My blog posts and YouTube videos mainly involve thought experiments. Some are from famous philosophers and some are of my own invention. In this post, I will put forward an argument for why I do this. Thought experiments are popular in philosophy, however, it is not a consensus that they give knowledge or they are persuasive to a skeptical audience. In popular media, I sometimes hear thought experiments being called ‘hypotheticals’. Unfortunately, more often than not this is done in a derogatory manner, such as “hypotheticals don’t prove anything” or “stop changing the subject with your hypotheticals”. Hence, I feel an article defending the use of thought experiments is needed.
The reason thought experiments are sometimes called hypotheticals is because they involve a story. And this story is usually an imagined, hypothetical event. But the story is not the only thing that is needed to make it a thought experiment. The story has a goal. The goal of the story is produce an anticipated answer to a question asked about the story. So, to make the story a thought experiment it requires a story and a question about the story the author thinks the audience will answer. This is why it is called a thought experiment. Similar to scientific experiments, there is a hypothesis; an observation; and a conclusion on whether the observation confirms or disconfirms the hypothesis.
Applying the science analogy to thought experiments. The hypothesis can be many things, but in philosophy thought experiments are often used in moral/ethical issues. The hypothesis will in this case be mostly in the form of ‘x is wrong’. The story will implicity reply that if x is wrong, then people will answer ‘x’ in my question about the story. This way, if people answer ‘x’ in the story, it confirms the hypothesis. And if they answer something else than ‘x’, this disconfirms the hypothesis.
A famous example is a thought experiment by Robert Nozick called ‘the experience machine’. In moral philosophy, there are some schools of thought that believe all that is required to live a good life is to maximise pleasure and minimise pain. This is called ‘the hedonistic view of the good life’. So, to challenge this, all that needs to be done is show a situation where although someone has lived a life of maximal pleasure and minimal pain, yet has not lived a good life, then the hedonistic good life must be wrong. This is because there must be something more than just pleasure and pain.
Nozick’s experience machine provides such an event. He asks us to imagine a machine that you may enter into. In this machine, you will experience all of your personal pleasures, preferences and desires and no pain. You will not know that you’re in the machine and can never leave it once you have entered it. Nozick asks us whether we would enter the machine? He anticipates that we obviously we would not, hence treats this as an argument that the hedonistic good life must be wrong. This is because it shows there is something that matters to us beyond maximising pleasure and minimising pain.
We can see how the thought experiment works as an argument in the following standard form:
- If the hedonistic good life is right, then we would want to enter the experience machine
- We would not want to enter the experience machine
3. The hedonistic good life must be wrong.
In these kinds of thought experiments, they are appealing to what our common intuitions are on how we should behave ethically. This is why Philosopher Daniel Dennett coined the term ‘intuition pumps’ to describe such thought experiments. These are the kind of thought experiments I usually focus on, so the kind of thought experiments I will be defending will be intuition pumps on ethical questions. There are many different kinds of thought experiments that demonstrate logical contradictions, or are focussed on scientific questions, metaphysical questions and so on. But these will not be addressed in this post.
Some common arguments against intuition pumps are as follows:
- Thought experiments beg the question
- Thought experiments are removed from reality (the ‘just’ hypotheticals argument)
- Thought experiments avoid the issue (changing the subject)
Thought experiments beg the question
The answer to the first objection is in my view the easiest to answer. The objection is that the stories in thought experiments already assume the answer the author is asking. This objection is a testable one. If a thought experiment is question begging, it must be impossible to answer the question in any other way than the anticipated answer. Is this the case with Nozick’s experience machine? No. One could easily raise their hand and say ‘sounds great, sign me up!” In fact, some people have answered this way (which leads to objection 4). So, there is nothing intrinsic about thought experiments that beg the question.
This does not mean that thought experiments never beg the question. Consider this thought experiment by Richard Swinburne arguing that there must be something non-physical about identity:
You have had a surgery where half of your brain has been removed and placed into two deceased bodies next to you. Let’s call them body A and body B. It is impossible for both A and B to be you, because this would be contradictory. A and B would be in different locations doing different things, so this would mean that I would be simulatenously be in location A and not in location A. And this is impossible.
