Excellent post Suzi! I had no idea the tradition of conceivability implying possibility went that far back. It explains why so many philosophers seem comfortable with it.
I've always been skeptical of this idea, for exactly the reasons laid out by the philosophers you cite. We may *think* we can imagine something with sufficient rigor that it's meaningful, but human minds are more limited than that. I sometimes wished philosophers had to take some programming courses, primarily so they could internalize how bad humans are at logic.
I think thought experiments are useful devices, but mainly for clarifying people's intuitions. The best ones challenge those intuitions, find where they start to break down or become contradictory. But most are simply telling a story that validates the author's intuitions. The ones people find compelling are the ones that validate their own intuitions. It seems like their main value is in the conversation they generate.
But I've never been wild about the word "experiment" being in the name. That implies they tell us something about reality. Daniel Dennett called them "intuition pumps", which I think is an improvement, although to me "intuition clarifiers" might be more accurate.
Anyway, looking forward to the rest of the series!
Thank you so much, Mike! I'm really glad you enjoyed the post.
It's fascinating how some philosophical ideas stick around for centuries, isn't it?
Haha -- programming has a way of humbling me about my logical abilities, almost everyday!
I agree with you on thought experiments -- they're fantastic for exploring our intuitions and assumptions, but you're right that calling them experiments doesn't fit well with the scientific definition. Dennett's term "intuition pumps" or, even better, your term "intuition clarifiers" works a lot better -- it focuses us on what they should be doing -- highlighting our intuitions and exposing poor thinking.
Imagine, for a moment, the rabbit hole I now proceed down!
Presumably Hume wanted amongst other things to rebut the Ontological Argument for the existence of God? Another seductive if pretty unconvincing thought experiment (asking people why it doesn’t work is often instructive - the problem of infinite regress isn’t always the first refutation attempted). Anyway, I’m off to see the rabbits!
I should name check the excellent books on thought experiments by Martin Cohen, “Wittgenstein's Beetle and Other Classic Thought Experiments” and Julian Baggini’s, “The Pig That Wants To Be Eaten: And 99 Other Thought Experiments”, both of which I have enjoyed as a non-philosopher. Looking forward to the series.
Thanks for jumping down this rabbit hole -- and for popping back up to share those great book recommendations!
Yes, Hume did take aim at the Ontological Argument! But there's a slight difference here that might be worth highlighting. Hume's point that Mary Shepherd criticised was a positive assertion that if something is "clearly" conceivable, then it "might" exist (it's possible). Shepherd’s point being that many things we initially believe to be conceivable are shown to be inconceivable once we gain more knowledge about them.
Regarding the Ontological Argument, I haven’t understood Hume's point to be that the concept of a perfect being is inconceivable. Rather, I understood his argument to be that mere conceivability cannot demonstrate actual existence. In other words, Hume argues that nothing can be proven to exist in reality solely based on our ability to conceive of it.
Thanks so much for mentioning those books by Cohen and Baggini. I've heard of them both, but haven't read either. I'll definitely add them to my reading list.
I'm glad you're looking forward to the series -- I'm looking forward to reading your thoughts.
Thank you for going to the trouble of drawing what I realise now is quite an important distinction. And thank you even more for doing this in such a kind way! I get a bit carried away with Hume in truth. All the best, John.
Fabulous. In college, many, many years ago, we had conversations on life, issues and being, that so resembled what you have so nicely stated that is deja vue all over again (as Yogi Berra might have said).
I’m very much in favor of well constructed thought experiments since they can illustrate the details of what a given proposal happens to practically mean. I’d suggest that philosophers who don’t like them, such as Daniel Dennett, might have this aversion because certain thought experiments display them to hold various supernatural beliefs. I only wish that John Searle’s Chinese room was as convincing for displaying this regarding Dennett, as I consider my own thumb pain thought experiment to be.
In any case I’m surprised that David Hume posited that conceivability suggests possibility in a physical rather than merely logical capacity. I certainly can’t support a notion like that! In the argument against I could add my own single principle of metaphysics. It observes that to the extent systemic causality fails, nothing exists to potentially understand. Here people or scientists in general ought to presume a perfectly causal world, and simply to preserve the potential for understandings to occur. Then those who’d entertain the notion that systemic causality does fail, would be free to posit things like a physical reality that’s only constrained by what’s possible to imagine — clearly a magical position.
I like a good thought experiment too! They can reveal underlying assumptions or beliefs we might not have realised we held.
Regarding Hume's position, your spidey sense is spot on. Hume was concerned with logical possibility, not physical possibility. Hume argued that if we can conceive of something without contradiction, it's logically possible, but this doesn't necessarily mean it's physically possible or actually exists in our world. I'm sorry if that wasn't clear in the article.
I agree with your point about the dangers of assuming a reality only constrained by imagination -- I like the idea of grounding our philosophical inquiries in empirical reality as much as possible.
Thanks Suzi, that sounds more like Hume to me. I may have gone through your post too quickly in that regard as I was trying to get a comment in yesterday before I needed to leave for work. Regardless I think I may have a reasonable way to straighten some of this business out.
It seems to me that lots of confusion and even error might be avoided by directly making epistemic versus ontological distinctions. Observe that each of the possibilities that we’ve been considering so far here have been epistemic rather than ontological. Thus each time we could add a “to us” qualification to the description — as in “logically possible to us”, “metaphysically possible to us”, “physically possible to us”, and “actually possible to us”. Perhaps some have been ignoring this unsaid element to put too much stock in the value of “conceivability”? Even if you can conceive of something, this shouldn’t be all that important since it merely suggests “epistemically possible” rather than “ontologically possible”.
Also consider what a diagram of ontological truth might be like. Here I see the largest circle to be of “logical truth”. Then there’d be two smaller circles within that have headings of “systemic causality” and “non-systemic causality”. Could these two smaller circles overlap as well? That would reflect the belief that we have an otherworldly soul which operates in a causal world. I’m not entirely sure this would fly logically however since “causal and non-causal” seems quite like “father and non-father”, for example. Regardless certain “many worlders” seem pretty close to asserting that anything which is logically true, should also actually exist in some of their theorized many worlds. Then of course many or most people believe that there is more to our world than causality alone, and regardless of any logical inconsistencies between both functioning causally and non-causally. Conversely like me, some people are perfect naturalists and so consider ontological reality to only concern systemic causality.
