|
Author
|
Topic: Speciation
|
nosivad
Member
Member # 767
|
posted 04. May 2004 16:30
Pim has suggested that Warren is not a "reliable authority". I agree. There can be no "reliable authorites" concerning a matter which has never been directly observed. To claim otherwise is ridiculous.
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 05. May 2004 10:39
It may be useful to note that authority and authoritative opinion are not the same as science and scientific opinion. An hypothesis whose validity is based on authoritative opinions or proclamations, no matter what type of credentials the proclaiming authorities possess, is still a subjective opinion and does not qualify as a true scientific hypothesis. An hypothesis is qualifies as a legitimate scientific hypothesis, at least IMO, if and only if 1) it can be objectively demonstrated that the hypothesis is logically consistent with known facts and 2) the hypothesis generates independently testable predictions which can be objectively confirmed or contradicted.
Regularly posting at ISCID are advocates of a number of very different and ultimately incompatible types or classes of hypotheses which address the subject of speciation. John Davison advocates some type of ‘intelligence and genetic determinism’ speciation hypotheses. Rex Kerr advocates, (I believe), some type of Darwinian/ flexible genetic determinism hypothesis. I advocate a ‘speciation controlled by purposeful intelligence/ intelligent teleology determinism’ type hypothesis. Some of the less frequent posters, again I believe, advocate other intelligent design or Darwinian type hypotheses.
Given participation by advocates of different types of hypotheses, it would be interesting to see which types of hypotheses actually meet what I consider the requirements for true scientific hypotheses and which are simply based on authoritative opinion. Based on my personal informal analysis, it would seem that John Davison’s type of hypotheses would have a difficult time accounting for the fact that there are so many existing species, many of which appear to have arisen relatively recently. Rex Kerr’s type of hypothesis would appear to have a difficult time accounting for the limited amount of direct evidence for currently occurring speciation. It is difficult to see how either type of hypothesis produces testable predictions. Advocates of both types of hypotheses are invited to present arguments supporting their position.
As discussed earlier, the ‘speciation controlled by goal directed intelligence’ type of hypotheses suggests that speciation involves a very simple developmental design that should occur frequently. However, under most conditions speciation is not the intelligent choice most compatible with long-term survival. This suggests or predicts that while speciation mechanisms or species boundary mechanisms should occur frequently, there usually will exist processes or mechanisms that suppress these species boundary mechanisms.
This is a testable prediction. Any change in a developmental process which prevents the development of fertile offspring from interbreeding of members of two subgroups would function as a species boundaries. Infertility and failure of developmental processes are relatively common phenomena. My type of hypothesis predicts that some of factors responsible for infertility or failure to could, if allowed to become dominant, function as species boundaries. If such boundary mechanisms can be identified, then it is possible to artificially create new species.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 05. May 2004 21:48
quote:
We do not need to know the actual processes or mechanisms involved in order create models and simulations of processing programs. Models and hypotheses of processing programs are generally based on observations of behavior. Such models are tested or validated based on their ability to predict behavior.
But if you give me N data points, I can fit it very accurately with an order N polynomial (or N-m splines of degree m), or with a multilayer perceptron with, oh, around N hidden neurons, or with a Fourier series with a maximum frequency of N/2, or (etc.). Outside of the places where the data fits, these methods diverge wildly--polynomials blow up to infinity, Fourier series oscillate forever, and MLPs become constant.
This--along with a glorious history of failed models--is why I don't trust phenomenological models (as they are called) very far. Especially when the model seems to imply that you can never get beyond phenomenology to physical mechanism.
Geneticists have known for a long time about partially penetrant phenotypes. This is an example of a one-genotype to multiple-phenotype mapping. What is the poor Darwinist to do? Well, they consider the survival value of each phenotype and the probability of each phenotype, and come up with a net survival value, on average, of having that phenotype. It's not really that hard in concept--it's just hard in practice because we don't always know what the rates of penetrance are, and what all the interacting effects are.
The bottom line, though, is that if through whatever mechanisms, there's sufficiently strong selection, you get evolutionary pressure. That this pressure exists for very many genes is easy to tell: you simply look at the mutants and see what is wrong. For many other genes, it is more difficult to tell. Are we missing some intelligent mechanism that is holding the genes in reserve? Are we simply not applying the appropriate environmental conditions in the lab that elicit a difference between mutant and wild-type? Is the system intelligently routing around the damage? We can't really tell.
So the mappings that are well-defined can be studied easily; those that can at least be classified and quantified can be studied with more difficulty; and those where we haven't much idea what is going on can't be studied much at all (which really isn't a surprise).
