Here’s what seems like innocuous and even noble advice for authors and referees: Try to ensure that papers cite members of under-represented groups.I’m not so sure. I’m concerned that this proposal, rather than helping to fix the problem, doesn’t understand the problem or take the problem seriously.Some background: In most academic fields, the practice is that you cite pretty much everything relevant to your work. In philosophy, people tend to cite themselves, their friends, John Rawls, and the papers they directly criticize. That’s one reason why citation counts and h- and i-indices in philosophy are much lower than in psych, poli sci, or chemistry.The epistemic concern with citing people from different demographic groups is that membership in such a group may
Jason Brennan considers the following as important: Uncategorized
This could be interesting, too:
Scott Sumner writes Facts about life expectancy (true and false)
Wolf Richter writes What’s Behind the Subprime Consumer Loan Implosion?
Steve Horwitz writes Hayek’s Modern Family for .99
Scott Sumner writes The China (manipulation) shock and the (mostly) Trump fiscal shock
Here’s what seems like innocuous and even noble advice for authors and referees: Try to ensure that papers cite members of under-represented groups.
I’m not so sure. I’m concerned that this proposal, rather than helping to fix the problem, doesn’t understand the problem or take the problem seriously.
Some background: In most academic fields, the practice is that you cite pretty much everything relevant to your work. In philosophy, people tend to cite themselves, their friends, John Rawls, and the papers they directly criticize. That’s one reason why citation counts and h- and i-indices in philosophy are much lower than in psych, poli sci, or chemistry.
The epistemic concern with citing people from different demographic groups is that membership in such a group may bias/influence/change how people reason about different topics, what information and evidence they have access to, which intuitions they have, and so on. Or they might not. These are testable empirical hypotheses. We don’t know a priori—and a priori arguments like you see in philosophy are at best mere speculation.
Economics and political science, unlike philosophy, take this issue quite seriously, and so there are a number of studies testing to see how different demographic factors influence what people think, all while controlling for confounds. You collect three sets of data: A) What people think, B) who they are, C) any confound you can think of. (For example, when testing how demographics affect voter preferences, you control for information. You also have to control for demographics when you want to know how information affects preferences.) Controlling for confounds is essential, a minimal condition for doing valuable work. (After all, we have plenty of examples in econ and poli sci where apparent demographic correlations turn out to be information effects or something once you run the controls.)
Further, this kind of work allows them to test which demographic factors are salient and which aren’t. That’s why, e.g., if a student says, “Economists favor free trade because they are rich, and tend to be white or Asian men,” the economist can answer, “That’s a good and worrisome hypothesis. So we checked. Here are various studies falsifying it. Here’s a study showing women are slightly more protectionist than men and this isn’t explained by information effects. Etc.”
If we care about improving the epistemic state of philosophy, we would study this carefully and see what effects there are…and aren’t.
It won’t be enough to just assume that when a self-identified member of group D writes a paper on Y, that their view on Y reflects their membership in D. Even if the author claims this is so, that’s not much presumptive evidence. The psych lit is quite clear that people are bad at assessing why they think what they think, especially on political issues. For instance, if you read Achen and Bartel’s Democracy for Realists, you’ll see lots of evidence that demographic factors place people into different political parties, but—importantly—not at all because these parties serve their interests or because of genuine ideological commitments. Boston Irish dudes like me usually identify as Democrats, but not because the Democrats are good for us or because most of us genuinely agree with the party platform.
In practice, the main effect of “trying to cite underrepresented groups” is likely to be that people will continually cite the known members of various salient groups they can easily think of. We’ll see availabilityhttps://whatis.techtarget.com/definition/availability-bias bias dominating both A) which “groups” people think need to be represented and B) which individual members of those groups get cited.
So, for instance, Rachel McKinnon is a famous trans-woman. I’d expect she’d benefit greatly from “trying to cite underrepresented groups” at the expense of other less famous transgender women who may have done better or more salient work on a given topic. (N.B. That’s not an attack on her. I’m not saying her work is bad or good. I’m just saying that she comes to mind more easily than other transgender women who may have done better work on any given topic.) Or we might pick “transgender woman” as the under-represented category because, thanks to current events, it comes easily to mind, when really on this topic the salient and important underrepresented group is something else.
How will we even know who belongs to what group? Should we go through various university webpages and try to guess based on names, apparent skin color, apparent gender presentation, and so on? Even at best, that is unlikely to help me identify members of other important groups, such as people from working class backgrounds, or the disabled. What if I think someone is white but he’s actually a Latinx who passes as white? Etc. There is no expansive or comprehensive public depository identifying which papers and books are written by which kind of underrepresented person.
Some proponents of this proposal say this last objection overblown. We should just cite people who write with a clear voice from various groups. But, other than in cases of explicit self-identification, I also don’t know whether people write in distinctly or easy to discern voices that reflect and indicate their background demographics. For instance, over the past month, I’ve read about a thousand papers as I’ve been compiling a list of texts for a new business ethics anthology. Other than from (the at best presumptive evidence of) first and last names, fame, or in cases where I know the authors, I have no idea what groups the authors belong to. For instance, I’m planning to includes pieces by Roland Fryer and Glenn Loury. I wouldn’t have known from the content of their writing that they are black. I know because I’ve met them or seen them on TV. I’m also including them because their work is exceptionally good, not because they are black.
Unless and until someone does the required regressions—including, necessarily, controlling for confounds—we don’t even know whether there are genuine demographic slants in how people think about philosophical issues. We don’t know which demographic categories matter and which don’t. We don’t know which things indicate a voice reflecting demographics and which are illusory.
Now, if we just did the “cite everything” method you see in most other fields, this would be moot. The only reason we’re even talking about this is that philosophers have bad citation practices other academic fields avoid. So, rather than try a lousy, bias-prone pseudo-fix for something we don’t even yet know is a problem, we could just insist authors use the proper citation methods you see in, say, chemistry.
If I were a cynical person who thought academia was ruled by self-interest, I might suspect all this is really just a ruse for certain people to increase their citation numbers. Good thing I’m too optimistic about people to think that way.