Re: [asa] historical versus experimental sciences

From: Cameron Wybrow <>
Date: Fri Aug 14 2009 - 17:11:46 EDT


I appreciate your knowledge of biological fact, but masses of further
examples aren't helping the discussion any. This may be indicative of the
difference in thought-patterns between philosophers on the one hand, and
biologists on the other. I'm trying to point out a general flaw in a line
of argument, and you are trying to answer me by drowning me in data.

If I recall, the news item that started this discussion was that there was
some fruit fly ravaging crops in California, or some other state. Then, if
I recall, you suggested that sometimes, if we don't know very much about a
species (e.g., if we didn't know much about the ravaging fruit fly in
question), we might get a clue about its nature, and therefore how to handle
it from an agricultural point of view, from its close evolutionary
relatives. Thus, if a close evolutionary relative employs a certain
pheromone, it might be that the ravaging fruit fly is sensitive to the same
one, and we might be able to use that knowledge as a method of control.

I have no objection to the basic logic of such an approach. The
difficulties I was posing were further along the line. If you don't know
very much about a species to start with (and we were talking about just such
a case), how are you going to determine its close evolutionary relatives?
You can't determine that X and Y are close evolutionary relatives unless you
know at least something about both X and Y. Now in the case of the fruit
fly invader, you've already suggested that we would have no direct knowledge
of its biochemistry and won't have sequenced its genome yet. So the
evolutionary classification you are talking about can't be based on either
of those things. On what is it based, then?

You've said repeatedly that shape alone is not a safe guide to evolutionary
closeness. You even gave an example of a weird-looking fruit fly that is
closer in evolutionary terms to a given fruit fly than many others which
look more like it. OK, I grant it. But where does that leave you? I don't
know anything about the biochemistry or genome of fruit fly X, so I can't
determine its evolutionary relatives based on those. I can't rely on form,
either, since that can be misleading. How, then, do I determine its
evolutionary relatives? Without that determination, the practical
application you're suggesting for Darwinian theory is dead in the water.

More arguments with detailed examples aren't necessary here. Just give me
*a list of the criteria* by which a biologist could deduce that two fruit
flies are close evolutionary relatives if the biologist had *zero* knowledge
of the first fruit fly's genome, biochemistry, and physiology. A list of
words or phrases will do, e.g.: "We could ascertain beyond a reasonable
doubt that the two fruit flies were close evolutionary relatives if they had
similar flight patterns, similar mating habits, a similar response to
high-pitched noises, or a similar diet." Once I have your list of criteria,
I will decide whether I need to ask further questions.


----- Original Message -----
From: "David Campbell" <>
To: "asa" <>
Sent: Thursday, August 13, 2009 6:24 PM
Subject: Re: [asa] historical versus experimental sciences

