Rationally speaking, is a presentation of New York City skeptics dedicated to promoting critical thinking, skeptical inquiry and science education. For more information, please visit us at NYC Skeptic's Doug. Welcome to, rationally speaking, the podcast, where we explore the borderlands between reason and nonsense. I'm your host, Masimo, and with me, as always, is my co-host, Julia Gillard. Julia, what are we going to talk about today?

Masimo Today we're going to talk about the thought experiment, the process of constructing an imaginary situation to help shed light on how things really are.

It's an indispensable tool in the philosopher's toolbox, and it's even got a long history in science as well. But while some thought experiments help clarify an issue, others are totally worthless and can even be misleading. So I'd like to talk about some examples of thought experiments in science and philosophy and ask what are they good for?

And what distinguishes a valid thought experiment from an invalid one?

Good question. By the way, before we go ahead, I'd like to say that the what you just said about five experiments, which is that some of them are, you know, interesting and bring up interesting conclusions or points and other on the other hand, they're misleading or downright wrong. Pretty much.

That applies to real experiments as well. Right. And if I would be kind of interesting to figure out whether it applies to real experiments for similar reasons, or is it because a completely different structure of the two?

I actually think there are some similarities in the way that real experiments and experiments can fail. But let's leave that as a bit of a teaser and get to that in a bit. For starters, maybe we should talk about the more successful thought experiments, because there have been some some pretty some pretty illuminating thought experiments that have been used. You brought up one of my favorites in the teaser for this episode, which was Galileo's thought experiment, showing that objects of different mass will fall at the same rate.

Right. So Galileo was faced, of course, with with Aristotelian physics. And according to Aristotle, you know, bodies would fall at a different speed depending on the on the mass. And we know today that that is false. And the first one to point out that that was the case was not allowed. And the interesting thing is they didn't do it by actually doing any experiment. I know that the mythology is that he actually went on the Leaning Tower of Pisa and dropped weights of different bodies of different mass and all that.

He never did.

But this is even cooler, right? Exactly. This is even better because you figured the whole thing just by thinking about it. And then many, many years later, during the Apollo 15 mission on the moon, the experiment is actually done in front of cameras and broadcast around the world. And sure enough, it turns out that Galileo was, in fact, right. And of course, the reason they did it on the moon is because there is no friction.

And so you don't have to deal with the interfering effects of the atmosphere. But essentially, Galileo thought in this way, said, well, look, there is there's supposed to be two bodies.

One is having the other one is, you know, light. And according to Aristotle, they should fall down on different different speeds.

Now, what happens if you join them somehow with a court or something? Now, from a U.S. perspective, you could think about the the combined body into different ways. On the one hand, you could say, well, the combined body is now heavier because it's got the heavy body plus the light body social issue fall faster, on the other hand, by the same logic. It's also true that the light body is going to drag, you know, slowly, more slowly the heavy body.

So it should be actually slowing it down. If you put the two things together, you get a contradiction, right? And if you get a contradiction, Galileo reasoned, then that that can't be true. And as a bonus, you got the real by reductio ad absurdum by pushing the argument to the to the extreme limit, you got the right answer, which is that the two bodies will fall. In fact, that the same and the same rate.

Now, the reason this is a particularly successful example of a thought experiment is because is one of those instances where we really did gain knowledge about the universe, empirical knowledge about the universe, but just thinking about it, because if you think about it, the Galileo's conclusion is not a matter of logical necessity.

You know, if the law gravitation were different, things would be working differently.

And he did not have any data. You know, he just made up the hypothetical situation is not only be studied out of empirical data.

So this is a great example where starting out with no data and using simply intuition and logical reasoning, the guy came up with something entirely new, which we still know today, 400 plus years later, to be true.

Yeah, I think that's one of the most exciting kinds of thought experiment where your conclusion is, is a new and accurate fact about the external world.

But there are other kinds of totally valid thought experiments that have a different purpose, I would say, especially the kind used in philosophy. And there I think the purpose is more along the lines of helping us clarify what we actually believe.

So, for example, pointing to inconsistencies in multiple things that we think we believe simultaneously. So, for example, if I think that I believe that there's a moral obligation.

