Even if there were such a thing as “the only logical argument” (there isn’t), a logical argument isn’t necessarily superior. We’re evolved to be emotional creatures. Plus, morality and ethics themselves aren’t logical in the sense that mathematics and science are.
I find your response here to be reasonable. That said, you probably won’t find many people over on Lesswrong (your link) saying that policy shouldn’t be based on statistics. Those folks are the tip of the spear for the effective altruism movement.
People view suicide differently from homicide: it’s very different situation when a shooter wants to die, plans and carries out their own death, to when the shooter attempts to murder others. Some even argue that ending one’s own life is one’s right, and others have no right to stop it.
Killings of criminals, especially by other criminals, are also viewed differently, and people view child murder as much worse than adult murder. […]
This is also a reasonable argument. But to figure out relative policy weight, we would need to assign values to suicide, to justifiable homicide in self defense, to police homicide, and to outright murder, and adjust our statistics to see the nature of the gun violence problem through that lens. How many suicides is as bad as one homicide? 10? 100? I’m not saying I know that answer, but you could throw a number at it if you like and we could run with whatever number you threw at it and have a reasonable discussion.
That discussion would, however, still end up being predicated on statistics. As I said in the prior reply, if police shootings were 10 per year or 10,000 per year we would weight that problem differently, which means the problem must exist on a continuum, that is informed by statistics.
Secondly you discuss only gun deaths, but omit non-fatal shootings or gun crimes in general. Even where injury doesn’t occur, the threat of being shot allows the criminal to commit far greater crimes than without.
If we’re going to dip to that level, then we also have to include instances of defensive gun use, including instances that don’t end up on the record because the crime was averted by the threat of being shot. And those numbers (google about) tend to be quite large in the USA. So there’s a counterfactual that must be considered. Folks on Lesswrong like discussing counterfactuals.
So if you want to go to this depth of analysis, we have to weight all the different kinds of gun deaths with different relative-horribleness-factors, and then also weight gun injuries with relative-horribleness-factors, and then weight all the crime that would have happened if the victim didn’t have a gun, and then weight all the crime that doesn’t happen in the street because someone might be carrying a gun, and run a very complicated cost benefit matrix.
And then, once we’d done that, we’d have to look at, say, the expected value for a buyback system (one idea) based on what kind of benefit we’d get from the collection scheme, at a dollar value.
It’s a complicated thing to jump into, and how you set those relative-horribleness-factors is completely a matter of opinion, and will tip the scales one direction or another.
I will say, however, that I stuck to the pure gun-homicide rate numbers and did a back of the envelope analysis on how many guns we’d have to buy back to save one life, based on Seigel’s research which is quite solid. The answer was $80 million worth of guns per single life saved.