> This discussion doesn't seem to be getting anywhere. Perhaps it would
> be more helpful to provide my revision of Dobzhansky's statement:
> Evolution is relevant to everything in biology, and everything in
> biology makes sense as a coherent whole in light of evolution. Even
> my research, which focuses on evolutionary questions, does not involve
> thinking about evolution all the time. As I do the lab work, I need
> to be focused more on doing the steps properly than on why. Carefully
> documenting the morphology and molecular features of the mollusks
> depends on factual reporting more than on evolutionary
> considerations-"hey, that mussel has a funny-shaped flap and so does
> that one, whereas those two have big bumps and a lot of little bumps,
> and those two have a thick stripe on the gill." But when I try to
> make sense of the pattern, I notice that they coincide with the
> patterns in a gene involved in how mitochondria make energy and in the
> DNA sequence for a spacer region between nuclear genes involved in RNA
> translation. The coincidence makes sense in light of evolution, but
> not in light of any non-evolutionary scenario that I have seen.
> Of course, such evidence does not rule out a model like Behe's in
> which there's a lot of evolution with some help as needed; the issue
> there is whether the evidence indicates a need for help, evidence of
> such having occurred, or a theological likelihood of it.
>> I did not say that Linnaeus was "committed to the fixity of species". I
>> did, however, imply that Linnaeus did not advocate an
>> evolutionary view or base his biological thinking on such a view. If I am
>> incorrect in this belief, please supply me with direct evidence that
>> Linnaeus supported or advocated macroevolutionary views and that they
>> affected the way he approached biological questions.
> I intended to give a bit of a caveat, not particularly important to
> the overall discussion. Linnaeus apparently did accept the idea that
> species within a genus might have some more or less evolutionary
> connection; there would have still been influence of ideas like the
> chain of being as well, so trying to fit him into a modern
> antievolutionary or evolutionary category would be anachronistic.
> Overall he would probably come out fairly similar to the evolution
> within a "kind" view found in some YEC and OEC circles. However,
> based on somewhat limited data, he assigned chimps to Homo. Whether
> or not this amount of evolution constitutes "macroevolution" or not is
> open to debate.
>> I did not say that "a priori" similar species ought to be chemically
>> similar. I said that if species are similar in a lot of ways there is a
>> good chance that they are biochemically similar as well. I granted fully
>> that this might not be the case, that the inference was only one of
>> probability, based on experience and common sense.
> But why? There are also superficially similar species that are
> chemically quite different. If you study them more closely, you can
> find other features that point to greater difference. How do you know
> what features are better or worse for predicting other similarity? On
> the other hand, closely related species can be quite different in a
> number of ways. Because we're built for walking and running, whereas
> they're built for knuckle walking and climbing, humans and chimps are
> quite divergent in many morphological features relative to some other
> groups of organisms. Features that obviously respond to external
> evolutionary pressures are likely to show convergence (e.g., shape of
> fast swimming shark, bony fish, ichthyosaur, and dolphin). One
> chemical similarity has no obvious a priori connection to many other
> chemical similarities. If you observe enough of them, you might have
> some empirical expectations, but no fundamental reason to back them
> up. Furthermore, there is a consistent pattern of decreasing, but
> still significant chemical similarity as you go farther and farther
> away across organisms. Why should a nematode be biochemically enough
> like humans to be a useful lab animal in studies relating to human
> health? It's quite true that the people working at this university on
> nematodes as model organisms don't care professionally about nematodes
> nor why they work as models, and they have a very successful and
> recognized research program. But if you ask the "why" type questions,
> evolution gives the answer.
>> In any case, as I made clear, Darwinian evolutionists cannot be sure that
>> the two evolutionarily related species will be biochemically similar,
>> either, so they have no advantage over the "common sense" approach in the
>> pheromone-similarity-hunt that we were discussing.<
> Not "no advantage", merely "no absolute guarentee of advantage".
>> I spoke not of oranges and tennis balls but of oranges and tangerines.
>> You
>> know that I do not suppose that oranges
>> and tennis balls are "alike" simply because they are round.
> But they are alike-they are round, similar in size, and sometimes
> similar in color.
>> Give me credit for enough brains to distinguish the living from the
>> non-living, please.<
> As an illustration of my inclination to pedantics, the exact
> delineation of living versus non-living is in fact quite problematic
> getting into things like viruses. More to the point, why should
> living things be more likely to be chemically similar than non-living
> things? Tennis balls and living things are both composed of large
> carbon-based molecules (the latter also have a lot of water). For a
> plant example, why are osage oranges less like oranges than lemons
> are? Osage oranges are round, with a bumpy surface, suggestive of an
