Making Sense - Sam Harris - #448 — The Philosophy of Good and Evil

00:00:00.000 Welcome to the Making Sense Podcast. This is Sam Harris. Just a note to say that if you're

00:00:11.740 hearing this, you're not currently on our subscriber feed, and we'll only be hearing

00:00:15.720 the first part of this conversation. In order to access full episodes of the Making Sense

00:00:20.060 Podcast, you'll need to subscribe at samharris.org. We don't run ads on the podcast, and therefore

00:00:26.240 it's made possible entirely through the support of our subscribers. So if you enjoy what we're

00:00:30.200 doing here, please consider becoming one.

00:00:36.640 Hi, I'm here with David Edmonds. David, thanks for joining me again.

00:00:40.060 Thanks for having me back.

00:00:41.480 So David, you have a new book, which I really enjoyed. It's titled Death in a Shallow Pond,

00:00:46.820 A Philosopher, a Drowning Child, and Strangers in Need. And you've written a kind of a short

00:00:53.400 bio of the philosopher Peter Singer, who's also been on the podcast several times, and

00:00:58.480 of the effective altruism movement that he has spawned, along with Will McCaskill and

00:01:04.260 Toby Ord, who've also been on the podcast several times. But it's also a great history of moral

00:01:10.580 philosophy in the analytic tradition. So I just want to track through the book, really, because

00:01:16.640 it's, I think the virtues of effective altruism, as well as the concerns surrounding it, are still

00:01:25.180 worth talking about. And I think just the core concerns of moral philosophy and how we think

00:01:33.060 about doing good in the world are really of eternal interest, because it's not at all clear that we

00:01:38.660 think about these things rationally or effectively or normatively in any other way. So, but before we

00:01:44.460 jump in, remind people what you do, because you and I have spoken before and you have your own

00:01:49.040 podcast, but where can people find your work generally and what are you tending to focus on

00:01:54.040 these days? Gosh, well, I had a double life as a BBC journalist and a philosopher. I've given up the

00:02:02.460 BBC bit, so it's all philosophy from now on. I've got a podcast called Philosophy Bites, which I make

00:02:09.460 with a colleague, a friend called Nigel Warburton. And yeah, I now write philosophy books. And I'm

00:02:15.160 linked to a centre in Oxford called the Uhero Institute, which is a centre dedicated to the

00:02:20.420 study of practical ethics, applied ethics. So yeah, those are the various strings to my bow.

00:02:26.300 So why did you write this book? And why did you take the angle you took here?

00:02:31.340 Oh gosh, I mean, there's some prosaic explanations for why I wrote the book. I just written this biography

00:02:36.520 of a guy called Derek Parfitt, who, as it happens, Peter Singer says is the only genius

00:02:41.180 he ever met. And so I was thinking, I had such fun writing that book, and he was such an extraordinary

00:02:47.360 character. I thought maybe I'll have a go at writing another biography. And Peter Singer is probably the

00:02:52.920 most famous philosopher alive today. So I wrote to Peter and said, how about I write your biography?

00:03:00.420 And he said, no, thank you. So then I thought I'd write a book about the history of consequentialism,

00:03:05.860 which interests me. And that would be a book that covered Bentham and Mill and Sidgwick and all the

00:03:12.460 way up to Parfitt and Singer. And then I was sort of daunted by the prospect of that. That was an

00:03:17.520 enormous task. And then I thought what I'll do is I'll cover those subjects just through one thought

00:03:23.520 experiment. I'd written a book about 15 years ago called Would You Kill the Fat Man, which was a very

00:03:29.340 similar kind of book. Again, it was a biography of probably the most famous thought experiment in

00:03:35.840 moral philosophy, which is the trolley problem. And Peter Singer's thought experiment, which we're

00:03:40.060 going to talk about, I hope, is probably the second most famous thought experiment in moral

00:03:44.580 philosophy. But I would say much more influential than the trolley problem. So anyway, that's what got

00:03:50.060 me into the subject. Yeah. Yeah. Well, I think we should start with the thought experiment, which

00:03:56.040 as I've spoken to Peter and other philosophers on this topic before, it'll be familiar to people.

