|
|
|
|
|
|
|
HELP | |
|
Author InterviewsStephen Jay Gould, From Brachiopods to BaseballDoug Brown, Powells.com
Stephen Jay Gould: It wasn't broad philosophical or political issues as I think many people assume. It really comes right out of an operational dilemma in paleontology. I had been trained, as Niles Eldredge had, in statistical methods for the study of subtle changes in evolution. Evolution at that time was defined as gradualism. The two were virtually equated; to see evolution meant finding gradualistic sequences, but every paleontologist knew that they had effectively never been found, and that was a frustration. As it turns out, the naive assumption behind this kind of statistical training was: maybe gradualism could be subtle, and the only way to see it is to do precise measurements because the eye can't pick up changes at this slowness. Indeed one of our thesis advisors, John Imbrie, who was probably the best statistically trained paleontologist, learned his biometrics for exactly that reason. He went out and did a dissertation on the Devonian strata of the Traverse group in Michigan, which is an unusually well preserved sequence. He studied some thirty brachiopod species and found that all but one of them had stayed stable. He ended up leaving paleontology, which is the irony; he didn't see what I think punctuated equilibrium saw - namely, that maybe you ought to revise your assumptions. Instead of recognizing stability as a frustrating negativity, consider it a signal about the main weight of the fossil record itself. So that's what it came out of: the hope that statistical study would finally enable us to get around this source of frustration by revealing gradual, definitive change; but that didn't work either. At some point, one had to face a radical alternative: maybe what you were seeing wasn't its imperfections; maybe it was in fact the signal about how evolution truly did work. And that at least had the salutary effect that now the main thing you saw, in a literal sense, was no longer the imperfection that blocked your access to the phenomenon you wanted to study, but was actually the proper expression of geological time and the phenomenon itself. The idea of punctuated equilibrium came from a very conventional motivation based on an operational frustration in the actual practice of empirical work. Doug: Do you think macro- and microevolution are different processes, or the same thing at different scales? Gould: Neither. The book I've been working on for ages - the galley's going to come out in February, The Structure of Evolutionary Theory, Harvard University Press, probably about 1300 pages - the main argument I make in there is that macroevolution is clearly not just the accumulative minutiae of ordinary microevolution. But it's not a separate process either; it's just that once you expand your purview into this very different scaled-up realm of geological time there are other kinds of phenomena that become important. The working out of microevolutionary phenomenon isn't the simple extrapolative mode. For example, in strict Darwinism, you're supposed to take the anagenetic changes - that is, the cumulative changes within a population - and just extend them. But punctuated equilibrium shows that for the most part those changes aren't even occurring; species are stable throughout their geological duration. Change occurs in geologically rapid, if ecologically glacially slow, episodes of speciation. Therefore evolutionary trends are reconceptualized as the differential success of certain kinds of stable species. The species as a stable entity becomes an atom of macroevolution in the same sense that the organism as a genetically stable entity through its lifetime is an atom of microevolution. So it's not as though there are stunningly new mechanisms of a genetic sort that you don't see at other scales; it's just that ordinary processes don't simply extrapolate. Doug: For living species, we have Ernst Mayr's reproductive isolation definition. How do you define a paleospecies? Gould: Probably the best inferences you can make to that same fundamental definition. After all, most living species aren't defined except in some theoretical sense by reproductive isolation, but in fact they are recognized morphologically. The vast majority of modern species have never had breeding experiments done [laughs]; they're just clearly distinct. If you find that the bugs in your backyard fall into definitely discrete packages, you assume that they remain separate in nature. I think there are enough inferential criteria. Clearly it is a frustration and sometimes you can't always tell that the entity is defined by its reproductive actions, which can't be directly tested in the fossil record; but they can be inferred with reasonable confidence. For example, suppose you're trying to determine the formation of something that looks morphologically different, different enough that you would on purely morphological grounds call it a species; and you want to know whether it's really separate, and whether punctuated equilibrium applies. What you can often do, and it works in the majority