The problem with philosophical thought experiments

James Wilson has an article up at Aeon, looking at the trolley problem and other ethical and philosophical thought experiments.  One of the things he discusses is the notion that many philosophers have, along with many fans of particular thought experiments, that they’re sort of like a scientific experiment.  It’s not that unusual for someone philosophically inclined to tell me that X is true, and cite a thought experiment as  evidence.

Many of you already know that I have serious issues with this view of thought experiments.  I don’t think a philosophical thought experiment tells us anything about the external world or reality overall, and the notion that they do is fairly pernicious.  It gives people misplaced confidence in a notion based on nothing but a concurring opinion from the author of the thought experiment.

In many ways, thought experiments demonstrate the power of narrative.  If you want to sell people on an idea, tell them a story where the idea is true.  A thought experiment does this.  The most memorable ones can even have characters with names in them.  Would Mary’s Room or the Euthyphro Delimma have the same punch if the key players weren’t named?

Now, some may make a comparison with all the Alice and Bob type descriptions used in physics.  But these narratives are almost always used in a pedagogical fashion, to get across a concept that has already been worked out mathematically, and may have empirical evidence backing it up.  In these cases, the narrative isn’t itself the main argument, it’s just a vehicle to get a concept across in a non-technical fashion.

There have been famous thought experiments in science that were used as arguments.  Schrödinger’s Cat comes to mind.  It’s original use was meant as a reductio absurdum, similar to Einstein’s “spooky action at a distance” argument.  But reality turned out to be absurd.

Anyway, philosophical thought experiments typically only have their narrative.  Does that mean they’re useless?  I don’t think so.  But we should understand their limitations.  All they can do, really, is clarify people’s existing intuitions.  That can be pretty useful, fulfilling the role of what Daniel Dennett calls “intuition pumps.”  But that’s basically it.

So an ethical thought experiment may tell us about people’s ethical intuitions (although even here, check out Wilson’s piece for many of the issues), but they don’t fundamentally tell us what those ethics should be.  Likewise the Chinese Room, Mary’s Room, or philosophical zombies, don’t tell us anything about their subject matter.  They only flush out people’s intuitions about those subjects.

Unless of course I’m missing something.

29 thoughts on “The problem with philosophical thought experiments

  1. “There have been famous thought experiments in science that were used as arguments. Schrödinger’s Cat comes to mind. It’s original use was meant as a reductio absurdum, similar to Einstein’s “spooky action at a distance” argument. But reality turned out to be absurd.”

    Quick question Mike: Is your conclusion that reality is absurd influenced by the thought experiments you referenced above, or were those comments meant to be “off the cuff” remarks with no correlation to your own metaphysical position that reality is absurd?

    Peace

    Liked by 1 person

    1. Lee,
      It is somewhat “off the cuff”, but it refers to the fact that the scenarios presented in some of those thought experiments, which where supposed to be too absurd to be reality, after further discoveries, have turned out to be far more plausible than the authors thought they’d be.

      Like

      1. Mike,
        “…… but it refers to the fact…”
        Your use the word fact was rather pernicious I would think. Spooky action at a distance is only “plausible” if information is not physical, something no materialist would endorse, so that’s a non-starter. Wave function and wave function collapse with its famous notion of superposition is a joke, along with all of the MWI fallacies.

        Personally, I can’t make a correlation to either superposition or spooking action at a distance being “plausible” if my metaphysical position was materialism. As a materialist, I wouldn’t see either theory as being objective in any stretch of the term, they are all subjective to the umpteenth degree. I can understand why one might be attracted to these notions, but there is no justification for doing so other than for the same reasons children and pack rats are attracted to shiny objects.

        Peace

        Liked by 1 person

        1. Lee,
          I’m not sure what a philosophical materialist would endorse. I’m often lumped in that camp, but my commitment is to evidence, and the evidence from quantum physics forces us to confront some issues. There are costs with all the interpretations. Either you have to live with non-locality, many worlds, or other seemingly absurd notions such as retrocausality or various levels of irrealism.

