8/15/2007

VOX DE BAIT #1

And we're off, with slinging salvos between Vox and PZ.

PZ thinks Vox's argument is, well, phony.
He claims that Vox has lifted a remark from Gingerich as to the amount of generations it might take to evolve an elephant from a mouse (Vox didn't specify sources in his original post), and he thinks that the argument depends upon a distorted understanding of how biologists measure rates of evolution. Whatever one thinks of this argument, I have to admit that I haven't heard this particular line of reasoning before. I wonder what Phil Gingerich might think about this? I think I shall ask him!

Vox, with an update to his original post, thinks PZ didn't read his argument very carefully and suggests that evolutionary experts should provide a prediction as to how long such a transition might actually take place, at the end of which time we should collect our Nobel Prizes.

At any rate, Vox used both a personal anecdote as well as this example as an explanation of the gnawing misgivings he has about evolution and natural selection. I'm still not sure what he thinks the actual status of these items shall be from his reply, but (no surprise) I think we can rule out any pan-adapationist orientation from Vox. Before I reply further, I think I would like to give Vox a chance to ask me any sort of follow-up question on this topic.

To be continued....

22 comments:

Anonymous said...

Scott - I appreciate the effort that you're putting into this, but I hope that you're not expecting him to actually debate evolution. He will use the following tactics:

1. Focus obsessively on mathematics, but place that focus on secondary issues that he hasn't fully understood. He will frequently accuse people (particularly scientists) of not being "accurate" in the way that economists are accurate (despite the serial failure of economics to predict much), while glossing over his own inaccuracies.

2. Where he can't raise "mathematical" objections, he'll usually resort to semantics to confuse the issue as much as possible. There will be extended trivial discussions over the precise meaning of a word, while he explains that he actually meant it in a slightly different way to how it seems. His meanings may or may not match the dictionary - that's optional for him.

3. Fixate on particular details in order to avoid having to discuss the overall theory, particularly where those details will make him look more informed than he actually is. Watch out for use of technical language in the same posts where he blathers on about how he's a layman and hasn't read much about this. He's covering the fact that a) he's read more than he says, but wants to make it look as if he's just smart, and b) he doesn't know as much as he needs to.

4. Use "humour" as a cover for his ad hominem attacks on people while avoiding addressing the substantive points that they raise. This will often happen when the person in question is better qualified than he is to comment on the issue.

5. Act like a pompous ass while drawing attention to his pompous assery in order to make people think that he's not serious about it. He is dead serious in his frighteningly high opinion of himself, and will use this as last resort when he's losing.

I recommend that you address only the substantive issues that he raises, which will be few and poorly-argued.

Nate said...

Ya know... as an interested member of the audiance... I believe this whole debate would be a great deal more focused and interesting if PZ would shut the hell up.

Anonymous said...

He's doing some fancy footwork here - but it's the usual sophistry.

He's making the implication that for a theory to be scientific, it must be predictive in the sense of predicting the future instead of if we know A and B and theory X is the case then we can predict the value of C. 'If we know the geological period we're digging in and the previous environment, and evolution is the case, we can predict what types of fossils we'll find if we dig here' is not the same as 'this mouse will evolve an extra leg on thursday'. Sometimes the future may be involved, but the model is time independent - the conclusion "given A and B and theory X, then we know C" can happen at any point.

It's a subtle difference if suitably obscured.

Of course, if science (or economics, for that matter) were that first kind of 'predictive' we'd all be fabulously rich. Or at least equally so.

His non-sequitur of mice and elephants should simply be countered with, yes, if the mice were microscopic, fastbreeding microorganisms, in a controlled environment and the environment could support the new larger breed you could do it in 20 years. Otherwise you'd have to factor in the longer generational period of mice (x20,000 say) and the ridiculously higher costs of controlling that level of population and its environment for that period, and you possibly could assuming some unforeseen environmental issue didn't bite you. Why you would want to is another question entirely.

Anonymous said...