Also I’ve been thinking about how many truly enlightening philosophical thought experiments I know of right now. Apparently not many. So I guess I generally agree with Dennett’s “intuition pump” attack against them. It’s not intuition that I like about good thought experiments though, but rather how they practically illustrate what a given proposal happens to suggest about other things. Einstein famously constructed these sorts of thought experiments. I consider them to be a relatively standard element of scientific exploration.
Hey Eric! Great point. I agree. This really is about whether or not what we can conceive (epistemology) tells us something about what is possible in reality (ontology). When we talk about different types of possibility (logical, metaphysical, physical, actual), we're implicitly referring to epistemic possibilities - what's possible "to us" based on our knowledge and reasoning abilities. But what we are really interested in is ontological possibilities -- what's actually possible in reality regardless of our knowledge.
I'm glad you bought up causality -- it's key in this debate. I've taken you to mean "systemic causality" as law-like causal relationships in the physical world. And non-systemic causality as possibly including things like free will or supernatural interventions? I think it was David Lewis who made an even stronger claim than "many worlders" in physics would assert -- he claimed that logical possibility implies actual existence. I think the many worlders would claim that existence is constrained by the laws of quantum mechanics.
On thought experiments... Just like most things, I think we like the ones that confirm our intuitions and we dislike the ones that don't.
Wow, some validation that many in academia appear to conflate “epistemic” for “ontological” forms of possibility! Few broken things ever get fixed without at least being acknowledged as such, so hopefully this is a positive step.
Going back through my blogging commentary since 2014 I see that I’ve used the term “causal” quite a lot to reference how things work. Of course the traditional term “materialism” seems limited to the term “matter”. Thus “physicalism” now seems more popular. But that leaves us with the task of defining “physics”. As a perfect naturalist I reduce this back to causality itself, and with a “system based” clause to exclude any outside influences. It seems to me that people often say they’re physicalists without really grasping the full implications. That should make it easier to cheat. Perhaps my clarification could help.
We can certainly use the term “law” in this regard as well, though to me it seems a bit epistemic Observe that scientists try to develop effective laws to explain observations. Politicians even pass or decree laws to constrain social function. The concept of “systemic causality” however is my attempt to conceptually get above mere human made rules.
Yes I’d say that under perfect systemic causality, ontologically there can be no freewill. I still think it’s useful for us to say that freewill exists epistemologically though. Here it should exist as a function of our ignorance — the less that’s understood about a given situation, the more freedom that a person may be seen to have. Conversely from a conceptual perfect understanding, no freedom should be seen. Thus even quantum dynamics should either be determinate from such a perspective, or magical given the failure of systemic causality.
I don’t mind “many worlds” in an epistemologically sense. So when someone says it seems like our world splits off into countless worlds each second given quantum function, I’m good with that. It’s when they posit that such other worlds actually exist that I accuse them of violating the tenets of systemic causality. Consider how this shouldn’t just be bajillions of actual worlds splitting off ours each second, but bajillions of worlds splitting off each of those each second as well, and so on resulting in infinite regress. I wonder if such an idea truly holds weight in academia today? Or am I simply mistaken about what’s claimed?
While we may tend to like thought experiments that conform with our own positions, I hesitate to presume that we all thus become uncritical. Some seem able to effectively criticize certain thought experiments even when they’re sympathetic with the ultimate message. That should at least be a sign that someone is still able to think effectively.
Fascinating! Ok, I’m trying to work through this model using Einstein’s “spooky action at a distance.” Because, in this case, Einstein essentially concluded that quantum entanglement was logically impossible since cause and effect couldn’t happen at the same time—especially not across a distance. It seems logically impossible because it didn’t adhere to our understanding of the physical universe at the time. BUT, as we now know, quantum entanglement is in fact an actual thing. So… would this be an example of something that is metaphysically possible but logically impossible? If so, how does that fit with the nesting model of possibility?
Or is this an inappropriate application of this lesson?
Great question! I had to think about this one for a bit.
I do wonder about the categorisation here. It seems to me that Einstein's claim was that it was physically impossible based on the understanding of physics at the time. I'm sure it seemed illogical to Einstein, but I don't think he meant logically impossible in the philosophical sense. Quantum entanglement seemed not physically possible, but it turned out to be physically possible after all.
*Types "/imagine prompt: Darth Vader existing without a body while simultaneously being Luke's father and not being his father, traveling above the speed of light inside a black hole while never leaving it."
Conceivability does not need to imply possibility to be useful.
And a hack towards imagining transparent iron: just imagine magnetic glass. Easy to do. Of course ignorance of the link between spin alignment and magnetism makes it easier but that is irrelevant. The body of knowledge we are ignorant of still vastly exceeds that which we know. We must not let this stop us
Thank you so much! I'm thrilled you're enjoying the articles.
I completely agree! Conceivability need not imply possibility to be useful. It's such an important idea, especially when we think about invention, innovation and creativity. After all, many groundbreaking inventions or theories started as what seemed like "impossible" ideas that someone dared to conceive.
Suzi—thank you for the thought experiment on thought experiments. A few comments:
In Hume’s “conceivability implies possibility” you point out this “allows us to explore counterfactual scenarios and potential alternative realities “. (Well said). Previously in the discussion on what is possible the diagram stops at ‘logical’ as an outer boundary. But what about counter-logical? This ‘illogical’ realm would be outside of the logical.
Then what of counter illogical? That is beyond logic/illogical and objective possibility? This counter-possible would be the ‘impossible’ realm —beyond the possible, or we might say the transcendent realm.
Each of the ‘possibilities’ in the diagram seem to be dependent upon this counterbalance concept: starting with the actual/physical (I’m not splitting these up because you haven’t taken up the details of the distinction yet)—counter actual/physical goes to metaphysical—counter metaphysical goes to logical only—(to extent this)—counter logical goes to illogical (beyond our current paradigm of logic and objective possibility)—and counter illogical/possibility goes to impossible (outside of our current understanding of possible—or ‘transcendent’). One could argue that counter transcendent goes to logical, metaphysical, and physical/actual. In this sense the existence of the actual, physical, metaphysical, and logical implies the existence of the transcendent. One might even say the physical etc. depends or relies upon the transcendent to be defined.
