Politics, Taxonomies and Censorship

Popline is the world’s biggest database on reproductive health, with about a third of a million articles. It’s funded by federal agency USAID, and managed by the Johns Hopkins School of Public Health. If you do a search in its database today under “abortion” you’ll find over 26,000 articles. Between February and yesterday, you wouldn’t have found any articles. What happened?

Back in February, staff at USAID (which has a reputation for enforcing the conservative anti-abortion views of the Bush administration) contacted Popline administrators to express concern about two (out of 26,000) articles they had found on the database which were about abortion advocacy. Popline reviewed the articles, decided they didn’t fit with the database’s collection policy, and pulled them. Okay, a slightly questionable decision in response to a heavy lean from your purse-holder, but what the hell, it’s only two articles.

But it seems the administrators didn’t want to be caught out like that again. So – one can only presume in the interests of ensuring future peace of mind – they took a decision of their own, to make “abortion” a stop-word. A stop word is a word that a search engine decides doesn’t exist. They were introduced to help search engines ignore non-meaningful terms like “and”, “the”, “of”. Now, it seems they can be usefully deployed to make whole concepts disappear. Well, not technically – because as database administrators helpfully pointed out in response to concerned enquiries from medical librarians, you could still search under other terms such as ‘unwanted pregnancy’ and ‘fertility control post conception’. Yeah right. As if those terms spring readily to mind when you’re looking for medical and social research on abortion. They had to be IT folks – I cannot imagine a more autistic response either to the initial problem or to the concerns subsequently raised by librarians. If they were library or information professionals, then they deserve to be cast out into the utter wilderness.

I don’t for a moment believe they were especially enthusiastic about pursuing the USAID agenda. This was not about being against abortion by removing all references to it (those 26,000 items – less two – are still there, remember?) This was about making sure they weren’t caught out again if pro-abortion articles snuck back in. Nobody – especially USAID – would be able to find them. The act was beautiful in its logic, and exhibits that curious mental separation one often encounters in bureaucrats and especially in IT professionals, between the action and the result.

It’s the same phenomenon described by Hannah Arendt in her study of Adolf Eichmann’s’ “banal” obsession with getting technical efficiency into the extermination process as if completely oblivious to the human impact of what he was doing. It would be invidious to suggest that the case is comparable in impact, but at root it’s the same blindness to consequences when an ethical responsibility is quite voluntarily subordinated to the convenience and attraction of a neat technical solution. “Hey, no censorship, the articles are (almost) all still there, it’s just a nasty inconvenient word we’ve removed”.

It was also a profoundly stupid move. Who could not notice the disappearance of results on a search term “abortion”? In fact, the library and research community took some time to react. It was only at the end of March that medical librarian listserves started discussing the mystery. One of them shared how one of their researchers had written to Popline to ask about the mysterious disappearance and got the following blithe, deeply naive reply, indicating the depth and breadth of the gap between administrators and the community they served:

“Yes we did make a change in POPLINE. We recently made all abortion terms stop words. As a federally funded project, we decided this was best for now. In addition to the terms you’re already using, you could try using ‘Fertility Control, Postconception’. This is the broader term to our ‘Abortion’ terms and most records have both in the keyword fields. Also, adding ‘unwanted w2 pregnancy’ in place of aborti*. We have a keyword Pregnancy, Unwanted and there are 2517 records with aborti* & unwanted w2 pregnancy.”

Understandably the library community erupted. By yesterday the New York Times was covering the story, and reported that the Dean of the Public Health School had ordered the database folks to reintroduce “abortion” into the English language, and was setting up an inquiry into how such a decision had been taken (which is how I know today that 26,383 articles had been “disappeared” in this way).

There are two sub-stories here about taxonomies as instruments of knowledge organisation. The first, most obvious one is picked up by Jens-Erik Mai at the University of Toronto, and very neatly summarised by one of his former students “Susan Beeswax”:

“Taxonomy is powerful in that it can make entire categories of information totally invisible, either unintentionally, or intentionally. The best taxonomies are largely invisible — that’s a sign of success (for some definition of success). When you notice the taxonomy, it’s because there’s a problem. Something is missing, something is categorized in the wrong place. Missing things & things in wrong places doesn’t on the surface seem like necessarily an awful thing, but because taxonomy — whether implicit or explicit — reflects world views, if that world view doesn’t recognize, say, abortion, or if it misses an important step in an escalation process in human services (child abuse, domestic violence, education, human resources, whatever), people and other assorted things people tend to think are important start falling through cracks”

It’s exactly this issue I discuss in my book, particularly in the case of Victoria Climbie where somebody did die horrifically because they slipped through the taxonomic cracks.

But the second, less obvious issue is about the governance of taxonomies. Pending the promised inquiry by Johns Hopkins, this incident looks and feels like an arbitrary but far-reaching decision by staff who had absolutely no idea about the ramifications of what they were doing. Taxonomies are powerful, sophisticated and political instruments. Shouldn’t they then be professionally run by appropriately skilled people working to codes of ethics and proper governance? When did taxonomy management get delivered into the lap of autistic IT administrators (or if I’m reading it wrong, librarians acting like autistic IT administrators)? How many more Poplines are out there?

1 Comment so far

Susan Presley

Wow—if I’d know someone other than me was going to read it, I’d have probably worked through that piece a little more carefully! smile

Posted on April 20, 2008 at 01:12 PM | Comment permalink

Page 1 of 1 pages

Commenting is not available in this weblog entry.

Comment Guidelines: Basic XHTML is allowed (<strong>, <em>, <a>) Line breaks and paragraphs are automatically generated. URLs are automatically converted into links.