It's halfway between a Swiftian 'modest proposal' and a callback to duelling laws. Simply implementing this as described would of course lead to all sorts of undesirable outcomes, but really it's just a combination of the fighting words doctrine and an expanded concept of imminence that incorporates the erosion of spatial obstacles to information.

Currently you shouldn't use deadly force absent a reasonable fear that it is about to be inflicted on you. But since it is entirely possible to send a threat like I know you live at [address] and I'm coming to [do crimes] and place the recipient in a state of continuous uncertainty, the very open-endedness of the threat manifests the fear of imminence in the target. Rather than making the victim devote resources to continually monitoring and assessing the behavior of the threatening party, we can simply expand the target's right of self-defense in proportion to the scope of the issued threat.

All well and good until I pose as someone else and make false threats to get them legally murdered

Could you try addressing the argument instead of constructing a magic trapdoor by imagining additional facts? Assume you can verify the authenticity of a threat, in case that was not clear from my earlier description.

That assumption was not at all obvious, if you want to talk about imagining facts.

To address the argument, it's not fair to assume that recipients of such threats would be either willing or physically able to respond. People with kids will not go around dueling people in the modern era. In my opinion, it's nothing more than a fantasy world. We have a justice system and law enforcement officers. Maybe they don't do a great job, but it should in no way fall to the individual, no matter how zealous they may be.

I'm not suggesting that it should fall to the individual, but that the individual have some latitude to defend themselves against a credible threat rather than being expected to just put up with it and hope for the best.

as much as I like the idea, I think it is a fair point to make. the original context for this discussion was threats made over the internet. it is actually kind of hard for a normal person to authenticate the identity of someone over the internet to a degree of confidence that justifies a killing in self defense.

The bulk of personal threats and violence takes place between people who know each other and is not so difficult to authenticate; likewise there are many ways and people willing to help identify the authors of internet threats.

The point again is not so much that 'this will solve the problem' as 'threatening people should have a significant cost' and this is one way of significantly increasing that cost without needing to appeal to state actors whose alacrity and integrity may be open to question.

This approach maximizes free speech while providing the targets of objectionable or harassing speech more options for responding to it.

you're right, this is a pretty bad failure mode. however, a similar (though much less severe) risk is incurred when we make universities responsible for handling threats. if someone poses as you on social media and makes death threats against another student, a court of law might find insufficient evidence to prove that you actually made those threats, but good luck showing the account wasn't yours under a preponderance of evidence ruleset.