The Loom

Text Versus Subtext

(Warning: this post contains some journalistic/blogging inside-baseball material.)

Back in the dark ages (otherwise known as the 1990s), writing about science felt a bit like putting messages in a bottle. I’d write an article, a few weeks or months later it would appear in a magazine, and a few weeks or months later I might get a response from a reader. In some cases, an expert might point out an error I made. In other cases, she or he might explain the real story which I had missed. The delay could make for some disconcerting experiences. The first time I met the late Stephen Jay Gould, to interview him for a book I was working on, I was still lowering myself into a chair when he began complaining about the cover headline to a story I had written about fossil birds over a year beforehand. I stared at him blankly for a while as I reached back into my memory banks to figure out what he was talking about.

It’s much better these days, now that people can hammer me with emails seconds after my stories is are published. Science is a murky, complex endeavor, and my job has never stopped feeling like an apprenticehip, as I learn from mistakes.

But this new arrangement comes with a downside. Some criticisms are unjustified, and instead of simply emailing me these complaints, people sometimes decide to publish them for all to see.

John Hawks, an anthropologist at the University of Wisconsin, has done just this. He has written a long complaint about an article I wrote for the latest issue of Discover. The issue celebrates the 25th anniversary of the magazine, and it contains a series of two-page spreads that take a look at different fields in science and where they’re headed. The editors asked me to contribute a piece on human evolution. I included an interview with Tim White of Berkeley, an essay on the growing role of scanning in studies on hominid fossils, and a large graphic showing how scientists used CT-scans to reconstruct the skull of Sahelanthropus, the oldest known skull of a hominid.

Hawks makes a series of complaints about the piece, but rather than sticking to the article itself, he tends to focus on the "subtext," which he alone has the mysterious power to read. For example, the subtext apparently says that "anything high tech must be better." I never made such a claim, and it would have been silly for me to add a disclaimer to that effect: "Warning–not all things high tech are better." Healthy skepticism is certainly a virtue, but Hawks is ignoring the fact that the entire issue is dedicated to promising new scientific developments. (Here’s an article on research on using lasers in art conservation. I suppose Hawks would complain that the article didn’t mention that lasers can also kill people.)

When Hawks does actually deal with the article itself, he makes some serious mistakes. He mocks the conclusion of my piece, in which I describe some new applications of fossil scans–such as reconstructing wounds, simulating hominids walking, and making the scans available online to other researchers who can’t see the originals. "So utopian," he sneers.

As evidence, he turns to my interview with Tim White, in which White talked about the importance of other kinds of technology to the study of human evolution–such as the global positioning system, advances in dating fossils. "No CT scans there," Hawks declares.

Hawks shouldn’t argue from the absence of evidence. Actually, White talked to me at length about the promise of CT scans, including some of the applications I mentioned in the article. It would have been redundant to include his comments. Hawks may not be impressed by scans, but he shouldn’t count White on his side.

So why isn’t Hawks impressed with scans? For one thing, scientists can make mistakes with them, producing recontructions as biased as any handmade reconstruction. "The principle of ‘garbage in, garbage out’ is everlasting," he says.

True, but so what? I remember the same argument being made in the 1990s, when some biologists were starting to reconstruct the tree of life by using computers to analyze DNA sequences and morphological features, rather than relying on a more intuitive sense of what evolved from what (a method known as cladistics). Critics warned that the cladists were just dumping bad data into their computers, and so their conclusions couldn’t be trusted. In fact, the cladists were producing testable hypotheses with explicit assumptions that anyone could challenge. Of course there are cases in which this approach may face problems (in comparing populations of the same species, for example, or species that can swap genes, for example). But that hasn’t stopped cladistic trees from becoming the standard for the field. The garbage-in-garbage-out complaint is equally beside the point when it comes to predicting the importance of scans to the study of human evolution.