Anonymous:
"He's making the implication that for a theory to be scientific, it must be predictive in the sense of predicting the future"

Not quite. Vox's precise question was "1. Do you understand the difference between a historical model and a predictive model? I" in which context he used the annecdotal examples of failed stockmarket and economic predictions.

For a theory to be scientific it must be testable, that is to say, it must in advance of experiment declare something precisely that can be compared with the results of experiment (or observation), i.e. it must make a testible prediction in some form.

Note, in the case of stockmarkets, Vox's annecdote was he had what turned out to be an historical model in that it backtested reasonably well, but soon failed in actual predictions of future performance.

I'm unaware of any scientific studies that demonstrate even "backtested" predictions of speciation or adpatation in the face of an ample dataset.

Anonymous said...

Vox's precise question was 1. Do you understand the difference between a historical model and a predictive model?

I realise this. But it's irrelevant. If he had, actually, "backtested it(his model) reasonably well", (like, say, by performing thousands upon thousands of tests by thousands of people against 150 years worth of collected data in several different fields) he probably would have noted that it didn't actually work all the time in the past and thus shouldn't be depended upon. Unlike evolution which has accomodated each and every datapoint it has come across.

It's a false analogy. "My model fell to bits in a strong wind, so therefore we must doubt evolution"

I'm unaware of any scientific studies that demonstrate even "backtested" predictions of speciation or adpatation in the face of an ample dataset.

I'm not entirely sure what you mean here. The fact of past speciations has been demonstrated many times, such as finding predicted new species and changed species after major environmental shifts such as ice ages.

Anonymous said...

predicted new species

By which I mean that there were new species, not that we knew what species they would be. There's a difference between knowing that random mutation has an effect and knowing what that effect will be ahead of time (which is why they call it 'random').

This also goes to Vox's "why do some populations evolve and some don't in the same environment". You can't plan a mutation, you can only select for it (e.g. it increases the chances of sucessfully procreating). Everything has mutations, but if it's not selected for, it won't have very much impact on the species at all.

Anonymous said...

Anonymous:
I realise this. But it's irrelevant. If he had, actually, "backtested it(his model) reasonably well", (like, say, by performing thousands upon thousands of tests by thousands of people against 150 years worth of collected data in several different fields) he probably would have noted that it didn't actually work all the time in the past and thus shouldn't be depended upon.

a) Backtesting of market trading models is done against the issue/price datasets (such as provided by the NYSE, NASDAQ, etc), which includes literally billions of trades back to each respective exchange's beginning. It is not tested against "people", and highly liquid markets have only existed for several decades, and even then their character (and the applicability of different models and trading strategies) changes rapidly. Vox's model (reputedly) worked for a few months and had tripled returns when he began to see signs of it failing at which point he (and his clients) abandoned it.

b) No trading models (even those that still work reasonably well in some cases, like RET) work "all the time". A "profitable" model these days, in the hands of a professional, is maybe 60-70 percent reliable.

So no, apparently you don't realize the relevance.

such as finding predicted new species and changed species after major environmental shifts such as ice ages.

I would like to see those "predictions" (not the observations, but the before the fact predictions) and the specific language or phrases in the evolution theory which made those predictions. Note that a scientist's intutitive expectation of a species existing or surving climate chnage is not the same as a theory itself making a precise, testible claim, in advance.

Anonymous said...

Ok, you got me. I'm not an economist and did not realise the precise meaning of backtesting. However, to use Vox's own terms, he backtested it reasonably well, which does lead one to suppose he could have backtested it better (I don't know, he's only really providing anecdotal evidence here - your guess is as good, if not better, than mine)

So the fact appears to remain - if he had a model which explained every documented event in the past he'd be making a much stronger point. As it is - it's more like Lamarkian evolution - it seemed to describe a lot of observations but when more were made it turned out be barking up the wrong tree. Again, whether this happens to data in the past or incoming data is irrelevant to its strength as a model.

the specific language or phrases in the evolution theory

You do realise a scientifc theory isn't a book, don't you - it's an explanation and can be expressed in numerous ways - even pictures.. There are, however, numerous books around to help you understand what evolution actually is, what it predicts, and how those predictions have been found to be correct. For example, when you reach a geographically separated area for the first time, you would expect to find new separate, but similar species. Happens all the time. Google Foja Mountains.