of cases, is to see whether the ancestor survived or not. As long as the ancestor survived and lives coextensively with its descendant, you know you have two entities that are not amalgamating. If the ancestor simply disappears, and then you see the descendant, you simply don't know. It could be punctuated equilibrium or the extinction of the ancestor. Or it could be a geologically rapid transition of the whole population, which is not punctuated equilibrium; it's still rapid transition. You're not going to be able to answer it in those cases. Doug: Do you think artificial life models like John Conway's "Game of Life" are useful for describing evolution, or are they too simplistic? Gould: It depends what you think they are doing. There are two different rationales utilized by people who do this kind of work. One rationale hopes by simplifying what in nature would be a horrendously complex, unresolvable situation, that you really can create an experimental situation. After all, that's what the experimental method is all about. You take things into a controlled laboratory setting so that you can reduce the number of causal variables. The other rationale, which is philosophically quite different - it's probably closer to the utility of this stuff - is people who say, "Look, I'm not saying this is a simulation of life as we actually know it; this is an artificial system in which we have the ability to test in an abstract sense what you can't in the world. We know the world doesn't work this way, but at least it lets us answer the question, 'What happens if I only vary...' I have an absolutely determinate system, because I made it, and it's running according to its rules, but we'd really like to know what happens if I only vary size, what happens if I only vary generation time, what happens?" In that sense I think it's been very valuable - as a set of abstractions, not as a claim to mimic what life is actually about. Doug: Folks like Stuart Kauffman, who also uses a lot of computer modeling, argue that life is an expected property of any planet that meets the minimum requirements. Do you agree, or do you feel that chance plays more of a role? Gould: What he's talking about is the basic biochemistry, the simplest forms of living systems. It's probably right. There's no proof of it, but the relevant paleontological observation runs contrary to what used to be the standard view; namely that life is very difficult, and the only reason it arose on Earth was that there was enough time to convert enormously improbable occurrences into expectations. But we now know that the oldest rocks that could contain life do, which does lead you to reverse that perspective. Under the right conditions, I suspect that the origin of something at protobacterial grade is an inherent property of biochemistry. But I think what's most interesting about Earthly life is that other things happened [laughs]. I think that's where you enter the realm of contingent pathways. Doug: Ward and Brownlee argued in Rare Earth that complex life is actually very uncommon. Gould: It probably is. In fact I had breakfast with Peter Ward today. He said the reason why he called the book Rare Earth - not Unique Earth, but Rare Earth - is that he's not fighting the notion that there's intelligent life elsewhere. But it's going to be very rare, even on that subset of planets that develop life at all, and that's probably right. Doug: Do you think birds are theropod dinosaurs, or did they arise from? Gould: That's really a linguistic rather than a biological point, and I think that it's much less theoretically interesting than has often been given out to the public. There's a popular misconception that one of the great and revolutionary discoveries of paleontology is that dinosaurs didn't die because birds are still in the trees. That's almost like saying cephalaspid fishes, or ostracoderms, didn't die because I'm talking to you now [laughs]. That's scarcely the point. It's not what in a vernacular sense we mean when we say that dinosaurs are dead. We mean that there aren't big things like Tyrannosaurus and Brontosaurus around any more, and there aren't. The fact that a very different lineage of small running dinosaurs evolved to birds doesn't mean that dinosaurs are still around. I think it's pretty clear that birds evolved from one small lineage [of dinosaurs]. But if you have this false transformational view, that apes changed into humans, then dinosaurs changed into birds and they're still around. But that's not the way to look at it. There's a big tree of dinosaurs. One little branchlet of that tree containing creatures that aren't very dinosaurian by our vernacular icon of the thing - which is something very big - well, one lineage evolved, and some of its members became birds, but that doesn't mean dinosaurs are still around. It means that the whole clade of dinosaurs didn't die out, as we used to think, but that one little lineage became birds. That's very interesting, and we want to know that, but a sparrow is not a Tyrannosaurus [laughs]. It just means that the clade of dinosaurs didn't become entirely extinct. One small part