomes in a population get mapped, a realistic one may eventually form in the near future. From what I understand there is a full court press on with humans and it will be interesting to see where this leads. As a practitioner you are obviously more in touch with what has been done and is presently available and where current research is going. So we would be all ears to hear your insight.jerry
February 8, 2008
February
02
Feb
8
08
2008
11:27 AM
11
11
27
AM
PDT
Bob -- I think it's more than just about Gene pool size. I.e. it isn't the number of traits per se, but (a) how complicated they are, and (b) whether the complex part of them were already implicitly encoded in the genome. For (b) you should check out Caporale's corpus of work.johnnyb
February 8, 2008
February
02
Feb
8
08
2008
11:13 AM
11
11
13
AM
PDT
From an engineering PoV my issue with metazoans from single-celled organisms is wet electricty. Electric currents transmitted as needed, inside a wet environment. No free electrons to use as we do with conductive wires. Metazoans require something bigger. Something that won't get lost with all the sloshing around. Ions are used (sodium & potasium). These aren't always present in the nerve but are allowed in as required via pumps on the outside of the nerve and pores to let the ions in (and out once the signal has stopped). (some poinsons work by stopping the pumps and closing the pores) But these ions are "dumb". So at the ends of the nerves are different neurotransmitters- molecular submarines with differing shapes that are pumped out the end of one nerve cell, across the synapse- a gap to far for an ion to travel- carrying the message for te brain. This submarine docks with the next nerve cell via opposite electric charges- the sub is - and the dock is +- the fit is tight and the message is transmitted. However that submarine just can't stay there, so there are other molecules that demolish the submarines and escort the parts back to the transmitting nerve so they can be recycled. And once the signal has went through the nerve all the sodium (or potasium) has to be pumped out and the nerve reset. How can we test for that type of system arising via an accumulation of genetic mistakes?Joseph
February 8, 2008
February
02
Feb
8
08
2008
10:22 AM
10
10
22
AM
PDT
Jerry - what would happen to your ideas if it could be shown that the original gene pool had less diversity? For example, what if I were able to show you a population started from a single individual, but with more diversity that could be carried by one individual? BobBob O'H
February 8, 2008
February
02
Feb
8
08
2008
10:03 AM
10
10
03
AM
PDT
Kairos - NFL is a very general set of ideas. Basically, this is a subset of NFL - it shows why one class of search algorithms is specifically bad for a particular class of necessary targets. "Has this been formally proved for biological evolution?" It does not yet have the rigor of a formal proof, but I don't see why it would need to be proven specifically for biological evolution. If evolution behaves the way Darwinists claim, then the mathematical model fits. If it does not, and instead works through jumps or not at all, then the ID'ers are right. So, you could show that evolution still occurs, but just at the cost of having it no longer be non-teleological.johnnyb
February 8, 2008
February
02
Feb
8
08
2008
08:47 AM
8
08
47
AM
PDT
"this is why natural selection can optimize and arrange existing pieces, but can’t create new ones. If a “piece” requires functionality that needs an underlying universal computer to process it, then the pathway to build it simply does not exist in the proper Darwinian line. Has this been formally proved for biological evolution?" This is why Behe said that ID research should test the limits of what millions/billions/trillions/ of reproductive events have produced. He pointed to Lenski's work with prokaryotes as an example of good ID research. Now Lenski's skin would probably crawl if he thought he was doing ID research because he probably thinks he is doing just the opposite. But Behe is pointing to his research as ID positive research. Similarly, with multi-celled animals we have trillions of reproductive events but without a definitely known starting point but in the future we should be able to estimate that starting point very well. If the genomes of various species of a genera or family (such as cichlids, birds, various types of mammals) get sequenced and