Descending again into my subtext, Hawks writes that "a read of the article gives the impression that every finding from this new advanced technology supports splitting hominids into several species." If I may indulge in a little subtext-divining myself, I think we’re getting somewhere now. Hawks is a long-time proponent of the idea that too many hominid fossils have been designated as separate species. It just so happens that a couple of recently published scans–one of Neanderthal children and one of the "Hobbit" brain–have been interpreted by the authors of these studies as supporting the idea that these fossils do not belong to humans, but to other species. But instead of directing his wrath at these scientists, Hawks directs it at me. In order to do so, however, he has to ignore the fact that I write about many other applications of scans that don’t support splitting hominids into several species.

Hawks is perfectly entitled to attack hominid-splitting (and on his blog he has done a great job of documenting new research that supports his attack). But I don’t appreciate him distorting my own writing to serve that agenda. It’s particularly unfair to do so when most people haven’t had a chance to read the article for themselves, and have to rely instead on Hawks’s misleading summary.

Update, Wed. 1pm: Another improvement on the dark ages: when I attack, the attackee can respond. John Hawks defends his post in the comments. I agree that CT scans of hominid fossils are not now being freely shared on the net. But I think there’s reason to be optimistic–see, for example, the Digimorph Project, which is building up a big database of scans of bones from living and fossil animals. Would it have been utopian to predict Digimorph a decade ago?

Comments

  1. #1 cats
    September 13, 2005

    Well, this is something almost every science writer may encounter. Science writing is to communicate information at best efforts. And this may inevitably ddissatisfy somebody, for it’s not a scientific paper (but even a paper can not be perfect).

    As for Hawks’ case, he may feel that you didn’t include his opinion. (yes, his “no CT scan here” is really ludicrous.) But it’s unrealistic to quote him every time.

    Anyway, cheers up, Carl.

  2. #2 Robert Karls Stonjek
    September 14, 2005

    …and they can hassle you with minor details and nit picking, like the grammar error in your second paragraph
    “after my stories is are published.”

    …not that I’d do anything like that :)

    Actually, being dyslexic has an advantage – as I don’t trust my own ability to spell I always leave it to the computer to check, and as I may also read over grammar errors I always get the text-to-speech engine to read it back to me. The surprising upshot is that this system is so effective that I find errors in other people’s work.

    We have a remarkable ability to error-correct as we read (but dyslexics don’t, so reading errors actually become compounded) so that we tend to correct our own errors and so don’t ‘see’ them. But in hearing a piece read back by another (person or computer, though a person may read over errors corrected on the fly as well) one tends to readily pick out the errors, as with the above example.

    Kind Regards
    Robert Karl Stonjek

  3. #3 John Hawks
    September 14, 2005

    Hi, Carl!

    I’m not sneering at you. Really. I’m a big fan. And financial contributor, since I have all your books (even At the Water’s Edge), and I frequently link over here to your posts!

    And no, I’m definitely not on the list of people to interview about CT scanning, although for Neandertals and genetics, I’m your guy.

    Of course, if I could *get* CT data, I would probably use it. That’s the “utopian” part. Everybody doing this work says how great it is that everyone else will have access to their data. But none of this data is available. Data from only seven fossils (all out of the ground for more than thirty years, I might add) are available for *purchase*, none are available to the public.

    Perhaps new data access guidelines at NSF will have an effect. But at the moment, consider the Sahelanthropus reconstruction that your article features. The CT reconstruction shows it likely had a vertical posture.

    But how many skeptics were allowed to examine the CT model? I happen to know, because I am one, the answer is zero. Yet, the simple *fact* of the model simply ends debate on the issue, even though nobody else can independently replicate it. If I wanted to write up the features of the skull that are *inconsistent* with vertical posture, I can’t — because someone else can keep these data away from me.

    So, I have grounds to complain about CT — from my perspective it has made the science worse. But do I have grounds to complain about your article?

    Well, reading over my post, I don’t think I said anything that bad about it. Surely it is true that your *text* says that not everything high-tech is good. But everything else that you wrote about high-tech said it *is* good. That’s the definition of a *subtext*. I don’t think I’m the only one who would find that subtext there: as you say, Discover was running articles that month about how great new technology is going to be for science. In other words, it was an *intended* subtext. It’s not my fault the editors wanted articles trumpeting tech changes, and it’s not your fault you wrote one.

