We in this case is defined as we who support (actual) tolerance. Because I'm making a prescriptive statement here, I am really speaking to those who believe they support tolerance, and explaining what that ought to mean in light of Popper's arguments. That is: the tolerant should be provoked to intolerance only in very rare cases, like when Peter Parker is provoked to try to kill someone, not like when Donald Trump is provoked to insult someone's appearance.

I don't think it makes terribly much sense to be a consequentialist with a value-metric that scores "tolerance". If one were, it might make sense to be extraordinarily intolerant for an incredibly long time in order to ensure even longer-lasting tolerance (depending on if and how you accounted for temporal discounting). However, as a practical matter we have great difficulty predicting even short-term consequences of our actions, so even if we declare consequentialism true-in-principle, outside of explicit discussions of philosophy, I don't take "I support tolerance" to mean "I am a tolerance-consequentialist".

Therefore, I view such statements as virtue-ethics claims: tolerance is a personal and societal virtue (this is what is being claimed by the person who advocates tolerance, unless they explicitly say otherwise). And because such a statement is relative--everyone is tolerant to some degree--an explicit affirmation of tolerance is equivalent to endorsing a belief similar to the one I describe above.

Otherwise, at the very least one isn't communicating well. If one says, "I believe in extreme tolerance!" and in practice what that means is, "You agree with everything I say or you're a Nazi and I punch you in the face", then you're perfectly justified in saying, "That's not what 'tolerance' means." (Then dodge, because you're about to be punched in the face, you Nazi!)

I haven't read Conscience. I really should, as I've read Haidt's The Righteous Mind. Pat Churchland is perhaps the clearest and most insightful thinker who I have ever met. But even the deeply insightful can make mistakes. Anyway, without the benefit of her detailed insight in this regard, my tentative impression is that pro-social behavior is deeply ingrained (i.e. "instinctive") in social animals, including among species who share a recent common ancestor with us, so I am very wary of arguments that depend heavily on the arbitrary nature of social contracts. Even in cases where society has a profound impact, I do not think it is justified to conclude that it has the only impact absent very clear evidence that this is in fact the case. Rather, I think the default hypothesis needs to be that there is an interplay between intrinsic and society-sculpted cognitive patterns, with the nature of the interplay not very deeply understood. Claims that something might be wholly cultural are therefore novel empirical hypotheses that need to be tested before we grant assent.

Even such things as "we eat dessert after a meal, not before" seems not entirely arbitrary when considering both nutrition and glucose regulation.