Contestable vs Question-Begging Arguments

They're not the same thing!

Clarifying the Charge of “Begging the Question”

As I wrote last year:

Sometimes people assume that an argument they personally find unconvincing is thereby “question-begging” or otherwise worthless. This is a mistake. A determined opponent can always just reject a premise; that’s inevitable. Arguments can’t force people to change their minds, so that isn’t a realistic expectation.

We do better to think of arguments as highlighting neglected costs (of rejecting the conclusion), and inviting those who nonetheless reject our conclusions to (i) seriously consider which costs they’re willing to accept (i.e. which premises to reject), and (ii) suggest any counterarguments that mitigate the apparent cost of their preferred move (or perhaps even show it to be a “feature” rather than a “bug”). In a successful dialectic, everyone leaves with a clearer view of the costs and benefits of the competing views on offer.

A question-begging argument is one that offers no such illumination. The conclusion is so transparently contained within the premises that there is no conceivably “neglected” consideration there to highlight—nothing that might, for example, help to sway a “fence-sitter” who was as-yet-undecided about whether to accept the conclusion. Any such fence-sitter would necessarily be just as undecided about the question-begging premise.

The common conflation of the contestable and the question-begging ties in with my perennial lament about bad referees using objection-minimizing norms rather than quality-promoting norms in their philosophical evaluations. By mistakenly charging virtuously contestable arguments with the vice of begging the question, bad referees prevent interesting critical work from being published, and instead incentivize authors to stick to “safe”, boring claims that are less likely to be contested. This is clearly bad for the discipline, and gatekeepers (both referees and the editors who rely on their reports) ought to make more of an effort to avoid this classic failure mode.

Testing for Vacuity

Recall that the problem with “question-begging” arguments is that they offer no substantive illumination. They are philosophically vacuous. So to assess whether a target work is indeed “question-begging”, one must ask questions like: (i) whether anyone could reasonably be swayed by it; or (ii) whether it highlights some previously-neglected cost or challenge that those who wish to reject the conclusion would need to grapple with. If the answer to either of these questions is yes—as it very often is—then it is not “question-begging”. It is, instead, merely contestable.

Remember that being contestable is a good sign in philosophy. If a paper weren’t contestable, that would seem another kind of vacuity or insubstantiality.¹ Good philosophy is substantive, which is why truly question-begging papers are not worth publishing or engaging with (there is literally no substance there with which to engage). But, as noted above, the practice of rejecting papers for containing contestable claims also promotes vacuity. In other words: false charges of begging the question are bad for the same reason that begging the question is bad. Both errors deprive us of interesting, substantial philosophy, and ought to be avoided for precisely that reason.

This suggests a couple more tests for false charges of begging the question: (iii) whether the paper is thought-provoking, and apt to prompt fruitful engagement (it’s hard to see how it could do so if it were truly insubstantial), and (iv) whether any narrowing of the paper’s ambitions, or reduction in the scope of its claims, would avoid the charge.

Here I want to pause to dwell on test (iv)—call it the removal test—for a moment. It’s obvious how making fewer, narrower claims could make a paper less “contestable”. Equally obviously, removing claims cannot make a paper more substantial. So if there’s a way to address the critic’s concerns simply by removing (or narrowing) the paper’s positive claims, then the critic’s concerns cannot actually be that the paper is vacuous (as the charge of “question-begging” would imply). They are simply contesting some of its claims, and misdescribing the nature of their objection in a way that egregiously exaggerates its force.²

“I see nothing of philosophical substance here!”

Protecting Contestable Philosophy

I think it’s usually bad for referees (and editors) to reject a paper on grounds that violate the removal test. Paying attention to this (so, critically checking for sufficient merits to publish, rather than sufficient demerits to reject) may be my #1 proposal for improving how we evaluate philosophical work.

As I wrote in that earlier post:

Given current norms, we all know that it can make a paper “more publishable” (i.e. referee-proof) to remove interesting ideas from it, because more content just creates more of a target for referees to object to. This is messed up. Good-seeking standards instead recognize that adding relevant valuable content is (typically) a good thing. Our evaluative standards should reflect this fact.

(Granted, there are possible cases in which extra claims just distract from the central point a paper is making, and are better removed for that reason. But those are rare, and I think authors are typically in a better position than referees to judge whether a contestable claim, or broadening of their argument’s scope, is worth including.)

Now, I don’t know how well most would manage it, even if they tried.³ But if referees would explicitly run through the four tests identified above before charging a paper with “begging the question”, that would surely go some way towards improving their verdicts.

1. The rare exceptions: ideas that are obvious only in retrospect. (I think effective altruist principles are like this, as I argue in ‘Why Not Effective Altruism’: not really reasonably disputable, but still important to write about because many haven’t yet paid close enough attention to notice.)

2. Example: this paper of mine was recently rejected from a top-5 journal (though I did appreciate referee 1’s glowing evaluation and helpful comments), with referee #2 dismissing it as “entirely question-begging” for relying on the contestable premise that deontic constraints entail decisive agent-neutral reasons (to prefer that a violating action not be done).

Note that my paper contains an entire section that argues at length against the standard alternative (on which deontic constraints are seen as purely agent-relative). It offers four independent reasons to prefer the agent-neutral conception, and finally stresses that: “One who judges the matter differently can simply treat the argument of this paper as instead targeting specifically agent-neutral conceptions of constraints, as found in Setiya (2018). The failure of that more specific view remains interesting and philosophically significant in its own right, especially since it otherwise seems to have so much going for it.”

I grant that my paper is (very!) contestable. But it is very clearly not question-begging or insubstantial, as can be seen by applying any of the four tests outlined in this post.

To apply the removal test in particular: their charge could easily be avoided by cutting half the content and narrowly framing the paper as an objection to Setiya’s view of constraints. Fortunately, I’m tenured, so I can resist the incentive to make my papers worse just to placate bad referees. But I do think this episode demonstrates a real problem with peer review in philosophy.

To be clear: this referee—and the many others like them—are effectively complaining because the paper is more substantial than it might have been, including additional interesting arguments that they simply happen to disagree with. That’s bad refereeing. But it’s also not unique to this particular referee: it’s bad refereeing of a kind that is utterly commonplace in our discipline. I get reports like this all the time on my more ambitious papers. The only solution (that I can see) is to try to raise awareness of the problem so that fewer referees (and editors!) make this same mistake in future.

Lesson: whenever a referee charges a paper with “begging the question”, apply these four tests to check whether they might instead simply disagree with some of the paper’s claims. And remember that first-order disagreement is a reason to publish a reply piece, not a reason to block publication of interesting new arguments that warrant consideration and engagement.

3. That’s one reason why I think editors should allow authors to respond before reaching their final verdicts as a matter of course. Referee objections are so often easily addressable, if given the chance, and it just wastes everyone’s time to force authors to restart from scratch at the next journal on their list. (I guess it’s a collective action problem, in that each editor skipping this stage of the process saves themselves a little time—one less page to read and consider before reaching this verdict—at the cost of plugging up the system for everyone else when they mistakenly reject a paper that they shouldn’t have.)