00:04:03.140 But I think we can't assume everyone has heard of it. So we should describe the thought experiment.

00:04:09.060 But before we do, perhaps we can discuss thought experiments themselves for a minute or two,

00:04:15.100 because even the act of entertaining them is somewhat controversial. What's the argument

00:04:21.940 against thought experiments? Give me the for and against what we're about to do here.

00:04:25.860 Okay. Well, thought experiments covers an enormous range of subjects. So there are thought experiments

00:04:32.440 in every area of philosophy. There are thought experiments in the philosophy of mind. There are

00:04:37.000 thought experiments in the philosophy of language. There are thought experiments in epistemology.

00:04:41.840 And there are thought experiments in moral philosophy. And the objections to thought experiments tend to

00:04:47.320 be directed particularly at thought experiments in the moral realm, I would say. So for example,

00:04:54.820 in the area of consciousness, there's a very famous thought experiment called the Chinese room.

00:05:00.740 There's another famous thought experiment when people argue about physicalism and whether everything

00:05:05.400 is physical. And that's a thought experiment called what Mary knew. And on the whole,

00:05:09.740 I mean, they are contentious and they're very heavily debated, but they don't arouse the kind of

00:05:15.980 suspicion, I think, that many moral thought experiments arouse. And the reason that moral

00:05:21.720 thought experiments arouse suspicion, well, there are many reasons, but one is people just say that our

00:05:27.240 moral intuitions are not built for weird and often wacky scenarios. They're built for normal life,

00:05:35.140 real life. And the problem with thought experiments is that they are often very strange, very artificial,

00:05:41.300 and so we shouldn't trust our intuitions. And I would say that was probably the main objection to

00:05:48.200 them. I mean, the response to that is there's a very good reason why they are artificial. The whole

00:05:53.860 point about a thought experiment is you're trying to separate all the extraneous circumstances and

00:06:00.900 factors that might be getting in the way of our thinking. And you're trying to kind of focus in

00:06:07.200 particular on one area of a problem. So you might have a thought experiment where there are two

00:06:12.980 scenarios which are different, except for the fact that one has a particular factor that the other

00:06:18.760 doesn't have. And the point is to try and work out whether that factor is making a difference or not.

00:06:23.400 And often you can only do that if you create a very artificial world, because the real world is

00:06:30.720 not like that. The real world is just full of music and noise and complications. And so the thought

00:06:37.980 experiment is designed to simplify and clarify and try and get at the nub of a problem.

00:06:43.400 Yeah. I mean, it's a kind of conceptual and even emotional surgery that's being performed. I mean,

00:06:49.140 you change specific variables and you look at the difference in response. And again, as you said,

00:06:58.360 this is often, it can often seem highly artificial or unlikely because you're looking for the pure

00:07:04.940 case. You're looking for the corner condition that really does elucidate the moral terrain.

00:07:09.960 I think we should describe both thought experiments here. I think because I think there's an analogy

00:07:13.860 between a common response to the trolley problem and what's happening in the shallow pond as well.

00:07:20.680 So before we dive into the shallow pond, I guess pun intended, describe the trolley problem case and

00:07:27.060 how it's used. Well, the main trolley problem case goes like this. You are to imagine that there is a

00:07:33.260 runaway train. It's careering down the track. There are five people tied to the track. And in the simple

00:07:41.520 case, you are on the side of the track and there's a switch and you can flick the switch and you can

00:07:47.780 divert the train down a spur. And unfortunately, on that spur, one person is tied to the track. So

00:07:55.940 the question is, should you turn the switch and divert the train away from the five to kill the one? 0.83

00:08:02.900 And that was invented by a woman called Philippa Foote in 1967. She was writing about abortion at the

00:08:09.520 time and it was in an article about abortion. And then 20 years later, an American philosopher called

00:08:16.340 Judith Jarvis Thompson comes up with another example. So this one goes like this. You ought to

00:08:21.360 imagine that the train is, again, it's out of control. It's heading down the track. There are

00:08:26.360 five people who once again are tied to the track. This time, there's a different way of saving them.

