 |
 |
|||||||||||
|
|
|||||||||||
|
|
|
|
A Tale of Two CitationsBy James DownardPart 2 Had Woodmorappe used the actual normalized value (or the criteria-consistent value that was still close to it), it would also have substantially narrowed the jump for the next group that represented taxa living six million years later. Now to go from Lucy the australopithecine to Phillip Johnson required less than 5 million years ... so we need to ask what is the "correct" rate at which animals are supposed to be accumulating change over even longer time frames? Does it say anything dire about the evolutionary process that clusters of animals separated by millions of years looked different? Only different in what ways? To figure that out, wouldn't it be necessary to discuss what it was that was changing? And that's where Woodmorappe got off the bus, because in no case did he ever explain what any of these features referred to. Indeed, he quite bluntly acknowledged that he would "not attempt to make any anatomical judgments" about any of the features being discussed. It is on this point that Woodmorappe's "normalizing" is completely unrelated to what happens in a real cladistic analysis (which Berlinski seemed to think his study was). To assess fossil relationships cladistically is precisely to "make anatomical judgments." Indeed, the whole point of the exercise is to rigorously search for the shortest (most "parsimonious") path through the full range of the available data. While some features might "reverse" fairly easily (the length of a thighbone, for example), others are more suggestive of deeper evolutionary processes. Woodmorappe even quoted paleontologist Zhexi Luo about the need to show how particular features acquired in competing models of descent were "biologically associated," but Woodmorappe swept this caution aside as smacking of "evolutionistic just-so stories." In other words, Woodmorappe was in principle disassociating fossil analysis from comparative anatomy and developmental biology -- the very disciplines essential to making sense of them as once-living forms. Only by this extraordinary exclusion was Woodmorappe able to so confidently decide that "the negative conclusions regarding evolution become all the more compelling." This Olympian seclusion from the grubby details of the fossils may have been part of its attraction for Phillip Johnson, who approached the reptile-mammal transition in exactly this superficial way in Darwin on Trial. Without mentioning even a single specific example of the transitional fossils he was supposedly discussing, Johnson obligingly excused himself from the drudgery of excessive detail in order not to "distract the general reader unnecessarily."[33] Johnson's dismissal of the reptile-mammal transition involved not only the wholesale ignoring of the available paleontological literature. He even managed to overlook the content of the few sources he did cite, which on their own explicitly contradicted half of his argument.[34] Such cavalier "scholarship" helps put Johnson over the top as the reigning Grand Master of the "meaningless concession" -- the tactic whereby one seemingly "accepts" inconvenient information, and then promptly thinks no more about it. This was how Johnson sidestepped the physical evidence for the origin of those two great vertebrate classes stemming from an ancient divide in the amniotic vertebrates: birds on the reptile diapsid side, and the mammals from the parallel synapsid branch. Birds and mammals being the only two vertebrate classes to have appeared in the last quarter of a billion years, Johnson has neatly sidestepped the bulk of recent macroevolution without even getting winded.[35] But lurking behind Johnson's glib recommendation of Woodmorappe to the selectively credulous Berlinski lies the rickety contrivance of those charts, which tried to do with the data something completely inappropriate to them. Each of those zeros and numbers in Sidor & Hopson's original set were simply a convenience of nomenclature, a necessity for the purpose of running them through appropriate cladistic algorithms. They didn't intrinsically represent a numerical condition that could be meaningfully lumped in the way Woodmorappe was doing (and which evidently impressed not only Johnson, but Berlinski, whose biz after all is math). For example, the "progressive" item no. 144 is a postcranial character. Sidor & Hopson's technical description read, "Humeral head shape: broad and strap-like (0), elongate oval (1), subspherical (2)."