Saturday, August 23, 2008

Were scientists biased in favor of large-scale evolution 150 years ago?

     I was asked roughly this question by someone I don't think was interested in an answer so much as trying to get ammunition for his claim of "conspiracy theory." He shall be nameless here. The fact is that I am not sure. I do know this. Scientists had reached a point (and rightly so) where "godidit" was unsatisfying as an explanation. But they are human and humans don't like to say "I don't know." So was opportunity for such a bias, particularly since no acceptable alternatives existed. Even I refuse to accept "godidit" as an answer. If I am amazed at the intricate design work on a building, I am interested in knowing how it was accomplished. A source of "Acme Construction" does not answer my questions.
     Closer to the heart of the idea, I don't see that we can conduct an actual test. The only "thought experiments" I can come up with involve taking direct observations over millions or billions of years. The direct observations would be necessary so that a missing animal would really be negative data. The only approximation we have are fossil digs and missings fossils are excused by the rarity of fossilization.
     I have seen much of the positive data and the idea is certainly plausible. But I am a stickler. A scientific theory must be subject to falsification attempts. In any of the "tests," absence of positive data, would not be considered negative data. Consider, the recent discovery of the centromere/telomere patterning on human chromosome 2 is supportive and serves as a plausible explanation why we have one fewer chromosome pair than other primates. But, suppose there were no such find on any chromosome pair; would that be negative data? Of course not. If it would, then the many variations in chromosome counts of animals would also be negative data, and, indeed, catastrophic to the hypothesis. If there were no such find, an easy explanation would be that mutations had washed out the patterns. Alternatively, one might suggest that any relevant genes had been transferred to other chromosomes and that the resulting unneeded pair had, through successive generations, not been transmitted.


Dave W. said...

You seem to be willing to make up ad hoc excuses for hypothetical biologists who hypothetically found no evidence of chromosomal fusion in humans. And so you seem to be saying that the fact that the prediction of fusion was not falsified is actually a strike against evolutionary theory, because (you argue) that the results didn't matter, biologists would have just made something up to explain whatever data happened to be found.

No. They would have created testable hypotheses about failures to find the fusion site. And then appropriate tests would have been run. If no testable hypotheses were found, or if all of them also failed, then "we don't know" would have been the only option left.

Fossils are not the only evidence we have for evolutionary theories. We've got bacteria, mosquitos, moths and a host of other species mutating right within our lifetimes. And, of course, there are these 29 bits of evidence, most of which have potential falsifications (many which don't are simply observations, anyway).

Pvblivs said...

Dave W:

     The claim is that all life forms share a common ancestry. There are, in fact, many varied chromosome counts in animals. This is not a hypothetical situation. It is an actual situation. Do they find sufficient fusion to account for all these variances? No. Do they consider this to falsify the "theory"? No. It can reasonably be surmised that the absence of such fusion in human chromosomes would not be accepted as negative data. If it would, the negative data associated with the general variations in chromosome counts would be catastrophic.
     Yes, we have actually seen small-scale evolution, making it an empirical observation. If we lived for millions of years, we might even be able to witness large-scale evolution. But bacteria becoming multicellular organisms has not observed in the lab.
     So far, "potential falsifications" for large-scale evolution have always turned out to be safe predictions. Something was only identified as a potential falsifier after it could be safely determined that it would not happen -- generally on the order of "we would have seen it by now." I have little interest in wading through to find more of the same. If you have a potential falsifier that is not / was not similarly safe, present it. I will check the context to verify the claim of "not safe" is accurate.

Dave W. said...

Of course nobody will find "sufficient fusion" to account for all variances in animal chromosome count, because some of it is due to chromosome duplication and some of it is due to simple chromosome loss. It's pretty clear that you want a step-by-step account of the evolution of humans from some pre-historic ancestor, and that's unlikely to happen because of the massive amount of time that's gone by, with time's attendant mutations. Instead, we infer the conclusion from an examination of all of the evidence at hand, including all the duplications, fusions and modifications that we can see going on today, coupled with the fact that there is no known limit to how far such mutations can go.

Just like you have done with your claim, "So far, "potential falsifications" for large-scale evolution have always turned out to be safe predictions." Surely you do not have a 100% inclusive list of all potential falsifications of evolutionary theory (there would be thousands of them). Instead, you have taken what you consider to be a healthy sampling of them, and inferred your claim from that data.

