I have sometimes seen arguments that fit this pattern, including on Less Wrong —
Your disagreement with me on a point of meta-level theory or ideology implies that you intend harm to me personally, or can't be trusted not to harm me if the whim strikes you to do so.
It seems to me that something is deficient or abusive about many arguments of this form in the general case, but I'm not sure that it's always wrong. What are some examples of legitimate arguments of this form?
(A point of clarification: The "meta-level theory or ideology" part is important. That should match propositions such as "consequentialism is true and deontology is false" or "natural-rights theory doesn't usefully explain why we shouldn't hurt others". It should not match propositions such as "other people don't really suffer when I punch them in the head" or "injury to wiggins has no moral significance".)
One mistake is overestimating the probability that the other person will act on their ideology.
People compartmentalize. For example, in theory, religious people should kill me for being unbeliever, but in real life I don't expect this from my neighbors. They will find an excuse not to act according to logical consequences of their faith; and most likely they will not even realize they did this.
(And it's probably safest if I stop trying to teach them to decompartmentalize. Ethics first, effectiveness can wait. I don't really need semi-rational Bible maximizers in my universe.)
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.