[36] From a categorical point of view the elements could just have easily been labeled A, B & C. But if they had been listed that way, how ever could Woodmorappe have "added" them (let alone "normalize" them)? An even more significant difference between a cladistic analysis and the juggling act Woodmorappe was playing comes from imagining what changes in the data would signify. Take two more features from Woodmorappe's "progressive" list. If a shallow indentation in the dentary bone (item no. 82) had remained absent in the gorgonopsid instead of being present, but the rear underside of the dentary thickened slightly into a trough adjoining the angular bone (item 84), the "normalized" value wouldn't have changed at all. Yet we would have just physically changed the features of the gorgonopsid ... it would be a different animal, and would therefore position differently when approached from a cladistic point of view.[37] But as seen from the creationist perspective, all these details are merely distracting minutia. And that's because, in a quite literal sense, antievolutionists do not actually conceptualize what intermediate forms or full transitional series would have looked like had macroevolutionary processes been in play. I know this to be explicitly true for Phillip Johnson (on the reptile-mammal transition), Michael Behe (concerning whale evolution) and Jonathan Wells (on Archaeopteryx), because I have made the point of asking them. One may apply this test as a virtual diagnostic for any antievolutionist. Put into prosaic terms, it cuts to the core of what it means to put a theory to the test. If you are insisting there are no Buick parts in the junkyard, you need to have some idea of what a Buick is supposed to look like. Otherwise, how would you know that you weren't already looking at Buick fragments? The failure of creationists to think through what sort of evidence they would accept, rather than simply tossing out whatever they refuse to accept, explains a lot about how Woodmorappe is able to ignore the implication of the data he is supposedly analyzing. We know that he contends that the synapsid chain fails to manifest "a unidirectional progression" toward mammalhood. But by that he is mushing all the fossil data together, not permitting lineages to branch. He is implicitly requiring all synapsids acquire these mammalian features en bloc, as if everything had to be directly related to later forms ... or none were. This is exactly the position Phillip Johnson took in Darwin on Trial: "If our hypothesis is that mammals evolved from therapsids only once (a point to which I shall return), then most of the therapsids with mammal-like characteristics were not part of a macroevolutionary transition. If most were not then perhaps all were not."[38] This is the equivalent of saying that you can't be related to a distant cousin (or have shared a common ancestry) unless each of the contemporaneous generations resembled one another exactly. Viewed in terms of the historical reality that exists outside the YEC six-thousand-year box, such reasoning is absurd. By definition the more taxa you explore the more likely you would be to include cousins rather than close "sister groups" (to use the terminology of cladism) related to the particular terminal form being investigated. There would be nothing special about taking mammals as that terminal position, however. If you focused on the lineage of gorgonopsids instead, mammals would be a definite out-group, and most of the synapsids wouldn't have been developing "gorgonopsid-like characteristics." Nor should they have been expected to. But even with the data Woodmorappe was selecting, we have to wonder whether any sort of distribution could have withstood his ad hoc conditions. To see this more clearly, we need only look at the second of his tables, drawn from Luo & Crompton's paleontological paper.[39] Here's Woodmorappe's summary of the matter, which directly followed the section quoted above on chart one:
To judge how "abrupt" these "discontinuities" were in reality would depend on knowing the actual chronology, which of course YEC believer Woodmorappe left out. Here is how the table looked in his article (again with my scholarly notations) but with the chronological yardstick attached. The table commences 270 million years ago: Quadrate skeletal
characters
a Luo & Crompton listed as "Gorgonopsia." b The "s" was lost from Luo & Crompton's "Probainognathus." c Read as Tritheledontidae. d Luo & Crompton had listed the specific taxon, Morganucodon. We have now moved 56 million years, and the horrible gaps he alluded to represented (surprise!) exactly the spots where the biggest chronological jumps were being taken between fossil samples. Because we're dealing with a much smaller block of data, though, it is easier to see how Woodmorappe's calculations weren't doing justice to the facts by taking a look at the raw data matrix.[40] The 14 characters in Luo & Crompton looked like this :