So, I ask you: why is inferrence from an imperfect data set okay for you, but not okay for biologists?

Besides, you want a not-safe prediction? Prior to 1954, it could only be predicted that life's evolutionary history when mapped using genetic markers would be the same as the evolutionary tree when mapped via phylogeny. The reason it wasn't a "safe" prediction is that nobody knew the genetic code or even the structure of DNA. If it had turned out that we could only build a completely different tree (or no tree at all), modern evolutionary theory would have been sunk. Turned out, of course, that genetic mapping builds almost exactly the same tree as phylogenic mapping does.

A single counter-example (like the above) is enough to falsify your inference, simply because you chose to use the word "always." If you'd like more, I could supply more, but really, it is your claim to support. So let's see the list that you've built of "safe predictions" made by evolitionary biologists, and make sure that said list includes every successful prediction ever made. Don't forget to supply a quantization mechanism for "safe prediction" so that there's no way for either of us to argue over what's "safe" and what's not.

Pvblivs said...

Dave W:

     Or you can give your single counterexample. After all, you said that that's all you need.
     My criticism is not that they use an imperfect data set. It is that the "theory" cannot possibly fail even if it is wrong. Although a big deal is made of the supporting evidence, its absence would predictably have been brushed aside. There are any number of alternatives that all assume large-scale evolution.
     From your own source: "Cladistic methods are often contrasted with 'phenetic' methods. Phenetic methods cluster and classify species based upon the number of identical characters that they share, that is, based upon overall similarity. Such methods can run into trouble with organisms like dolphins and tuna, which have many superficial similarities. These organisms, however, are not closely related and should not be classified together if one expects classification to reflect phylogeny." In brief, methods of building a tree that didn't match were discarded. Indeed, if nothing matched the gene tree, genes could be discarded as mechanism. It would also likely mean that there had no relation between genes and features. That was more a test of whether genes were plausible as mechanism than of whether large-scale evolution was real.
     So, as a recap: genes as a mechanism for features and a plausible mechanism for evolution have been tested. If the results of experiments had gone another way, genes would have been ruled out. If that's what you mean by "modern evolutionary theory" then you are correct. But then we are talking past each other, because I am talking about the overarching concept of large-scale evolution even if mechanisms are unknown.
     Now, a good question is is that overarching concept a sacred belief? Well, did you just tell me that I would have to hunt down every single paper published on the subject of macro-evolution before I could suggest that it looked like it wasn't being put to the test? Hint: "So let's see the list that you've built of 'safe predictions' made by evolitionary biologists, and make sure that said list includes every successful prediction ever made." [Emphasis added]
     By the way, my claim is a non-scientific generalization. It is possible that there exists a successful unsafe prediction I don't know about. But I would expect such a discovery to be made very public. Impressive looking positive data are made public when the corresponding absence would be met with a round of "ho hum."

Pvblivs said...

Dave W. said...

Okay, your problem isn't with inference.

You say that the absence of confirming evidence for evolution would have been predictably brushed aside. Upon what data do you base such a prediction? Do you have any empirical data that confirms your prediction? Do you define "brushed aside" in such a way as to make your prediction falsifiable?

Your summation of cladistics is lacking in an important respect. There are an infinite number of possible classification schemes. We could begin with an animal's primary coloration, such that fire ants and cardinals wind up in the same high-level group (akin to "kingdom"), while Argentinian ants and crows are both in a different group (because they're black, not red). Does it make any sense to build the "tree of life" along such lines, though? That's the problem with classifying tuna with porpoises, and it's a problem that was examined and discarded long ago: where a creature lives or what it eats or even the fact that it has "fins" doesn't make sense for a high-level classification once you start looking at details or at internal anatomy. "Hey, these seals, which have fin-like things and a flappy tail and eat other fish, so we thought they were fish, have a rib cage and a heart and hair like a dog does." And classing squirrels with birds because they both live in trees stops making sense when you add more characteristics, too. Attempts at making a sensible hierarchy of life began before Linnaeus came up with his still-partly-in-use naming scheme, which he published nearly 100 years before the publication of Origin. In other words, long before the idea of common descent was taken seriously, cladograms to support it had already been developed by people who spent a long time arguing over what made sense and what didn't in terms of how creatures should be classified. Once you start looking at the details, and not just one or two characters, other phylogenic trees make as much sense (and are as useful) as a random classification scheme.