Swinburne asks us whether we would end up in body A or body B or neither (died). He argues we can’t know. And since we can’t know, there must be some non-physical substance connected to our identity.
This thought experiment does beg the question. It begs the question because it assumes there is something beyond the physical to be transferred. A brillant counter thought experiment by QualiaSoup https://www.youtube.com/watch?v=RS4PW35-Y00 illuminates this by showing that the same identity problem can occur with a car battery. He asks us to imagine a car battery split in two and placed in separate cars. Even though we cannot know which car has the original battery, this would not lead us to conclude that there must be some non-physical substance connected to the car battery. To agree with Swinburne that the identity problem applied to minds assumes a non-physical substance, would require to already accept that there is a non-physical substance. This is what makes the thought experiment question begging.
Even though this demonstrates that some thought experiments can beg the question, it does not mean there is anything intrinsic about thought experiments that do this. And this is what is required if we want to dismiss thought experiments as a philosophical method.
Thought experiments are removed from reality (the ‘just’ hypotheticals argument)
This objection has been addressed by Adrian Walsh. Walsh describes this objection as the ‘objection from modality’. He calls it this because some philosophers will lay the charge that many thought experiments are “modally bizarre”. This means that a story that is far different from our own does not tell us anything about the world we actually live in. Walsh points out that many of our moral beliefs are grounded in some general principle. For example, I will not torture someone because of the general principle that it is always wrong to torture. So, if we do believe in such a principle, it should apply in all possible worlds. If this is true, then challenging this principle by asking someone to imagine a world different to our own is completely legitimate.
Returning to the torture example. A popular thought experiment that challenges the principle that it is always wrong to torture is known as “the ticking-time bomb”. This thought experiment asks us to imagine a man who has publicly bragged about having a bomb hidden that will destroy the world and will detinate in 1 hour. You capture this man who knows where the bomb is hidden.
The thought experiment assumes we would see torturing the man as at least morally permissible, or in some circles morally obligatory. We know without a doubt he planted the bomb (publicly bragging), time is running out and the consequences are extraordinary. Therefore, this provides a counterexample to the principle that it is always wrong to torture.
This is a bizarre story. We have only experienced comparable situations in Hollywood. Nonetheless, as Walsh notes, appealing to the bizarre nature of the thought experiment is irrelevant. It is irrelevant because if you truly believe that torture is always wrong, then you should have no trouble answering the thought experiment.
Walsh mentions that the issue some bizarre thought experiments possess is not in respect to how bizarre they are, but the conclusions they attempt to draw from them. The torture thought experiment only argues that there are some cases where torture can be morally permissible. This thought experiment does not tell us whether torture is always permissible, or even in most cases.
This does highlight that we should be careful about what we are trying to argue for when putting thought a bizarre thought experiments. If we are trying to make a case for torture in our current conflicts, we would need to demonstrate that it is analogous to the ticking time bomb. This is where use of a bizarre thought experiment can be illegitimate. Hence, objecting that a thought experiment is bizarre, removed from reality or ‘just a hypothetical’ is only legitimate if it does not lead us to the conclusion it is trying to make.
Thought experiments avoid the issue being discussed (changing the subject)
An objection similar to the ‘just hypotheticals’ argument is that thought experiments are changing the subject. Such an objection may be expressed as “stop giving me hypotheticals and stay on topic!” This suggests that any thought experiment is fundamentally irrelevant to the actual case being discussed.
Walsh also addresses this criticism, whilst also conceding that some thought experiments can change the subject. The example Walsh gives is a discussion concerning whether women should have full decision making regarding terminating a pregnancy. In the discussion, someone puts forward a thought experiment where a man can become pregnant via an artificial womb. Walsh points out this does change the subject because this has nothing to do with how pregnancies occur in our current day. However, like when begging the question was being considered, just because some thought experiments can change the subject does not mean all of them do. Walsh says that as long as the thought experiment remains relevant to the context of what is being discussed, it is not changing the subject.
In conclusion, when someone dismisses thought experiments or hypotheticals on the basis of being ‘just hypotheticals’. It can be helpful to mention why appealing to imagined situations are perfectly legitimate as long as they address the question being asked.