          Of course, you could simply fold your hands and say we just don’t understand it. Many people do take that stance. Simply use the mathematics for other scientific theories and technology (“shut up and calculate!”), and assume all the other issues are unknowable. I can respect that epistemic humility.

          On the other hand, if you are going to venture into metaphysical arguments about reality at this level, they should be compatible with the data we have.

          Like

      2. I’m kinda with Lee here. I don’t think there are any “discoveries”, in the empirical science sense, that prove superposition is a physical state. I think that idea comes from the intuition that the math sets the reality, as opposed to simply maps to it. I know there are discoveries that say the math works, but that does not necessarily mean the math is reality. And it’s that same intuition that gives you a multiverse.

        *

        Liked by 1 person

        1. I think the double slit experiment (and many variations of similar experiments), with the wave of a single particle interfering with itself, demonstrates that the wave has causal effects. For me, that’s enough to say it’s real.

          Scientists have been able to isolate ever larger molecules and keep them in a superposition. And Bell’s inequality theorem and subsequent confirming experiments force us to confront either non-locality, or pay the cost to get locality back.

          It’s also hard to see how quantum computing works without the superpositions being real states.

          Liked by 1 person

          1. It’s real, but it’s not a wave and it’s not a particle. Under certain circumstances it does things that are described by wave type math, and under other circumstances it does things that are best described by particle type math. Keeping things in a “superposition” can be another way of saying keeping things from interacting with other things.

            I admit I don’t understand Bell’s inequality theorem or how quantum computing works, but I’m pretty sure that if I try really hard I can shoehorn them into my preconceived understanding of things. 🙂

            *
            [I’m pretty good at that]

            Like

          2. Well, it’s pretty hard to argue with, “quacks like a duck, moves like a duck, but isn’t a duck,” type arguments. It’s a bit like the old scholarly joke that the Iliad wasn’t written by Homer, but by another poet with the same name.

            On putting superposition in quotes, I would just note that quantum computing requires that the qubits at least behave like they’re in a superposition of spin states. Without that, they’re just expensive classical computers, and Google’s recent results seem to show they’re more than that. So more duck like behavior from the non-duck.

            Like

          3. I’m not sure what your duck argument is saying, because what we have is something that looks like a duck and sniffs like a rabbit and quacks like a duck and looks like a rabbit. Does that mean it’s both a duck and a rabbit at the same time? Or does it mean that it’s something else which sometimes looks like a duck and sometimes looks like a rabbit, depending on how you look.

            I have no problem believing quantum computers work differently from classical computers. But I don’t see how saying certain parts have to be in a superposition is different from saying certain parts have to be kept from interacting with the environment. Kinda like computing with dominoes. You have to set them up in a certain way and make sure nothing in the environment knocks one over before you start.

            *

            Like

          4. The main point is that quantum computing requires that the understood dynamics of superposition and entanglement work. It’s what allows Google’s 53 qubit computer to hold 2^53 basis points (well, minus error correction and other overhead) and complete a task in 200 seconds that would require the largest classical superclusters 2.5 days. (In principle, a 300 qubit computer would have more basis points than there are atoms in the observable universe.)

            We can’t rule out that the underlying reality isn’t radically different from what the data and math show. That’s true anywhere in science. What we can say is that the data and math work, and that classical rules don’t.

            Liked by 1 person

          5. The only thing the double-slit experiment demonstrates is that there is an inference pattern caused by some “thing”. It does not demonstrate, nor does it prove that the interference pattern is “caused” by a particle interfering with itself. The notion that a particle is a particle when it is loaded into a propulsion device, and then somehow changes to a wave when the propulsion device propels the particle, and then when this wave passes through the slits, this wave somehow changes back into a particle again is absurd. It’s analogous to asserting that my 150 grain bullet is a bullet when I load it into my 30-06, but once I fire the gun, the 150 slug turns into a wave; and then once this wave hits the target, it somehow turned back into a 150 grain slug. I mean, it has to; I can’t see it, and besides, on a really windy day, I cannot predict with any degree of certainty where the 150 grain slug will hit the target; and I can’t see the wind either, so it must be interfering with itself.