Anonymous said...

A quick read of wikipedia on backtesting and I think there's a better metaphor to be made.

We know why it rains, how rain occurs, we understand the water cycle. We can tell when it flooded in the past and when ice ages occured and we are certain that there was rain before there were human beings. But if you want me to say what the weather is going to be like a year from today - I couldn't tell you.

This, howeever, does not invalidate meteorology.

Anonymous said...

However, to use Vox's own terms, he backtested it reasonably well, which does lead one to suppose he could have backtested it better

That is the relevant distinction between historical models (market trading algorithms based on past results) as opposed to predictive models (fluid dynamics predicting future airfoil performance under virually any condition) which actually predict with near certainty future results.

You do realise a scientifc theory isn't a book, don't you

Well, evolution's "predictive theory" certainly isn't, agreed.

By contrast go research the predictive ability of thermodynamics, fluid dynamics, molecular chemistry, electromagnetism, relativity, even some quantum effects are genuinely, accurately predictable.

Genetics, genomics, etc have made truely astonishing advances in technology, especially in diagnostics, hybrids, and medicine.

To that extent they are certainly highly beneficial scientific endeavors and they have made great strides in understanding how "DNA works" and how to observe it and tweak it.

But to the extent that same technology is useful in studying evolution, does not equate to evolution theory itself (arguably more like evolutionary hypothesis) being testable, much less predicatble to the same degree as other sciences are testable and predictable.

There are, however, numerous books around to help you understand what evolution actually is, what it predicts, and how those predictions have been found to be correct.

Again, I'd like to see actual peer-reviewed scientific papers (not popularizations or for-profit textbooks) elaborating the precise evolutionary theoretical prediction and the observations after the fact confirming them. I don't question the intutive expectations of scientific researchers, just the specificity of the predictions in the theory. Again compare what astronomners can do to predict gravitation lensing and black holes for example, versus evolution.

This, howeever, does not invalidate meteorology.

Agreed. But meteorology isn't making claims to have proven to the exclusion of other hypotheses what the earth's early weather was or how it changed. Meteorology constrains itself to giving probabilities a day, week, two weeks, three months in advance.

It doesn't mean evolution as a field of funded research shouldn't proceed. It does mean that its claims to being "proven", "testable", and "predictable" ought to be considerably more mitigated.

(Gotta go tend to other priorities, be back much later.)

Anonymous said...

Correction (ggrrrr):
However, to use Vox's own terms, he backtested it reasonably well, which does lead one to suppose he could have backtested it better

I intended to have cited:

if he had a model which explained every documented event in the past he'd be making a much stronger point. ... Again, whether this happens to data in the past or incoming data is irrelevant to its strength as a model.

and then my response:

That is the relevant distinction between historical models (market trading algorithms based on past results) as opposed to predictive models (fluid dynamics predicting future airfoil performance under virually any condition) which actually predict with near certainty future results.

Anonymous said...

That is the relevant distinction between historical models (market trading algorithms based on past results) as opposed to predictive models (fluid dynamics predicting future airfoil performance under virually any condition) which actually predict with near certainty future results.

But this is exactly the point. You seem to be wanting evolution to be reduced to an equation. It's not. Mutation A + Selection Pressure B = Morphology C isn't going to happen except in the most basic sense - there's too much else going on for that level of precision in every element of the process. Different groups might find different ways to deal with the same problem. Evolution isn't going for the best solution, merely one that will permit survival and replication. Sometimes it doesn't even find it it time, hence extinction.

Again, evolutionary theory doesn't have to predict the future to be a
successful theory, it merely has to account for the available data - which it does - and also, where data is missing, predict what that data is when it turns up - which it also does.

Here's a list of some specific predictions that have been met, with references:

Asking for it to magically overcome complexity to give us a roadmap into the future is missing the point.

