Let’s start with the first question, rephrased: If I am uncertain if I’m epistemically superior to a person P with respect to issue S, can I still cancel P for their position on S?  

I think the answer is No (since righteously punishing someone requires a reasonable belief in the punished’s wrongness or culpability), and this answer implies the unrighteousness of most actual cancellations (since they involve much epistemic uncertainty).

However, some may object that there’s a shortcut to judgmental superiority: expert testimony. Suppose expert E, who is epistemically superior to both me and P, says that P is wrong about S. Does learning this fact (along with facts about E’s superiority) lift the epistemic barrier to my canceling P over S? I think the answer is No. This is because experts who agree with you aren’t common in real-life cancellation contexts, and even if they do exist, you may not learn about them, and even if you do learn about them, you’re seldom justified in believing in their superiority (due to confirmation bias and practical barriers), and even if you’re justified in believing in their superiority, you’re seldom warranted to assess to what extent does their position contradict that of P on S (without sufficient expertise yourself); therefore, you’re almost never justified in the belief that a relevant expert contradicts P on S. But even if you’re justified in this belief, your having expert testimonial evidence on S doesn’t make epistemic uncertainty go away. What if P also has evidence in favor of their position? What if another epistemically superior expert E* agrees with P on S against you and E? What if there are a hundred experts, each with their own take on the issue (and no meta-analysis is available)? So I conclude: You should (probably) be uncertain if you’re epistemically superior to a person you want to cancel, and you should (probably) refrain from cancellation for that reason.