            Here’s what everybody forgets: Space is not “no thing”. Space has structural and qualitative properties just like every other physical object in our universe. Just because physicists have not isolated those physical properties does not mean those properties do not exist. It is clear, based upon the evidence of the double-slit experiment, that whatever is causing the interference pattern, that “some-thing” is quantum, because it is destroyed by a measurement device. Based upon this empirical evidence, I think it would be safe to posit that space is that “some-thing” causing the interference pattern because space is quantum.

            Peace

            Like

          6. Lee,
            The problem is reality is absurd, at least by human standards. So using absurdity alone to reject interpretations that match the data eliminates all of them. If you can really come up with a rigorous interpretation that avoids all absurdities, there’s a Nobel waiting for you.

            Like

          7. Reality is not absurd Mike, not even by so-called human standards. What is absurd, are the delusions humans derive from our experiences. No need to address that issue, for one can go shopping for any intellectual construct that is out there on the open market. I’m going to close this out with a posting I listed on broad speculations:

            One’s culture is one’s world view. SOM has been the prevailing world view for over twenty-five hundred years (2,500). Add to that paradigm the advent of computers, programs and algorithms; and one now has the perfect storm for ignorance. According to this paradigm, the mind is seen as an organic computational mechanism and everything that is not a mind is an algorithm. It’s a nice, neat, tidy little package with holes in it large enough to drive a semi-truck through.

            What I really don’t understand is the resistance to a cogent argument, an argument that offers a much simpler assumption, an assumption into which one can fit more data. Explanatory power rocks, party on Sabine Hossenfelder.

            Peace

            Like

  2. You are spot on about such thought experiments. I solved the Trolley Problem quickly when presented it in a somewhat Gordian Knot kind of solution.

    To alert the people in danger, who are not looking at the oncoming trolley which is between me and them, I reach into my shoulder holster, pull out a gun and fire two shots into the air. The people turn at the sound of the shots and see the trolley and evade it. Ta da!

    Hypothetical problem, hypothetical solution.

    There is an adage which states “When presented with a choice of one of two actions, always take the third.” This was clarified by an example of an elderly lady shopping for produce who asks the grocer “How much are the cantaloupes?” And he says “Two for one dollar fifty nine cents.” She asks “How much for just one?” And he says “Eighty cents.” She says, “I’ll take the other one.”

    Liked by 3 people

    1. That’s an excellent point! It matches points Wilson makes in his piece about the scenarios often not being realistic.

      I know as someone in management, I often have various options presented to me. But I’ve learned not to assume that those are the only choices. Often there are ones the presenter hasn’t considered, or doesn’t want to call attention to, that may be better overall than the choices they list.

      Liked by 1 person

  3. As we have discussed before, totally with ya on philosophical thought experiments. I quite agree they are different from the analogies and metaphors often used in science to communicate complex or purely mathematical ideas into more common contexts. (Funny timing. The post I have set to publish tomorrow is about metaphor.)

    I do think some of them at least form a nucleus for thought and discussion. Searle’s Giant File Room springs to mind. I agree we tend to argue them from our own perspective, but when others argue the same idea from a different perspective, it can get one thinking. OTOH, I’ve never found any value in p-zombies, for example. To me that one has a fatal begging the question error. I see it as little more than science fiction.

    Liked by 1 person

    1. One thing I did forget to mention in the post, is that in addition to clarifying people’s intuitions, they can also show contradictions in those intuitions, and that can definitely spur discussion. I think as long as we take them as that kind of starting point, they can be useful.

      On p-zombies, I agree, although they do reveal what a lot of people think about consciousness, that it’s something in addition to all the functionality. But the idea that just the existence of the concept somehow proves something is clearly question begging and circular, but people who buy the p-zombie argument often bristle at the suggestion.