In the laboratory, evolution does exactly what you'd expect it to, up to and including speciation. All the data we've found and continue to find fits within its possibilities. What more do you need?

I can't help but think you might have grabbed the wrong end of the stick about what the theory of evolution actually involves. Perhaps if you gave me an idea about what kind of prediction you expect evolution to make I'd be able to see where you're coming from more clearly.

Scott Hatfield . . . . said...

I think Billy has provided us with a falsifiable hypothesis or two...:)

Anonymous said...

Anonymous:

You seem to be wanting evolution to be reduced to an equation.

No, I want evolution to make consistently reliable, testable predictions, if it is to be considered established science. Further, evolution is often equated with having been "proven", in which case I expect it to stop being revised.

I don't much care if a consistently reliable, testable, predictive evolution theory does so with formulas, algorithms, logic trees, or goat entrails. It is to be judged on the merits of its results.

Again, evolutionary theory doesn't have to predict the future to be a successful theory,

It does if it wants to be accorded the status commensurate with its claims. Evolution didn't stop, it is ongoing, right? A theory that only works in hindsight is of marginal consequence.

it merely has to account for the available data - which it does - and also, where data is missing, predict what that data is when it turns up - which it also does.

Oh, please. There is much in the fossile record unaccounted for by evolution and the missing transitional species ought to have been found by now. And then there is the matter of it being revised to account for new data, such as the recent finding that Homo habilis did not evolve into Homo erectus, and that evolution isn't as linear as was once presumed.

It's ok that the theory be revised. Revisions are good science. But as long as the revisions are ongoing and predicted transitional species missing, it isn't reliably predictable and it certainly isn't proven. And its testability will never be any better than the specificity of its predictions.

Here's a list of some specific predictions that have been met, with references:

There is not enough information in that article to discern whether those were predictions specifically required by evolution theory itself or were the intuitive anticipations of evolution scientists. There is a difference. Further, making some predictions is not sufficient. To be classified as predictive (let alone proven) its predictability must be applicable to (i.e. account for) most (if not all) species and traits most of the time.

Fluid dynamics, for example, doesn't make a few correct predictions. It is right in nearly every circumstance regardless of surface material and ambient fluid conditions.

Perhaps if you gave me an idea about what kind of prediction you expect evolution to make I'd be able to see where you're coming from more clearly.

At the high end of range consider Einstein's theory of special and general relativity. When originally published it made numerous specific, testable predictions: mass-energy equivalence, gravitational mass will bend light, time dilation and length contraction, equivalence of trajectories of bodies falling in a gravitational field, etc. It has never been revised, it continues to be proven correct with ever increasing precision. Unproven as yet is only the prediction for gravitational waves.

At the low end of the range consider meteorology. As poorly understood as it is, it still manages to make generally usable short term probabilities of large scale and local weather, and it is getting better.

Were evolution theory to be granted comparable weight and credence, I would minmially expect it to give probabilities, for example, of how the Avian Flu virus is likley to mutate next.

The objection raised, no doubt, will be how mutations are random but patterns recognizable, selection pressures are unknown, don't have enough real time data on all the current flu variants and their environmental conditions, etc.

Those problems aren't much different than what meteorology faces. Chaotic movement of air and water at microscopic levels but patterns recognizable at larger scales, undetermined impact from solar and human sources, not enough real time data, etc. And yet those scientists have for years been rolling up their shirt sleeves and figuring out what they needed to know, what they can approximate, and what they can ignore to make their science predictable, and set about to do it.

I see no reason to give evolution theory a by. Rather, it needs to step up.

Ideally, I would expect evolution theory to predict theoretical, testable consistent phylogenies showing the evolutionary common ancestry for archaea, bacteria and eukaryota, and then phylogenies for eukaryota including all the main phyla and showing common descent through to homo sapiens. Logistically, technologically this seems feasible; genomes can be mapped and phylogenies computer generated. It would seem what is lacking is an evolution theory to organize the results.

Anonymous said...

If you're bringing up lack of transitional fossils as an argument you're willfully ignoring bucketloads of evidence - take Scott's post today for an example where you are completely wrong.