And once genetic cladograms began to be made, late last century, had they been vastly different than the phylogenic trees that had been around for ages, there would have been a few possibilities: one, the phylogenic trees were wrong (for example, maybe air-breathing came before backbones?). Two, the genetic trees weren't actually genetic (DNA would have had nothing to do with heredity). Three, common descent was a pipe dream. To say that such problems would have been "brushed aside" is to say that if the genetic cladograms disagreed with our phylogenic cladograms, we would nevertheless still be teaching that backbones came before air breathing, DNA is the genetic material, and common descent is true, all with proof positive that at least one of them is wrong.

I'm pretty sure you don't want to hear it again, but that would (as <NAMELESS> indicated) require a conspiracy among scientists for the last 50 or so years.

Instead, we know that Linnaeus' classifications were at least a little wrong, and instead of brushing aside those problems, they have been fixed. Hey, science advances by correcting past mistakes, not brushing them aside. Einstein's General Relativity didn't "brush aside" Newton's gravitation, it fixed it (at least, outside of black holes).

Now, is any "overarching concept" in evolutionary biology a sacred belief? Certainly not to me. I couldn't care less if it were all wrong, just so long as it were no longer promoted as being correct after evidence of its wrongness came to the surface. The reason I asked you if you had cataloged all successful evolutionary predictions is not because I would accept only that as evidence (due to the sacredness of evolution), but only because you chose to use the word "always," which suggests that that's the standard of evidence that you hold yourself to, and which suggests that you have done a massive amount of analysis of evolutionary experiments, probably taking years of your life. You now admit that it was a "non-scientific generalization," which means little more to me than that you're either willing to throw away your skepticism when it suits you, or you're sloppy with your words. Take your pick (or both).

You're clearly rejecting the massively "unsafe" pre-1954 genetic predictions as an example of this "overarching concept" you mention, so I have to ask you to be more specific. All of the big concepts in evolutionary theory have names, so please use one instead of making everyone else guess. However, since we've got tons of experimental data on all the various forms of selection and mutation, and you seem to have a thing for fossils, so I can only imagine it's the big "deep-time" concept of common descent, but that's what the genetics data supported. So I'm at a loss as to how to continue until you nail down your position. I'm not going to play the game where I offer some more "unsafe" predictions and you say "I didn't mean that sort of thing" to all of them in turn (like you did this last time around).

August 24, 2008 10:11 PM

Pvblivs said...

     I said he would be nameless here; and I meant it. There doesn't need to be an active conspiracy. Social pressures when a group is giving the same answer, especially when it's the one the instructor gave, will suffice. I provided a link to a study regard such social pressures. In the study, all of the pressures were passive. The subjects were not laughed at by the confederates and were not challenged by the confederates in a direct effort to get them to agree. Even then, there was a high degree of conformity. Simply citing the posts directed at me could, by the standard that you appear to be applying, be called a conspiracy theory. Just my quoting what you and the other have said could easily result in you saying that I was theorizing that you were conspiring against me. So, no, it is without merit. You don't want my definitions or my thoughts. You want me to give up speaking on it in frustration. How do I know? Because I have already stated that I will answer "questions" under the taint of the accusation. Making the accusation here in earnest would be pointless. I cannot be convinced of it as it looks like a deliberate attempt to pressure me to conform. On the other hand, it could be very useful in an effort to frustrate me and possibly shut me up.
     I submit that if you truly believed that I was engaging in a conspiracy theory, that you would not tell me so, as you would consider me blind to the fact and possibly that I would dismiss you as "part of the conspiracy." I have been answering questions as to why I believe that large-scale evolution is held sacred in the scientific community. I have also been taking seriously and responding to items posted as evidence that I am incorrect. If you actually thought you had evidence, "conspiracy theory" is the one thing you would not mention as it is the one thing that will make me think that you are being dishonest.
     On the other hand, if you think that I am right, but want to ingratiate yourself with the majority so that you are not similarly ridiculed and shunned, going with the flow and saying "conspiracy theory" would be just the way to go about it.
     That's the irony of it. I don't have a reason to believe in an active conspiracy among the scientific community. Simply put, I wouldn't be able to tell. However, at the level of dealing with me, there is a conspiracy of sorts, and it's not hidden. Ultimately, a group of people woking together toward a common goal is a type of conspiracy. When the goal is considered worthwhile, it is seldom actually called a conspiracy. Ordinally when someone believes in a conspiracy and calls it as such, he thinks that the group is working together toward a malevolent purpose. Since I have stipulated that the really believe in large-scale evolution, we can rule out the possibility that I think there is a malevolent goal.
     So, perhaps you would like to say it like this: "I agree with <nameless>. Even though you say that you think scientists believe in evolution and are just holding it as a sacred belief, I think you really believe that they don't believe in evolution at all and are plotting to scheme people out of their money." (Substitute whatever plot you think is appropriate.) That is what it would mean for me to be engaged in a conspiracy theory as the term is normally used -- well when it is not used as a distracting "don't pay him any attention" device. There is your accusation out in the open. Would you care to substantiate it?