> overgrown underripe orange.
>> Further, it was *you* who said in your last post that morphology (absent
>> knowledge of the genome) is still the main factor driving evolutionary
>> classification. And you *also* said that in the case of the fruit flies,
>> evolutionary classification would be useful in predicting the
>> biochemistry.
>> So your argument implies that in many cases predictions about
>> biochemistry
>> would in fact be ultimately based on morphology. Now you are saying that
>> morphology is unreliable as a guide to probable biochemistry? So then the
>> evolutionary biologists would be making an unreliable guess about the
>> pheromones? Then why do you think evolutionary biology would be so useful
>> in such a case? Your logic escapes me.
> Because evolutionary thinking helps us identify which morphological
> and molecular features are likely to be better indicators of overall
> similarity.
>> So, tell us: since we are talking, ex hypothesi, about a fruit fly
>> species
>> for which we *don't* have detailed biochemical or genetic data, and since
>> bodily form can't adequately distinguish the various fruit flies for
>> evolutionary purposes, how on earth did you determine that our unknown
>> fruit
>> fly invader, X, was evolutionarily more closely related to weird-looking
>> type A
>> fruit fly than to the more visually similar type B fruit fly?
>> Evolutionary
>> classification must be based on *something*, and we've just ruled out
>> genomics, biochemistry and morphology in this particular case. And fruit
>> flies have no bones to be preserved, so the fossil record will be of no
>> help. So tell me how your evolutionary biologists could have classified
>> this new invasive fruit fly well enough to know what its evolutionary
>> "close
>> relatives" would be. I sense smoke and mirrors here.
> Based on evolutionary studies, we know what features are better
> indicators of overall similarity. Morphology is quite useful, but it
> has to be examined in detail, with awareness of the evolutonary
> patterns in the group, in order to figure out what are the best
> features to look at. Actually, insects have decent fossil records-the
> outer layer is tough, and by flying they can get into amber, lakes,
> lagoons, and other good places to get fossilized. Nevertheless, I
> don't think the fossil record gives us a real good phylogeny of fruit
> flies. "These two flies look similar to me" is not going to be very
> reliable, given that I am not an expert in fruit fly classification,
> but someone who has carefully studied the morphology could give a
> reliable suggestion of close relatives to study.
> Of course, someone could look up current classification of fruit flies
> and see what species are considered to be most similar, without
> bothering to consider the evolutionary work that lay behind the
> classification. But the classification had to come from somewhere.
>>> But to know what the gene does and how it's supposed to work,
>>> evolutionary data may be quite handy. In turn, this will help tell us
>>> how to deal with the underlying genetic problem. Of course, the first
>>> task is stopping the cancer, which depends on medical knowledge
>>> (although discovering a cancer drug in species X should prompt study
>>> of its close relatives to see if there are related useful compounds).
>>> Cancer itself is something of an evolutionary situation, in which the
>>> cancerous cells are evolving ways to dodge the body's controls while
>>> the body tries to respond. Remembering that the cancer (or any
>>> pathogen or parasite) is continually evolving is an important part of
>>> maintaining medically relevant treatments. Cf. the Doonesbury cartoon
>>> where the doctor asks the creationist patient whether he wants the
>>> original drug or the ones that are effective against the strains that
>>> have evolved resistance to the first drug.
>> This paragraph equivocates on the meaning of "evolutionary situation".
>> What
>> you are calling evolutionary here is microevolution, which everybody
>> grants.
> Not exactly. Wells claims cancer is not caused by mutations, for
> example. Moreover, the distinction that you are making between
> microevolution and macroevolution is dubious. On the one hand, it is
> true that one is not automatically obliged to accept full common
> descent, etc. just because evolution clearly operates on some lower
> levels. On the other hand, there is no fast line between the everyday
> evolution within populations and the origination of new major kinds or
> features. Furthermore, the line between "microevolution" and
> "macroevolution" is drawn by ID, YEC, etc. at all levels, from "no
> species can evolve into another species" to "full common descent, but
> with some help from fine tuning or relatively rare miraculous
> insertion of information." Thus, when someone claims to accept
> "microevolution" but not 'macroevolution", what they mean is
> ill-defined and probably more often than not includes rejection of the
> sorts of speciation events that are regularly observed, given the
> vigorous disparaging of "evolution" on the part of YEC and some ID.
>> That cancer cells "evolve" in the way you are describing is
>> non-contentious.
>> And I grant your practical point, i.e., that doctors have to take into
>> account the shifting behaviour of the cancer cells in their diagnosis and
>> treatment. But that does not mean that they have to believe, e.g., that
>> mammals evolved from reptiles via Darwinian means. It means only that
>> they
>> must grasp the genomics and biochemistry of cancer cells.
> Then we do agree that evolution is useful and relevant to medicine.