In the example of Darth Vader—the logical possibility is that he is Luke’s stepfather. Thus he is Luke’s father—the only one Luke has known as ‘Father’, and he is not Luke’s Father—not his biological father. Or perhaps Luke was the product of multiple fathers whose genetics were blended to create the Y chromosome haploid. Thus Darth Vader is Luke’s father (one of them) and is not Luke’s ‘father’—the sole source of his Haploid DNA. These are logical arguments within fiction writing for probabilities—stories are based upon what is probably not what is possible.
Story writing (imaging) is where the unintelligible becomes actual fiction—the substance of imagination. Take Mary Shepherd’s deciduous tree blooming in December example. If a southern hemisphere deciduous tree is moved to a greenhouse in the northern hemisphere (with heat, light, and nutrients appropriate) it will thrive in December and January. The ‘unintelligent’ becomes actual with a bit of imagination and creativity technology. The calendar being a non-universal ‘law’ of physics for flowering vegetation even on this planet.
That said, “the nature of things” is bound by time, location, culture, light (a whole lot of physics) which is the definition of ’thingyness’. The nature of non-things, or things beyond things is the realm and substance of concepts—which is one factor in imagings and the ‘conceivable’.
Here Peter Van Inwagen’s concept of the inconceivable transparent iron is actually a question of perspective scale. We don’t need to get to structural details of condensed matter physicists. Transparency is a function of scale. In an electron microscope of iron (NASA has published one from a moon rock). There is a lot of space. If we go deeper beyond this view the space between the quarks, between the protons, neutrons, and electrons in iron grow larger and larger. In a sense it is the space between these that makes iron ‘ironie’. At this scale iron is transparent. And at the scale of the entire universe iron is beyond ‘apparent’—it is therefore trans-parent (so to speak). My point is the statement ‘transparent iron’ is the concept (not inconceivable)—the argument is simply one of level of possibility, not impossibility.
So I wonder, what exists beyond conceivability? How can we break the bounds of concepts (and the possible) without the substance of thought and thought experiments?
Wow, that's quite a journey through the realms of possibility and beyond! I love how you've extended the idea of "counter" concepts all the way to the transcendent.
Your point about transparent iron is fascinating. You're right -- so many "impossibilities" are really just a matter of perspective or current understanding.
You raise a great point about what exists beyond conceivability. We could ever know. How could we use thought to go beyond the limits of thought itself?
Hi Suzi. Love the questions you ask in your comment and I hope this finds its way into your research and writing someday. I look forward to reading your thoughts on this (yes, irony of thoughts on the limits of thought fully accepted). We could wonder if AI will be an effective tool (someday?) for us to push on this boundary—if it exists. Thanks to you as always for pushing our thinking.
Another beautiful piece, Suzi. You make me think about Kants understanding of the imagination, here linked to conceivability. Kant might argue that our ability to conceive of a flower even in the absence of a flower is imagination. On this view even the scientist who values sensory data cannot function in the absence of imagination. Do out of body experiences exist? Are they made of subatomic particles? In sociology—which focuses on abstract structures like economies, family, occupation, religion, education, etc. (all included in the three-part structure of ontology we/you discussed earlier)—what is the function of conceivable but impossible? Transparent iron exists at the level of quarks and what not. What about transparent schools?
You're so right! -- even in empirical sciences, imagination plays a crucial role - it's what allows us to form hypotheses and design experiments.
Your comment about "conceivable but impossible" in sociology got me thinking. How do we define "possible" and "impossible" when we're talking about things like economies or education systems?
One quibble: Anthropologists and sociologists consider themselves empirical scientists in that the rely on observations that are potentially verifiable by a disinterested third person. These fields don’t operate on fantasy, which is the only option for one claiming that something impossible logically is truly possible. Prudence is right. The notion of conceiving as possible or impossible based on the physical sciences (my term for what you are calling empirical science) is wrapped up in logic and physical reality. But it’s that third space as Popper called it, it’s not physical but it is real to us—Van Gogh, cave paintings, law of supply and demand, a school. Here, I think, empirical evidence requires a disciplined scientist with as you say nuances.
I completely agree about anthropologists and sociologists being empirical scientists. When I say "empirical sciences," I'm usually casting a pretty wide net. I tend to think anyone who uses observation and evidence-based reasoning is doing 'science', which definitely includes social sciences and anthropology.
Your point about the "third space" is great -- we navigate between the physical world and our conceptual frameworks. And its in this realm, things like economies or educational systems exist as real phenomena, even if we can't touch them the way we can touch a rock.
I'm curious about how you see physical phenomena informing this third space.
When sociologists or anthropologists (or even ethicist) are forming hypotheses or interpreting data, they seem to be engaging in a form of disciplined imagination -- but how much is that imagination informed by the physical world? They're not operating on fantasy, for sure, they are often conceiving of possibilities that haven't been directly observed yet, but I wonder how much imagination is influenced by past observations of behaviours and other physical phenomena.
This reminds me of the debate about whether science can inform morality (but let's not open that can of worms!).
I think it’s useful to distinguish between imagination and conceivability. Conceivability means without logical contradiction, it’s a disciplined imagination. It’s a way of clarifying our concepts.
And the other thing is that while conceivability may not reliably tell us what is true, it is reliable in ruling out the false. So, when you say it doesn’t tell us “anything” about whether something is possible in the real world, it can tell us that square circles and married bachelors don’t exist. It rules out the impossible as false, rather than establishes the possible is actual.
Thinking about that in the context of the p-zombie thought experiment, it’s not suggesting zombies actually exist, it’s arguing “against” the idea there is an identity between mind and brain. If there was such an identity, there would be a contradiction in the concept of p-zombies.
Yes, the word "clearly" is doing a lot of heavy lifting in Hume's maxim. Thanks for pointing out the distinction between imagination and conceivability -- it's an important one.