You're not going to find every conceivable transitional fossil. Fossilisation is a rare event - it doesn't happen to everything that dies. But what we do have is more than sufficient to realise the truth of evolutions claims, unless we wish to be spuriously sceptical to the point of obtuseness.

Thgere's an abundance of evidence of the kind of predictions you want also. I provided you with some. Merely claiming that 'the scientists expectations' in some way coloured the results or made them less valid is a very poor argument. You can see the predictions - you can see why they would be the case if evolution is the case, and you can see that they were fulfilled. Your interpretation is conjuring faults out of nothing for the sake of it.

And you fall into your own trap comparing it to meterology - yes, some predictions on a larger scale can be made, and we can do the same via evolution. Selected for large size - average sizes within the population will grow. EVERY TIME. You can take it to the bank. Evolution is fact. That doesn't necessarily mean we can identify the precise mechanism ahead of time - there might be different mechanisms in separate groups undergoing selection. That's how evolution works, and all it needs to do to be a fact is do it every time, regardless of whatever spurious level of detail you personally require in anticipation.

Please look up the meaning of theory in the scientific sense. You don't have a very good grasp of it.

Anonymous said...

Further, evolution is often equated with having been "proven", in which case I expect it to stop being revised.


You're getting confused here between the fact that evolution occurs and the 'Theory of Evolution' which explains how it happens.

Evolution is a fact - it occurs every day - it's been observed countless times. To state otherwise is willful perversity in the face of mountains of evidence from all manner of sciences.

The theory of how it occurs, and the exact details of the precise mechanisms involved, are still being worked out - it's a huge field, after all. This is why Gould can propose punctuated equilibrium, and if it turns out to explain the data better, it will be accepted - not as proven (a theory is never 100% proven and scientis would make that claim) but as the best currently explanation available. But none of that changes the hard fact that evolution demonstrably occurs

Anonymous said...

Anonymous:

You're not going to find every conceivable transitional fossil. Fossilisation is a rare event - it doesn't happen to everything that dies. But what we do have is more than sufficient to realise the truth of evolutions claims, unless we wish to be spuriously sceptical to the point of obtuseness.

Regardless of the frequency of fossilization, within the existing fossil record "transitionl forms" between genera or family ought to be greater represented by at least 1 or 2 orders of magnitude.

As per Raup (1982 "Mass Extinctions in the Marine Fossil Record" and 1979 "Conflicts between Darwin and Paleontology") there are approximately 800 familes, 4-5 thousand genera, and 250,000 species in the fossil record.

The fossil record is reasonably intact and accurate:

The quality of the fossil record of the vertebrates:
"The third new approach, comparing phylogenetic and stratigraphic evidence to assessing the adequacy of the fossil record, has been applied with considerable success to vertebrates. This approach is based on the observation that phylogenies (cladograms founded on morphological characters and molecular phylogenies) are constructed independently of geological evidence. The order and timing of splitting events in phylogenies may then be cross-compared with stratigraphic evidence on the order and timing of the appearance of groups in the fossil record in order to assess the degree of congruence. Good matching of the data sets implies that both the fossil record and the phylogeny are probably good, while a mismatch implies either a misleading fossil record or an inaccurate phylogenetic hypothesis.

The phylogenetic congruence assessments have indicated that there is no evidence that vertebrates have a fossil record that is either any better, or indeed any worse, than that of any other major group of animals. In addition, there is no evidence that the record of continental (that is, terrestrial and freshwater) tetrapods is worse than that of marine echinoderms or fishes. These kinds of assessments have proved highly fruitful, and they provide a sound quantified answer to the old cry of 'the fossil record is pretty incomplete and uninformative'."

The article Scott Hatfield cited (Taxonomy, Transitional Forms, and the Fossil Record) defines a transitional form as "Based on the above discussion, a transitional form is simply a fossil species that possesses a morphology intermediate between that of two others belonging to different higher taxa."

