I can imagine myself acting as if I’m without consciousness but I can’t imagine my consciousnessless twin acting like me with or without consciousness. Physicalism with a three-level ontology (to include institutions and museums as “real”) makes sense, and I can’t conceive of a P-Zombie.
I am somewhat amazed that Chalmers’ syllogism is taken seriously at all. It has a feel of the witch’s test in Monty Python. Nonsense predicates will produce nonsense conclusions.
The sense (illusion?) of consciousness is a state of matter or information or some undiscovered system. It may reside in a material form, as an emergent property of the brain (and body) or be interpreted or transduced by the organ from some universal consciousness.
But whether it is separate or one in being is a latter-day Arian conspiracy, and the syllogisms will parallel it unless and until we find an observable system that illuminates a key distinction.
For example, we may discover that the “intelligence” of LLMs or their successors is not inherent in the processors, but that those processors (here to include the software models and the silicon) are physicalizing consciousness that is inherent in the world in the way math is. The value of an alternative physical model is to allow distinction that removes ambiguity.
Hi Matt! You raise an interesting point about the validity of Chalmers' argument. Our current philosophical debates about consciousness might be as unresolvable as ancient theological debates. We need concrete, observable evidence to make real progress. But this is where we so often reach a stalemate. We can't agree on what that evidence would look like because what we are looking for is influenced by our starting assumptions.
I'm curious about your idea that consciousness is universal or an emergent property. Do you lean towards consciousness as a fundamental property of the universe or an emergent one?
I feel like there is something to be learned from the current AI work, even if it isn’t “conscious”
I used to be a staunch physicalist, and more and more I am open to the brain as an extraordinary instrument for accessing information that may not be wholly contained within it. I am not yet panpsychic.
What I do feel is that if we are talking about AI and the embodiment of intelligence and potentially consciousness in a machine, we have to:
A. Allow for a more thoughtful discussion of consciousness in most lifeforms
B. Consider that our consciousness is not emergent but compressive - an efficient model of something larger, like math as an algorithmic compression of physics or LLM as a compression of human expression.
Wow, going for the big one up front. I like it! And this is excellent Suzi! I particularly like the discussion of s-zombies.
Definitely, for a physicalist, the whole argument is circular. We have to assume non-physicalism for p-zombies to be conceivable, so the purported conclusion carries no wait.
But it seems like P&¬Q has a deeper issue, even for non-physicalists. If we can have P without Q, then Q makes no difference. It has no causal effects in the world. Even our discussions of it can't be affected by it. This version of Q has long struck me as fragile and brittle. A causal version of it seems much more robust and in line with out intuitions. But it also brings us back into physicalism and something accessible to science.
I also very much like the point that an epistemic gap doesn't necessarily imply an ontological one. Or that a *current* epistemic gap doesn't imply a forever unclonable one. It seems like making these assumptions leads to a lot of philosophical problems, problems that disappear if we just take a more cautious stance about what we know.
Thanks Mike! -- I'm glad you enjoyed the discussion on s-zombies!
Your point about Q having no causal effects if P can exist without it is compelling. It's a good point, I think, you have raised before. It's the interaction problem in substance dualism and epiphenomenalism. How can a non-physical mind interact with the physical world?
Unless we are an eliminative materialist, I think we all agree there is an epistemic gap. The debate, as I see it, is whether that gap in knowledge is closable, and whether that gap in knowledge implies a gap in reality.
Thanks Suzi. The epiphenomenal issue has been on my mind a lot lately. It's come up in a number of conversations (including ours). And it seems like it will be an issue in many of these thought experiments (Mary's room, inverted spectrum, etc.).
I'm actually skeptical about the epistemic gap. I'm a functionalist, but being a functionalist seems to involve an eliminative stance toward non-functional aspects of consciousness, at least for me. That isn't to say we don't have a lot of gaps in our understanding, but they seem scientifically tractable. Of course, it's always possible I'm missing something.
I must be missing something: if a zombie is defined to not experience consciousness, then how can anything about consciousness be implied by being able to imagine that something without consciousness exists?
Chalmers is using a specific argument -- the conceivability argument. This argument claims that we can know things by just thinking about them logically, without needing empirical evidence.
The conceivability argument suggests that if we can coherently imagine or conceive of something without logical contradiction, then it is at least possible in some sense. This philosophical approach attempts to derive conclusions from what we can conceive in our minds.
In the case of p-zombies, the claim is that we can conceive of beings physically identical to us but lacking consciousness. If this is truly conceivable without contradiction, the argument suggests it must be possible. And if it's possible, it implies consciousness is not necessarily tied to physical facts, challenging physicalism.