Woodmorappe specifically claimed that "9 of the 14 quadrate characters used by Luo & Crompton were likewise reversing," but that wasn't strictly true. He was including no. 9 as a reversing trait, when you can clearly see there were no reversals there at all, only some data missing at the far end.[41] Because there were fewer data points and still fewer missing slots, Woodmorappe decided not to "normalize" the values first and managed to do his sums better, with only the "9" for item 19 being off. It should have read 5, the same as the cynodont above. Or rather "6" ... if he had not excised the progressive trait no. 9. When you stop treating these as abstract numbers and look at what they represent, however, the picture once again starts looking different from Woodmorappe's collage. Of the eight traits actually "reversing," two of those (one quarter) were doing this only in the Gorgonopsia (columns 2 and 7), a taxon living early in the sequence (and not held to be on the direct line of descent to mammals anyway). In other words, the farther back you cast your data sweep, the more likely you would be to net related cousins, who may very well have gone their own evolutionary paths that mirrored some of the changes that would occur later on in other taxa. Had Woodmorappe not been in such a hurry to convince people as gullible as Phillip Johnson or David Berlinski, he might have informed his readers what these traits referred to. Item 2 involved "Curvature of the contact facet of the posterior side of the dorsal plate: flat or nearly flat (0); convex (1); concave (2)." And no. 7 concerned "Dorsal margin of the dorsal plate: with a pointed dorsal process (‘dorsal angle') (0); rounded (1)."[42] Now what do all those ten-dollar words mean? Translating from Woodmorappeze into English, what he was implicitly saying about these jaw details was that he could dismiss the gentle shift in curvature of one bone and the rounding of another solely because the gorgonopsids had developed this same feature in their shared anatomy. The scale of what was going on was even more evident in one of the traits that seemed to be prone to reversal, no. 10. It related to "Articulation of the pterygoid to the medial margin of the quadrate: the quadrate ramus of the pterygoid contacting the anterior face of the medial margin of the quadrate (0); the posterior end of the quadrate ramus of the pterygoid is laterally overlapped by the medial side of the quadrate (1); no articulation (2)." As illustrated in Luo & Crompton's paper, this involved how close the end of the pterygoid bone was getting to the adjacent quadrate bone. All of the taxa shared the same anatomy (as well they ought, since they were so closely related in an evolutionary sense). And because the animals concerned were very small (mouse-sized), the actual difference between "contacting" and "overlapping" and no "articulation" at all involved the tiniest of anatomical shifts, less than a millimeter.[43] Viewed in terms of what is known about the developmental process, where the existence of these similar anatomical features implied a shared genetic blueprint -- and concomitantly a common repertoire of possible gradual variation -- Woodmorappe's cavils are preposterous. Creationists are forever complaining how "macroevolution" cannot account for the origin of new organs or body plans, but that's not what's happening here. No new bones were added in this activity, only how observable shifts in them made sense when seen as the steps in living lineages. The idea that the later forms couldn't be naturally related because of what had happened in the gorgonopsids 16 and 32 million years earlier helps us understand why there are so few creationist paleontologists. It requires them to do exactly what their stilted methodology is least suited to, letting all the evidence stand together. Woodmorappe's harping on the "abruptness" of the jumps (without paying attention to the vast time frame in between, or what features were doing the jumping) sounds even more like an arbitrary dodge when we try to imagine whether any possible data set could satisfy his finicky criteria. Imagine if Luo & Crompton's data set had looked like this:
Now every one of the characters would have been progressive -- no reversals -- and each trait change would have accumulated in a totally gradual way across that 56-million-year period. (You may notice that trait number 5 had actually done that in the original set.)[44] So how "abrupt" would this situation be?
