Sunday, April 15, 2007

Non-archaeological book review
In this month's American Scientist. It's a review of two books on the epidemiologist John Snow, famous for what many have called the first scientific epidemiological study of disease transmission:
. . .Snow focused on a sudden eruption of cholera within a single densely populated neighborhood. He showed that use of water from the Broad Street pump was a common factor in almost all of the cholera deaths and also that nonuse of that water was a characteristic of two groups (workhouse residents and brewery workers) that suffered little from the disease. In likening the behavior of the apparent cholera agent to a living thing, Snow is often listed as a pioneer of the germ theory.


The review is about two books on Snow, which the reviewer thinks tend to gloss over the complexities of the scientific milieu in which Snow operated and how his views eventually came to be accepted:
The form is as follows: a protagonist, an outsider representing truth and virtue (qualities that are linked through some unexplained dynamic of reciprocity), takes on entrenched intolerance.


I've blogged on this scenario before, what I usually call the Hollywood view of science. I admit I have something of a negative bias against these sorts of explanations of How Science Really Works, probably from my early days of being a young, green scientist reading S.J. Gould with rapt attention (note: when in academia, one learns rather quickly that quoting Gould as some sort of scripture doesn't impress too many actual academics). Gould wrote on this a lot, mostly taking on the Hollywood stereotype, which he generally ascribed to some sort of textbook phenomenon that he gave a name to, but which I can't remember right now (how's that for a Dickensian sentence!). He would trace certain textbook presentations of how a scientific revolution came about, going farther back into the literature and showing how and when the story first developed and was then passed on uncritically throughout many editions and authors until the story became more legend than anything else. So we end up with a story of the bold truth-seeking revolutionary scientist going up against and eventually prevailing over the establishment thinkers of the day -- mired in their old-fashioned ideas about how the world ought to work -- while the hero becomes the beacon of yet another new age of reason.

One wonders how this script came about, and whether it's a later last-couple-of-centuries Western concept or somethng basic to human nature. I tend to think the latter since we have this kind of thing all over the literature. Prometheus bringing fire and wisdom to mankind -- and being rather brutally punished for it -- springs immediately to mind. Come to think of it, there were a couple of Star Trek episodes devoted to something of a debunking of the mythology of a man. They both had to do with Zefram Cochrane who invented the "warp drive". In the original series, they found him marooned on some planet and tried to convince him to come back to civilization, but he refused after hearing what sort of a hero society had made out of him. The second was one of the Next Generation movies where they went back in time and found him to be a hard-drinking womanizer rather than the brilliant and upstanding scientist everyone on the crew thought he was.

Anyway, back from sidetrack:
The failure to explain how Snow is relevant to us reflects a broader cognitive failure, jointly of historical analysis and the representation of epidemiological reasoning. The chief historical fallacy is presentism. Retrospectively, the story is so simple: good versus bad, truth versus error. Our post-Koch conviction that Snow was on the right track makes it seem as if his arguments should have been enough for his contemporaries too, had they only been honest. Both authors struggle to label those who disagreed with Snow.

. . .

They characterize the therapeutics of Snow's age as errant quackery uninformed by experience or theory. Some of it was. But medical theories were rational (if, in retrospect, partial or erroneous); doctors developed, and shared experiences; and, for many, their commitment to patients or to science cost them their lives. Snow was not the only hero.


This "presentism" is very common in popular histories of science, taking what we know now and projecting that back on Snow's contemporaries. How could they not see the truth of Snow's ways and the error of their own? Why, it's so obvious! They must have had some selfish and sinister reason for refusing to listen to reason. But as Hamlin argues, and Gould did this remarkably well also, they had reason for believing the way they did. As Kuhn has argued, existing paradigms are in place for a reason: they seem to work. This also gave them reason to question Snow's methods, some in terms of their own paradigm, others that are stll true today. For example, Hamlin notes that even in our own day, Snow's hypothesis had some serious methodological difficulties:

First, many, even most, who presumably drank bad water did not get the disease. . . .Second, since the drinking of bad water had long preceded the epidemic, it could not be its cause.


So, not quite the slam dunk that we are led to believe.

As I said, I'm more or less biased in this regard, and though I have some experience in clinical trials and epidemiology, I'm no expert, so take my comments and the reviewer's with appropriate doses of NaCl. But the article at the link is free (as are several others in the current issue) so read the whole thing yourselves.

UPDATE: Also part of my bias, and I think some explanation for it: In our first-year theory classes with RC Dunnell, we spent a lot of time on the New Archaeology and its contrast with the Old Archaeology, the culture historians. Dunnell made a point of defending the culture historians against the charges of the new archaeologists that the former were atheoretical and "unscientific" (which they, of course, were, explicitly). There's much to be said for those "old archeologists" who developed a number of methods, notably seriation, that were highly quantitative and methodologically rigorous, if developed more or less inductively rather than through deduction like we usually decribe the scientific method (though I suppose one could do a whole blog on THAT).