This method of argumentation is controversial, as it relies heavily on our intuitions and ability to conceive of scenarios completely and without hidden contradictions. Critics often question whether we can truly conceive of such scenarios without sneaking in hidden assumptions or overlooking logical inconsistencies. And, there's also questions about what it means to think about something logically, without needing empirical evidence. Is that even possible?
I have deep issues with the Chalmers argument. As you mentioned in your post, it seems to assume the existence of the non-physical in the first place. Also, many seem to consider non-humans as having no consciousness, yet cannot even define it, let alone discuss it in terms of non-physicality!
I didn’t study philosophy; my training is in mathematics. What I like about your post is that it makes me want to learn logic in a deeper way in a less mathematical framework. What I mean is to learn it as it applies to language other that mathematics. I find myself viewing things more abstractly and in symbols rather than natural language.
That's an interesting point of view -- I've always considered philosophy to be a lot like mathematics -- logic seems to be mathematics in language form. If you find any good resources that help with your learning -- pass them on! I'd be interested, too.
I think that mathematics differs from most philosophy mainly due to languages: the language of mathematics is formal and is built theorem by theorem using a formal logic, whereas analyzing a concept that is communicated by other languages seems to pose interpretation issues because of the ‘fuzzy’ nature of informal languages. I had a undergraduate course in philosophical logic and I found that while the approach is the same, the proper dissection of the language necessary to construct an argument properly depended on how the structure of a given statement or concept was interpreted.
I will gladly share whatever I find that helps clarify my understanding of all this.
Two things struck me about this article: first in time and impact, that Chalmers is a charlatan attempting to prove a religious thesis by bogus logic, and second, that a discussion of emergence would cast considerable light on the question (the discussion of H2O almost got there, but the emergence of flow when enough water molecules are present would have been more to the point, imo). And yet, as Douglas Hofstadter discussed at length in *Gödel, Escher, Bach*, the question of "something more vs nothing but" seems like an irreconcilable question of faith.
Robotics is soon going to present this question much more forcefully, as there will exist robots indistinguishable from humans in every observable way short of dissection who will be programmed to experience pain (yes, they can already do that) as an urgent danger signal to a central processor as well as pleasure (with attendant responses). Chalmers will say they aren't conscious, but the robots will argue that they are. Most people will believe them because of that thing about ducks: if it walks like one, quacks like one and looks like one then it must be one.
I like how you raised emergence and the future of robotics. Your point about robot pain adds an interesting twist. And, I agree, many will take the duck approach. But, perhaps, only if they are very very human-like.
There's an interesting phenomenon called the "uncanny valley" that might complicate this. As robots become more human-like, people tend to find them creepy or unsettling when they're almost, but not quite, human. This psychological effect might make us more skeptical of robots' claims to consciousness, even if they behave very similarly to humans.
What do you think? Do you think the uncanny valley effect would influence how we perceive robot consciousness?
Strangely, I have considered this question quite a bit in the context of sexbots. Right now, the sexbots are not in the uncanny valley, but some robots are on the edge of the other side. There have been some extremely important developments there - skin, most of all maybe. They can now implant 3d printed sensors by the thousands if not millions in nanotech based skin.
And they can put sexbots on skeletons identical to humans, although the motion (balance) is still a little iffy. So that leaves muscles in the skin still unsolved, I think. But I think it won’t be long before they get that, too. So my conclusion is that they’ll soon cross the uncanny valley in every respect. There are already processors that can reliably pass the Turing test.
And there will be a ton of money in sexbots both for the sex industry and also for the spying/government industry. To answer your specific question, yes, I think the uncanny valley would make a huge difference in our perception of consciousness - look how we can't even credit crows or dogs with self-perception. I think this is really just a human bias masquerading as science.
On crediting crows or dogs with self-perception. It wasn't that long ago that only half the crowd would raise their hands to the question of who thinks animals are conscious. But, in a recent public seminar I attended, almost all the hands went up (even for insects). So, I wonder whether the perception is changing.
I think I don't understand what is meant by *conceivable*. At least, I am confused.
You posit an s-zombie that can run twice as fast as me. Is it conceivable? It certainly is to me.
But then you go on to say that it's not conceivable because if it's physically identical to me, it's not physically possible to be twice as fast as me (muscles, sinews etc). This seems to me to sneak in the meaning of *possible* to say that it's not *conceivable*. That's cheating, no?
I can conceive of flying elephants even if they are not possible.
Separate point: if it's possible to conceive of a mysterious non-physical substance that gives me consciousness (but which there is no evidence of), surely it's possible to conceive of a go-faster-ness substance (which there is no evidence of) too?
