There Is Exactly One Good Objection To The Drowning Child Argument

But it won't let you off the hook completely

Peter Singer’s drowning child argument is probably the single most influential thought experiment in human history. It’s led to lots of people giving money to effective charities, and also lots of people feeling guilty about not giving more. Fortunately for those who don’t want to give all their money away, there is a not totally implausible response. Unfortunately, it won’t let you entirely off the hook.

Singer’s argument goes: imagine that you came across a child drowning in a pond. To save the child, you’d have to jump into the pond. But if you did so, you’d ruin your $4,500 suit and get all wet and muddy. Nearly everyone has the intuition that you’d be obligated to save the child.

But then Singer asks: how is this different from failing to donate to effective charities? The intuitive explanation of why you should save the child is that we have an obligation to prevent very bad things from happening if the cost of doing so is small in comparison. When one donates to effective charities, a life is still saved and the cost is the same as when one pulls a child from a pond. The best charities can save lives for just a few thousand dollars. So what’s the difference? If we’re obligated to ruin a $4500 suit to save a child in a pond, why aren’t we obligated to pay $4,500 to save the lives of far-away children?

There are lots of answers people give that don’t hold up. Some say the difference is proximity. The child in the pond is closer to you. But how physically near someone is to you doesn’t affect your obligation to save them. Imagine that you were in a train traveling away from someone. You could press a button that would save the person’s life, but it would cost a few thousand dollars. Your obligation to press it wouldn’t weaken as the train got farther away from them. Learning that the train moved unexpectedly quickly, so you’d traveled farther than you thought wouldn’t weaken the moral case for pressing the button.

Similarly, imagine in the drowning child case that the child is very far away, but you have really long arms. The child is on the other side of the world, and you could pull them out of the pond with your long arms, but doing so would ruin your expensive watch. It still seems like you’d be obligated to do that. If a child was trapped in a pit, would you have no duty to pull them out, so long as the pit was deep enough to make them far away from you?

Lastly, imagine that there was a child drowning 1,000 miles away. To save them, you’d have to jump into a pond to press a button at the bottom, but doing so would ruin your expensive suit. Surely you’d be obligated to do that, even though the child was far away. And surely the strength of your reason wouldn’t depend on how far they were. It isn’t as if you’d be obligated to do it if they were within a 500 mile radius, but not a 1,000 mile radius.

(Claude didn’t do a great job drawing continents. They all look like potatoes).

Similarly, people sometimes suggest that the difference is that we can see the child. We are more personally connected with the child drowning in front of us than the distant child. But even if you were blind, and consequently unable to see the child, you’d be obligated to pull her out of the pond. Similarly, you’d be obligated to jump into the pond and ruin your suit to save a child who was two miles away, even if you couldn’t see her.

Sometimes people suggest that the difference is that the child is one of your own countrymen. But you’d be obligated to jump into the pond to save a foreign child, so that can’t be the difference. If I saw the child was Canadian, that wouldn’t absolve me of an obligation to save them.

Others suggest that the difference is that in the drowning child case you’re the only one who could save the child, while anyone can give to charity. But if there were other people sitting by idly watching and doing nothing while the child drowned, you’d still be obligated to pull them out. This is also an odd defense; it would be surprising if everyone could defend their own inaction by noting the inaction of everyone else.

The core problem with these responses is they focus on morally irrelevant features of the situation. What matters in saving someone’s life is the cost to you and the benefits to them. The reason to save the child is that you can prevent a death at small cost. Features like how close the child is to you or whether you can see them are paradigm examples of morally irrelevant features. If we intuitively think about what determines whether you should save someone’s life, factors like how far away they are just don’t seem to matter.

While donating to effective charities may not emotionally strike us the same as pulling a child out of a pond, in terms of the costs to you and the benefits to others, it is the same. And those are what matters. While you can try to come up with a gerrymandered account to fit our intuitions about both of the cases, it will inevitably end up caring about something that isn’t ultimately morally important—or is only morally important to a small degree.

As I say, these are bad responses to the drowning child argument. But there is one good one, developed by Travis Timmerman. It dampens the extreme Singerian position, but it still means we should give more than we tend to.

Timmerman’s paper is bravely titled Sometimes there is nothing wrong with letting a child drown. In it, Timmerman grants that failing to donate is like failing to pull children out of the pond. But he argues that we’re not really required to always pull the child out of the pond.

Imagine if rather than a single drowning child, there were endless drowning children. Every day you walk past the pond and see more children. There are always more than you could possibly save. Each time you save one, it cost you money.¹ In such a case, would you be obligated to spend every waking hour and every last cent pulling children out of the pond? A plausible intuitive answer is no. We’re obligated to do a lot to help others, but not spend our entire life doing it.

I don’t know if this response is ultimately satisfactory. I think there are some serious objections which I’ve laid out here that might mean we should give more than Timmerman suggests. But it’s not a crazy response. If we’re not obligated to spend all our time and money saving the children from the pond, this would be why. If you find the drowning child argument too crazy to accept, then this is the best response by far.

But suppose you buy this response? Are you off the hook? Can you simply go on with your life, never paying another thought to donations to effective charities? No! If there were endless drowning children, you wouldn’t be obligated to spend all your time and money saving drowning children, but you’d be obligated to spend some of it saving them.

If there were endless drowning children, then saving them ought to be a significant part of your life. If you never saved any drowning children, then you’d be failing in your duties. If you’d save a single drowning child if there was just one, then after learning there are many, you should still save at least one.

How much are we required to give? I don’t know that there’s some precise obligatory amount. But you’re obligated to give an amount that isn’t trivial. If children are drowning all the time, pulling them out ought to be one not totally trivial part of your life. A reasonable middle-ground is giving something like 10%.

So if you buy the best objection to the drowning child argument, you still ought to take the Giving What We Can pledge and give away 10% of your income to effective charities. And, in fact, the case for it might even be stronger than for pulling the children out of the pond at personal cost. It seems very likely that taking the pledge makes people’s lives better and more meaningful. If you knew that, say, pulling 100 children out of the pond wouldn’t just save 100 children but would also make your life happier and more meaningful—that it wouldn’t even be a cost—then you’d be obligated to do that.

In practice, it doesn’t really matter whether you buy the Timmerman response. Even if Singer is right about our obligations, basically no one ends up following them and giving the full amount Singer says they ought to. In practice, whether you think Singer is right or Timmerman is, the thing to do is give away some amount of your money to effective charities—the more the better! 10% is a reasonable heuristic, though more is always better.

Timmerman claims that sometimes there is nothing wrong with letting a child drown. But even if you buy it, even if you are sometimes allowed to let a child drown, nearly everyone errs in letting too many children drown. Even if the best objection to the drowning child argument is right, we generally ought to give much more than we do.

As I was writing this post in a public space, a small child who I’d never seen before waddled into the room I was in. Hearing him laughing and babbling in the adorable way that little children do, the common practice of doing approximately nothing about the thousands of children dying daily seemed so grotesque.

1. Timmerman imagines a case where you discover this endless trove of drowning children when on your way to the bank, where with each moment spent saving children, you lose thousands of dollars. But in that case, saving the child is a one-off opportunity, which is somewhat different from normal life. So I think a better analogy is if there are endless drowning children but each time you pull them out, doing so costs you money.