To save someone's life, if it would be a small cost to myself, and I also believe that I'm not doing a moral wrong by not giving away a lot of my income to charity, then a thought experiment such as the sort promoted by Peter Singer could highlight that inconsistency and encourage me to resolve it in one direction or another.

That's right. So the experiments for experiments like the trolley dilemma and in philosophy, in ethical philosophy are interesting, not because they want to.

Their point is not to prove their particular position, but it's to explore how, you know, what what are the assumptions that are embedded into our reasoning on the trolley dilemma, as most of our listeners probably know about the situation where it really is going down the tracks and it's out of control and you have a chance to say five people that are about to be hit by the trolley if you pull a lever.

But if you pull the lever, you're going to hit another innocent person on a different track. Now, when posed in front of that kind of situation, most people would say that, yes, they would pull the lever. But when you present them with a different situation, that it's that it's only formally different, at least from an ethical perspective, which is instead of pulling the lever, they would actually had to push somebody in front of the tracks to block the trolley.

Most people say they wouldn't do it. And the two situations are, from an ethical perspective, identical, because in both cases you're saving five people and you're killing one. But the different reaction most people have highlights the fact that there is an important emotional component to our ethical reasoning, because the only difference there is that instead of doing these rather impersonal thing or pulling a lever, you are actually doing the very personal thing or pushing somebody else.

Right. And I actually wanted to talk about this thought experiment because I think it highlights a potential danger with thought experiments, and that's that they can tempt you to specify conditions in your experiment that would never actually occur in the real world.

And so it's not clear to me that you can trust the conclusions that the experiment with its imaginary hypothetical conditions leads you to. So, for example, the thought experiment specifies that when you push this man, they always say the fat man because because a lot of people try to say, well, why don't I just jump on the tracks to save the kids? And so the question says, oh, no, no, you wouldn't have enough math. So you have to push this fat man the only way to save the kids.

And if you push this guy, then that will definitely kill him and will definitely save the kids.

But but the problem is, in real life, you you could never be sure of that. You could be confident that pushing the man would kill him and would stop the train. But you can't be 100 percent certain.

And so I think there's a reasonable case to be made that that sliver of doubt about what could actually happen is enough to justify staying out of it. You could push the man and the train could kill him and still go on to kill the kids. And so there's some wiggle room there.

And I think this this is also the problem with a much more common and a weighty thought experiment cited in politics, which is the ticking time bomb scenario, which people bring up when they're they're arguing that torture is sometimes justifiable.

And they say, well, if you knew that torturing this one, this one captive, was the only way to find out the location of a bomb that was going to go off soon and kill thousands of innocent civilians, then, well, then is is torture still wrong? And and phrased like that?

It is a tricky question. But but I think one of the best arguments against the unexpected is that in real life, you never know that torturing the men will definitely save those people's lives. And so, again, maybe that bit of doubt is enough to justify refraining from Georgia.

I'm going to I think that's a very good point. But I'm going to slightly disagree with your take on this. And I'm going to try to make a distinction.

I think that the point of those thought experiments, such as the trolley dilemma is in fact to explore people's intuitions or people's reasoning or people's assumptions about, say, moral reasoning. It's not to actually predict what people would do.

And in practice, just like you did the experiment Galileo did.

The point of that one wasn't to predict what would actually happen if you did drop two different, you know, bodies of different mass from the Tower of Pisa and in fact, had actually done that experiment. It probably would have failed because of the interference of of air, which would cause friction on the bodies. So in other words, those kind of experiments in which you're referring to these simplifying assumptions and very strict conditions, that's no different from in a failed experiment in physics.

And in fact, a lot of physics, even even at sort of basic introductory level, we study physics and we talk about frictionless surfaces or, you know, or things that are ideal gas, things like that.

Well, ideal gases don't exist and friction surfaces don't exist either.

But the reason those experiments work is not because they predict the behavior of actual surfaces or bodies or gases, is because they lead you to think about how under control conditions those things would actually work.

So I think that the situation is analogous in the case of our experiment. Now, let me get to your example of the ticking bomb scenario, because I think that one draws out a distinction that is important there. If the question were really purely a matter of what our moral intuitions are as opposed to being a question of actual policy. I think that question is a valid question.

You know, what do you think? If you were faced with that sort of situation, what would you do?

The trick is that, of course, as soon as you answer that question, I said, well, OK, maybe under those are those very special conditions. I think that it would be justifiable to do whatever is possible to save a million people or something like that.