I get ya Ragged Clown. “This seems to me to sneak in the meaning of *possible* to say that it's not *conceivable*. That's cheating, no?” Suzi knows about the ambiguity that traps the meaning of “imagine” in a loop with “conceive.” I’m with you. I can “imagine” an s-zombie but I can’t “conceive” one. The distinction I think is “I can believe it could be real.” This doesn’t apply to s-zombie. But I can imagine s-zombies like flying elephants
Conceivability and possibility are a minefield. Thinkers disagree about what is conceivable and they even disagree about what conceivable means.
Terry is right, I was (trying) to use 'conceivable' in a stricter philosophical sense, meaning "logically possible without contradiction," rather than simply "imaginable." So, I take it to mean that, to conceive of something without contradiction means we can think it through in detail without running into logical inconsistencies -- so it's more rigorous than just imagining something in my mind.
Your flying elephants example is a good one -- I can imagine them, but they might not be logically conceivable in the stricter sense.
But this brings up a broader discussion about the idea of 'a priori' knowledge. The conceivability argument supposes that we can know things through reason alone, without empirical evidence.
But we might want to question what this means. What does it mean to think about something logically, without needing empirical evidence? I might want to say that a square circle is illogical. But can I really make this claim without empirical evidence? When we are asked to conceive of something -- to think about whether it is logically consistent -- how much information about the world are we willing to add to premises before it becomes 'a posteriori'? And is it even possible to think about something logically, without needing empirical evidence?
I'm working on an article that digs into these sorts of issues.
Back in my younger days, I used to play a game called robocode where you'd write some computer code to make a robot fight a battle against other robots. Robocode was 100% software, but you could certainly build a similar game with physical robots made of aluminium and rubber but controlled by the same software. Each robot would be identical except for a few lines of software code.
Some folks would build a complex robot that would take all of the inputs (phenomena?) from its radar and build a model of the world (consciousness?) that it would use to make decisions about where to fire its gun and how to best fight its battle.
Other folks would write simpler code that would make the robot react directly to its inputs. For example, a robot might detect a robot to its right and it would turn to its right and fire its gun. No complex model (no consciousness) would be required. The behaviour of the two types of robot would be similar. Someone who was not experienced with the workings of the game would not be able to tell them apart.
Could we build a zombie that simply reacts to its inputs like the second set of robots with no complex model of the world? Is the difference between me and a p-zombie analogous to the difference in a few lines of software code in our robots? Is the computer code physical? Are the rules stored and executed in 85 billion neurons different from computer code?
Another commenter suggested that Chalmers and Co seem to be assuming a world where consciousness is not physical and his thought experiments rely on that assumption rather than proving it. Take away the assumption and the experiment proves nothing.
Oh I love this! This is a fantastic thought experiment.
I'm particularly fascinated in the way you have defined consciousness here. Is consciousness simply a complex model of the world, or is there something more to it?
I'm working on an upcoming article that touches on this issue, so I'd be very interested in your thoughts once it's done.
My guess is that the brain takes all these sensory inputs and past memories and different parts of the brain think about different aspects of the memories and the phenomena separately. There is some kind of voting scheme where each part of the brain shouts “my thoughts are the most important!” and the ‘winning thoughts’ spread across the brain until every part is thinking about the same thing. Consciousness is the illusion that this coherent thinking is something new and special.
Tying this back to my previous post about robocode, the consciousness has a model of the world that combines many elements from different segments of the brain and the whole brain has access to this model.
I can imagine a machine that acts like me but is not conscious. It might be beyond today's capabilities but I see nothing in theory that would prevent the construction of such a machine. Any motor action that I can do could be generated by a system of pulleys and levers. Any argument that a machine that acts like me must be conscious should be able to identify exactly the observable activity that is impossible for an unconscious machine to perform.
I find more interesting actual p-zombies - somnambulists and the walking concussed. They appear to be completely conscious but, in fact, are not. They show brain activity predominately in the cerebellum and brain stem, but not so much in the cortex.
I agree that reality is more interesting than hypothetical scenarios, but such people are not P-zombies. This is part of the problem; often when people try to imagine p-zombies, they actually imagine something completely different. If they are distinguishable from normal people by observation or brain scans or dissection **or literally any other possible method** they are not p zombies and completely miss the point of the thought experiment.
I think you’re right that something that superficially looks and acts like you, good enough to fool most people most of the time, could exist without having subjective experience. Ie a highly convincing android. I used to think such an android would have to be conscious; since the rise of LLMs I’m much less sure. It may be possible to fake the appearance by using a different kind of design.
But that also is not a p zombie because you could tell which is which with a metal detector or an autopsy etc. The p zombie thesis asserts not that a machine could pass the Turing test etc, but that it is literally impossible to tell if a person is a zombie or not BY ANY MEANS WHATSOEVER. If it is true, I could never have a good reason to believe that I am not the only non-zombie to ever exist.