Dave W. said...

Apparently, the c-word is such a hot-button for you that you focus on it to the exclusion of any other questions put to you. Let's see if I can help clear the waters, and perhaps allow you to describe the evidence and definitions I asked for.

The Asch study had people conforming about a third of the time, and was not done explicitly on scientists. Assuming the results held for a group that is trained to argue for what's true (or our best approximation of it), we would still have two out of three scientists loudly proclaiming that some part of evolutionary theory must be false had the genetic data been drastically different than the phenotypic data. Asch is not supportive of the situation we have now, especially when Asch also found that group size was irrelevant, and that conformation was less likely when someone else besides the subject disagreed with the majority.

Second, in science, conforming to the norm does not generate rewards. Newton, Darwin, Wegener, Einstein, Gould and Marshall (for just a few examples) aren't remembered for saying what everyone already knew, or thought they knew. Nobel Prizes aren't handed out to scientists who maintain the status quo. Even run-of-the-mill grant applications won't be approved for research that simply reguritates scientific "dogma." Scientists are in competition with one another, and being able to show that some "overarching concept" of evolutionary biology is wrong - even a little bit - will nearly guarantee fame. Furthermore, many people become scientists simply because they've got a passion for knowledge, a drive to find the truth regardless of what that truth is or what it means to the status of previously held conclusions.

And so, when one examines your proposal in light of the scientific context, one must conclude that any scientist who "brushes aside" inconvenient facts in deference to a declared "truth" does so in contradiction to his/her own economic, social and intellectual self-interests, and because Asch's results do not suggest more than a 33% conformance rate, one is forced to conclude that you are proposing a massive, malicious and active conspiracy among scientists to repress the truth (regardless of your intent). Passive conformity simply cannot explain the phenomenon you claim to be witnessing, and so must be rejected as a hypothesis. That is why people think you're offering a conspiracy theory, whether you are or not.

So back to the meat of the discussion... The only "overarching concept" that hasn't been seriously challenged since before Darwin's day is that evolution occurs. It's a simple matter to see that once upon a time there were no dinosaurs, then there were dinosaurs, and now there are no dinosaurs again. Obviously, something changed. Evolutionary theories are attempts to explain those observations, and it is those observations that are pretty much the only part of the entire biological landscape of ideas that have not been overturned in the last 200 years or so (predating Darwin's birth, even). And if those observations are what you think are "sacred beliefs," then you'll have to include geologists (at least) in with the evolutionary biologists as perpetrators of this allegedly passive agreement to not rock the boat.

Pvblivs said...

Dave W:

     It's true that Asch's study only generated a 33 percent conformance rate, when the group answer was obviously false and when the group answer did not come from someone expected to have greater knowledge. It also did not include members of the group directly pressuring the subjects (which is the only purpose I can see for the "c-word.")
     Science has advanced to a point (I think you'll agree) that further progress requires a significant trust in the judgement of predecessors. Otherwise, it would be mired in perpetual repetition of the work already done. And since the consensus answer is actually plausible, A high rate of agreement would be expected. Your prediction that (based on the results of the study) two-thirds of scientists should be vehemently opposed to evolution would require that it was not plausible and that they had not come to trust it before being admitted into the scientific community.
     "[O]ne is forced to conclude that you are proposing a massive, malicious and active conspiracy among scientists to repress the truth"
     No, one is not justified in coming to such a conclustion, let alone forced. If I were saying "The idea is false. An experiment in <wherever> showed that evolution and common descent cannot occur above (some level) but they destroyed the documentation" that would be a conspiracy theory. What I have actually stated is that I do not how to conduct an experiment capable of falsifying the idea (if it is false) without setting up controls over millions or billions of years. I don't think the data could be collected within our current abilities. So, no one is actively repressing data. Finding it is beyond our abilities. Similarly, there is no Nobel prize in challenging evolution unless one can come up with an experiment and controls in which a reasonably possible outcome (reasonably possible under the assumption that evolution is false) could not be reconciled with the idea of evolution. If such an experiment existed (even though I can't conceive of one within our abilities) I would expect someone to try it -- if only for the shot at fame. I have been presented with no such experiment. I conclude, then, that there is none. (The other alternative, that someone tried to conduct such an experiment but was blocked, would be a conspiracy. But I make no such assumptions.)
     "The only 'overarching concept' that hasn't been seriously challenged since before Darwin's day is that evolution occurs."
     And that is the only thing that I am identifying as a sacred idea. Genes are challenged. Passing on of acquired traits is challenged. But "evolution is true" is not challenged.