00:08:31.900 You are standing on a footbridge. You're standing next to, in the original article, it was a fat man.

00:08:40.120 Now for modern sensibilities, it's a man with a heavy rucksack. So you're standing next to a man

00:08:45.440 with a heavy rucksack. You can push the man with a heavy rucksack over the footbridge. And because

00:08:51.280 that rucksack is so heavy, or in the original case, because the man is so fat, he will stop the train

00:08:57.220 and so save the five people. And he would be killed in the process. And the puzzle that Judith

00:09:05.220 Jarvis Thompson asks us to grapple with is that she seems to think that in the first case, you should 0.94

00:09:12.340 turn the train to save the five and kill the one. But in the second case, you shouldn't push the fat

00:09:19.180 man or you shouldn't push the man with a heavy rucksack to save the five at the cost of the one.

00:09:24.040 And that was her intuition. And it's been tested all around the world. It's been tested on men.

00:09:29.900 It's been tested on women. It's been tested on the highly educated, on the less educated. It's 0.99

00:09:34.060 been tested in different countries. And on the whole, almost everybody thinks that in the first

00:09:40.280 case, it is right to turn the train. And in the second case, it's wrong to push the fat man or the

00:09:45.720 man with a heavy rucksack. And so the puzzle in this thought experiment is to explain why. Because in

00:09:50.820 both cases, you are saving five lives at the cost of one.

00:09:54.800 Yeah. And to be clear, the dissociation here is really extreme. It's something like 95%

00:10:02.240 for and against in both cases. But the groups flip, right? So in the case where you just have

00:10:09.820 to flip a switch, which is this kind of anodyne gesture of you're touching something mechanical

00:10:14.440 that diverts the train onto the other track, killing the one and saving the five, 95% of people

00:10:19.740 think you should do that. And when you're pushing the man from the footbridge, fat or otherwise,

00:10:25.880 something like 95% think you shouldn't do that because that would be a murder. And I often have

00:10:32.980 thought that there's a kind of a lack of homology between these two cases because at least in my

00:10:37.720 imagination, people are burning some fuel trying to work out whether pushing the fat man really will

00:10:44.720 stop the train, right? There's kind of an intuitive physics that seems implausible there. But leaving

00:10:48.580 that aside, I think the big difference, which accounts for the difference in behavioral or

00:10:54.680 experimental result, is that when people imagine pushing a person to his death, there's this up

00:11:02.620 close and personal, very affect driving image of actually touching the person and being the true

00:11:12.880 proximate cause of his death. Whereas in the case of flipping the switch, there's this mechanical

00:11:18.800 intermediary and you're not, you're not having to get close to the person who's going to die,

00:11:23.240 much less touch him. And that seems to be an enormous difference. And this is often put forward

00:11:29.120 as a kind of an embarrassment to consequentialism because, you know, the consequences on the surface

00:11:34.400 seem the same. We're just talking about body count. There's a net four lives that are saved.

00:11:38.680 So they should be on a consequentialist analysis, the same case. But I've always felt that this,

00:11:43.760 and I'm sure we'll cycle back to this topic a few times because I think it's important to get this

00:11:48.340 right. I've always felt that this is just a specious version or at least an incomplete and

00:11:53.660 unimaginative version of consequentialism or what consequentialism could be and should be,

00:11:59.400 which is to have a fuller accounting of all the consequences. So if in fact, it is just

00:12:03.660 fundamentally different experientially for a person to push someone to his death and to

00:12:08.820 flip a switch. And if it's different to live in a society where people behave that way versus the

00:12:14.560 other, well, then that's part of the set of consequences that we have to add to the balance.