Somnambulists and the walking concussed may not meet the strict definition of p-zombie but they effectively do lack phenomenal consciousness. Yet they appear to be conscious. Somnambulists can walk, talk, eat meals, and have sex. Yes, a brain scan would reveal a physical difference, but that difference shows us that consciousness really is physical. So, I don't see it as completely irrelevant to the thought experiment.
“Yes, a brain scan would reveal a physical difference, but that difference shows us that consciousness really is physical.”
Well, exactly.. But zombism asserts that that isn’t actually evidence of anything, at least not of consciousness. Because those brain scans are just the correlates of consciousness, or rather correlates of the sounds people make with their mouths when they claim they are conscious. And as we'll all heard, correlation doesn't necessarily establish causation. According to the thought experiment, all that could happen exactly the same without anyone actually being conscious. Or at least such a world is “conceivable” therefore it’s “metaphysically possible” therefore…ghosts? Or something.
If you can’t tell, I think it’s nonsense. I can conceive of being able to run twice as fast as I actually can with all the laws of physics of being the same as easily as I can conceive of p-zombies, but I don’t think imagining fantasy worlds is a useful method for learning the truth about reality. I know non-physicalists would take “ghosts” as an insulting straw man, but the way I see it, if you posit “something else” you need to specify what that something else is and provide evidence for it. The weight of evidence seems to suggest that consciousness is a property of brains; brains are arrangements of atoms that owe their powers to the functional organization they have due to their evolutionary history, and nothing else is necessary. It’s not enough to just say “I don’t understand how that could work, and you can’t explain it to my satisfaction, therefore magic.” Occam’s razor would seem to suggest that it’s just something we don’t understand yet. Yes as previously, it was inconceivable how combinations of “mere inert matter” could have the miraculous powers of living tissues, hence the theory of vitalism; the belief that some mysterious non-physical force was needed to animate inanimate matter. I think consciousness will turn out the same. I could be wrong. Maybe it will only make sense when we have breakthroughs that turn our understanding of physics inside out. Maybe such discoveries could vindicate some form of dualism. But that’s all just speculation. Regardless, I don’t think we’ll get any closer to the truth by imagining six impossible things before breakfast.
Excellent and thought provoking presentation. It does remind me of the trapped prior. I don’t think it proceeds to any new conclusion without an inclination to believe one way or the other at the outset. It is enjoyable though and refines our reasoning in showing us where the holes are. Thank you.
I had to think about it for a while, but I finally decided p-zombies are *not* conceivable. I don't know that it comes from a claim to understand the physical world, but from the observation that much in this world exists *because* we have experience. Why would zombie world ever have music, sports, movies, and theatre? Why would they even have aspirin? For me, a key question with thought experiments is how did they come to be? I cannot conceive of any evolution in any physical world that could lead to zombie world. Thus p-zombies are not conceivable to me.
I don't accept that one can will an entire world into existence by fiat. Human language is too facile at generating logical gibberish. The various forms of the Liar Paradox show how possible it is with language to make inherently impossible statements.
The second challenge seems to have legs, too. An epistemic gap *now* doesn't imply we won't close it later (as the H₂O example shows). I don't think we know enough now to make any claims about an epistemic gap. We're only in the foothills Mind Mountain. The gap speaks only to our ignorance now, not what we'll discover.
Even if the gap remains — if there proves something ineffable about consciousness — that doesn't deny physicalism. Science already understands there are areas inaccessible in principle. A full account of consciousness might be included, but that doesn't deny physicalism.
Even the third challenge works for me. Exactly so. Chalmers *assumes* zombie world lacks magical nonphysical Q, but physicalism insists that consciousness is physical, so what can zombie world lack?
As an aside, SF author Robert Sawyer wrote about p-zombies in his novel "Quantum Night". Turns out 4/7 of the population are p-zombies. Sawyer's character asserts, though, that Chalmers's zombies should be called philosopher's zombies because they are anything but philosophical (which is another count against their conceivability). The characters are Canadian, so they call them "p-zeds", which I liked.
I *am* prone to forgetting a "not" now and then, but in this case I didn't. Sawyer's character felt they should be called philosophER'S zombies rather than philosophICAL zombies. Because they *are* the former, but not the latter.
I still have no idea what the difference is between logical possibility versus metaphysical possibility. If Chalmers is talking about logical possibility then everything turns on the definition of consciousness…a sticky issue.
Water is still something like “life liquid” to me. H2O is simply the scientific designation. Nothing about the prescientific understanding of water has been negated by it.