Dave W. said...

Science never requires trust in one's predecessors, because any test you execute using their conclusions as your premises necessarily tests those conclusions, all the way back to whatever basic observations were made in the first place, that led to an entire field of discovery.

If two of your predecessor's conclusions, from their experiments, logically suggest a new hypothesis which you test, and the test fails to validate the hypothesis, then it is clear that one or both of the earlier conclusions was wrong (either that or the logic or test were flawed, but we're assuming competence among researchers, are we not?). Those conclusions, likewise, rested upon other premises, which are also called into question in the failed test, and so on back to the most-basic evidence, whatever it is. Somewhere in amongst all those premises and conclusions is a mistake, and the first urge of many researchers facing a failed experiment is to find the mistake, correct it, and use that new knowledge to suggest new experiments.

Whether or not such a mistake would do in all of evolution would, of course, depend on the mistake. I offered a few possibilities in a prior post. The first would require nothing more than a re-write of textbooks. The second would require dropping all DNA-based research as fruitless, and starting over in our search for the mechanism of heredity. The third would pretty much be the death knell for most of modern evolutionary theory. But all of them would be serious enough that any scientist who tried to "brush aside" the negative results would be a laughingstock, while his peers who accepted the results and ran the tests to learn where the theory went wrong would be accepting Nobel Prizes. Any other scenario, with a positive disproof of an evolutionary basic like common descent, would indeed require a massive, forceful conspiracy to implement and maintain (for what motive, I couldn't begin to guess).

Yet no matter what the mistake, the fact would remain that evolution happens. That part is not hypothetical: we can see organisms change in the lab, and we can see that they changed over millions of years in the rocks. Even if all of today's theories were to vanish, those direct observations would still be begging for an explanation. The most-basic "overarching concept" of evolution is that the porportions of various heritable characteristics in populations of organisms change over time. And most people only need to look at a couple generations of family photos, and see that children are not clones of their parents, to understand that in its most-basic sense, "evolution is true."

Your claim is not unlike "nobody ever tests to see if gravity is true." But they do. Every rocket that successfully reaches orbit is a test of our theory of gravity, as is every time your GPS successfully locates you at the right spot on the Earth (Einstein's General Relativity - our current theory of gravity - is required to do time corrections on the fast-moving GPS satellites). Nobody has yet formulated a rigorous proposal for why these small-scale tests are not valid on the scale of galaxies or clusters, or even the whole universe. And where we can test them on large scales - by observing how big things move - Relativity has done remarkably well in terms of its successful predictions.

Similarly for evolution, there has not been any rational argument put forth to date as to why we should doubt that the mechanisms we can successfully test over decades will work on larger scales over millenia. The fossil record is analogous to instantaneous snapshots of galaxies, and we infer that between the snapshots, things have changed (the galaxies due to gravity, the animals due to evolutionary mechanisms like genetic drift and selection). So every dug-up fossil gives us further confirmation of the "overarching concept" of evolution (or it gives us new data with which to refine the theory even more than it is now).

Finding a Devonian Bunny, on the other hand, would throw the theory for a big loop.

Oh, hey, before I forget again: where is the evidence that led you to the conclusion that negative results to evolutionary experiments "would predictably have been brushed aside?"

Pvblivs said...