As soon as you met that in response to a thought experiment for an ideal situation, the person that typically prompts you with that kind of question draws the conclusion that therefore it is acceptable policy to to conduct, you know, engage in torture and then actual real situations.

But that wasn't the situation that is covered by the thought experiment.

There's a very important distinction between, you know, trying to probe somebody and push to the limit somebody's moral intuitions. And it's a completely different thing to use those moral intuition and to say, well, therefore, in an actual policy matter and the real conditions here as well, what I'm going to do, that's that's, I think, a very different situation. It would be very much like a scientist who works on the theory of of ideal gases to try to use that theory to predict a non equilibrium thermodynamics.

It wouldn't work.

So and that's a very good practical reason to avoid actually using these thought experiments in policy decisions.

I just meant that if the conclusion that someone's gut reaction is irrational or contradictory in those cases may not be justified because their gut reaction may be based on the fact that in real life they would they it would be better to refrain just because you can't have absolute certainty that that's all I meant.

And we should also probably talk about two of the other big problems that can arise in thought experiments, which is I would say first, the problem that they can fool us into thinking that just because we can conceive of something, then it must be possible. And second, that just because we can't conceive of something, then it must not be possible. Right. And you give a good example of the latter in the eternity's or in an article that you wrote for Philosophy magazine, but about the I think the Lucretius who argued that the universe must be infinite because if you threw a spear at the border, the boundary of the universe, and it went through, well, then it couldn't be the end of the universe.

And if it bounced off, then there must be something on the other side of the because there is a boundary between.

Right and something else. That's right. That's actually apparently the earliest recorded thought experiment and the sporting effort.

And it's a great no, it's a great example because you're right, it does reach a conclusion that turns out to be wrong, because today we we think modern physics thinks that the universe can both be bounded and infinite. You know, just think of asteroid, for instance, or even in in one dimension to the mention, just a circle. Obviously, that's a that's a figure that is doesn't have an end. And yet it is, in fact, you know, confined to a particular it's bounded.

So it was definitely wrong.

But it's an interesting spin both in terms of sort of historical importance and also because it started out a tradition of other thought experiments that eventually went wrong, including in physics. I mean, Newton, for instance, made a proposal, a famous experiment involving buckets and rotating spheres and things like that to show the space is absolute.

Well, it turns out he was wrong and he wasn't Newton.

He wasn't exactly the, you know, somebody who didn't know anything about physics. Einstein, on the other hand, proposed his famous thought experiment about what would happen if we if we were to travel, you know, try to catch up with a light wave and that show. And in fact, there was essentially the beginning, the foundations of the basic idea of relativity showing that, in fact, in the frame of reference in which the observer would find itself would be in accordance to a relativistic view of the universe and not Newtonian the other universe.

So again, we're bringing example examples where thought experiments worked and examples where they didn't work.

But so far, I think it seems to me fair to say that that's true also for actual experiments. And and, you know, if far expand, it is an interesting question to ask what kind of thing a thought experiment is.

And there are several theories in philosophy of science about what experiments and where they come from, two that are kind of opposite extremes and are worth considering because they sort of encompass the whole the whole spectrum are due to James Robert Brown on the one hand and to John Norton on the other hand.

So according to James Robert Brown, if thought experiment is essentially a way in which we perceive a platonic reality and by platonic reality, it means.

That that the the laws of the universe, as well as mathematical constructs such as numbers, are really out there independently of the human mind, meaning that when we when we do when we think about mathematics and fundamental physics, we actually discover things. We don't invent them.

Now, we could do a whole separate show about Platonism, especially Platonism in mathematics, because it's a really interesting situation.

But it sounds like a bizarre notion until you start thinking about experiment that we started with Galileo's figuring out something about the real world without new data. And and, you know, and we don't know and actually no actual experiment. That's an interesting example, because what what was he what was he getting those ideas from? If if if in some sense the answer wasn't already out there, it simply certainly didn't make it up. It's not something it's not a construct of Galileo's mind.

It's something he discovered it was out there. It's a it's an actual structure characteristic of the physical world. So how did you discover it?

By not having any empirical realized it.