I agree with Mike that consciousness should make a causal difference, but I don’t think that necessarily means phenomenal experience as such is open to third person observation. But I think in terms of a more classical understanding of consciousness as including desires, motivations, qualities like color and scent, emotions, ideas, morality, beauty, will, a stream of experience with both a background and intentionality that we experience and can’t help but assume about each other in our everyday lives. To call these things acausal seems absurd. I am commenting on this post because I’m interested in the topic, not because my particles are creating the illusion of my interest. If you believe nothing else exists outside of what physics currently allows, if you believe in a deterministic universe, this phenomenal “inner” world is going to be a very hard problem indeed.
Yep, good point! The physicalist faces a triple challenge here. They have to explain how subjective experiences could arise from physical processes. They need to grapple with the consequences such a view has on free will. And, as you pointed out, they must consider the implications of 'no free will' on the very necessity of conscious experiences. If our actions are determined, why do we need consciousness at all?
"Conceivable implies possibility"—if something can be imagined, it means there is a possibility for it to exist, even though it may not necessarily exist in our reality, but only in a possible world that might differ from ours.
This can be understood as follows
{ ... It is all about the question, "🧩 What is exist?" Exist is "what limit our perception"
📍"Exist" limiting in such a way that it changes our state of consciousness ... }
Thus, as long as something can alter or limit our perception, then it exists for us, albeit in different degrees of reality (more realistic, less realistic, and so on).
However, this knowledge does not provide a solution to the zombie thought experiment in terms of proving the existence of metaphysical (non-physical) consciousness. Why?
REBUTTAL. Because "physical & consciousness" can be considered two different things, as we can think of the things always have to be paired as "if physical exists, then non-physical exists", in the sense that one accepts the principle of pairing, "every empirical reality is paired," But it does not necessarily follow that if there is something physical, that the non-physical side doesn't have to be a consciousness. Why? Because we can't even imagine "consciousness in the imagined zombie," except merely imagining expressions of consciousness.
The notion that "existence" does not always refer to something physical, but rather to anything that affects the limits of our perception, remains true. However, the zombie thought experiment has not yet provided a definitive solution as to whether consciousness is non-physical.
Physicality and consciousness may not automatically be paired, and we cannot fully imagine consciousness in zombies, because we can only imagine its expressions, not the consciousness itself.
This shows that the debate on the relationship between physicality and consciousness remains open, and while the zombie argument is intriguing, it has not definitively proven that consciousness must be understood as something non-physical.
To properly assess this case, reasoning beyond mere cause-and-effect thinking is required—reasoning that transcends into absolutes.
I won’t delve too deeply here, but will merely offer an outline of this densely packed discussion.
Perhaps over time, I will gradually include more details. For now, this will suffice.
Good writeup Suzie, but I was confused by the last section.
You object that Chalmers begs the question by “defining Q as something that can be absent while all physical facts (P) remain the same”. But this isn’t the definition of Q, Q is just qualia (“an arbitrary phenomenal truth”).
If you agree we’re forced to say Q is non-physical, then you’ve encountered the force of the p-zombie argument. But that Q is non-physical isn't assumed anywhere in the argument. P&(not-Q) just describes a p-zombie. If you agree with P1 and P2, then yes, “we must conclude, therefore, that Q cannot be physical”.
Why is this such a big problem, why does it make physicalism false? Well, who is reading this, talking about this? Non-zombies. The existence of your qualia isn’t something you can plausibly deny.
And you ask who has the burden of proof? Anyone who wants to argue their position on the question. If we want to say zombies aren’t conceivable, then we need to demonstrate the contradiction.
Notice that for the s-zombie this is easy to do, we only need to appeal to the physiology of muscles etc. Running isn’t something over and above the movement of the physical parts. But that sort of explanation fails with qualia. Hence, the hard problem for the physicalist.
Chalmers seems to want us to consider whether consciousness is either physical or not physical. So, when he asks us to consider a world physically identical to ours in every respect but then asks us to imagine this world not having 'something' that is in our world, he seems to want us to consider whether that 'something' is physical or that 'something' is non-physical.
If the 'something' is non-physical, a physically identical world without the 'something' is conceivable (the zombie world).
But if the 'something' is physical, the concept of a physically identical world without that 'something' is self-contradictory.
By including the entire physical world in his thought experiment, Chalmers leaves no room for the 'something' to be physical. He removes that option.
I'm struggling to see this any other way. What am I missing?
There’s room in the first premise. If consciousness is physical, then yes, P&(not-Q) will be contradictory. Which is just to say, p-zombies will be inconceivable. So, the physicalist just needs to point out the contradiction.