Conceivability does seem to work better in the negative sense. We can't conceive of a square circle or a married bachelor so we rule them out. But I wonder about these examples. They are ruled out by definition -- a square is defined by having four sides and a bachelor is defined as an unmarried man. But I wonder about other sorts of claims. Like Einstein's claims about quantum entanglement. He thought that "spooky action at a distance" was not conceivable, and therefore impossible. But he turned out to be wrong.
It seems like there might be some room for different types of conceivability as well as the different types of possibility, so the relationship between the two could get quite nuanced.
That's an interesting take you have on the p-zombie thought experiment. I've always seen Chalmers' argument as primarily making a positive claim about conceivability. His second premise being -- "If it is conceivable that there be zombies, it is metaphysically possible that there be zombies".
But your point has be thinking.
I wonder if we might say that Chalmers makes a positive claim about conceivability to set up a scenario that then allows for the kind of logical analysis you're describing.
I think of it this way, logical possibility deals with the a priori, the concepts we’re using. If we say quantum entanglement is impossible, the first thing to do is clarify what those words or concepts mean. That will always depend on our current level of knowledge, which is (probably) wrong, or at least incomplete. But there's no alternative to using our best knowledge.
Einstein was dealing with physical possibility and the assumption nothing travels faster than light. If that’s true, then spooky action is impossible.
It’s my understanding that Saul Kripke first used zombies to argue against identity theory (an a priori claim). Chalmers used zombies to argue if consciousness doesn’t logically supervene on the physical (ie zombies are conceivable), there can be no reductive explanation. In a physicalist dominant environment, that’s radical, because it means physicalism is false.
We talk about these ideas all the time in literary circles when it comes to genre, suspension of disbelief, and historical progression. Great charts and breakdowns, and yet nothing quite gets at the empirical is-ness of things. That’s what keeps me reading and being.
It's the tension between our conceptual frameworks and the raw, experiential nature of reality that keeps us exploring, isn't it? Always more to ponder, always more to discover.
I agree, our imaginations are deeply rooted in our experiences and the language we use to describe them.
Your point about AI potentially imagining concepts beyond our comprehension is intriguing. I'm not entirely convinced that AI, at least as we currently understand it, can truly come up with novel ideas in the way you suggest. Maybe.
Current AI systems, even the most advanced ones, are essentially processing and recombining existing information in sophisticated ways. They're not updating their 'ideas' in physical experiences the way biological minds do. This makes me wonder whether truly novel ideas - ones that go beyond the recombination of existing concepts (if such a thing exists) -- might need to be tested in the actual world.
What is the benefit to considering metaphysical possibilities that aren’t logically or physically possible? Logical possibilities seem useful to work towards our understanding of the physical world and pushing technological limits. However, metaphysical possibilities that fail to be logically or physically possible seem like a metaphorical hamster wheel without much of an end except mental gymnastics. I'm sure there is a benefit just can't put my finger on it?
Great question. I've been scratching my head with similar questions myself.
I think the problem is that metaphysical possibilities are not clearly defined. It seems we are permitted to include possible worlds here where the laws of physics are different from our own, and we can also include possible worlds where the laws of physics are exactly the same as ours. So we might have a metaphysical scenario that is logically possible (it doesn't involve any contradiction in logic) but is not physically possible because it doesn't align with our laws of physics. But we could also have a metaphysical scenario that is logically possible and doesn't involve any contradiction in our laws of physics -- but it's still not possible in our world given the current state of the world.
For example. I might like to say that it could have rained where I live this morning. That doesn't violate logic or the laws of physics in general. But given that this morning (where I live) there were no clouds in the sky, so, can I really say that it could have rained this morning?
When philosophers argue -- if I can conceive that it could have rained, so it's metaphysically possible that it could have rained -- I think we are allowed to question the grounds on which this conceiving is taking place.
That said, considering metaphysical possibilities, might have some use in prediction and scenario planning. Exploring possibilities, including those that might seem to violate our current laws of physics or the present state of the world, might be useful for understanding, anticipating novel situations, and innovation.
I do wonder how much the utility of this approach depends on our stance on determinism, though. If we lean towards a more deterministic view of the universe, where future states are entirely dictated by present conditions, the value of considering the metaphysically possible but physically implausible scenarios might be limited. On the other hand, if we believe in a less deterministic universe, exploring a wider range of metaphysical possibilities could prove more fruitful for prediction and innovation.
Really appreciate such a detailed and thought-provoking reply, thanks, Suzi!
Considering prediction and planning scenarios for innovation and development is a really good argument. Pushing the limits and trying for the impossible seems like a great motivating factor for inventors and engineers.
It seems to me philosophers in recent years have left out one kind of possibility Analogical. Isaac Newton belonged to a group of philosophers who were toying with the idea that perhaps Nature could be treated as a cipher. Some ciphers are breakable. If we treat Nature as cipher and succeed in breaking it, we know what Nature is about. Well, Newton and his fellow natural philosophers succeeded in making a beginning in this enterprise. Which enterprise continues to this day. A possibility became a reality many years ago. Perhaps this is the reason why we forget that it was once nothing more than a possibility?
I love how you've brought up the concept of analogical possibility and tied it to the historical development of science. We often forget how radical and "merely possible" some ideas seemed before they became established. It makes me wonder what analogies we're using today that might seem outlandish now but could lead to major breakthroughs.
Excellent post Suzi! I had no idea the tradition of conceivability implying possibility went that far back. It explains why so many philosophers seem comfortable with it.
I've always been skeptical of this idea, for exactly the reasons laid out by the philosophers you cite. We may *think* we can imagine something with sufficient rigor that it's meaningful, but human minds are more limited than that. I sometimes wished philosophers had to take some programming courses, primarily so they could internalize how bad humans are at logic.
I think thought experiments are useful devices, but mainly for clarifying people's intuitions. The best ones challenge those intuitions, find where they start to break down or become contradictory. But most are simply telling a story that validates the author's intuitions. The ones people find compelling are the ones that validate their own intuitions. It seems like their main value is in the conversation they generate.