00:12:20.680 And I think it is, I mean, I think it is obviously different and that's what's being teased out in the

00:12:25.180 experiment. So I now recognize that we should probably define consequentialism in order to continue

00:12:31.680 to this conversation. So anyway, I just lob that back to you and perhaps respond, but also give us

00:12:38.320 a presi on consequentialism. Well, consequentialism is the theory that what matters purely are the

00:12:45.440 consequences. So in these two trolley cases, as you say, the consequences of flipping the switch and

00:12:53.240 pushing the man with the heavy rucksack are the same. If you accept the hypothetical example,

00:12:58.060 which is that one person dies and five people are saved. So if you're a pure consequentialist,

00:13:04.520 it looks like there's no difference between these two cases. So there are dozens of these

00:13:10.860 trolley cases in philosophy. There were dozens of scenarios which involve runaway trains. There

00:13:16.280 are tractors. There are all sorts of things going on in these trolley cases. And it's been given a jokey

00:13:21.180 title, which is trolleyology. The study of these trolley cases is trolleyology. And they've studied

00:13:26.180 precisely the thing that you bring up. So the question is, is really the difference just a sort

00:13:32.100 of emotional difference about pushing the fat man as opposed to turning the switch? So they've tested

00:13:38.400 that and they've come up with a very ingenious way of testing it. So what they do is they ask people

00:13:44.000 the following scenario. Imagine that the man with the heavy rucksack is on the footbridge, but this time

00:13:50.580 you're standing next to a switch. And if you turn the switch, the man with the heavy rucksack will

00:13:56.500 fall through a trap door and will plummet to the ground. And once again, we'll stop the runaway train

00:14:03.360 from killing the five people. Now, if you are totally right about this, what you should get.

00:14:10.440 I see where this is going. I'd forgotten all these iterations here. And I think that it definitely

00:14:16.040 dissects out the up close and personal touchy feel part of it. But what it doesn't change is

00:14:22.260 the fact that the man himself is being manipulated, right? So you're not manipulating the train,

00:14:28.820 you're manipulating the man. And the man is becoming the instrument. His murder is the instrument

00:14:35.300 rather than the effect of the flipping the switch. I think that does seem somehow a crucial difference.

00:14:41.020 Right. So, but that's not a consequentialist difference, right? So what it is, I would just

00:14:46.940 say it is if in fact, I mean, just imagine being these two, in one universe, you flip the switch

00:14:52.320 as 95% of people think you should. And you feel while it was not, not pleasant to do,

00:14:58.860 your conscience is totally clear. In another universe, you flip the switch to the trap door

00:15:04.500 and watch this man fall to his death and stop the train. And you've, you can scarcely live with

00:15:10.160 yourself because of the, you know, the psychological toxicity of having had that experience. That's

00:15:16.580 part, for whatever the, I mean, we can talk more about the reasons why there is a difference there.

00:15:21.020 And I'm, I'm happy to hear all your thoughts on that matter. But if there just is in fact a

00:15:25.180 difference, you know, albeit maybe only in 95% of people, that's part of the consequences. And you

00:15:31.800 can imagine the ripples of those consequences spreading to any society that would make policy of a sort that

00:15:38.960 would, you know, enshrine one behavior as normative versus the other. Right. So I mean,

00:15:43.100 this is what, I mean, there are all kinds of strange examples that are hurled at consequentialism

00:15:48.540 at that seem to be defeaters of it, which always seem to me to be specious. I mean, one, one you

00:15:54.380 actually deal with in the book, which is perhaps the most common one, which is the doctor who

00:15:59.240 recognizes he's got five patients who need organ donations, and he's got a perfectly healthy person

00:16:05.020 in his waiting room, just waiting for a checkup. And he decides to euthanize this person and

00:16:10.540 distribute his organs to the waiting five, saving a net four lives. That seems on, you know, on this

00:16:16.600 narrow focus on body count to be acceptable on a consequentialist analysis. But of course it's not

00:16:21.760 because you have to look at the consequences of what it would be like to live in a society where

00:16:26.540 trust has so totally eroded because we know at any time, even by the doctor who purports to

00:16:33.380 have our wellbeing at heart, we could be casually murdered for the benefit of others. I mean,

00:16:38.900 no one would want to live in that society. It'd be a society of just continuous terror and for good

00:16:43.700 reason. So anyway, that's just my pitch that I've yet to hear, I mean, perhaps you can produce one in