There’s an excellent comment here from Alex Popescu. He frames the issue as one of structure and situates it in the ongoing work in philosophy. We can then think of physics as structural and consciousness as intrinsic. We have the same physical/non-physical contrast, but it’s informative exactly what we mean. (An added bonus, it doesn’t trigger anyone’s ideological commitments).
Chalmers does talk about physics explanations being limited to structure and function. If you see the issue this way, that the language of physics is limited to describing structure and function, and consciousness is intrinsic, then it at least shows how intractable the problem is.
I think we can put forward a more charitable version of the conceivability arguments, which doesn’t require that we beg the question against physicalism. On this version, it’s not that Q inherently requires a non-physical truth, it’s that it requires an assumption that Q is a non-structural truth. You did briefly mention structure at one point, but I don’t see it anywhere in your actual discussion of the conceivability arguments.
If you read Chalmers original article from 1995 (https://consc.net/papers/facing.pdf) on the hard problem (and subsequent work) as well other phenomenal “realists” like Goff, it’s obvious that they take the issue to be one of structure.
Reformulated this way, we would see that the conclusion follows from the assumption that if physical states are purely structural, and phenomenal states intrinsic (aka non-structural), then it’s not possible that the latter could be reduced to the former.
Why think this? The thesis that physical truths have to be purely structural follows from the work done on epistemic structural realism. Basically, if we accept the microphysical thesis (that all physical truths are equivalent to or reducible to microphysical truths) then all physical truths have to be captured in the mathematical vocabulary of physics (so it will be structural).
Could we deny this? We could by, for example, denying the microphysical thesis (we can see now the reason why Chalmers adopts this premise). Chalmers and co. do of course consider the implications entailed by rejecting the microphysical thesis elsewhere but reject it on the grounds that a) either such intrinsic physical truths explain consciousness- in which case it seems that we are led to some kind of weird physicalist alternative like Russellian monism- which people like Goff tend to endorse, or b) they do not explain consciousness, in which case they are irrelevant, and the hard problem still goes through.
What about the assertion that phenomenal truths have to be non-structural? This, I take it, is the very reason for formulating the argument in terms of a conceivability-possibility divided. The assumption that phenomenal truths are intrinsic is supposed to follow from the fact that we conceive of them that way. As Chalmers puts it, the qualities that we find in experience, like redness, don’t appear to be structurally decomposable-they are presented to us as being simple properties of objects like apples, as though they existed in a non-physical edenic environment. This of course is contrary to the findings of modern science, which tells us that redness is a complex property, structurally decomposable into EM waves and so forth.
Of course this just pushes the argument one step back, namely that we have to further assume that our phenomenal insight is reliable. That when we have an experience of a phenomenal quality as intrinsic, we are therefore justified in thinking that it actually is intrinsic. To support this premise, Goff and co. have argued for a thesis of ‘revelation’ (see here: https://assets.zyrosite.com/Aq2ZN1NGQMILGL8l/real_acquaintance_and_physicalism-dWxM5kOPGWCzLaBG.pdf) wherein we are justified in assuming that our introspective capacities are reliable.
In short, if you grant the above, then the conceivability arguments are certainly not circular! There’s a reason that proponents of phenomenal realism have spent a lot of time justifying the reliability of our phenomenal insight (in opposition to the illusionists), and a reason that they have framed the argument as being about a structural-intrinsic divide from the very beginning.
Hence, the focus of many physicalists (who have engaged in the literature) on denying the reliability of our introspective powers, like the illusionists do. The alternative approach, to deny the purely structural nature of physical truths, has largely fallen out of favor as a workable approach I gather.
To sum up, the conceivability arguments are only circular if you assume that there is no justification for the premises of the arguments, but obviously that’s not the case, otherwise nobody would take them seriously to begin with!
Yes I think that’s the main component of the debate (whether phenomenal qualities are structural). That said, I’ve spoken with Chalmers via email in the past on this exact issue and he insisted that some version of the conceivability argument should still go through even if phenomenal states were purely structural. That is, he thinks there remains an explanatory gap between physical states and phenomenal states- even if both are conceived of as being purely structural. So make of that as you will.
Others in the literature do think that the explanatory gap, and all the associated arguments (e.g. conceivability, knowledge) just reduces to a structural-phenomenal gap though. I’m thinking of Torin Alter in particular. What I will say is that the main aim of these arguments is to point at something which proponents think that phenomenal states have but physical states lack. What that something is, of course, has to be spelled out in some detail in order to justify the premises of these arguments. In the absence of such justification, it’s probably best to just view such arguments as intuition pumps and nothing more.
So far the focus of most phenomenal realists (including Chalmers) has been on the structural-phenomenal gap, but it’s also possible to try to point to some other attribute. I would say that in the absence of the structural-phenomenal gap, defending such arguments becomes much much harder.