But I've never been wild about the word "experiment" being in the name. That implies they tell us something about reality. Daniel Dennett called them "intuition pumps", which I think is an improvement, although to me "intuition clarifiers" might be more accurate.
Anyway, looking forward to the rest of the series!
Thank you so much, Mike! I'm really glad you enjoyed the post.
It's fascinating how some philosophical ideas stick around for centuries, isn't it?
Haha -- programming has a way of humbling me about my logical abilities, almost everyday!
I agree with you on thought experiments -- they're fantastic for exploring our intuitions and assumptions, but you're right that calling them experiments doesn't fit well with the scientific definition. Dennett's term "intuition pumps" or, even better, your term "intuition clarifiers" works a lot better -- it focuses us on what they should be doing -- highlighting our intuitions and exposing poor thinking.
Imagine, for a moment, the rabbit hole I now proceed down!
Presumably Hume wanted amongst other things to rebut the Ontological Argument for the existence of God? Another seductive if pretty unconvincing thought experiment (asking people why it doesn’t work is often instructive - the problem of infinite regress isn’t always the first refutation attempted). Anyway, I’m off to see the rabbits!
I’m back (for a moment)!
I should name check the excellent books on thought experiments by Martin Cohen, “Wittgenstein's Beetle and Other Classic Thought Experiments” and Julian Baggini’s, “The Pig That Wants To Be Eaten: And 99 Other Thought Experiments”, both of which I have enjoyed as a non-philosopher. Looking forward to the series.
Hi John!
Thanks for jumping down this rabbit hole -- and for popping back up to share those great book recommendations!
Yes, Hume did take aim at the Ontological Argument! But there's a slight difference here that might be worth highlighting. Hume's point that Mary Shepherd criticised was a positive assertion that if something is "clearly" conceivable, then it "might" exist (it's possible). Shepherd’s point being that many things we initially believe to be conceivable are shown to be inconceivable once we gain more knowledge about them.
Regarding the Ontological Argument, I haven’t understood Hume's point to be that the concept of a perfect being is inconceivable. Rather, I understood his argument to be that mere conceivability cannot demonstrate actual existence. In other words, Hume argues that nothing can be proven to exist in reality solely based on our ability to conceive of it.
Thanks so much for mentioning those books by Cohen and Baggini. I've heard of them both, but haven't read either. I'll definitely add them to my reading list.
I'm glad you're looking forward to the series -- I'm looking forward to reading your thoughts.
Thank you for going to the trouble of drawing what I realise now is quite an important distinction. And thank you even more for doing this in such a kind way! I get a bit carried away with Hume in truth. All the best, John.
Great post, looking forward to the rest.
Thank you!
Fabulous. In college, many, many years ago, we had conversations on life, issues and being, that so resembled what you have so nicely stated that is deja vue all over again (as Yogi Berra might have said).
thx
Respectfully.
Thanks David!
It's wonderful how some conversations are timeless staying with us through the years, isn't it?
I’m very much in favor of well constructed thought experiments since they can illustrate the details of what a given proposal happens to practically mean. I’d suggest that philosophers who don’t like them, such as Daniel Dennett, might have this aversion because certain thought experiments display them to hold various supernatural beliefs. I only wish that John Searle’s Chinese room was as convincing for displaying this regarding Dennett, as I consider my own thumb pain thought experiment to be.
In any case I’m surprised that David Hume posited that conceivability suggests possibility in a physical rather than merely logical capacity. I certainly can’t support a notion like that! In the argument against I could add my own single principle of metaphysics. It observes that to the extent systemic causality fails, nothing exists to potentially understand. Here people or scientists in general ought to presume a perfectly causal world, and simply to preserve the potential for understandings to occur. Then those who’d entertain the notion that systemic causality does fail, would be free to posit things like a physical reality that’s only constrained by what’s possible to imagine — clearly a magical position.
I like a good thought experiment too! They can reveal underlying assumptions or beliefs we might not have realised we held.
Regarding Hume's position, your spidey sense is spot on. Hume was concerned with logical possibility, not physical possibility. Hume argued that if we can conceive of something without contradiction, it's logically possible, but this doesn't necessarily mean it's physically possible or actually exists in our world. I'm sorry if that wasn't clear in the article.
I agree with your point about the dangers of assuming a reality only constrained by imagination -- I like the idea of grounding our philosophical inquiries in empirical reality as much as possible.
Thanks Suzi, that sounds more like Hume to me. I may have gone through your post too quickly in that regard as I was trying to get a comment in yesterday before I needed to leave for work. Regardless I think I may have a reasonable way to straighten some of this business out.
It seems to me that lots of confusion and even error might be avoided by directly making epistemic versus ontological distinctions. Observe that each of the possibilities that we’ve been considering so far here have been epistemic rather than ontological. Thus each time we could add a “to us” qualification to the description — as in “logically possible to us”, “metaphysically possible to us”, “physically possible to us”, and “actually possible to us”. Perhaps some have been ignoring this unsaid element to put too much stock in the value of “conceivability”? Even if you can conceive of something, this shouldn’t be all that important since it merely suggests “epistemically possible” rather than “ontologically possible”.
Also consider what a diagram of ontological truth might be like. Here I see the largest circle to be of “logical truth”. Then there’d be two smaller circles within that have headings of “systemic causality” and “non-systemic causality”. Could these two smaller circles overlap as well? That would reflect the belief that we have an otherworldly soul which operates in a causal world. I’m not entirely sure this would fly logically however since “causal and non-causal” seems quite like “father and non-father”, for example. Regardless certain “many worlders” seem pretty close to asserting that anything which is logically true, should also actually exist in some of their theorized many worlds. Then of course many or most people believe that there is more to our world than causality alone, and regardless of any logical inconsistencies between both functioning causally and non-causally. Conversely like me, some people are perfect naturalists and so consider ontological reality to only concern systemic causality.
Also I’ve been thinking about how many truly enlightening philosophical thought experiments I know of right now. Apparently not many. So I guess I generally agree with Dennett’s “intuition pump” attack against them. It’s not intuition that I like about good thought experiments though, but rather how they practically illustrate what a given proposal happens to suggest about other things. Einstein famously constructed these sorts of thought experiments. I consider them to be a relatively standard element of scientific exploration.