Dave W:

     What led me to believe that negative results would be brushed aside. The beginning was the fossil digs themselves. While these are not the only things said to confirm evolution, they are the most visible to the general public. They continue to stand out in my mind. When scientists find a fossil they are looking for, it makes headlines. It is considered powerful positive data. But if they don't find it. If they find nothing (which is the most negative result that can be expected) then it is simply noted that fossilization is rare. Now, I'm not going to challenge the assertion that fossilization is rare. But the result is that it becomes "confirm or inconclusive." There isn't any result that can qualify as negative. And this is what I am calling a "brushing aside." Some things are brought up as "poetential falsifiers" (e.g. bunnies mixed with dinosaurs.) But if we were going to find that, it would probably have been found already. In fact, I have failed to find any reference to that as a potential falsifier earlier than that condition holding.
     "Your claim is not unlike 'nobody ever tests to see if gravity is true.' But they do. Every rocket that successfully reaches orbit is a test of our theory of gravity, as is every time your GPS successfully locates you at the right spot on the Earth (Einstein's General Relativity - our current theory of gravity - is required to do time corrections on the fast-moving GPS satellites)."
     Interesting. Understand my perspective is that 1> such a failure would only an idea within gravity was false (of course gravity is directly observed) 2> these are not actually meant as tests of the theory (although the early launches might have been.) The relativistic theory of gravity is sufficiently trusted that, unless every possible mechanical failure was ruled out first, a possibility of error in the theory wouldn't be given a first thought.
     Yes, when things go unexplainably wrong, scientists do take apart the old theories. But under normal circumstances, they assume that what has worked in the past will continue to work.
     "Similarly for evolution, there has not been any rational argument put forth to date as to why we should doubt that the mechanisms we can successfully test over decades will work on larger scales over millenia."
     But that's just it. It is trusting an extrapolation. Right now, the idea is academic. No vaccine will fail if it turns out that the extrapolation fails. I just don't see a way to put the extrapolation to an actual test -- well, not anything we can actually do. If we fail to find a particular transitional form, it could be because the form never existed or because it existed but never fossilized. If we could establish "never existed," that would be a disconfirming event. So, if we had a way to check for "never existed," we would have a test for large-scale evolution. But we don't have a way to check for that (at least not that I am aware.)

Dave W. said...

So it is inference ("extrapolation") that you have issues with, after all!

If I go out and count swans one day, perhaps I see 49 white swans and one black one. I infer (extrapolate) from that data that 98% of all swans are white.

How do I test it? By continuing to collect data on swan coloration. Perhaps the real number is 97.4% or 99.6%, I can't ever be sure unless I count every swan in existence, and I will never know that I've done that. Perhaps deep in some forest that I simply cannot afford to trek into there is a large population of all-black swans which seriously break the inference, or at least require an additional explanation (perhaps they're so genetically different they shouldn't be called 'swans'?).

Anyway, the only way to test the inference is to keep counting swans. Likewise, the only way to test our inference to common descent and evolution is to keep digging up fossils and to keep running lab tests, and [i]when[/i] we find things that don't match one of our previous theories exactly (it happens all the time), we either modify the theory or discard it.

And of course, if I go out to count swans but don't find any, that doesn't say anything about my inference, just like not finding a fossil says nothing about evolutionary theory. Now, if I go looking for a fossil in certain strata and find it in different strata, that's a failure of the prediction I used to identify the likely strata to search. (For example, all T. rex fossils we've found so far have been between 65.5 and 68.5 million years old, so we infer that that's when they lived, but a single fossil turning up in 70-million-year-old strata would break that inference and force us to make a new one. Perhaps a trivial example, but "bigger" inferences work the same way.)

Similarly, we infer that gravity works the same way all over. If you've got a perfectly functional GPS unit and it sees perfectly functional GPS satellites but your position as displayed by the unit is wrong by a significant amount, then either Relativity is wrong at your location (disproving the idea that it is universally applicable) or you've discovered a gravitational anamoly. Either one would be a mind-bender to a lot of physicists, who would flock to the site to figure out what's going on. Brushing it aside would not be an option.

The point is that an inference (from a decently large data set) is justified until data comes along which contradicts it, so simply collecting more data is a method of testing the inference, because new data which doesn't fit will require modification or discarding of the inference. And every use of an inference as a premise for further testing (or as a technology) is also a test of the inference.

And inference is the best we can do. Ever. Even with a test-tube experiment in the lab, if it worked five, ten, fifty or a hundred times, we can only infer that it will work the 101st time, or outside the lab.

Pvblivs said...