For what it’s worth, I’m a physicalist who too believes that phenomenal states are complex states that can be fully structurally decomposable. But I arrived at this position after spending most of my life on the other fence.
The water-without-oxygen and philosophical zombie examples aren't quite analogous. We can imagine oxygen-free water only by conveniently ignoring water's chemical composition - a fundamental physical property. But Chalmers' zombie argument is more complex. It asks us to consider a world physically identical to ours in every respect, yet devoid of consciousness.
By definition, this zombie world includes all physical properties and physical phenomena. So we can't simply disregard inconvenient facts as we did with the water example. Instead, we're forced to rely on our preconceptions about consciousness to determine whether such a scenario is truly conceivable. This seems problematic to me.
The zombie argument's strength hinges on our ability to imagine a world that's physically indistinguishable from ours, yet lacks subjective experience. But can we really do this without sneaking in our assumptions about the nature of consciousness? I'm skeptical.
It strikes me right off the top, that I can't imagine a creature that acts just like me, but is not conscious.
I find it difficult to conceive of such a thing too.
I can imagine myself acting as if I’m without consciousness but I can’t imagine my consciousnessless twin acting like me with or without consciousness. Physicalism with a three-level ontology (to include institutions and museums as “real”) makes sense, and I can’t conceive of a P-Zombie.
I am somewhat amazed that Chalmers’ syllogism is taken seriously at all. It has a feel of the witch’s test in Monty Python. Nonsense predicates will produce nonsense conclusions.
The sense (illusion?) of consciousness is a state of matter or information or some undiscovered system. It may reside in a material form, as an emergent property of the brain (and body) or be interpreted or transduced by the organ from some universal consciousness.
But whether it is separate or one in being is a latter-day Arian conspiracy, and the syllogisms will parallel it unless and until we find an observable system that illuminates a key distinction.
For example, we may discover that the “intelligence” of LLMs or their successors is not inherent in the processors, but that those processors (here to include the software models and the silicon) are physicalizing consciousness that is inherent in the world in the way math is. The value of an alternative physical model is to allow distinction that removes ambiguity.
Hi Matt! You raise an interesting point about the validity of Chalmers' argument. Our current philosophical debates about consciousness might be as unresolvable as ancient theological debates. We need concrete, observable evidence to make real progress. But this is where we so often reach a stalemate. We can't agree on what that evidence would look like because what we are looking for is influenced by our starting assumptions.
I'm curious about your idea that consciousness is universal or an emergent property. Do you lean towards consciousness as a fundamental property of the universe or an emergent one?
I feel like there is something to be learned from the current AI work, even if it isn’t “conscious”
I used to be a staunch physicalist, and more and more I am open to the brain as an extraordinary instrument for accessing information that may not be wholly contained within it. I am not yet panpsychic.
What I do feel is that if we are talking about AI and the embodiment of intelligence and potentially consciousness in a machine, we have to:
A. Allow for a more thoughtful discussion of consciousness in most lifeforms
B. Consider that our consciousness is not emergent but compressive - an efficient model of something larger, like math as an algorithmic compression of physics or LLM as a compression of human expression.
Just musings
Wow, going for the big one up front. I like it! And this is excellent Suzi! I particularly like the discussion of s-zombies.
Definitely, for a physicalist, the whole argument is circular. We have to assume non-physicalism for p-zombies to be conceivable, so the purported conclusion carries no wait.
But it seems like P&¬Q has a deeper issue, even for non-physicalists. If we can have P without Q, then Q makes no difference. It has no causal effects in the world. Even our discussions of it can't be affected by it. This version of Q has long struck me as fragile and brittle. A causal version of it seems much more robust and in line with out intuitions. But it also brings us back into physicalism and something accessible to science.
I also very much like the point that an epistemic gap doesn't necessarily imply an ontological one. Or that a *current* epistemic gap doesn't imply a forever unclonable one. It seems like making these assumptions leads to a lot of philosophical problems, problems that disappear if we just take a more cautious stance about what we know.
Thanks Mike! -- I'm glad you enjoyed the discussion on s-zombies!
Your point about Q having no causal effects if P can exist without it is compelling. It's a good point, I think, you have raised before. It's the interaction problem in substance dualism and epiphenomenalism. How can a non-physical mind interact with the physical world?
Unless we are an eliminative materialist, I think we all agree there is an epistemic gap. The debate, as I see it, is whether that gap in knowledge is closable, and whether that gap in knowledge implies a gap in reality.
Thanks Suzi. The epiphenomenal issue has been on my mind a lot lately. It's come up in a number of conversations (including ours). And it seems like it will be an issue in many of these thought experiments (Mary's room, inverted spectrum, etc.).