of it became a very different kind of creature that has become very successful. In fact, there are probably more genera of birds now than there were ever genera of dinosaurs. So in that sense the clade persists at high species diversity; but still, that's why I mean ostracoderm fishes don't exist just because mammals are around. There are more mammalian species than there ever were ostracoderm fishes. Doug: Are you still agnostic about cladistics? In Dinosaur in a Haystack, you used that phrase. Gould: It's not a question of agnosticism. It's good to have cladograms, because cladograms offer the best hypotheses about actual branching orders, based on inferences of characteristics in common. The issue is whether you ought to, in some rule-bound or automatic sense, go from the cladogram to classification. That's very important. There are good reasons why you might not want to do that, including these kind of misunderstandings about dinosaurs and birds [laughs]. In fact, cladists will say, "I'm defining by branching order, and by branching patterns the system of dinosaurs never disappeared." But dinosaur as a morphology did. Doug: On a related note, I was wondering if you knew that Sue, as in Tyrannosaurus Sue, is here in town at our science museum? Gould: [Surprised] They let it leave Chicago? Doug: Yeah, believe it or not. I think it's touring the country. Gould: I guess its monetary value was too great [laughs]. Doug: What did you think of the trial and the impounding of the fossil? Gould: I don't want to sound cynical or anything, but you know the Tyrannosaurus in New York and Pittsburgh are really pretty darn good. So-called Sue might be somewhat more complete, but the notion that's so hyped in popular sources - because dinomania is so pervasive - that somehow Sue is a great scientific or revolutionary advance, is complete nonsense. It's just a better specimen of a creature we happen to be quite interested in. That I understand perfectly well, but the notion that somehow the purchase of this thing was enormously important in some theoretical sense for the development of science is nonsense. It's a cultural icon, more than anything else. It's a lovely specimen, I'm glad it exists, but I think it's scary that the world of professional paleontology has become more like the world of art curators. I'd hate to be an art curator because it's so caught up in high commerce. If you're an art curator you spend perhaps half your time going to fancy cocktail parties talking to people you have no interest in, in the hopes that they'll give you their money someday [laughs]. That's not what I want most professional paleontologist curators of great public museums to be doing. It's a capitalist world, and if that's what the market will bear, then that's what it's going to bear. But I think it's very unfortunate. Doug: What do you think of the general concept of private collecting and ownership of fossils? Do you think they should be in museum collections? Gould: As long as we have the economic system we have, I don't see how you can ban private collecting. It would be like banning private ownership of great paintings and books. I think the danger is when these things become so monetarily valuable that private collectors see them mostly in terms of their worth and their profitability. That's when you can get distortions. That's when people are tempted to put them in bank vaults, or sell them to the highest bidder, where they'll go out of public hands; or out even of the possibility of scientific study. That's a terrible thing. In the past this really hasn't been a problem because fossils did not have high monetary value. Most private collectors were in it out of personal passion. And since these things didn't have enormous monetary value, most private collectors were very happy to have scientists study them and look at them. They weren't seen primarily as sources of profit for a family, or a firm, or anything. The private collectors who did it were passionate about what they did, and were for the most part cooperative with scientists. There's a long tradition of cooperation. That is very much threatened when the worth becomes so artificially high because of supply and demand, classic issues of capitalist economics. That's a great shame, I think. Doug: Are you still teaching at Harvard, or are you just doing research and writing? Gould: They don't pay you if you don't teach, that's for sure [laughs]. Doug: What courses are you teaching? Gould: Same as always. I've taught this course on the history of Earth and life pretty much since I started; of course the content changes enormously. Doug: Do you still do your writing on the old Smith-Corona? Gould: I still use a typewriter. Doug: Do you use email at all? Gould: Hell, no. It's not just idiosyncrasy either. There's definitely method in what is my madness of avoiding email use. I'd be getting hundreds of email messages a day I really don't want to read. Life is very short, and the struggle that I have is not to be in touch with people; it's not to be in touch with people, so that I can have time to do my own work. And some of the trivialities...before