Wellyopolis

August 23, 2007

The AHA's muddled approach to classifying research

The American Historical Association (AHA) has a confused way of classifying its members' research interests. That confusion causes members some consternation when the AHA then proposes to eliminate certain categories.

The root of the confusion is that the AHA's taxonomy attempts to fit several dimensions of historical research into a one-dimensional list. The AHA claims that since members "can select up to three categories that interest them," a category that lacks support from at least five members must "lack a significant constituency" in the membership. Last year the AHA tried to eliminate "psychohistory," this year they propose to potentially eliminate various periods of Japanese history, New Zealand history, Sudanese history, and British Columbia because fewer than five members listed these categories as a research interest. In response I've gone and changed my membership profile to now include New Zealand history, perhaps to save it from the chopping block.

Apparently, some of the reasoning behind eliminating categories is to save space on the printed form that members fill out to indicate their research interests. In the age of the Internet it scarcely seems necessary to have printing technology dictate intellectual decisions about classifying the areas of historical research. A favorite rhetorical trick of historians is to say "What if your students said that?" Well, what if your students used similar logic? This is the AHA saying that it is revising intellectual distinctions because it ran out of notepaper. What if your students said that?

The AHA is at pains to point out that categories can be added back into the taxonomy of research interests, if enough members request it. This is hardly the way to organize a classification system. To be meaningful a classification system should be stable over time, not subject to the whims of membership votes. If a category is significant as a field of study it should be left in the taxonomy, so that if and when it re-emerges with more support that change can be accurately tracked.

The biggest confusion in the AHA's research taxonomy is its approach to combining geographical, chronological and topical classificiations into one dimension. A basic principle of classification is that you should not combine separate aspects into one. Take my own research, for example. A compact description of what I do is economic and business history of New Zealand and the United States between [about] 1860 and 1960. I think I have a relatively coherent research agenda, but the AHA's muddled research taxonomy doesn't allow me to adequately describe it in the three categories I'm allowed to choose.

How the AHA comes up with its categories I don't know. Under United States history you're allowed to choose various time periods, including in the modern era "1877-1920," and "since 1920." Every time I see this it mystifies me. I understand that 1877 is significant as the end of Reconstruction, but that surely privileges a particular master narrative of American history over any alternatives. And "since 1920"? You can make any sort of argument you like for the significance of 1920, but 1930 is pretty significant too, as well as 1932, and 1912, and well, you get the idea ...

The AHA should revise its research taxonomy entirely. It should allow members to separately


  • Select several countries they are interested in. If countries have important sub-national divisions, either historic or contemporary, they can be selected too. At present the AHA's categories for Canadian research are a mystery as well. You can select British Columbia or Québec as separate subjects, but Ontario must fall under the generic "Prairie Provinces." I know only a little about Canadian history (it's glorious!) but Ontario is a little different than Manitoba, principally on account of the city of Toronto and its environs.
  • Select the time period they are interested in. This could be specified as a starting year and ending year. It would then be trivial for analysis to combine similar years into categories and centuries for tabulation. More interesting information on the overlap in research time would result from this approach rather than the rigidly specified "1877-1920" approach taken at the moment.
  • Select several topical approaches they are interested in. The present list of topics is an alphabetically ordered soup of themes and methods. Comparative, economic and psychohistory are arguably methods, mixed in with themes or topics like labor, immigration, and the mysterious "General Studies". The latter is about three times more popular than New Zealand history, for what it's worth.

The AHA might reasonably respond that this would lead to an explosion of information. It is not clear whether the AHA's membership taxonomy is meant to be a survey of the core topics its members are prioritizing in their research right now, or a broad survey of the wider field members set their research in. The way it's set up at present you have to choose between different levels of specification. For example, say you're a Canadian historian and you're presently working on a topic in British Columbia. Which do you choose? The obvious answer would be both, with the choice being presented as the hierarchy of precision that it is. The current system forces a trade-off between narrow and broad visions of research without indicating what the system is designed to elicit.

The geographic and chronological categories reflect this confusion. Take those U.S. history time periods again. Why the 43 years between 1877 and 1920 deserve their own category while the modern era (87 years and counting) and the colonial period (about 200 years) are implicitly all of the same is not clear. Within these eras, few people are likely to be doing research on the whole period specified by the AHA. It can't possibly be a fair representation of what people are doing. If your research spans several of these artificial eras, what are you meant to do? My dissertation dealt with the period from 1850 to 1940. I could use all three of my apparently generous allocation of category choices to describe only my time period. If you do American Indian history your chronological divisions are Pre-contact, Colonial, 19th century, and 20th century. Perhaps there's a long 19th century concept in here that covers the 1776-1800 period, but that's not stated. And why does American Indian history work in neat centuries while the rest of American history pivots on wars and presidential elections?

Similar absurdities in chronological specificity are apparently present in the Asian history section. In the Chinese history section if you study Taiwan you're explicitly constrained to doing it after 1949, but if you study Hong Kong and Macao there's no time period. I can sort of divine the reasoning here, but it's not really clear. Moreover, "and"? What if you only studied one of Hong Kong or Macao?

The same error appears where "Australasia and Oceania" are specified as options. I appreciate the AHA specifying New Zealand as a separate option, but in case they don't know Australasia implies Australia and New Zealand. As it stands, it's the Australian historians who could complain they're not fairly represented. Moreover, while many historians of the Pacific islands (Oceania) have to consider Australasia, the converse is not true. Most Australian and New Zealand historians work in blissful ignorance of the Pacific islands, the islands being merely convenient transit points on the way to San Francisco or Vancouver by boat or telegraph. You could also be an historian of Oceania's interaction with the United States or other Pacific Rim countries without studying Australasia much, if at all. Lumping Oceania in with Australia and New Zealand misrepresents that field too.

In short, the taxonomy is a mess. It would be better to start over than keep building on the existing foundations.

Posted by eroberts at August 23, 2007 4:44 PM