I mean, we frequently we have a lot of information rattling around in our minds. And frequently there are logical dependencies between different statements. But we don't realize that until we think hard about it. I don't think that that's as mystical as some philosophers seem to think it is. Now, I don't think this has anything to do with mysticism, but.

But things are weird, right? But your objection goes, in fact, in favor of the other view. The one by Jonathan. Jonathan thinks of experiments as if he is as a kind of if then reasoning. So he thinks of our experiments as something that starts out with certain implicit or explicit empirical premises, which we get from previous observation or experience.

And then we construct a set of if then situations where we explore by deductive reasoning what follows from those premises.

I assume that that's a more palatable view of what field experiments are.

And if you think about it, it is very similar, at least in my opinion, to the way in which a very different kind of activity works in science. And I'm referring to a computer simulation. So a computer simulation is something that in in a qualitative fashion, really works a lot like a thought experiment of the if then type that Naughton is talking about, because, you know, computer simulation, you start out with certain premises that you used to build in to build the program or whatever it is you're simulating.

Those premises typically come from from knowledge that is external to the simulation itself is not generated by the simulation. So those are premises may come out of, you know, knowledge of the laws of physics or certain assumptions about economics or empirical data or whatever.

And then what you do is you write a computer program that is essentially a set of if then statements and feedback loops that by deductive reasoning, quantitative deductive reasoning draws out the consequences of those assumptions. Right. A computer submission doesn't produce new data.

I mean, it's true that we often call it call the output of a computer simulation data, but there their data in a very different sense than the empirical information that we had about from real experiments about the real world. Right. These are data that are actually generated by this, by the simulation, and they are illogical and in fact, necessary consequence of the premises of that simulation. So the idea is the same in experiments. According to Norton, what we do is very much like a qualitative simulation in our mind.

We start out with certain premises and then we build a certain kind of deductive reasoning. And of course, just like in a computer simulation, there are fundamentally two things that can go wrong with a thought experiment. Either the premise could be wrong. So the reason it can be perfectly fine.

But you start out with wrong premises and of course, is anybody that is taking logic why no one knows if if you if you use deductive reasoning starting, I would Bronk premise is the conclusion cannot does not in fact follow or vice versa.

You can start with the correct premises or a set of correct premises, and then you can make a mistake in the actual reasoning, which will be the equivalent of making a mistake in the programming of the simulation, in which case, of course, also you cannot trust the output, but in this case you cannot trust the output for a very different reason. So if you want to take issue with the results of a failed experiment, you would go about the same way in which you go about if you want to take issues with the results of a computer simulation, you do question the premise of your question, the actual program.

I like it at the it's not quite as charming as chucking a spear at the edge of the universe, but something more precise, I wanted to make sure to bring up a comment from a reader of our blog that I thought was particularly perceptive.

It has to do with the one of the earlier waves that I mentioned that experiments can go awry in which they us and thinking that just because we can conceive of something, that it must therefore be possible. So commentor says, it seems to me that most of the time when we believe we can conceive of something, we're simply fooling ourselves. We have an image of an object or a scene in which we can use the term in a seemingly meaningful manner.

But in fact, we have no precise idea at all of what we're talking about.

So assuming we can conceive of something doesn't seem to amount to much. And I think I totally agree with this.

And I think a great and a common example that we've brought up in previous episodes is the philosophical zombie, the idea that we can imagine a creature who's an automaton whose brain processes or whose robot brain processes information from the world exactly the way ours does, and who acts and behaves the way we do, but is not at all conscious.

And as I believe you and plenty of other smart people have pointed out, this this might just be fundamentally inconsistent. So even though we think we can imagine, it might not actually make any sense.

I think I think we actually we may not be conceiving of things when we think we're conceiving of them. Like, I actually think we can our brains can fool us into thinking that we're imagining some things, though, like I could even something as stark as a square circle.

I feel like a lot of people could say, oh, I can sort of imagine that. But maybe what they're imagining is just a circle with, like, the word square floating around in their mind, you know, and that's not actually imagining it because it's logically impossible right now.

As you pointed out, I don't have a particular opinion of the zombie thought experiment in philosophy mind, which is due to David Chalmers.

On the other hand, there are, in fact, some interesting experiments, even in science, that are about cognitive ability.

So whatever fundamental physicist, for instance, talk about what the universe will look like if if the physical constants were slightly different from what they are, which is a typical way of reasoning about the foundational foundational physics. Well, that's perfectly legitimate.