there was email, you wouldn't bother to express them or you would wait for the couple of times a year you see somebody. I don't need to express every thought I have to all my colleagues, and engage them in colloquy [laughs]. I need to be able to when I need to; but for that the US mail is still running a pretty efficient service, and there's telephones, and there's plenty of other ways to be in touch. Doug: What do you like to read? Gould: Depends on where I am. Mysteries on airplanes is different from scientific literature. Doug: Are there any books you've read more than once? Gould: The Origin of Species [laughs]. Doug: What are some of the big questions you would like to see answers to? Gould: [After a moment's thought] I think I'll pass on that one, just because it's too long to get into it. Doug: What's your favorite critter? Gould: Well, certain things are definitely overexposed, of which dinosaurs are a prime example. Dinosaurs aren't enormously interesting, they weren't a dominant group among vertebrates for a long time... Doug: And invertebrates are a much larger part of the fossil record. Gould: The fossil record is ninety or more percent invertebrates. I like snails because that's what I work on, and I like Cerion because that's where my expertise is. I'm mostly interested in creatures that can teach us things about how evolution works. And dinosaurs aren't, for the most part, great at that. As I said earlier, I was trained in statistical methods, and you can't do a lot of fine scale statistical analysis on dinosaurs because sample sizes aren't good enough; the stratigraphic resolution isn't fine enough. Before all this interest in Tyrannosaurus, for example, all our knowledge of Tyrannosaurus was based on two skeletons: namely Pittsburgh and New York, and fragments and pieces. You can't do much of a statistical analysis on a sample of two [laughs]. Now there are quite a few more, but still it's not like collecting snails where you can get hundreds for each stratigraphic level of the sequence. With that you can really do some work. Doug: In Full House you talked about how many baseball statistics are lower on average today than they used to be because the players are actually improving. Gould: The range of variation has been lowered. With hitting, I'm not saying hitting has been lowered; I'm saying that the average batting average has always been around .260, and has been converged upon, so the variation on both sides of the unchanging mean has been declining through time. That standardization is an expression of increasing excellence of play. Doug: Do you think there are any longstanding records that will never be broken because play has gotten better? Gould: There are certain ones that won't be broken, but they're uninteresting. They're fairly obvious. Nobody's going to match Cy Young's record of winning 511 games because nobody pitches that much any more and people aren't expected to finish games. Cy Young was pitching every third or fourth game, and he was expected to finish, and therefore get the decision in each one. That record will not be broken, but I don't know whether that reflects anything interesting about changes in styles or excellence of baseball. It probably reflects the sociology of baseball. It's played differently. I think the unbreakable records are mostly the ones that flow from changes in habits of play. The records that are based on prowess will probably be all someday be broken. I'm not sure if Ty Cobb's lifetime batting average will be broken, but that's part of the decrease in the range of variation. Doug: Have you been following Bonds? Do you think he's going to make it? Gould: Well, I didn't think he was going to. The thing about Barry Bonds that's interesting is that he's a very streaky hitter, and I was pretty confident that he wouldn't come close because it's very characteristic of his performance to have hot streaks and very cold streaks. This year he didn't have any extended cold streaks. He's always had the capacity to break that record, if he could be consistent throughout the season. I don't say that critically. There are just some players who are very consistent and some who are streaky. I don't think it's because they're not trying. He's got four more games, so it looks like he's got a very good shot if someone will throw him strikes [laughs]. But nobody wants to go down in history as the person who gave up number 70 and 71. He's not going to see that many strikes. Ah,
the clarity of hindsight. Barry Bonds did in fact break baseball's
record for most home runs in a season, finishing with not 71 or
even 72 (either of which would have been plenty impressive from
a guy who hadn't previously hit more than 49 in one season) but
73. Rest assured, Gould's position among the Harvard faculty remains
secure despite the failure of his statistical analysis on this
matter.
|
|
| |
|
|
Don't Miss
It's hard to have a discussion
of popular natural history and evolution literature without mentioning
Stephen Jay Gould. The Alexander Agassiz Professor of Zoology
and Geology at Harvard, his essays have been collected in a number
of books, including 