Hey Eric! Great point. I agree. This really is about whether or not what we can conceive (epistemology) tells us something about what is possible in reality (ontology). When we talk about different types of possibility (logical, metaphysical, physical, actual), we're implicitly referring to epistemic possibilities - what's possible "to us" based on our knowledge and reasoning abilities. But what we are really interested in is ontological possibilities -- what's actually possible in reality regardless of our knowledge.
I'm glad you bought up causality -- it's key in this debate. I've taken you to mean "systemic causality" as law-like causal relationships in the physical world. And non-systemic causality as possibly including things like free will or supernatural interventions? I think it was David Lewis who made an even stronger claim than "many worlders" in physics would assert -- he claimed that logical possibility implies actual existence. I think the many worlders would claim that existence is constrained by the laws of quantum mechanics.
On thought experiments... Just like most things, I think we like the ones that confirm our intuitions and we dislike the ones that don't.
Wow, some validation that many in academia appear to conflate “epistemic” for “ontological” forms of possibility! Few broken things ever get fixed without at least being acknowledged as such, so hopefully this is a positive step.
Going back through my blogging commentary since 2014 I see that I’ve used the term “causal” quite a lot to reference how things work. Of course the traditional term “materialism” seems limited to the term “matter”. Thus “physicalism” now seems more popular. But that leaves us with the task of defining “physics”. As a perfect naturalist I reduce this back to causality itself, and with a “system based” clause to exclude any outside influences. It seems to me that people often say they’re physicalists without really grasping the full implications. That should make it easier to cheat. Perhaps my clarification could help.
We can certainly use the term “law” in this regard as well, though to me it seems a bit epistemic Observe that scientists try to develop effective laws to explain observations. Politicians even pass or decree laws to constrain social function. The concept of “systemic causality” however is my attempt to conceptually get above mere human made rules.
Yes I’d say that under perfect systemic causality, ontologically there can be no freewill. I still think it’s useful for us to say that freewill exists epistemologically though. Here it should exist as a function of our ignorance — the less that’s understood about a given situation, the more freedom that a person may be seen to have. Conversely from a conceptual perfect understanding, no freedom should be seen. Thus even quantum dynamics should either be determinate from such a perspective, or magical given the failure of systemic causality.
I don’t mind “many worlds” in an epistemologically sense. So when someone says it seems like our world splits off into countless worlds each second given quantum function, I’m good with that. It’s when they posit that such other worlds actually exist that I accuse them of violating the tenets of systemic causality. Consider how this shouldn’t just be bajillions of actual worlds splitting off ours each second, but bajillions of worlds splitting off each of those each second as well, and so on resulting in infinite regress. I wonder if such an idea truly holds weight in academia today? Or am I simply mistaken about what’s claimed?
While we may tend to like thought experiments that conform with our own positions, I hesitate to presume that we all thus become uncritical. Some seem able to effectively criticize certain thought experiments even when they’re sympathetic with the ultimate message. That should at least be a sign that someone is still able to think effectively.
Fascinating! Ok, I’m trying to work through this model using Einstein’s “spooky action at a distance.” Because, in this case, Einstein essentially concluded that quantum entanglement was logically impossible since cause and effect couldn’t happen at the same time—especially not across a distance. It seems logically impossible because it didn’t adhere to our understanding of the physical universe at the time. BUT, as we now know, quantum entanglement is in fact an actual thing. So… would this be an example of something that is metaphysically possible but logically impossible? If so, how does that fit with the nesting model of possibility?
Or is this an inappropriate application of this lesson?
Great question! I had to think about this one for a bit.
I do wonder about the categorisation here. It seems to me that Einstein's claim was that it was physically impossible based on the understanding of physics at the time. I'm sure it seemed illogical to Einstein, but I don't think he meant logically impossible in the philosophical sense. Quantum entanglement seemed not physically possible, but it turned out to be physically possible after all.
Great answer. Thank you.
"Is Everything We Imagine Possible?"
Let's see:
*Opens Midjourney*
*Types "/imagine prompt: Darth Vader existing without a body while simultaneously being Luke's father and not being his father, traveling above the speed of light inside a black hole while never leaving it."
Hmmmm, Midjourney says "Maybe"? (https://cdn.midjourney.com/4cc7efb6-6a0d-4d69-b59c-9fb31a8a945a/0_2.png)
Another fun read, Suzi!
Brilliant! Although, shouldn't Darth Vader look more like spaghetti? Or, I guess the Force is stronger than spaghettification!
Sith-ification trumps spaghettification. It's the way of the universe.
I am becoming such a fan of this author.
Conceivability does not need to imply possibility to be useful.
And a hack towards imagining transparent iron: just imagine magnetic glass. Easy to do. Of course ignorance of the link between spin alignment and magnetism makes it easier but that is irrelevant. The body of knowledge we are ignorant of still vastly exceeds that which we know. We must not let this stop us
Thank you so much! I'm thrilled you're enjoying the articles.
I completely agree! Conceivability need not imply possibility to be useful. It's such an important idea, especially when we think about invention, innovation and creativity. After all, many groundbreaking inventions or theories started as what seemed like "impossible" ideas that someone dared to conceive.
Suzi—thank you for the thought experiment on thought experiments. A few comments:
In Hume’s “conceivability implies possibility” you point out this “allows us to explore counterfactual scenarios and potential alternative realities “. (Well said). Previously in the discussion on what is possible the diagram stops at ‘logical’ as an outer boundary. But what about counter-logical? This ‘illogical’ realm would be outside of the logical.
Then what of counter illogical? That is beyond logic/illogical and objective possibility? This counter-possible would be the ‘impossible’ realm —beyond the possible, or we might say the transcendent realm.
Each of the ‘possibilities’ in the diagram seem to be dependent upon this counterbalance concept: starting with the actual/physical (I’m not splitting these up because you haven’t taken up the details of the distinction yet)—counter actual/physical goes to metaphysical—counter metaphysical goes to logical only—(to extent this)—counter logical goes to illogical (beyond our current paradigm of logic and objective possibility)—and counter illogical/possibility goes to impossible (outside of our current understanding of possible—or ‘transcendent’). One could argue that counter transcendent goes to logical, metaphysical, and physical/actual. In this sense the existence of the actual, physical, metaphysical, and logical implies the existence of the transcendent. One might even say the physical etc. depends or relies upon the transcendent to be defined.