Because were including even more character traits, the "abruptness" gets even worse! Some of the steps now shift by 4, 5 and even 6, double the 2 and 3 of the actual data set. And if we wanted to, we could have magnified this still further if the polarity convention for trait no. 4 had been flipped, and given a zero for the initial state rather than one.[45] You might think that the increments could be reduced if they were distributed across the whole range rather than progressing steadily in each trait, as done in the above chart. Apart from how arbitrary the whole idea is (as though animals could be expected to coordinate their variations, like subway riders plowing up parallel escalators) it turns out that such a smoothing process won't work. The reason is that there are simply too many trait shifts in Luo & Crompton's chart (a minimum of 27) to allocate to too few taxa (only ten). Since there can be no state changes in the initial row (because it is the base line) the best you could expect is three changes per taxon ... which is exactly the range seen in Woodmorappe's own example. The problem here reminded me of the strict rules Arnold Schoenberg used for his modernist serial musical compositions, where you couldn't use a note again until all the chromatic scale had been played through. For long pieces you're going to get a lot of repetitive chords and dissonance. Let's start ever so conservatively, and try to keep all the observed progressive features just as they were above, adjusting the "reversing" ones so that they fill out the missing "Schoenberg" slots as smoothly as possible. That actually proves not to be possible, though I managed to get close enough, needing to slightly adjust only one trait (in this case, no. 5). Here's how this list looked:
Although everything is still progressive, you still get "abrupt" jumps of 3 and 4, again inevitably as an artifact of the data set:
Now in the foregoing case the 2 state of character trait 11 was seen as the oddity, and the transition smoothed out from 1 to 0. As described by Luo & Crompton, that trait concerned "Articulation of the quadrate with quadrate ramus of the epipterygoid: absent (0); present, the ramus abutting the edge of the medial margin of the quadrate (1); present, the ramus contacting the anterior surface of the dorsal plate of the quadrate (2)."[46] If we aren't paying attention to the actual anatomical features, and think like a "design" advocate obsessing on raw numbers rather than a biological reality, we could take 2 as the end product and redesign the taxa accordingly. If we are also not so fussy about retaining the progressive traits seen to have happened in the actual animals, we could just as easily come up with a chart like this:
Still relentlessly progressive and gradual, except we have now only magnified our Schoenberg problem:
The stages again show 4 and 5 jump increments, with even one "reversal," where the Trithelodontidae are seemingly more "mammal" than the mammals! That strange circumstance was due to trait no. 4 being designated by a 1 to 0 value, which meant it had to drop a notch to "progress" to the final zero condition. That there can be curious "reversals" to the numeration even when there are no "reversals" going on should be a clue as to much elbow room Woodmorappe has to stage-manage the data, once you divorce yourself from the obligation of paying attention to what all the numbers are referring to. Just how far out on the analytical limb Woodmorappe is capable of going here may be seen in his third chart, one based on data from a 1994 anthology chapter by Zhexi Luo. Chronologically, the chart begins with Thrinaxodon 250 million years ago, but the taxa were listed according to how they related in the cladistic analysis. Thus the fairly primitive Sinoconodon appears before full mammals in a taxonomical branching sense, even though its fossil specimen happens to date 200 million years ago (about six million years after the Trithelodontidae and Morganucodon).[47] For this reason we'll leave the chronology aside to focus on the cladistic structure. Again with my own notational asides, table 3 appeared in Woodmorappe's article thus: Dental and cranial
characters
a Luo had listed Thrinaxodontidae here, not the specific taxon. b Luo had listed Probainognathidae here, not the specific taxon. c Read as Tritheledontidae. Woodmorappe's commentary on this chart (which picked up from the passage on table 2 quoted above) clearly aimed for the jugular:
