Perhaps you could at least give some examples of arguments for which the necessary response is violence.

Possibly, but only if you roll a natural 20 on your necromancy check. You are replying to a comment that was posted in 2007 - and on a different blog! Perhaps not the ideal place to play the "my position is the default - you are the one who needs to supply all the evidence!" game. Or, then again, perhaps it is the ideal place!

A strong claim but without any evidence to back it up.

It is not even a question for which demanding evidence makes sense - at least without specifying more clearly what kind of observations of the world you are considering. One could assume that you mean "give me evidence that the consequences of responding to arguments with violence can positive" - but then you have already lost to Caledonian's position. When you are looking at consequences, "argument" and "violence" are just two different kinds of power. Occasionally the latter is to be preferred to the former.

The only way "there's a flat law against meeting arguments with violence, anywhere in the human world" was going to hold was if it stayed purely in the ideological realm. And a Traditional Rationality ideology realm more than a Bayesian Rationality one. "Arguments" can, at times, be a greater epistemic rationality violation than mere violence.