Most of the 250,000 species in the fossil record are not transitional forms and, commensurately, many of the transitional forms found are between species, i.e. species variations, or "speciation". Plausibly, transitions between taxa of different species (e.g. Hyracotherium to Equus) can be interpreted or construed from the fossil record.

But how many of those 250,000 specimens in the fossil record are transitional forms between families or genera?

The problems arise when "evolution" asserts as fact that one genera or family transitioned into a different genera or family. The evidence cited by Miller's ASA article for amphibians transitioning to reptiles and mammal-like reptiles transitioning to mammals is speculative based on some similarity in skeletal appearances. Were the evidience more precise and conclusive, there would not be so much difficulty identifying (and recataloging) to which family specimens belonged.

As noted above, with 800 familes and 4-5 thousand genera in a reasonmably intact and accurate fossil record, there ought to be more than a few dozen proposed "transitional forms" showing the common descent of different genera or familes and not just the variation within one species. Were "macroevolution" proven or a "fact" we would expect to see a couple thousand "transitional forms", on average one "common ancestor" for every two genera and a few intermediate transitional forms showing the evolution of the "common descendants" for both genera.

That doesn't necessarily mean we can identify the precise mechanism ahead of time - there might be different mechanisms in separate groups undergoing selection. That's how evolution works, and all it needs to do to be a fact is do it every time, regardless of whatever spurious level of detail you personally require in anticipation.

Not actually knowing what mechanism might be involved and how they'll operate is your first clue that that you don't know how evolution works. You have observations of speciation, which are explainable by genetic mutation and natural selection. What you don't have is concrete evidence and a well understood mechanism that accounts for all the changes (differences) between families and genera. "Evolution theory" purports to be the theory of common descent which arguably begins with a common ancestor from which, there are many mechanisms and differences unaccounted for. Perhaps at a future time "Macro-Evolution" will have detailed, precise explanations for all of them and additional evidence of macro transitional forms in the fossil record. But today, that is a non-fact.

Evolution is a fact - it occurs every day - it's been observed countless times. To state otherwise is willful perversity in the face of mountains of evidence from all manner of sciences.

Speciation is observed. yes it occurs every day - it's been observed countless times. But evolution theory claims much more than just speciation and extrapolating "evolution theory" from mostly speciation observations is not proof. And even the mechanisms of speciation are not well understood. If they were well understood, the selective pressures could be measured and the possible genetic mutations could be ranked, and various probable 'next phenotypes' predicted, of Avian Flu for instance.

Macro-evolution is not a "fact", nor (by the standards of all other scientific disciplines) does sufficient evidence exist to prove the theory of Macro-evolution (or Neo-Darwinism, or whatever the buzzword dujour is) in absence of testible, predictable results.

The theory of how it occurs, and the exact details of the precise mechanisms involved, are still being worked out

Especially then in light of ongoing revisions and incomplete data and tests (which is understandable), its advocates ought to stop claiming it is proven scientific fact (which is not understandable).

(FYI, my participation is going to be sporadic and intermittent over the next few days)

Anonymous said...

Correction (sigh):

But how many of those 250,000 species in the fossil record are transitional forms between families or genera?

Anonymous said...

But evolution theory claims much more than just speciation and extrapolating "evolution theory" from mostly speciation observations is not proof.

Aside from the fact that "evolution theory" is not just based on speciation observations you are neglecting (again) the fact that Evolution Theory isn't supposed to be proof. It's supposed to be (and I'm sorry for all caps, but you don't seem to get it) THE BEST EXPLANATION FOR ALL THE EVIDENCE. As I said - look up what a theory actually is, rather than what you expect it to be.

Also, you are shifting goalposts - "where are the transistional fossils, oh, there they are, well why aren't there more of them?" Well, possibly because the selection pressures aren't a consistant factor - for example, an ice age starts relatively quickly and then lasts for...an age.

If you have a better explanation for the available evidence - let's hear it.

Anonymous said...

Anonymous:

Evolution Theory isn't supposed to be proof.

You didn't parse that sentance correctly. Observations of speciation are not proof for evolution theory.

