Here is an interesting item from John Holbo at Crooked Timber:

No, I don’t mean: arguing fair. I think it should be ab homine. A moving (irrationally) away from the man. It’s a fallacy.

Here’s the context. Matthew Yglesias and Jonathan Chait have a diavlog in the course of which Chait takes the scrupulous high-road position that, when it comes to charges of racism, you really have to be slow to accuse. He rolls out the standard fair-play-in-debate considerations: if the person is saying something wrong, but not explicitly racist, you can just point out the wrongness, without speculating, additionally, that they said the wrong thing out of racism. There is, he implies, no real loss in not being able to delve into dark motive.

But here’s the problem with that. In an environment in which creative and speculative accusations of bad motives are, otherwise, flying back and forth in free and easy style, a social norm against accusing people of one sin in particular is actively misleading. It inevitably generates the strong impression that this bad motive – out of the whole colorful range of diseases and infirmities of the mind and spirit – is an especially unlikely motive. Which, in the sorts of cases Chait and Yglesias happened to be discussing, is not true. So, contra Chait, an inconsistent semi-norm against ad hominem arguments encourages an ab homine error that may be less angry (that’s not nothing) but is significantly more confused tha[n] what excessive – but even-handedly excessive! – hermeneutics of suspicion would produce.

I seem to remember talking about something like this before.  I'd call this an error of excessive scrupulousity.  Philosopher types fall prey to this one out of an overabundance of charity: Sure this argument really blows, but maybe there really is a good one in here somewhere.  (Sometimes, I think, philosopher types do not want to trouble their beautiful minds with silly arguments, so they just deny their existence or refuse to discuss them).

Nonetheless, I still think one ought to be especially careful in attributing motives, including racist ones, to other arguers.  In the first place, those motives are hard to know (and therefore easily disputed); second, they are hard to define (and therefore easily disputed); third, and most importantly (and tragically) they are come across as illegitimate ad hominems (and are therefore easily disputed).  The ease of dispute of imputation of racism places a heavy practical burden on the accuser.  Does one really want, in other words, to go through the necessary evidence in order to make a point likely to be only tangentially related to the discussion at hand (even if true)?

So this is mostly a pragmatic objection to Holbo's point.  Unfortunately, what makes these sorts of accusations difficult (even if true and relevant) is the deeply entrenched presence of the fallacy fallacy.  This is the view that the very criticizing of someone else–especially in accusing them of fallacies–is itself a kind of fallacy.  The rules of our dumb discourse prevent legitimate criticism.  The only thing that counts, I still maintain this week, is consistency.  This is why Pat Buchanan, despite his Hitler apologetics, is constantly on TV.  He's consistent.