[dionidium]

The argument goes like this:

There is probably a line. But you don't know where it is and neither do I. You and I might agree that X is to one side of that line, but if we ban that behavior, then we have initiated a process that we might call line-discovery -- the search for the line that X was to one side of -- and line-discovery is highly prone to outcomes that result in bans on content from the other side of that line. So we don't want to engage in line-discovery, even though there are obvious examples of things to one or the other side of the line.

You may think you can ban the obvious things without ultimately engaging in line-discovery, but, the argument goes, you are mistaken. You will ultimately find yourself doing line-discovery.

You start out with obvious-sounding prohibitions on racism and hate speech, but eventually you're arguing about, say, whether it's racist to report on polling showing that violent protests are unpopular. [0]

And that's because banning any speech always leads to line-discovery.

So it comes down to a question of which scenario is worse:

A. You ban obviously bad stuff while accepting some risk of banning things that aren't actually over the line.

B. You privilege all content to avoid that outcome.

Some people are outraged by this framing and think it's obvious that you would want to risk banning some behavior to the right side of the line if it means eliminating the most obnoxious speech. But, basically, that is not obvious to everyone, no matter how many times they are reminded that there is some really bad stuff out there. [1]

[0] https://reason.com/2020/08/27/protests-violence-david-shor-k...

[1] Interestingly, this is really not so different from the argument about evidentiary standards for punishing criminal behavior, except in that case the politics are flipped. There conservatives would rather risk punishing some innocent people if it means the absolute worst actors are guaranteed harsh punishment, but liberals think it's worth risking some amount of literal rape and murder in order to prevent punishing the innocent. So I think, actually, both sides are entirely capable of seeing this from the other side; they just don't want to.

[wpietri]

Do you feel this is somehow responsive to my point? Because I'm not seeing it.

[dionidium]

Yes. I am addressing the second-order effects of each motivation.

Let's grant that the harms of the kind of speech you're worried about are exactly the same in either case. [0] Platforming "racist harassment" because of a political commitment to free speech implies that other forms of controversial speech will get the same treatment, preventing the kind of line-discovery I described in my previous comment.

"Platforming racist harassment because you just kinda like racist harassment" leads to who knows what. All we know about that person is that they like racist harassment. Maybe other stuff gets banned. Maybe not. Either way, it's unlikely to be in service of avoiding harmful second-order effects.

So that's an enormous difference between the two motivations. In the first case the position is in defense of an ethic of open dialogue and an attempt to prevent second-order effects that are harmful to that dialogue.

In the second case -- who knows.

It seems to me that the first motivation is much more likely to prevent the kinds of second-order effects I'm worried about and that distinguishes it from the second one.