In the example of Darth Vader—the logical possibility is that he is Luke’s stepfather. Thus he is Luke’s father—the only one Luke has known as ‘Father’, and he is not Luke’s Father—not his biological father. Or perhaps Luke was the product of multiple fathers whose genetics were blended to create the Y chromosome haploid. Thus Darth Vader is Luke’s father (one of them) and is not Luke’s ‘father’—the sole source of his Haploid DNA. These are logical arguments within fiction writing for probabilities—stories are based upon what is probably not what is possible.
Story writing (imaging) is where the unintelligible becomes actual fiction—the substance of imagination. Take Mary Shepherd’s deciduous tree blooming in December example. If a southern hemisphere deciduous tree is moved to a greenhouse in the northern hemisphere (with heat, light, and nutrients appropriate) it will thrive in December and January. The ‘unintelligent’ becomes actual with a bit of imagination and creativity technology. The calendar being a non-universal ‘law’ of physics for flowering vegetation even on this planet.
That said, “the nature of things” is bound by time, location, culture, light (a whole lot of physics) which is the definition of ’thingyness’. The nature of non-things, or things beyond things is the realm and substance of concepts—which is one factor in imagings and the ‘conceivable’.
Here Peter Van Inwagen’s concept of the inconceivable transparent iron is actually a question of perspective scale. We don’t need to get to structural details of condensed matter physicists. Transparency is a function of scale. In an electron microscope of iron (NASA has published one from a moon rock). There is a lot of space. If we go deeper beyond this view the space between the quarks, between the protons, neutrons, and electrons in iron grow larger and larger. In a sense it is the space between these that makes iron ‘ironie’. At this scale iron is transparent. And at the scale of the entire universe iron is beyond ‘apparent’—it is therefore trans-parent (so to speak). My point is the statement ‘transparent iron’ is the concept (not inconceivable)—the argument is simply one of level of possibility, not impossibility.
So I wonder, what exists beyond conceivability? How can we break the bounds of concepts (and the possible) without the substance of thought and thought experiments?
NASA electron microscopes of iron crystal: https://science.nasa.gov/resource/scanning-electron-microscope-view-of-iron-crystal/
Hi Dean!
Wow, that's quite a journey through the realms of possibility and beyond! I love how you've extended the idea of "counter" concepts all the way to the transcendent.
Your point about transparent iron is fascinating. You're right -- so many "impossibilities" are really just a matter of perspective or current understanding.
You raise a great point about what exists beyond conceivability. We could ever know. How could we use thought to go beyond the limits of thought itself?
Hi Suzi. Love the questions you ask in your comment and I hope this finds its way into your research and writing someday. I look forward to reading your thoughts on this (yes, irony of thoughts on the limits of thought fully accepted). We could wonder if AI will be an effective tool (someday?) for us to push on this boundary—if it exists. Thanks to you as always for pushing our thinking.
Another beautiful piece, Suzi. You make me think about Kants understanding of the imagination, here linked to conceivability. Kant might argue that our ability to conceive of a flower even in the absence of a flower is imagination. On this view even the scientist who values sensory data cannot function in the absence of imagination. Do out of body experiences exist? Are they made of subatomic particles? In sociology—which focuses on abstract structures like economies, family, occupation, religion, education, etc. (all included in the three-part structure of ontology we/you discussed earlier)—what is the function of conceivable but impossible? Transparent iron exists at the level of quarks and what not. What about transparent schools?
Hey Terry!
You're so right! -- even in empirical sciences, imagination plays a crucial role - it's what allows us to form hypotheses and design experiments.
Your comment about "conceivable but impossible" in sociology got me thinking. How do we define "possible" and "impossible" when we're talking about things like economies or education systems?
Thanks for sharing, you really made me think.
One quibble: Anthropologists and sociologists consider themselves empirical scientists in that the rely on observations that are potentially verifiable by a disinterested third person. These fields don’t operate on fantasy, which is the only option for one claiming that something impossible logically is truly possible. Prudence is right. The notion of conceiving as possible or impossible based on the physical sciences (my term for what you are calling empirical science) is wrapped up in logic and physical reality. But it’s that third space as Popper called it, it’s not physical but it is real to us—Van Gogh, cave paintings, law of supply and demand, a school. Here, I think, empirical evidence requires a disciplined scientist with as you say nuances.
I completely agree about anthropologists and sociologists being empirical scientists. When I say "empirical sciences," I'm usually casting a pretty wide net. I tend to think anyone who uses observation and evidence-based reasoning is doing 'science', which definitely includes social sciences and anthropology.
Your point about the "third space" is great -- we navigate between the physical world and our conceptual frameworks. And its in this realm, things like economies or educational systems exist as real phenomena, even if we can't touch them the way we can touch a rock.
I'm curious about how you see physical phenomena informing this third space.
When sociologists or anthropologists (or even ethicist) are forming hypotheses or interpreting data, they seem to be engaging in a form of disciplined imagination -- but how much is that imagination informed by the physical world? They're not operating on fantasy, for sure, they are often conceiving of possibilities that haven't been directly observed yet, but I wonder how much imagination is influenced by past observations of behaviours and other physical phenomena.
This reminds me of the debate about whether science can inform morality (but let's not open that can of worms!).
Thanks for this mental workout!
Great article.
I think it’s useful to distinguish between imagination and conceivability. Conceivability means without logical contradiction, it’s a disciplined imagination. It’s a way of clarifying our concepts.
And the other thing is that while conceivability may not reliably tell us what is true, it is reliable in ruling out the false. So, when you say it doesn’t tell us “anything” about whether something is possible in the real world, it can tell us that square circles and married bachelors don’t exist. It rules out the impossible as false, rather than establishes the possible is actual.
Thinking about that in the context of the p-zombie thought experiment, it’s not suggesting zombies actually exist, it’s arguing “against” the idea there is an identity between mind and brain. If there was such an identity, there would be a contradiction in the concept of p-zombies.
