Editing Openai/6973e373-2b8c-800d-84b7-d7001cfc6bac (section)

===== - We classify the function of a behavior, regardless of whether it was conscious, intentional, or sincerely believed. =====
* We care about what it does, what it buys, what it reinforces, what it costs, and what it predicts.

So instead of:
* “They’re intentionally trolling,”

you can say:
* “They’re behaving in a way that is effectively indistinguishable from trolling in its impact on discourse.”

Instead of:
* “They’re selfish,”

you can say:
* “Their participation is effectively extractive: low epistemic effort, high imposed cost, low repair.”

This is how you avoid mind-reading while staying firm.