> The evolution of mammals from reptiles does have some consequences for
> medicine, but much less immediate-more on the "why is the body
> arranged this way", "why are this and that linked" type of questions
> than "how to address this infection or cancer"
>> And by the way, creationists are not so dumb as you make out. They
>> understand perfectly well the fact of drug resistance.
> Not dumb, but many are as careless as that, denying all "evolution".
> The quality of popular antievolutionary material is generally
> appalling, and few in the general public have given serious thought to
> the immediate practical importance of evolution in pathogens.
>> I am no creationist (in the narrow sense that you mean)
> Not me-it's Trudeau's usage.
>> In both these examples, you avoid specifying how we determine(d) what was
>> "closely related". How did you determine, for example, that the parasite
>> was "more closely related" to plants than animals, without employing
>> genomic
>> or biochemical data (which is forbidden by your original claim in the
>> fruit
>> fly case), and if morphological data is not reliable?
> The work on Toxoplasma, including work by people at Messiah College,
> involves genomic data. But before the complete genome was sequenced
> (it has been now), the similarities that they spotted suggested that
> there would be additional similarities. More importantly, they
> suggest that Toxoplasma will overall be more biochemically similar to
> plants than to us, which suggests that herbicides that don't harm us
> may be worth investigating to see what they do to Toxoplasma, even if
> we don't have the molecular mechanisms fully characterized (either of
> Toxoplasma or the chemical).
>> What you need to show me is that evolutionary classification gives some
>> predictive power *that is not directly derivable from the criteria you
>> used
>> to do the classifying in the first place*. I claim (though I am willing
>> to
>> change my view if evidence is provided) that ultimately evolutionary
>> classification in itself has little or no predictive power, but depends
>> on
>> other comparisons (morphological, genomic, physiological, etc.), which in
>> themselves do have predictive power. In a word, evolutionary
>> classification
>> is an interpretive gloss of little or no technological value.
> Instead of writing long posts to the ASA list, I ought to be working
> on, among others, a paper on Juga, a genus of freshwater snails from
> the northwestern U.S. I have DNA sequences from several populations
> (one or two mitochondrial genes). Classification has been based
> mostly on shell sculpture, but there are some problems (changing
> sculpture with growth; is weak sculpture closer to smooth or to
> strong, etc.). Species and subgenera are currently treated as
> widespread. However, when I analyze the gene sequences using an
> evolutionary model, several patterns stand out:
> Although things with similar sculpture are often close in molecules,
> relatively smooth shells turn up in many different groups (other
> features, such as soft anatomy, provide useful characters).
> Similarity in sequence correlates closely with where you are, although
> the complex geologic history means that sometimes the connections
> match past rather than present connections. (E.g., at a large scale,
> they are more similar to east Asian forms than eastern North American
> forms).
> Molecular differences generally delimit populations from relatively
> small areas. In turn, these group into several (usually multispecies)
> groups that are also geographically circumscribed.
> Thus, based on evolutionary considerations, I can predict something
> about unstudied Juga populations. I am trying to get DNA from some
> additional populations south of the ones I have data for. If they
> finally cooperate, Juga from the American River will probably
> genetically most similar to the smooth-shelled forms from the Pit
> River and immediately south of it (Feather headwaters, Battle Creek)
> rather than to those to the northwest or those to the west of the
> Sacramento system.
> There's nothing conspicuously different about the lifestyle of species
> that occur in the American River versus the Klamath or Umpqua or
> Columbia, or even some of the eastern North American forms; some of
> the eastern North American ones can be confused with the western ones
> in shell form, too. Why should the American River taxa be more
> similar chemically to the ones in the Pit, except for evolutonary
> reasons?
> More fundamentally, the request "What you need to show me is that
> evolutionary classification gives some predictive power *that is not
> directly derivable from the criteria you used to do the classifying in
> the first place*. " is fundamentally impractical for anything, not
> merely evolution. Nothing predicted by any theory is not ultimately
> derivable from the data that are used to generate the theory.
> Theories provide general frameworks for data, and often point to
> unexpected implications, but they are based on the data, and one could
> always say "here's a pattern in the data that I expect to find in
> other things" without explicitly formulating a theory.
>> I agree with you about the need for theory in engineering. But in
>> engineering, the theory is directly relevant to diagnosis and practical
>> solutions. Biochemistry is like engineering as well. That is why
>> biochemistry is so useful to medicine.
> Although biochemistry is based on chemistry and ultimately on atomic
> theory, in practice there's relatively little direct use of atomic
> theory in biochemistry. There's a lot of empirical data involved, and
