Discussions take place among people.
People are internally relevant beyond externally fixated objects by natural laws.
Therefore, discussants must have free will.
To deny free will is to discredit the value of participating in discussion.
Therefore, discussants must respect free will.
I’m not sure how even considering determinism or fatalism is a respectful notion because if those ideas ring true, then it denies the faculty by which discussion is relevant.
Now granted people could consider structure before agency. That is people could consider in terms of rational choice that some outcomes are likely or guaranteed since free willing players pursuing self-interests with information will execute strategies.
However, the idea of people internally being out of control begs the question as to how people are interacting in the first place (nevermind how it can be anticipated that people will interact respectfully into the future). Perhaps people are acting according to biological imperative, but then a discussion would be no different from coding a computer. Furthermore, the anticipation of discipline from the other side would be unreliable. That is biological imperative could “take offense” from semantics and syntax that aren’t offensive. People could claim they’re provoked into violence because they’re out of control such that the other side has to endure the consequences whether it likes to or not.
I suppose the exception here would be emotivist compatibilism. People could be emotionally influenced in what they say. For example, someone’s emotions could dispose someone to being hyperjudgmental and focused. Someone could become unable to take things lightly or think creatively over making jokes. Instead, someone could take everything seriously such that someone actually comes off as autistic in not understanding when others are teasing. On the other hand, some people become so emotionally riled that they can’t stop making fun of others such that some people come of as psychopathic.
There’s still a problem here though. Even if someone understands that others are teasing, that doesn’t mean someone will necessarily be able to tease back. Likewise, if someone’s emotionally riled, that doesn’t mean someone will have anything else to say besides mockery. Therefore, testing the falsifiability of (at least two forms of) emotivist compatibilism would be unreliable.