So, it turns out that Bryan Caplan isn’t a bleeding heart libertarian either.

Unlike Will, though, Bryan’s problem isn’t with the libertarianism of BHL, but with its bleeding heart. Bryan, you see, doesn’t think we have much in the way of obligations to strangers beyond simply leaving them alone. That might sound like a radically libertarian doctrine, but Bryan thinks that is really just a matter of “common-sense ethics.”

That sounded quite wrong to me, so I asked Bryan what he thought of the sort of “shallow pond” examples made famous by Peter Singer. To me, these cases seem to present exceptionally clear examples of situations where we do have a positive moral obligation to strangers.

Yesterday, Bryan blogged his response (UPDATE: fixed link). Bryan starts off by questioning whether we really have an obligation to save the drowning child. After all, we’d praise someone who rescued a kid in that way as a hero. So perhaps this shows that they are going above and beyond the call of duty?

This strikes me as pretty implausible. To have a duty to X is for it to be wrong for one not to X. And it’s pretty clear to me that someone who walked passed the drowning child without a very good excuse would be acting quite wrongly. We might call the person who saves the child a hero, but I think that’s a bad way of distinguishing duty from supererogation. The way to make that distinction is not to look at how we react to people who do the good thing, but to look at how we react to people who fail to do it. And someone who fails to save a drowning child is going to trigger a very different set of reactive attitudes than someone who fails to perform a paradigmatic supererogatory action like, say, volunteering for the Peace Corps.

But is it an enforceable duty? Bryan asks:

Would it be morally permissible to point a gun at a person if he fails to rescue the Drowning Child voluntarily?  Much less clear.  To pull the trigger?  Even less so.

Why, oh why, does it always have to be about guns for libertarians? Yes, I know that in some ultimate sense, every law is backed by the threat of violence. If you break the speed limit and are sent a fine, and don’t pay it, and resist when the cops show up at your house, and resist very effectively when they try to physically force you into their car, then eventually they very well might take out their gun. But that just. doesn’t. mean. that posting a speed limit sign is the same thing as pointing a gun at you. Or even the moral equivalent of doing so. So yes, I’m quite willing to say that people have an enforceable duty to rescue drowning children. Whether pointing guns at the recalcitrant is morally defensible or advisable is an entirely different question, though to sate your curiosity I’m sure we could come up with a suitably fancy thought experiment in which I’d be willing to do it.

Which leaves us with Bryan’s final, and I think strongest, point. Peter Singer wants to use the drowning child case to argue that we have a moral obligation to give lots and lots of money to poor people in the developing world. The idea is that their poverty puts them in a desperate situation analogous to the drowning child, and our (relative) wealth puts us in the position of the passer-by who can help at relatively little personal cost.

But as my old mentor David Schmidtz forcefully argued  (along with many others), there seem to be a lot of morally significant differences between the case of the drowning child and the case of international poverty. And one difference – the difference that Bryan highlights here – is that while being in a position to pull a drowning child out of a pond is a bizarrely rare occurrence, being in a position to rescue someone in poverty is not. Indeed, we’re all in that latter position right now. And if you head on over to Oxfam and save somebody by, say, donating $500 right now, you will still be in that position with respect to all the other starving children in the world, and all the rest of the money in your bank account.

So what does this mean? Well, one way of putting the point is this. It’s reasonable to think that people have a moral obligation (even an enforceable one) to save any drowning children they come across because the expected cost this obligation imposes on any given individual is vanishingly low, while the expected social benefits are high. If, however, we held that people had the same obligation to rescue starving children abroad as they do to rescue drowning toddlers in ponds, then the costs to individuals would be immense. We can live a rich, normal life being fully committed to rescuing every single drowning child who crosses our path. A commitment to rescue every starving child in the world, in contrast, would consume our life.

If this line of argument works, it counts against an expansive personal duty to rescue. But…and here’s the rub for a libertarian like Bryan…it looks like it might actually count in favor of certain kinds of redistributive public policies. Here’s why. The problem with an expansive personal duty is basically one of fairness and free riding. Since most individuals will likely not comply with such a duty, those who do will be left with an unreasonably large share of the overall burden. If, on the other hand, everyone could be counted on to do their share, then each individual’s share might not be unreasonably large. And that’s just what coercive redistributive policies do. By greatly reducing the free rider problem, they can (in principle) devote serious resources to problems of poverty without imposing an unduly large burden on any particular individual. Think about it – on a purely domestic level, how much of your tax dollars go to programs that are actually designed to redistribute wealth to the poor (as opposed to, say, the elderly)?

So here’s the takeaway. The idea that individuals have no positive duties toward strangers is entirely implausible. And nothing in Bryan’s argument gives us any reason to doubt that those duties are real obligations, and in some cases even enforceable obligations. Moreover, part of Bryan’s argument actually counts against viewing those obligations as individual, private duties and in favor of viewing them as collective duties that should be coercively enforced. In other words, Bryan’s given us no reason here to oppose institutionalizing the duty to rescue in the form of a state-funded minimal social safety net.

Obviously, a lot matters on the details, and nothing in the argument presented here addresses other arguments that might be made against the welfare state, or against particular forms and mechanisms by which the welfare state might be instituted. But, in principle, I don’t see anything in Bryan’s argument that should keep him from joining us in the BHL camp. I’ll be sending him his t-shirt soon.