Thanks Prudence!
Yes, the word "clearly" is doing a lot of heavy lifting in Hume's maxim. Thanks for pointing out the distinction between imagination and conceivability -- it's an important one.
Conceivability does seem to work better in the negative sense. We can't conceive of a square circle or a married bachelor so we rule them out. But I wonder about these examples. They are ruled out by definition -- a square is defined by having four sides and a bachelor is defined as an unmarried man. But I wonder about other sorts of claims. Like Einstein's claims about quantum entanglement. He thought that "spooky action at a distance" was not conceivable, and therefore impossible. But he turned out to be wrong.
It seems like there might be some room for different types of conceivability as well as the different types of possibility, so the relationship between the two could get quite nuanced.
That's an interesting take you have on the p-zombie thought experiment. I've always seen Chalmers' argument as primarily making a positive claim about conceivability. His second premise being -- "If it is conceivable that there be zombies, it is metaphysically possible that there be zombies".
But your point has be thinking.
I wonder if we might say that Chalmers makes a positive claim about conceivability to set up a scenario that then allows for the kind of logical analysis you're describing.
I think of it this way, logical possibility deals with the a priori, the concepts we’re using. If we say quantum entanglement is impossible, the first thing to do is clarify what those words or concepts mean. That will always depend on our current level of knowledge, which is (probably) wrong, or at least incomplete. But there's no alternative to using our best knowledge.
Einstein was dealing with physical possibility and the assumption nothing travels faster than light. If that’s true, then spooky action is impossible.
It’s my understanding that Saul Kripke first used zombies to argue against identity theory (an a priori claim). Chalmers used zombies to argue if consciousness doesn’t logically supervene on the physical (ie zombies are conceivable), there can be no reductive explanation. In a physicalist dominant environment, that’s radical, because it means physicalism is false.
This is really helpful. Thanks Prudence!
It’s useful for physical science, yes, and that is important, but may need adjustment to be useful in the social sciences.
https://terryunderwood.substack.com/p/a-world-without-words
I love everything about this line --
"Fragments model chaos"
We talk about these ideas all the time in literary circles when it comes to genre, suspension of disbelief, and historical progression. Great charts and breakdowns, and yet nothing quite gets at the empirical is-ness of things. That’s what keeps me reading and being.
Hi Nick!
It's the tension between our conceptual frameworks and the raw, experiential nature of reality that keeps us exploring, isn't it? Always more to ponder, always more to discover.
After Darth's sex change...
Silicon is a hard dark gray metal. Mix it with oxygen and...
The human imagination is bound only by our past experiences, fodder for the fertile mind. That and the words to describe them.
AI is imagining concepts, worlds and events beyond our own, and soon, beyond our comprehension and perhaps, beyond our possible.
I agree, our imaginations are deeply rooted in our experiences and the language we use to describe them.
Your point about AI potentially imagining concepts beyond our comprehension is intriguing. I'm not entirely convinced that AI, at least as we currently understand it, can truly come up with novel ideas in the way you suggest. Maybe.
Current AI systems, even the most advanced ones, are essentially processing and recombining existing information in sophisticated ways. They're not updating their 'ideas' in physical experiences the way biological minds do. This makes me wonder whether truly novel ideas - ones that go beyond the recombination of existing concepts (if such a thing exists) -- might need to be tested in the actual world.
What is the benefit to considering metaphysical possibilities that aren’t logically or physically possible? Logical possibilities seem useful to work towards our understanding of the physical world and pushing technological limits. However, metaphysical possibilities that fail to be logically or physically possible seem like a metaphorical hamster wheel without much of an end except mental gymnastics. I'm sure there is a benefit just can't put my finger on it?
Great question. I've been scratching my head with similar questions myself.
I think the problem is that metaphysical possibilities are not clearly defined. It seems we are permitted to include possible worlds here where the laws of physics are different from our own, and we can also include possible worlds where the laws of physics are exactly the same as ours. So we might have a metaphysical scenario that is logically possible (it doesn't involve any contradiction in logic) but is not physically possible because it doesn't align with our laws of physics. But we could also have a metaphysical scenario that is logically possible and doesn't involve any contradiction in our laws of physics -- but it's still not possible in our world given the current state of the world.
For example. I might like to say that it could have rained where I live this morning. That doesn't violate logic or the laws of physics in general. But given that this morning (where I live) there were no clouds in the sky, so, can I really say that it could have rained this morning?
When philosophers argue -- if I can conceive that it could have rained, so it's metaphysically possible that it could have rained -- I think we are allowed to question the grounds on which this conceiving is taking place.
That said, considering metaphysical possibilities, might have some use in prediction and scenario planning. Exploring possibilities, including those that might seem to violate our current laws of physics or the present state of the world, might be useful for understanding, anticipating novel situations, and innovation.
I do wonder how much the utility of this approach depends on our stance on determinism, though. If we lean towards a more deterministic view of the universe, where future states are entirely dictated by present conditions, the value of considering the metaphysically possible but physically implausible scenarios might be limited. On the other hand, if we believe in a less deterministic universe, exploring a wider range of metaphysical possibilities could prove more fruitful for prediction and innovation.
Really appreciate such a detailed and thought-provoking reply, thanks, Suzi!
Considering prediction and planning scenarios for innovation and development is a really good argument. Pushing the limits and trying for the impossible seems like a great motivating factor for inventors and engineers.
It seems to me philosophers in recent years have left out one kind of possibility Analogical. Isaac Newton belonged to a group of philosophers who were toying with the idea that perhaps Nature could be treated as a cipher. Some ciphers are breakable. If we treat Nature as cipher and succeed in breaking it, we know what Nature is about. Well, Newton and his fellow natural philosophers succeeded in making a beginning in this enterprise. Which enterprise continues to this day. A possibility became a reality many years ago. Perhaps this is the reason why we forget that it was once nothing more than a possibility?
I love how you've brought up the concept of analogical possibility and tied it to the historical development of science. We often forget how radical and "merely possible" some ideas seemed before they became established. It makes me wonder what analogies we're using today that might seem outlandish now but could lead to major breakthroughs.