I don't think there is a limited amount of information on currently occuring speciation. We calculated somewhere (whether in this thread or elsewhere) that at the upper end we had on the order of 60 speciation events per year; and speciation has been at an uneven rate (presumably at least in part due to filled vs. open niches), and our niches seem decently well filled now. And we do see what looks like incipient speciation or recent speciation events: Drosophila in "evolution valley" in Israel for example, or the green warbler ring species around the Himalayan plateau. Maybe with a more careful analysis it could be shown that rates of speciation are significantly off track from what would be predicted based on historical rates, but in a back-of-the-envelope calculation, I don't see a problem.
IP: Logged
|
|
nosivad
Member
Member # 767
|
posted 06. May 2004 07:15
Rex Kerr
You speak of "selection pressure" as if it were a reality. Where is the evidence for "selection pressure"? It is a mandatory consideration for the Darwinian model, yet it cannot be clearly documented except in microorganisms. Even there, preadaptation may be at work. Someone once wisely observed - "Animals are not struggling for existence; most of the time they are sitting around doing nothing at all!" Your comments also suggests that evolution is going on at present which is very much in doubt, as I and many others have maintained. What we are actually witnessing at present is the antithesis of evolution - rampant extinction. Of the literally hundreds of species that have disappeared in recent times, not one replacement can be documented. We are observing only the products of evolution, the terminal twigs of the evolutionary tree and it is being severely pruned by a drastically altered environment. If there was ever a test of the Darwinian scheme, it is right now. It is a dismal failure.
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 06. May 2004 10:02
I fully agree that fitting data to a model, by itself is not a sufficient criterion for building a useful model. This is also why so called descriptive hypotheses are considered to be of minimal value. The most common scientific criterion for a sound model is 1) fit available data and 2) produces reliable verifiable predictions.
The causal process or causal pathway criterion you mention is useful in special circumstances. The criterion asserts in effect that we can establish the validity of a cause and effect relationship that acts over time and distance if identify and predict a continuous chain of events from the original cause to the end effect. As an example, we predict where a cannon ball will land by knowing the speed and direction at which it leaves the mouth of the cannon. We confirm a cause and effect relationship because we can calculate or predict the intermediate points or path of the cannonball.
Essentially none of the evidence or examples used in evolutionary biology fit the causal process criterion. What is called evidence for the neo-Darwinian theory consists of a limited number of data points that fit simple models that use biologically unrealistic simplifying assumptions. The fact that these simple models do not produce reliable, non-trivial, testable predictions leads some to question the usefulness of these models even if they can be modified to fit a wide range of data.
Quote Rex: So the mappings that are well-defined can be studied easily; those that can at least be classified and quantified can be studied with more difficulty; and those where we haven't much idea what is going on can't be studied much at all (which really isn't a surprise).
It would just as accurate to say that the evolutionary change data that is logically consistent with simplified neo-Darwinian models and assumptions get reported as evidence for the theory, and the change data that is logically inconsistent with the simplified models gets reported as not yet explained.
Evolutionary change events are complex and the data available on any such event will always be incomplete. I am proposing that given recent advances in computer science, it is always both possible and practical to fit available data on developmental events and processes and evolutionary change events and process to some member or subset of members of the class of goal directed algorithms or programs. I am also proposing that it is always possible and practical to formulate predictive hypotheses from this type of model. This I suggest goes well beyond what can be accomplished using neo-Darwin type models.
With respect to speciation, you appear to be arguing that the data currently available is not incompatible with or can be fit to your type of hypothesis involving slow gradual speciation. John argues the available data is compatible with his evolution has stopped type hypothesis. In science, IMO, such disagreements over interpretations of data should be resolved using testable predictions. I have offered my predictions. Can your types of hypotheses produce predictions which could resolve the issue?
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 06. May 2004 11:09
SPECIATION IS A REDUNANT TRAIT.
Speciation is easily understood from the observation that the major part of multipurpose genomes consists of redundant genetic elements. Such elements do not convey any reproductive advantage or disadvantage, and thus can easily be lost from the genome or inactivated. Speciation has been directed by such redundant elements. Maybe they directed semi-meiotic mechanisms. Since they are not subject to selection they are also easily inactivated (neutral theory). They must have left their traces in the genome, hoever, and should therefore be found in the organism as inactive DNA elements, such a as retroviruses or pseudo retroviral DNA elements, that are usually referred to as ´junk´.
This view predicts that the relics of such elements must be found in the same location in the DNA of related organisms. It also explains why speciation has ceased. This is GUToB.
peebee
[ 06. May 2004, 11:23: Message edited by: peter borger ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 06. May 2004 13:15
What kind of predictions are you looking for? Things like ring species and apparent incipient speciation (populations that are largely but not completely reproductively isolated) have been predicted and found.