00:16:49.060 this conversation, but I've yet to hear a real argument against consequentialism that takes all

00:16:54.920 consequences, all things considered into account. Right. So in your hospital case where somebody's

00:17:01.780 bopped on the head and their two kidneys and their two lungs and their heart are used to save

00:17:07.480 five patients. So you're obviously right that if that, as it were, got out, then that would be

00:17:14.800 terrifying for everybody. You would never go and visit Auntie Doris in the hospital because you'd

00:17:20.340 think, well, there's a risk that when I go and visit Auntie Doris, the same thing is going to happen

00:17:25.120 to me. Of course, what the philosopher does is they then create a hypothetical example that

00:17:30.340 just a one-off case. Yeah, it's a one-off and nobody finds out about it. And the person has

00:17:34.700 got no friends and blah, blah, blah. But again, the response to that is, well, we can't really

00:17:39.860 imagine that, you know, our intuitions aren't really coping with that really kind of cocooned example.

00:17:46.580 We're imagining that this news is going to leak out. In the trolley case, I think it's much more

00:17:54.160 complicated. It is true that people would find it more difficult to live with themselves by pushing

00:18:01.260 the fat man or by dropping the fat man through the trap door. But the question is why? And I think the

00:18:08.140 explanation is that people have one very powerful non-consequentialist intuition. And it goes

00:18:16.220 something like this, although they don't articulate it and they're very puzzled by this

00:18:21.540 thought experiment. If you put the following to them, they think, yes, this explains my intuition.

00:18:27.580 So imagine that you push the large man from the footbridge and the large man is wearing a rubber

00:18:34.720 suit. And instead of dying, he bounces off the track and he runs away. So what's your reaction to

00:18:41.720 that case? Your reaction to that case is that's not good because the whole point of pushing him over

00:18:47.980 was that so he got in the way of the train so that he would save five lives. Now imagine that in the

00:18:54.860 first case, the train is going along and it's going to kill the five people and you flick the switch and

00:19:00.540 it goes down the spur. Now imagine the person on the spur is able to extricate themselves from their

00:19:07.080 ropes and able to run away. How would you feel about that? Well, you'd feel absolutely delighted.

00:19:14.040 And why do you feel delighted? Because you haven't killed the five and you haven't had to kill the one.

00:19:20.220 So the difference between the two cases, and this comes back to the doctrine of double effect that

00:19:24.580 goes all the way back to Thomas Aquinas, is as you hinted at earlier, in the fat man case,

00:19:30.240 you are using the fat man as a means to an end. And that's not the case with the spur case. Another

00:19:37.380 way of putting that is you intend to kill the fat man when you push him over the foot. You want to 1.00

00:19:43.340 kill him. Well, you need him to get in the way. You don't intend to kill the person on the spur. 1.00

00:19:48.740 Hmm. Well, yeah, it's interesting. Well, it's, um, I'm not sure I totally buy that. That all turns

00:19:56.320 on there being an important difference between acting in a way where it seems there's a hundred

00:20:02.680 percent chance of killing a person, but still there being, it being true to say that you don't intend

00:20:08.820 to kill the person. I think it's the key distinction is the distinction between intending

00:20:14.500 and merely foreseeing. So it's the distinction in the Geneva Convention between attacking a

00:20:21.560 munitions factory. It's a collateral damage issue, right? Exactly. Exactly. Exactly. So

00:20:27.080 attacking the munitions factory, knowing that a hundred civilians will die, but this munitions

00:20:31.980 factory is so important to the enemy's war effort that the attack on the munitions factory is justified,

00:20:38.360 even though you know that a hundred people, a hundred civilians will die. It's the difference

00:20:42.480 between that and intentionally targeting those 100 civilians. So, yeah, I think I, I misspoke

00:20:48.780 a moment ago. I do clearly see that distinction. I guess it's the, um, let me see what's, what's

00:20:55.240 bothering me about this. Well, perhaps it'll come out just in further discussion here around, uh, the

00:21:00.480 other thought experiments. Well, let's talk about the shallow pond and kind of fill in more of this