> often the system of interest is too complex to adequately reduce down
> to individual atoms.
>> More generally, for a science to have technological applications, it must
>> be able to predict *future* events.
> Not merely that, it must predict future events of interest to
> technology. Evolution can provide prediction about future events (as
> well as about future discoveries relevant to past events, e.g. new
> fossils or the results of new biochemical or morphological studies),
> but systems of interest are rarely simple enough for straightforward
> prediction, in contrast to many cases with physical laws.
>> I agree with your complaints about many of these things. However, the
>> hard
>> fact is that none of these things seems to apply to electrical theory,
>> Newtonian physics, etc., etc. The general public accepts 95% of the
>> science
>> which is taught in high schools without reservation.
> There is less direct objection to most other areas of science
> (exceptions include postmodernist fringe denial of any objective truth
> as well as buying into incorrect alternative "science"), but still a
> generally rotten grasp of them and a willingness to accept bogus
> science such as astrology.
>> And while I grant that
>> religion plays a role in resistance to evolutionary theory, the truth is
>> that physics and chemistry teachers are simply much better at answering
>> questions -- and I'm talking about scientific questions, not theological
>> questions -- to the satisfaction of students than biology teachers are
>> (at
>> least, when the biology teachers are teaching Darwinian evolution).
> But many students are taught to be a priori dissatisfied with
> evolutionary explanations and are not particularly paying attention to
> the merits of the answers. Physics and chemistry get attacked, too,
> when they are inconvenient for young earth or other favored claims.
>> Darwinian explanations are always looser, less precise, with very little
>> quantification (how many mutations would it take? how many years would it
>> take? -- the biology teacher doesn't have the slightest idea), and the
>> proposed mechanisms are of a broad, vague, non-specific kind ("drift",
>> "mutation", "natural selection"), making them extremely elastic, and
>> thus,
>> when combined, capable of explaining virtually any possible outcome (X,
>> the
>> direct opposite of X, and every outcome in between), so that the theory
>> seems oddly slippery.
> Drift, mutation, and natural selection have very specific definitions
> and several well-formulated mathematical models. However, the
> situations of interest for evolution are inherently more complicated
> than an introductory physics or chemistry problem.
>> One thing that might help would be for evolutionary biologists to stop
>> crowing about how great their theory is, and put in caveats showing
>> intellectual humility and caution.
> As a recent Science News pointed out with regard to physics, this is a
> general problem in science (not to mention human endeavor generally).
> Steve Gould commented favorably on our paper's disinclination to
> assert that the results of our molecular phylogenetic analysis were
> the authoritative last word, though in reality that reflected as much
> our morphological backgrounds as any particular humility or, at that
> point, enough experience with the vagaries of molecular analyses to
> know what's solid or not. (Campbell DC, KT Hoekstra, and JG Carter.
> 1998. 18S ribosomal DNA and evolutionary relationships within the
> Bivalvia. In P. A. Johnston and J. Haggart, eds. The Bivalvia: an Eon
> of Evolution-Paleobiological Studies Honoring Norman D. Newell.
> Calgary: Univ. Calgary Press, p. 75-85.; Gould's comment is in the
> preface to the volume)
>> (predictive *in the engineer's sense*, I mean; not in the weak sense that
>> we
>> would expect fossils like Tiktaalik, but in the strong sense that it can
>> tell us what rabbits placed in the Amazon jungle will evolve into a
>> hundred
>> thousand years from now, as astronomers can tell us where Mars will be a
>> hundred thousand years from now).
> Although 100kyr is not enough to have significant uncertainty in the
> position of Mars, it could be just about anywhere (apart from close to
> the Sun on a galactic scale) in 4 billion years, including in the Sun,
> collided with Earth, or ejected from the Solar System. The long-term
> prospect of a gene with a known frequency in a population, with a
> known degree of beneficial or negative selection, can be calculated as
> well as the prospective position of Mars. Rabbits already live in the
> Amazon and haven't changed all that much, so they probably won't do
> anything dramatic over the next 100kyr-even if significant climate or
> other environmental change occurs, there are rabbits in various
> habitats, so the general body plan works OK. (Whether this particular
> species is adaptable or not is an open question-several related taxa
> are severely imperiled). Also, 100,000 years is not all that long,
> evolutionarily. Make Mars or Antarctica, etc. habitable, add rabbits
> but not a full suite of mammal types, and give it a few million years,
> and the situation is much more interesting and hard to predict in
> detail (not to mention de rest of de body).
> --
> Dr. David Campbell
> 425 Scientific Collections
> University of Alabama
> "I think of my happy condition, surrounded by acres of clams"
> To unsubscribe, send a message to with
> "unsubscribe asa" (no quotes) as the body of the message.

To unsubscribe, send a message to with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Fri Aug 14 17:13:10 2009

This archive was generated by hypermail 2.1.8 : Fri Aug 14 2009 - 17:13:10 EDT