If you pick a scenario and make a prediction, I'll attempt to make a prediction for the same scenario based on modern evolutionary biology. (I say "attempt" because I am not actually an evolutionary biologist by training, so my knowledge is a bit limited.)
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 07. May 2004 04:29
Lets pick the scenario of the evolution of winged stick-insects.
What would contemporary evolutionary theory predict? It predicts that wings evolve only ones and that all winged stick insects are directly interconected through common descent. Wings do not get lost and reappear.
And what does the GUToB predict? Since I have set up GUToB I know what it predicts. Because the stickinsect MPG cryptically contains the information for wing development it is expected that due to non random shuffling of genetic elements wings can developed several times independent of common descent.
GUToB wins.
Peebee
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 07. May 2004 06:59
Quote Rex: What kind of predictions are you looking for?
In order to meet the testability requirement, a scientific hypothesis needs to generate a whole series of predictions. Also, the predictions generated must depend on the publicly available expression of the hypothesis. Predictions based on the personal opinions or judgments of the advocates of the hypothesis are of questionable value.
Rex, if you advocate a gradual speciation approach, then your hypotheses would need to define either a speciation gradient or specific speciation steps. Your hypothesis, to qualify as predictive, would also need to provide some description of the process responsible for moving speciation along the gradient or for the step changes. These definitions should be sufficiently precise to make it practical to identify groups undergoing speciation, identify the stage or degree of speciation which has occurred, and finally the hypotheses should be able to predict measurable speciation changes.
Both neo-Darwinian and Davison’s approaches are based on genetic determinism. This means, I assume, that the presence or absence of a speciation boundary depends on definable, identifiable genetic differences. Both types of hypothesis should therefore predict that certain types of artificially induced genetic changes will generate a species boundary. Both types of hypotheses should also predict that certain types of artificially induced changes will not generate species boundaries or speciation.
All hypotheses of evolutionary change must recognize the existence of speciation and species boundaries. All hypotheses, it seems, must therefore predict that speciation or species boundaries can be artificially created by inducing certain changes. All evolutionary change hypothesis must also predict that speciation will not be produced by specific types of changes. It would seem to me that predictions which define how speciation can and can not be artificially induced, would provide a good basis for chosing between different evolutionary hypotheses.
IP: Logged
|
|
nosivad
Member
Member # 767
|
posted 07. May 2004 07:13
I agree with Peter Borgher. The key word here is cryptic. I have used the terms latent or prescribed but the bottom line is that the necessary information was present in the genome and needed only to be derepressed probably exactly as is the case in ontogeny. One thing that remains unknown is the fate of this prescribed information during evolutionary history. Is it still around in highly evolved forms or is it lost once expressed? I am inclined to agree with Schindewolf and Grasse that evolution is, or more accurately was, an irreversible process. I have offered a cytogenetic explanation for irreversibility in the Manifesto. It is based on the extremely low probability that a chromosome alteration, such as an inversion, which involves two breaks, is likely to be reversed. The chromosome would have two break in exactly the same two places to be restored to the original configuration. Point mutations, which the Darwinians regard as significant, are of course reversible. That fact alone indicates they have played no role in evolutionary change. Apparently there is or was an enormous storehouse of developmental information in the chromosome. The role of chromosome restructuring was to release new combinations of features to be expressed in new life forms. This places an entirely new interpretation on what has erroneously been described as "convergent evolution". Unrelated forms simply have used the same preformed blueprints to produce similar morphologies. It is beautifully illustrated with the remarkable similarities to their placental counterparts that characterize marsupial moles, squirrels, mice and especially saber-toothed cats. These genotypes and phenotypes obviously were independently produced in both the marsupial and placental mammalian lineages. I would not be surprised if there were marsupial bats. To assume that the environment (natural selection) was involved is without foundation. These considerations obviously involve intelligent design. Blueprints require a draftsman. For a further discussion I refer you to my paper "Ontogeny, Phylogeny and the Origin of Biological Information" Rivista di Biologia 2000, Vol 93 pages 513-524, a version of which is available on my home page.
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 07. May 2004 14:04
Actually, the appearance-disappearance-reappearance of winged stick-insects (GUToB) scenario has been scientifically proven and waspublished in Nature, 16 january 2003.
The consequences should have been obvious for any unbiased thinker. Apparently the Darwinian claptrap continues.
In addition, breakpoints in chromosomes may well be used more than once. A close up look of the primate genomes demonstrates that man shares the same configuration as orang utan for chromosome 10(12), while chimp and gorilla have a distinctly different pattern. So, a random, irreversibale shuffling mechanism is unlikely. The same can be observed when we look into the chromosomes of chimp, human and bonobo: Non-random shuffling. Bonobo shares several chromosome configurations with man not present in chimp. I am (not) sorry to say it but the chromosomes proof neodarwinian theory is dead. Why propagate the claptrap?
peebee
[ 07. May 2004, 14:36: Message edited by: peter borger ]
IP: Logged
|
|
nosivad
Member
Member # 767
|
posted 08. May 2004 16:07
"We have long been seeking a different kind of evolutionary process and have now found one; namely, the change within the pattern of the chromosomes.... The neo-Darwinian theory of the geneticists is no longer tenable."