Well, that settles that, doesn't it? Except all of this depended on a confluence of the bad numerical and analytical habits seen in the earlier charts coming to a ridiculous head. Once again, we have to get the categorizations out of the way first. Because this "third database is intermediate in size between the first and second," he dealt with the two columns differently: "the progressive and reversing characters are treated the same as in Table 2, whereas the Mammalness Index is computed the same as in Table 1." Woodmorappe's note 23 explained, "Owing to the fact that many character polarities were missing for Adelobasileus as well as Kuehneotheriidae, both were omitted from Table 3. However, had they been included, the trends shown in Table 3 would not have been altered to an appreciable extent." His note 27 added how "One additional character (No. 39 in Luo, Ref. 10) was rejected because its character-polarity was unknown for too many taxons. This left a total of 28 progressive characters and 53 reversing ones available for the present study." Other than these caveats, Woodmorappe didn't specify which items fell in which category, but a study of the original data indicated he had misfired on several. The only missing data in no. 39 were the two slots represented by "Adelobasileusas well as Kuehneotheriidae" (000011?111111?) -- a trait which overall looked decidedly "progressive" even without the missing blips.[48] Part 3[33] Johnson (1991, 75-78, 173). [34] For example, Barbara Stahl (1985, 410-411) flattened the idea of mammalian polyphyleticism that Johnson (1991, 77, 173) lamely sought to revive (and in the very chapter of Stahl's book that Johnson claimed to have based his argument on, to boot). Johnson also attempted to defuse the implication of the mammal jaw evolving from its reptile-like ancestors by accepting the paleontological facts, but only as "a narrow point" to be flicked aside at the first sharp turn. Cf. the astonished Gould (1992, 120). The peculiar contortion required to get from a basal amniotic jaw layout to the different mammalian one not only is seen in the fossils ... it is physically confirmed in the embryology of living mammals. As this point has been noted by Hopson (1987, 18), a source known to Phillip Johnson, either the Berkeley lawyer was too obtuse to recognize its significance ... or simply elected to ignore it. By the way, the only other major antievolutionist who has taken a whack at the reptile-mammal transition is Woodmorappe's hero, Duane Gish (1995, 150-157. But since one of the sources for Gish (1993, 163) also mentioned the jaw embryology, McGowan (1984, 139), Gish is in the same boat as Johnson over the suspicion of data suppression. See Gould (1990) for some history on the jaw transition, and Müller (1996, 129-131) or Rowe (1996) on the technical details. Kenyon (1994, 178), Pennisi (1999, 577), Shigetani et al.(2002), and Koentges & Matsuoka (2002) re Depew et al. (2002) track the progress of research into the genetics of vertebrate jaw evolution. [35] See Johnson (1991, 79; 1993, 208-209; 1997, 51-52). [36] Sidor & Hopson (1998, 272). [37] Items 82 & 84 were "Dentary masseteric fossa: absent (0), present (1)" and "Posteroventral part of dentary: confluent with lower border angular (0), thickened and angular in trough (1)," Sidor & Hopson (1998, 272). The dentary bone enlarged in the synapsids until it became the sole jawbone of modern mammals, leaving the adjacent bones to be appropriated for hearing. For example, Benton (1993, 112) notes that "the angular bone traveled to become the mammalian ectotympanic, a C-shaped ossicle that holds the eardrum taut." Cf. note [34] above on creationist distance from the anatomical details. [38] Johnson (1991, 77). The point he returned to later concerned the polyphyleticism issue, which involved his ignoring the content of Stahl (re note [34] above). [39] Luo & Crompton (1994). [40] Luo & Crompton (1994, 360). [41] Woodmorappe's note 21 specified 1, 3, 5, 8 & 14 as the progressive traits in Luo & Crompton. [42] Luo & Crompton (1994, 373). [43] See Figures 3, 5, 7 & 8 in Luo & Crompton (1994, 346, 348, 350-351), concerning Thrinaxodon, Probainognathus and Massetognathus. The overall changes to the quadrate bone are neatly shown by Luo & Crompton (1994, 366-367). [44] Item 5 involved "Lateral margin of the dorsal plate: straight (0); flaring posteriorly (1); flaring and rotated posteromedially (2)." [45] Trait 4 involved "Shape of the trochlea: cylindrical (0); trough-shaped (1)." The trochlea is a ridge on the quadrate bone; some variations are illustrated in Luo (1994, 106). [46] Luo & Crompton (1994, 374). Illustrations of the relevant skull details may be seen in Luo (1994, 103). [47] See Sidor & Hopson (1998, 259-260) for the temporal range of the taxa. [48] Luo (1994, 127) described the trait (which concerned whether certain nerves went though a single opening or were separated) as "Internal acoustic meatus: without separate foramina for cochlear and vestibular nerves (0), with separate foramina (1)." Luo et al. (1995) and Luo (2001) explore the details of the shifting therapsid-mammal ear. 
|
|