So the question arises in this case is one that has been discussed recently in philosophy science, which is the relationship between cognitive ability and possibility.

And you know, it they're not exactly the same thing. Clearly, we can conceive of things that actually not possible, but of course, we can conceive of things that are possible and is and we shouldn't limit ourselves to not, you know, not doing that. But when we need to be careful about what we draw out of as a conclusion of these experiments for experiments. Let me give you an example from philosophy, which I thought it was very amusing.

And it's exactly one of these situations where the conclusion of the experiment hinges on the unconceivable to do. And here's two major philosophers who have very different ideas about what's conceivable. So Frank Jackson famously proposed this experiment, these experiments, to show, in his opinion that there must be something wrong with physicalism.

A physicalism is the idea that the mind and consciousness and reduce only two physical characteristics.

And so he said, OK, I imagine that there's this woman named Mary who grew up in an environment, an experimental environment that is on purpose. It's kept black and white or shades of grey. There's no colours anywhere to be seen. So she has never had an experience of color. Throughout growing up, she has been told about color by science. The scientists are running the experiments and it's obvious that the scientists in this case would be engaging in something very unethical behavior.

But nevertheless, you know, human subjectivity is rather of you cannot go to jail for the experiment, I suppose.

But so so the idea that Jackson put forth was that, well, look, then suppose that somebody then at some point opens the door and Mary steps out of the of the confines of an experiment and for the first time experiences color.

She would have no idea of, you know, what to expect.

In other words, the experience would be very different from whatever she could possibly have learnt and has learned about color, which shows that there must be something more than just the physical descriptions and physical understandings of color.

OK, I don't buy that that even at that point. But let's assume that you buy the do something fun, in fact, about this situation.

When Daniel Dennett, another philosopher of mine, responded in this way, said that this, quote, experiment, in fact, you propose a thought experiment that starts out just like Jackson's.

But then when Mary leaves the lab, she says, ah, color perception is just as I thought it would be.

In other words in other words, the physical description given to her by scientist prepared to exact her exactly for the Dayaks of, you know, sensorial experience.

And what you take what you get out of this experiment is based on your assumptions. Do you think that the physical descriptions of a particular. Kind of experience is the same as or is as good as the actual experience or not, depending on what you conceive of the relationship between physical experience and knowledge about the physical experience, you draw completely different results from the experiment.

Yeah, you've made this point before, and I thought it was very well taken. And I think the original thought experiment illustrates another big danger thought experiment, which is just that if you tell a convincing and believable enough story, then you know you can.

It's very easy to accept the conclusion without questioning the logic of the thought experiments.

Unless you're somebody smart, like then it actually comes up with a counter for an experiment. So that's that's the whole point, right, of these things. You put them out there so that people can actually look at the assumptions and the reasoning and point out where the floor is.

And there are so many that experiments that we're not getting a chance to bring up. But but I understand that you have a few more waiting for us in the picture. So let's hurry up and move on to the rationally speaking, PEX.

Welcome back. Every episode, Julie and I pick a couple of our favorite books, movies, websites or whatever tickles our irrational fancy.

Let's start with Julia Spicks. Thanks, Massimo. My pick is a book called 50 Great Myths of Popular Psychology. I've been enjoying this book a lot because I think like many people, I'm very vulnerable to urban legends and and spurious citations are spurious anecdotes from the world of popular psychology.

Just because they make such good sound bites that I often don't put in the hard work of actually checking them the way I said, because they sound so good and they make for a good retelling. But so, you know, some of the myths in this book, I was well aware remits, but others I actually believe to be true or at the very least had heard and had not rejected out of hand when I'd heard them.

So, for example, I thought it was quite possible that just as I'd always heard, expressing your anger or aggression will help dissipate it and that you'll then become less angry.

But apparently this is really not the case. For more than 40 years, apparently studies have been showing that encouraging the expression of anger actually increases your aggression. So one study showing this had people pound nails and do various other activities to to express their aggression after someone had insulted them and then found that the people who had done that were actually more angry later on when when measured in a follow up. So that one of the other fun things about this book is just reading about all the clever experiments that the researchers divide to to test these popular believes, popularly believed myths.