      Liked by 1 person

  4. You should change your title to “the *biggest* problem with philosophical thought experiments.”

    I don’t think Euthyphro is a thought-experiment. It’s just as you named it: a dilemma. Actually a multi-lemma, but with the third and fourth options (what the gods love is unaligned, or only accidentally aligned to the good) not advocated by anyone.

    Mary’s Room is a lot like those Alice and Bob physics thought-experiments, in my view. It is a pedagogical device. Its purpose is to remind us of the Hard Fact (aka “Hard Problem” to those who mistakenly think it is a problem). Or did you call it the Hard Truth? I forget.

    Liked by 1 person

    1. I definitely agree with your implied point that thought experiments have other issues. Often, the logic in them is simply confused. I sometimes wish philosophy required all students to take some programming classes, so they could get an idea of just how bad humans really are at logic.

      I think I did use the “hard truth” label at some point, although I don’t remember where. But I agree. It’s not a problem, but a stark fact of reality. Our external models of how our minds work may never completely reconcile with our internal ones, at least not to the satisfaction of those troubled by that issue, but that’s just an epistemic limitation of the system’s ability to model itself.

      Like

  5. Hi Mike,

    I think I’d agree with you, generally. I skimmed the article you linked to, and I felt like the thought experiments were really “set-ups” in a way. I think a good analogy can be very instructive–can simplify and communicate the fundamentals of something in a way that the absence of an example cannot–but it’s not a good experiment if the outcome is imprinted in the set-up. And some of those felt that way. More like arguments than open-ended thinking.

    For me thought experiments are invaluable to many aspects of my life. I guess they’re not “philosophical” thought experiments, but they do provide insight into reality I think. They at least reveal to me what I might have already known, but not known that I knew it, without the intensity of thought that comes with a good question. My examples are largely related to understanding physical processes. When I played a lot of soccer, and was trying to learn a new skill on the ball–taking free kicks, for instance–it helped tremendously to do experiments in my mind on how the foot and shoe would interact with the ball in a certain condition. I would try it, see the result, and go back to work. Thought experiments could also be very insightful for creating different scenarios that may occur in a game, and studying how various responses might play out. I think my ability to perform these thought experiments was a differentiator in my development on the field.

    Another place thought experiments are helpful is in the engineering I do. Good thought experiments save lots of time. You can quickly test extremes of a situation mentally to understand what dynamics may be at work, and then return to the more complicated difficulty with insight into some of the underlying phenomena. It helps steer you through the solution space more efficiently.

    I guess those aren’t “philosophical” thought experiments. But you used the word thought experiment several times without that qualifier, so I wasn’t sure if you were saying all thought experiments are highly limited, or just the philosophical ones. I’m thinking a “good” philosophical one might explore a scenario from different vantages to see how it played out if one notion was true as opposed to another. It strikes me that this could be done, but I agree that some of the ones described in that paper feel more like arguments than experiments.

    Michael

    Liked by 1 person

    1. Hi Michael,
      I definitely was focusing here on philosophical thought experiments. Obviously if we get too broad with this, we can end up talking about any form of forethought or deliberation.

      Many of the other types you listed have one important thing the philosophy versions lack, a reality check. If you simulate various soccer techniques, you’re going to learn how accurate those internal deliberations were when you try them. I think that’s why scientific and engineering thought experiments can often work, because they’re either the start of an empirical and/or mathematical investigation, or they’re being used after that stage for education.

      Of course, a philosophy thought experiment has a reality check as well, how people react to it, whether it resonates with their intuitions or challenges them. But in general I don’t think they really work to convince anyone. As you note, the set up too often is obvious. They’re really just arguments for the author’s point of view.

      I do agree that a thought experiment with premises we all agree on but with unexpected conclusions could be a powerful thing, the kind of thing we see happen with Einstein’s ponderings of train and elevator scenarios. Unfortunately the philosophical version of that seem pretty rare.