00:21:05.100 picture. And, uh, I think we'll, we'll cycle back on whether consequentialism has any real retort

00:21:11.800 because I, because you said a moment ago that this is a, this was a non-consequentialist, uh, intuition

00:21:16.980 and, um, my deep bias here, I'm, I'll be happy to be, um, disabused of it, but my deep bias is that

00:21:24.360 when you drill down on, on any strongly held intuition that pushes our morality around and we

00:21:30.580 can't shake it, it is either at bottom, some intuition about consequences, about, you know,

00:21:35.660 what it would mean to live in a world where this kind of rule was repeated. So it's kind of a rule

00:21:39.680 consequentialism rather than an, an act consequentialism per se, or we just have to bite the

00:21:45.020 bullet and admit that, okay, this is a, an illusion. It's some kind of moral illusion, right? So, um,

00:21:50.980 I mean, there, there's so many things that we could care about as we're about to see and, and,

00:21:55.440 and magically don't care about. And it's, um, it is inscrutable that we, even when they're pointed

00:22:01.240 out, we don't feel differently. I mean, the one that always comes to mind for me is, you know,

00:22:05.640 if we just changed our driving laws just slightly, I mean, just to slightly inconvenience ourselves,

00:22:10.780 I mean, that we made the speed limit 10 miles an hour lower on every street in the nation. I mean,

00:22:16.320 so just speaking of, of America here, where we have 40,000 traffic deaths a year reliably, uh, and I

00:22:23.040 don't know how many people are maimed, but you know, 40,000 people are killed outright based on

00:22:27.100 how badly we drive. If we just reduce the speed limit by, you know, let's say 10 miles an hour,

00:22:32.860 we would save thousands of lives. I think there's no question of that. And it's just the real,

00:22:37.740 the only real consequence. I mean, I'm sure you could, maybe a few people would be inconvenienced

00:22:42.080 in a way that might prove fatal, but it certainly wouldn't, it would be massively offset by the number

00:22:47.240 of lives saved. The real consequence would be that it would be less fun to drive, right? Or,

00:22:52.820 or we could actually, I mean, even to make it more, um, inscrutable still, we could put governors

00:22:58.020 on all of our cars that, you know, so for whatever car, you know, from a Ferrari on down could never

00:23:04.540 exceed the speed limit, right? You could drive however you wanted, but you could just never

00:23:08.100 drive faster than the speed limit. That's technologically feasible. No one would want

00:23:12.120 that no matter how many lives it would save because it would be less fun to drive. Uh, somehow we want to,

00:23:18.000 we want to somehow carve out the possibility of driving faster than the speed limit. Uh, at least

00:23:22.540 sometimes. And yet when you talk about that body count, nobody moves from that point to the obvious

00:23:30.540 conclusion that we're all moral monsters for so callously imperiling the lives of everyone,

00:23:37.480 including our own, really. I mean, we, there's no identifiable victim in advance. That's part of the

00:23:41.460 problem, I think. But I mean, there's thousands of people, 40,000 people are guaranteed to die this

00:23:46.540 year in America based on the status quo. How is this acceptable and how are, how are we not

00:23:53.320 monstrously unethical for accepting it? Uh, and somehow the, the sense that there's even a moral

00:23:58.780 problem here evaporates before I even get, can get to the end of the sentence. Yeah. So 40,000 is a lot

00:24:03.740 of people. I think there were 58,000 killed in the whole of the Vietnam war, right? So that's,

00:24:08.300 that's, that's a big figure. And oddly in London, in much of London now, they've reduced the speed

00:24:14.380 limit to 20 miles an hour. If you'd like to continue listening to this conversation, you'll

00:24:19.360 need to subscribe at samharris.org. Once you do, you'll get access to all full length episodes of

00:24:25.220 The Making Sense Podcast. The Making Sense Podcast is ad-free and relies entirely on listener support.

00:24:31.200 And you can subscribe now at samharris.org.

Making Sense - Sam Harris - December 08, 2025

#448 — The Philosophy of Good and Evil

Episode Stats

Summary

Transcript