And it's also striking to realize how many of these statistics and studies that I heard about, they keep sort of rattling around the popular psyche or actually completely false or fabricated, like, for example, the the study, which you probably heard about. I've heard about it a million times in which showing the power of subliminal advertising. So this guy in a series of movie theaters and I think New Jersey had subliminal messages, flashing ads for Coca-Cola and for popcorn in movie theaters.

And the result was that sales of Coke and popcorn skyrocketed during the six weeks that this experiment was going on. Turns out the guy made it all up, but it's just like become permanently embedded in our collective psyche.

The story's too good to be true. Yeah, exactly.

And so you think that to some extent, the the problem here is that a lot of pop psychology and self-help books and all that sort of stuff, they're not actually based on psychological research, which is out there, because a lot of people actually look into these things.

They're just based on intuitions. Right? Somebody just turned down saying, well, it makes sense to me that we should, you know, deal with this way with anger.

And so it starts out writing books about how to deal with anger without actually even bothering to check in the literature. I mean, one of the things that struck me some time ago is really interesting is that there is actually research looking into different kind of talk therapy and how effective they are on the long term to change people's behavior.

And it turns out that the only one that we know empirically is effective is cognitive behavioral therapy.

Right now, the interesting thing to me was that then therapists were told this and for the large part, they completely ignore it responding that well. But I know it works because I see it working on my on my clients, thereby entirely missing the point of a controlled experiment in science. And this is actually does a harm to people.

At least you're wasting your money if if you keep going to a therapist who refuses to take into consideration empirical evidence. Right.

And for example, one of the myths that has persisted is that people who are sexually abused as children are all or almost to a one suffer from severe psychological problems as adults. And certainly some of them do. But the statistics are much weaker than you would actually have thought. And so just as you said, when when psychologists and therapists were told this, they insisted that it didn't match what they saw on their practice. But, of course, the only people they see in their practices are the ones who have psychological problems.

So it's not random sample, right? Exactly. What's your pick mathematically? So my pick is John Norton's website, and in particular a subset of it called Goodis. Jonathan is a philosopher at the University of Pittsburgh. And we'll have the link for our listeners to on our website and incidentally, hosts a really interesting and well put together Center for Philosophy of Science. Anyway, Jonathan is the same generator that I mentioned earlier about you, is the one that has the theory that thought experiments are equivalent to if then kind of sort of a deductive reasoning.

Oh, the one I liked.

Yes, the one that you'd like. And so Nortons website has a really interesting set of thought experiments that are listed. And there are some of them are actually shown in animated fashion and then they're discussed in depth. And the first one, just to give you an example that comes on on the list. Is the cheesing the light experiment that's Einstein's famous phone experiment, which Einstein himself in, in his autobiographical notes tells tells us is what got him started on the general theory of relativity and so on and so forth.

And if you look at our website, at Northern's website, there is an animation of our export experiment is supposed to work. But then what's interesting is that Einstein thought that this failed experiment clearly indicated that the theory of either the idea that which was dominant at the time, the idea that space was not empty, but it was filled with this thing, either the US to say the full name for the ether that fills space is in fact that this experiment clearly shows that the right to be wrong, but not on.

His analysis is really interesting because it goes to three the three reasons why Einstein thought this experiment would succeed in debunking the ether theory.

And then he gives the either theories to reply to each one of those three objections, which actually and the reply actually is very interesting and convincing. And then Norten goes on and say why Einstein turned out was in fact right, but not for the reason that he thought it was right. So it makes for a really interesting read on the website. Again, is that Jonathan Goodis at the University of Pittsburgh?

Well, it's very cool. This is all these animations and pictures about Einstein and relativity. And apparently I'm scrolling down the page now and there is there's a separate section arguing that time actually exists, which is very comforting to me. It might come as a shock, but yes, it does exist.

All right. We're out of time. So we're going to wrap up this episode. This concludes another episode of rationally speaking. Join us next time for more explorations on the borderlands between reason and nonsense.

The rationally speaking podcast is presented by New York City skeptics for program notes, links, and to get involved in an online conversation about this and other episodes, please visit rationally speaking podcast Dog. This podcast is produced by Benny Pollack and recorded in the heart of Greenwich Village, New York. Our theme, Truth by Todd Rundgren, is used by permission. Thank you for listening.