Thread by @swarthmoreburke

A relatively long time ago I wrote a piece about how the Library of Congress subject headings were so broken that they should just be discarded and we should start over. swarthmore.edu/SocSci/tburke1…
1/

That got a bit of attention back then, partly because I was ruder than I had any right to be in the original piece. To some extent, support for expert authority within libraries and catalogs is now so far in the rear-view mirror of institutions that it's a moot issue. 2/

But, you know, we still have those subject headings, and increasingly I feel as if most of my heuristics for keyword searching in discovery-driven searches for fields I don't know well are suffering from diminishing returns. 3/

So I sometimes try the old subject headings just to see what happens. Well, in our local catalog here at @swarthmore, especially not much, because it's not a huge research collection. But even in the biggest search spaces, often not much. 4/

I want to give an example of how badly the headings still function. I had a student today come to talk about doing a thesis next year on Chinese immigrants in Mexico (my colleague who works on Latin American history is on leave). 5/

She mentions Jason Oliver Chang's book Chino, which I think sounds amazing, and describes it some to me. history.uconn.edu/faculty-by-nam… 6/

When she leaves, I look it up, just to get a sense of how big a subfield of Mexican or Asian-American history there is dealing with this specific historical subject matter. 7/

Here's the LC subject headings I saw in Tripod:

Chinese -- Mexico -- History -- 1810-
Racism -- Mexico -- History -- 1810-

8/

For the first one, (Chinese--Mexico--History), Chang's book is the only result that returns. 9/

Huh, I think. I wonder whether to go off into a much bigger catalog. But I stick to Tripod for a minute. I try my old trick of seeing the people who like this book also like on Amazon, and sure enough there's a really closely linked book that turns up. 10/

It's Grace Delgado, Making the Chinese Mexican Global: migration, localism, and exclusion in the U.S.-Mexico borderlands. That also sounds great and intensely relevant to what the student is interested in. 11/

Here's the first LC subject heading for that.
Chinese -- Mexican-American Border Region -- Ethnic identity -- History -- 20th century. That returns Delgado's book and a book review of another book. 12/

That book review (but not the book) leads me to a third book, Robert Chao Romero, The Chinese in Mexico. The first LC subject heading for that?

Chinese -- Mexico -- History
13/

Now that subject heading returns real discovery gold: it's essentially the "right" subject heading for this search. You have to strip the dates off the first subject heading I got for Chino to get it. 14/

Which is as it has always been, and it's always one of the things we teach students to do. But there's still something so wrong about the way specificity is structured in the whole system. 15/

That level of specificity in terms of expert bibliographic control should be hidden behind an expert level interface. In a discovery based search--especially for undergraduates--it's deeply misleading. 16/

The subject metadata associated with publications should be in that sweet spot where scholarly conversation actually happens, where research works are referencing and talking to one another. It should never sever that conversation. 17/

Those three books are so obviously in the same historiographical space--though simultaneously, yes, in several others. A good subject hierarchy should always preserve or recognize that clustering. So, maybe, still? burn the catalog? If we had the time and money to make a new one.