Richard B. Goldschmidt, The Material Basis of Evolution 1940
Note that there is no provision or necessity for the inclusion of new information in this concept of evolutionary change.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 09. May 2004 16:59
Speciation is hypothesized to occur via allopatric, parapatric, or sympatric methods.
In allopatric speciation, populations are isolated by geography/environment; adaptation to their environments along with genetic drift eventually render the two groups behaviorally and (typically more slowly) reproducively incompatible with each other. We don't know enough to predict in advance what mutations to find, but we can predict that the longer two groups have been isolated, and the smaller the effective population sizes of the isolated populations, the more likely it is that reproductive isolation will have occurred. The isolation may be due to a collection of point mutations plus compensating mutations, or due to an accumulation of chromosomal rearrangements.
In addition, strong selective pressures on one or both groups is likely to increase the chance of accidental reproductive isolation as a byproduct of adaptive gentic changes. To identify groups undergoing speciation, we therefore look for isolation and strong selective pressure; when we find strong selective pressure, we expect that the genes important for fitness are the same ones as those maintaining reproductive isolation.
Sympatric speciation, in contrast, occurs when one population finds itself at a fitness minimum, with two different available niches. This tends to favor extremes of variation within a population. The fitness of each extreme is enhanced if they reproduce only with each other, so you expect selection for traits that either behaviorally or mechanically prevent mating. (Producing sterile offspring is not helpful.) You thus expect strong assortive mating phenotypes with weak (or perhaps no) sterility/inviability of hybrid offspring early in the separation, followed by increasing (drift- or adaptation-induced) hybrid sterility as time goes on.
Parapatric speciation is essentially like allopatric speciation, except the populations aren't actually physically separated--the ends of the range are just too far for gene flow to keep the two ends compatible with each other. Animals with long, narrow ranges are particularly likely to undergo parapatric speciation, and in cases where the long, narrow range loops back upon itself you may find a ring species, where the two incompatible ends of what was once one species intermingle over part of their range. However, traveling around the ring, you may find no breakpoint where one species becomes the other; it's a gradual accumulation of differences across the range that eventually results in assortive mating or hybrid sterility or whatnot. This mode of speciation can collapse back into one species if the ends are wiped out, but it will be promoted into two distinct species if the middle of the range is wiped out.
It is important to note than in all of these cases speciation can be symmetric but need not. If the effective population size of the ancestral species is large, it will tend to stay about the same, and the new species will appear to be a different offshoot. If the population sizes are both small, however, neither may end up looking identical to their ancestor (especially if the ancestral niche has been lost due to environmental changes).
In fact, evidence of all three of these modes of speciation has been found. Sympatric speciation has been the most controversial, I think, but there is pretty good evidence of sympatric speciation underway, and no obvious way to prevent it from becoming a full speciation event. Note also that (as far as I am aware) these modes of speciation were predicted first on the basis of evolutionary theory, and then researchers went out to look for evidence that speciation was or was not occuring in these ways.
[ 09. May 2004, 17:00: Message edited by: Rex Kerr ]
IP: Logged
|
|
nosivad
Member
Member # 767
|
posted 09. May 2004 18:37
Rex Kerr
You have presented the standard Ernst Mayr interpretation of speciation. You have properly used the word hypothesized. I have rejected this explanation on the grounds that it has not exceeded the varietal level. Even if it did, and I am not willing to concede that it has, such a gradual mechanism could never be extended to the origins of the higher taxa. It is obvious that we are at an impasse. I have presented my hypothesis and of course I stand by it. The best evidence against the Darwinian model are Darwin's finches of the Galapagos. They started out as several species and are now probably one. The differences in bill size are phasic and reversible. That alone is proof that they do not represent true evolution which is never reversible. The several scientists on whose work my own securely rests have done a far better job of exposing the inadequacy of the Darwinian scheme than I ever could. To continue to pretend they never existed is a hallmark of the Darwinian literature. It is inexcusable.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 10. May 2004 00:10
As far as I know, there are six genera of finches currently (recently?) recognized. I seem to recall you claiming that the difference between species and genera was unmistakeable. Could you clarify what it is that you are claiming when you say "one species"? Do you mean, for instance, that the tree and ground finches are actually the same species, even though they were classified as separate genera?
IP: Logged
|
|
|
Topic Closed
Printer-friendly view of this topic
