So, for example, adding qualifications can show uncertainty and humility, yet adding "often works" is longer than "works" both from a character limit and a typing it on a phone keyboard.
Another example, saying "that's annoying" can be quicker than "that annoyed me" or more so "that annoyed me and may annoy others like me."
As much as I try to catch myself taking these linguistic shortcuts, I still may give in to the quickness of them, especially when I'm in a rush, and sometimes, especially online, conversations can move so quickly they put me into a heightened rush state.
Anyway, I'm grateful you pointed out this element of efficiency and scalability. I think these conversations can scale better by taking a little extra time to communicate than by always taking the shortcut. Maybe akin to how code can become more legible, and thus easier to maintain and scale, if the developer takes the time to more precisely name things. The name may be longer but may be more precise and prevent future problems.
I was accused of saying something I didn't say at a bar, by a black woman I was talking with. She complained to the bartender who, without having heard any of the conversation, called me a racist and threw me out.
I was very distressed by this. On hearing the story, a friend brought up the fact that bartenders don't have the time or training to resolve every dispute between customers, so they just make snap decisions that are often wrong. He also said that in days gone by, the black person would have been the one who got thrown out in a dispute, and so this is a form of restorative justice. I found this a helpful idea, because it implies that people still aren't treating each other any more equally or listening to both sides, or weighing things wisely or on the merits; and that's what needs to be fixed if we want to have any kind of discourse at all.
But "modern discourse just can't scale" is a bit of a defeatist mindset. Surly there is a solution out there for reducing polarization.
Well, there's compassion. But that's always been rather difficult, and seems to have fallen out of fashion...
Maybe you meant a technical solution though.
From that perspective, canceling/labeling heresy could potentially be thought of simply an act of automation. People identify reoccurring patterns of arguments, classify their adversary, and use a cached response. Thus, instead of spending the costly time and energy to respond to every argument in detail, you tag, use the appropriate function, and move on. And because modern society is so polarized, those functions tend to be fairly absolute- who wants to get dragged into another argument they’ve had before?
The problem is that modern discourse just can’t scale.