the gene pools of these species show only a reduction in the gene pools of the species from the estimated gene pools of the original ancestors, such research would obviate the creativity of the modern theory of evolution. If it shows that the gene pools of all the species in these families do not exhibit any creative elements but only reductions in the gene pool from some prior larger set, then this is ID supportive research. Such research is going on in the biology labs today all over the planet and like Lenski, the researchers would probably be upset to know that they are doing pro ID research. Because they are vitiating the power of Darwinian processes. As this research progresses, the natural question to ask is where did the original gene pool come from when variation and natural selection over millions of years cannot create it. While they may not point to ID as the answer, we will have a smile on our faces at their predicament.jerry
February 8, 2008
February
02
Feb
8
08
2008
02:02 AM
2
02
02
AM
PDT
#5 Johnnyb Interesting ideas. It seems to me that this line of reasoning is quite similar to what Dembski did for NFLT. Isn't it? The ultimate reason why complex creatures are tough to build comes from Wolfram’s Class 4 Complex Systems. In order to be capable of universal computation (and in fact, anything remotely interesting), a system must be a Class 4 complex system. However, these systems are by their nature chaotic, and therefore cannot be built through natural selection (the chaotic nature of the Class 4 systems prevent selection from working). This is why natural selection can optimize and arrange existing pieces, but can’t create new ones. If a “piece” requires functionality that needs an underlying universal computer to process it, then the pathway to build it simply does not exist in the proper Darwinian line. Has this been formally proved for biological evolution?kairos
February 8, 2008
February
02
Feb
8
08
2008
12:13 AM
12
12
13
AM
PDT
Paul - The ultimate reason why complex creatures are tough to build comes from Wolfram's Class 4 Complex Systems. In order to be capable of universal computation (and in fact, anything remotely interesting), a system must be a Class 4 complex system. However, these systems are by their nature chaotic, and therefore cannot be built through natural selection (the chaotic nature of the Class 4 systems prevent selection from working). This is why natural selection can optimize and arrange existing pieces, but can't create new ones. If a "piece" requires functionality that needs an underlying universal computer to process it, then the pathway to build it simply does not exist in the proper Darwinian line.johnnyb
February 7, 2008
February
02
Feb
7
07
2008
10:44 PM
10
10
44
PM
PDT
Paul, You say that because you're a young earth creationist. :-) Sorry, couldn't help myself. I read Bill's article on Bulverism before I read yours and absolutely couldn't resist trying it on for size at the first opportunity. It's a timesaver, that's for sure! DaveScot
February 7, 2008
February
02
Feb
7
07
2008
08:06 PM
8
08
06
PM
PDT
The formula in comment 2 should be: M^3 x FT - LF + FAM (fortuitous accumulation mechanism= metazoan My apologies for any confusion :)Joseph
February 7, 2008
February
02
Feb
7
07
2008
01:27 PM
1
01
27
PM
PDT
Bevets, this is pretty cool. 'Seems that without even observing the evolution of an irreduceably complex system, Ann Gauger has falsified the unfalsifiable ID. Not bad work!bFast
February 7, 2008
February
02
Feb
7
07
2008
12:44 PM
12
12
44
PM
PDT
Volvox*- a colony of cells in which we see basic differentiation, is basically the starting point- although it seems that the volvox can only be traced back to some 75 mya. However the extrapolation would be some other volvox-like colony. The story would be that "primitive" volvocine - the gametes- received some magical mystery mutations. So we have 3^m x FT (father time) next we have to factor in some culling effect, along with some accumulation effect- 3^m x FT - LF(less fit) + MAM(magical accumulation mechanism) = metazoans * see D. Kirk Seeking the Ultimate and Proximate Causes of Volvox Multicellularity and Cellular DifferentiationJoseph
February 7, 2008
February
02
Feb
7
07
2008
12:34 PM
12
12
34
PM
PDT
John Lynch is suggesting there has been a coverup over Ann Gauger's paper given at the conference.bevets
February 7, 2008
February
02
Feb
7
07
2008
11:20 AM
11
11
20
AM
PDT

Leave a Reply