      Like

      1. Good clarification. I was going to say something about the value of thought experiments for testing the limits of our understanding and models, and for guiding further exploration, but it seems you meant to be critiquing thought experiments which have no foreseeable “reality check”. Of course, that which is currently unforeseen can change. I would not be surprised if someday our understanding advances to the point that we all generally agree that p-zombies fall within the purview of our reality checks.

        I think it’s also worth noting that there’s value in understanding our intuitions. The trolley problem has been used extensively by experimental psychologists to learn about those intuitions without supposing they’re learning about some meta-ethical reality, but rather with the intention of informing our understanding of human nature.

        Liked by 1 person

        1. Thanks Travis.

          I actually wan’t just criticizing thought experiments with no foreseeable reality check, just the practice of taking them as evidence in and of themselves. If someone tried to do that with a physics thought experiment without the actual experimental test, I’d be critical there too, although if the physics one at least had rigorous mathematical extrapolation behind it, I’d be more sympathetic.

          I definitely agree there’s value in clarifying our intuitions, as long as we’re clear that’s what we’re doing. It’s particularly neat if the thought experiment can flush out contradictions in those intuitions or other unexpected outcomes.

          Like

  6. I published an article about thought experiments after being involved in a philosophy festival where I was part of a panel of people who reviewed the Oxford professor Timothy Williamson’s book “Doing Philosophy.” Since I had already blogged about all 100 of the thought experiments in Julian Baggini’s book “The Pig That Wants to Be Eaten and 99 other Thought Experiments”, I was naturally in favour of using thought experiments to do philosophy. As was Williamson. Here’s a very short justification from that article:

    ——-
    Why do [thought experiments] count for arguments of reason too? How come, as Timothy Williamson asks in Doing Philosophy, “philosophers get away with just sitting in their armchairs and imagining it all?”

    The reason is that our imagination is an incredible tool that has been honed to a fine edge over billions of years of evolution. Evolution is usually characterised as a series of trials and errors, but ones that are done blindly by Mother Nature. And until very recently, that’s how all life on Earth adapted and survived. But now that we know about this, we humans can conduct those trials and errors with a bit of wise foresight and consciousness. Scientists carefully plan their trials and errors all the time, but there are some places where it’s impossible for scientists to go. As Williamson says:

    “Imagination is especially useful when trial and error is too risky. … Imagining is [also] our most basic way of learning about hypothetical possibilities. … Only the dumbest animals would not think about [these]. … Thought experimentation is just a slightly more elaborate, careful, and reflective version of that process, in the service of some theoretical investigation. Without it, human thought would be severely impoverished.”

    Williamson is right. Over thousands of years, some of the best thinkers in history have churned out mountains of these trials and errors of the imagination, and they have the power to fundamentally change the way we navigate the world. They’ve certainly changed me.
    ——

    See here for my full article, as well as other’s from the panel, and an introduction and set of responses from Williamson:

    https://www.thephilosopher1923.org/williamson

    Liked by 2 people

    1. I do think thought experiments have their place. But as good as our imagination is, it has limits. And we frequently are unable to imagine the limits of our imagination. If we can’t have a reality check, then a thought experiment has some uses, but in and of itself, it proves nothing, except what our intuitions are. That can be useful.

      Other than that, all we can get from them are hypotheses, hypotheses that we shouldn’t accept as conclusive until or unless we can actually test them, get a true reality check. I get how frustrating that can be for many topics, because the reality check is frequently not possible. I’m not a logical positivist. I do think there is value in thinking about those issues.

      But as soon as someone crosses the line into concluding they know truth from only these exercises, they’re putting a faith in human imagination and logic that history simply hasn’t borne out.

      Like

      1. I don’t disagree with that. But I could say the same thing about empirical knowledge. As soon as someone claims truth from that, they’re putting faith in human faculties and logic that history hasn’t born out. And one of the ways I “know” that is from thought experiments like Descartes Evil Demon and the Gettier Problem. No empirical observations will tell you the limits of empirical observations. You need thought experiments to see that. Many are hypothetical intuition pumps alone, as you say, but we’d need to talk about specific experiments to get more specific in our criticisms.

        Like

Your thoughts?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.