THE BEST EXPLANATION FOR ALL THE EVIDENCE.

Science does not grade on a curve. It is pass/fail. A "best effort" is not the passing criteria.

That's not to say the course can't be retaken, but until the minimum criteria are met consistently, it is neither proven fact nor comparable to other scientific theories.

you are shifting goalposts

You raised the issue. I was answering.

If you have a better explanation for the available evidence - let's hear it.

The false premise in there is that the evidence is sufficient (todate anyway) to support any scientific explanation at all. Speculation and hypothesis is easy. "science" is considerably more exacting and demanding. That's not to say "evolutionists" should stop looking or theorizing. Rather they should be more circumspect about what they allege is fact or proven, and testable/predictable.

Anonymous said...

Observations of speciation are not proof for evolution theory.

No, NOTHING IS PROOF OF A THEORY. THEORIES ARE NOT PROVEN. You are simply not using these words correctly. a layperson's guide

It's part of the large amounts of evidence for evolution. Add it to all the other evidence and resistance to the fact that evolution occurs is being willfully perverse. The theory is still collecting evidence - not that evolution occurs but how it occurs.

Science does not grade on a curve. It is pass/fail. A "best effort" is not the passing criteria.

Did you miss the part where I said no scientist ever considers a scientific theory 100% proven. Even Dawkins has said this. You are arguing against an argument no-one else has made. Your quest for "proof" is blinkering you to what is actually being said. Please try harder.

Anonymous said...

Anonymous:

No, NOTHING IS PROOF OF A THEORY. THEORIES ARE NOT PROVEN.

Strictly speaking, mathematical theories are, and that would include proving the predictive and explanatory math that underlies many natural science theories.

You are simply not using these words correctly.

Fine. Strictly speaking then, substitute the word "confirmation" for "proof", and special and general relativity are largely "confirmed" while evolution theory remains largely "unconfirmed".

None of which adds one iota of predictibility or testability to evolution theory, does it. Nor removes any predictibility or testability from the theories and examples I cited above, does it.

The theory is still collecting evidence - not that evolution occurs but how it occurs.

That is your a priori unscientific bias operating.

You admit that evolution (the phenomena) is caused by unexplained mechanisms yet you insist without "confirming" what those mechanisms are or how they work, that they occur. You can not "confirm" a particular mechanims is the cause of an occurrance without confirming how the mechanism operates. They are co-requisites. Science can postulate a mechanism exists, define it, and set about searching for "confirmatory" evidence, but genuine science won't insist that mechanism actually occurs until science has confirmed both what is the mechanism being sought and the confirmed evidence is explained by that mechanism.

There is a vast distinction between a phenomena such as time dilation (confirmed) or genetically transitional forms for families or genera (unconfirmed) and the mechanisms which cause it such as constant c across relativistically moving reference frames (confirmed) and evolution theory (unconfirmed).

And yet you would accord evolution theory a special status that unconfirmed phenomena be accepted as confirming evidence for its ill-understood mechanisms.

Your excuse will be that because there is ("unconfirmed") evidence that some species have changed you extrapolate that the same ("unconfirmed", and "not understood") mechanism is the reason that entire families and genera likewise change, again in absence of "confirmed" fossil evidence of most familes or genera having biologically/genetically descended from common ancestors. As previously noted, a few similaries in skeletal structures of "unconfirmed" transitional forms does not "confirm" the existance of the few thousand such transitional forms that ought to be in evidence, let alone the mechanism that explains how these missing transitional forms (haven't yet) occured.

Did you miss the part where I said no scientist ever considers a scientific theory 100% proven. ... You are arguing against an argument no-one else has made.

Nor do they ever consider them 100% "confirmed" either. And as I said above, Einstein's theory of general relativity predicted gravitational waves, was as yet "unconfirmed", and that fluid dynamics predicts with "near certainty" in "nearly every circumstance". Those were not "100%" arguments, were they.

You are arguing against an argument no-one else has made.

I never said 100% "proven" (or "confirmed") did I. That is your repeated strawman, not mine, isn't it.