     I realize that an inference is the best that we can do. My "problem" (with calling it a theory) is that (because of our limitations) the worst that can happen in the tests is an inconclusive result (e.g. a particular fossil is not found.) It is a little like I have a coin that I postulate always comes up heads when flipped. But sometimes the coin rolls under the counter and I can't see whether it came up heads or tails. Even worse, for some strange reason I cannot tell the difference between tails and rolling under the counter. To use my analogy. I have simply not seen any test where a result of "tails" could be distinguished from "rolling under the counter." I am quite aware that you have many results of "heads."

Dave W. said...

Nonono, the worst that could have happened is if we were to have found the fossils in random order. If that had happened, there would be no mention today of "common descent," because it would make no sense in light of the data.

The fact that all of our raw observations support the inference is simply a demonstration of how strong the inference really is. Potential falsifications (like the Devonian Bunny) abound, but never actually appear. That means the correspondence between our explanation and reality is very, very good.

Because the problem with your coin analogy is that your hypothesis (always heads) doesn't match the assumed reality (head or tails). Evolution, on the other hand, is an attempt to create a match between "what is" and "what we know." It's not an attempt to verify something we don't see, it's an attempt to explain what we do see.

Try a different analogy. Fred is always honest. Our hypothesis is that Fred always stops by McDonald's before entering his office for work. If Fred is sick or on vacation, and so doesn't go into work, his failure to go into McDonald's on those days is not a failure of our inference. Our only test is that when Fred shows up in his cubicle on works days, we ask, "Hey Fred, did you go to McDonald's again today?" Surely it is not a problem with either our inference or our method of testing if he never says "no" (remember, he's always honest). It simply means that our hypothesis is correct, in every known case.

What if one day, he says "no?" Our hypothesis fails! And he says, "no" for seven days in a row. Perhaps we would devise a different test, and go across the street and look at the McDonald's restaurant ourselves. Perhaps there is a sign on the door that says, "remodeling for two weeks." And sure enough, four work days later, Fred starts answering "yes" again to our standard query. We're forced to change our hypothesis, to "Fred always stops by McDonald's before entering his office for work, so long as the nearest McDonald's is open," and we can once again say that our hypothesis is correct in every known case.

Then, the next time Fred says "no," we can be quick on our feet and make a prediction and test it: "Oh, the McDonald's across the street was closed?" we ask Fred. "Yes," says Fred, "some guy got shot there last night." How "safe" is it to make such a prediction?

Anyway, the point is that if the data we find always supports our theory, it means that we have a good theory. 200 years ago, we didn't have as much data, and so the inference was weaker and people argued about which theory explained the data better. But people kept on collecting data and it became clear that there really is only one inference that makes sense (that correctly maps hypotheses to reality). Perhaps the problem you're having is simply that the tests you want to think up have already been done, and the answer was positive support of evolutionary theory. That we're forced to dream up more and more silly-sounding falsifiers (like Devonian Bunnies) is nothing more than a consequence of the theory's strength.

By the way, if a coin always flips heads when we can see the results, the natural hypothesis is that the coin has two heads. Whether it rolls under a counter or flips tails suddenly becomes moot, doesn't it?

Pvblivs said...


     I admit that I am using an imperfect analogy. I am to convey the idea of being able to see a positive result but being unable to distinguish a negative from a neutral result. Presumably, if certain transitional forms never existed, it would falsify evolution. But we can't see that. There is no way to distinguish the absence of a form from a failure to fossilize. If some transitional form really never existed, we can't know that. We can already determine that many species will be forever unknown because they did not fossilize.

Dave W. said...

You're right: we can't distinguish not finding a particular fossil from that fossil having never existed, just like with our simple tests on Fred, we can't distinguish him being out sick from him going to McDonald's but then getting hit by a bus as he crossed the street to come to work. With Fred, our tests can become more sophisticated (we could ask more and more varied questions), and we can with little effort find testable alternatives to explain his behavior.

But what testable alternatives exist to common descent? Better yet, what plausible alternative exists to common descent?

Of course the absence of a competitive theory does not mean that evolution is correct, but the 'gaps' in the fossil record - at least for critters with hard parts - just keep getting smaller as we dig up more and more of them, which (as I said) strengthens the inference.

Look at it this way: we've got evidence that animals exist, and we've got evidence that animals just on either side of any particular gap existed, and we've got evidence that animals change over time. The natural inference is that animals of some sort filled that gap, especially when we have zero evidence in favor of any other possible gap-explaining mechanism, plausible or not.