    But as a scientist writing about my science, I have to give my own opinion (that’s why I’m a blogger and not a journalist), and I don’t think it was unfair to point out the agenda. I have done so with Discover articles before, and I’m a subscriber so I have some right (there’s that financial contribution again!).

    I do mention the theme of the issue right in the first paragraph of my post. The article doesn’t give any information at all about why anyone would think CT scanning isn’t the cat’s pajamas; I give positives and negatives. And what’s the use of having a blog if I can’t be snarky once in a while? The idea that everyone will be happy when they have CT scans at last *is* utopian; the printed text of Tim White’s interview *doesn’t* mention CT scans. Perhaps it’s sin by omission, but whose fault is that? Most of the other interviews in the issue (this you didn’t mention) *did* focus on just the same topic as the article they accompanied.

    So, just to sum up: I am still a fan of Carl Zimmer. I do not wish to rack up quotes in future Carl Zimmer articles. I do not think my post was unfair; it is more balanced on the positives and negatives of CT scanning than the original article. And my complaint is not that the CT interpreters are splitters and I’m a lumper; it is that the CT interpreters are acting like high priests of technology who don’t deign to allow their work to be replicated independently.

    I think that a survey of people in the field would show this point of view is widely shared.

  4. #4 Artedi
    September 15, 2005

    Check my lonely spanish blog, please

  5. #5 Sandra in Dallas
    September 16, 2005

    In his Sept 14 comment above Hawks says CT scans have made science worse. However, what he really is decrying is the lack of public access to the scans. There is a huge difference and it puzzles me that he is beating the wrong drum, and so obviously.

    In any case, I am always eagerly awaiting the next Carl Zimmer book.

  6. #6 Martin Brazeau
    September 18, 2005

    I’m currently working in a lab that is undertaking CT scans of various fossil lobe-finned fishes and early tetrapods. We are getting excellent results with the technique. I find this “garbage in, garbage out” comment to be rather peculiar. While I see how a person could “fake up” their three dimensional model, it would take a lot of work to make corresponding data appear in the scan slices. Besides, if somebody questioned your work, it would simply be a matter of scanning the specimen again (though it might cost them a pretty penny). The results are reliable, repeatable, and checkable.

    This method is becoming a major component of my Ph.D. thesis. There are aspects of morphology (such as the presence of positional relationships of certain bones) that one’s bias cannot easily interpolate into the modeling process. The information obtained, however, gives us very important clues into the morphological evolution and functional morphology of these animals. I think that eschewing this kind of promising technology is scientifically irresponsible. New technology might not be perfect, but ignoring the virtues it does offer does not help science.

    Compare CT with the serial grinding methods of yesteryear (of pre-Dark Ages (sensu Zimmer) time): fossils were actually ground up, one slice at a time, each slice was drawn, blown up, and traced into a sheet of wax. The result, was a larger-than-life wax model of you fossil, and no original fossil left over. This method offers higher resolution than CT, but comes with its own set of problems!

  7. #7 Scott Maxwell
    October 4, 2005

    Hmmmmm……do I detect a faint whiff of the old
    Chris Stringer vs Milton Wolphol(sp) controversy ???
    God I hope so. The field has been very bland since those two went head to head. Mr Hawks is obviously in the Milton camp. Are you a closet
    Stringer fan Carl ? If so, let the barbs and personal attacks begin. I will buy tickets for a front row seat.

  8. #8 Scott Maxwell
    October 4, 2005

    Hmmmmm……do I detect a faint whiff of the old
    Chris Stringer vs Milton Wolphol(sp) controversy ???
    God I hope so. The field has been very bland since those two went head to head. Mr Hawks is obviously in the Milton camp. Are you a closet
    Stringer fan Carl ? If so, let the barbs and personal attacks begin. I will buy tickets for a front row seat.