I'm actually skeptical about the epistemic gap. I'm a functionalist, but being a functionalist seems to involve an eliminative stance toward non-functional aspects of consciousness, at least for me. That isn't to say we don't have a lot of gaps in our understanding, but they seem scientifically tractable. Of course, it's always possible I'm missing something.
I agree. The gap, to me, seems like a gap in our current understanding. But this gap in understanding is closable (at least in principle).
I must be missing something: if a zombie is defined to not experience consciousness, then how can anything about consciousness be implied by being able to imagine that something without consciousness exists?
Great question, Jack!
Chalmers is using a specific argument -- the conceivability argument. This argument claims that we can know things by just thinking about them logically, without needing empirical evidence.
The conceivability argument suggests that if we can coherently imagine or conceive of something without logical contradiction, then it is at least possible in some sense. This philosophical approach attempts to derive conclusions from what we can conceive in our minds.
In the case of p-zombies, the claim is that we can conceive of beings physically identical to us but lacking consciousness. If this is truly conceivable without contradiction, the argument suggests it must be possible. And if it's possible, it implies consciousness is not necessarily tied to physical facts, challenging physicalism.
This method of argumentation is controversial, as it relies heavily on our intuitions and ability to conceive of scenarios completely and without hidden contradictions. Critics often question whether we can truly conceive of such scenarios without sneaking in hidden assumptions or overlooking logical inconsistencies. And, there's also questions about what it means to think about something logically, without needing empirical evidence. Is that even possible?
I have deep issues with the Chalmers argument. As you mentioned in your post, it seems to assume the existence of the non-physical in the first place. Also, many seem to consider non-humans as having no consciousness, yet cannot even define it, let alone discuss it in terms of non-physicality!
I didn’t study philosophy; my training is in mathematics. What I like about your post is that it makes me want to learn logic in a deeper way in a less mathematical framework. What I mean is to learn it as it applies to language other that mathematics. I find myself viewing things more abstractly and in symbols rather than natural language.
Thanks so much for sharing your work!
That's an interesting point of view -- I've always considered philosophy to be a lot like mathematics -- logic seems to be mathematics in language form. If you find any good resources that help with your learning -- pass them on! I'd be interested, too.
I think that mathematics differs from most philosophy mainly due to languages: the language of mathematics is formal and is built theorem by theorem using a formal logic, whereas analyzing a concept that is communicated by other languages seems to pose interpretation issues because of the ‘fuzzy’ nature of informal languages. I had a undergraduate course in philosophical logic and I found that while the approach is the same, the proper dissection of the language necessary to construct an argument properly depended on how the structure of a given statement or concept was interpreted.
I will gladly share whatever I find that helps clarify my understanding of all this.
Yes, language does seem to get in the way a lot, doesn't it? Philosophy so often has to deal with the fuzziness of everyday language.
Two things struck me about this article: first in time and impact, that Chalmers is a charlatan attempting to prove a religious thesis by bogus logic, and second, that a discussion of emergence would cast considerable light on the question (the discussion of H2O almost got there, but the emergence of flow when enough water molecules are present would have been more to the point, imo). And yet, as Douglas Hofstadter discussed at length in *Gödel, Escher, Bach*, the question of "something more vs nothing but" seems like an irreconcilable question of faith.
Robotics is soon going to present this question much more forcefully, as there will exist robots indistinguishable from humans in every observable way short of dissection who will be programmed to experience pain (yes, they can already do that) as an urgent danger signal to a central processor as well as pleasure (with attendant responses). Chalmers will say they aren't conscious, but the robots will argue that they are. Most people will believe them because of that thing about ducks: if it walks like one, quacks like one and looks like one then it must be one.
Hi Jack!
I like how you raised emergence and the future of robotics. Your point about robot pain adds an interesting twist. And, I agree, many will take the duck approach. But, perhaps, only if they are very very human-like.
There's an interesting phenomenon called the "uncanny valley" that might complicate this. As robots become more human-like, people tend to find them creepy or unsettling when they're almost, but not quite, human. This psychological effect might make us more skeptical of robots' claims to consciousness, even if they behave very similarly to humans.
What do you think? Do you think the uncanny valley effect would influence how we perceive robot consciousness?
Strangely, I have considered this question quite a bit in the context of sexbots. Right now, the sexbots are not in the uncanny valley, but some robots are on the edge of the other side. There have been some extremely important developments there - skin, most of all maybe. They can now implant 3d printed sensors by the thousands if not millions in nanotech based skin.
And they can put sexbots on skeletons identical to humans, although the motion (balance) is still a little iffy. So that leaves muscles in the skin still unsolved, I think. But I think it won’t be long before they get that, too. So my conclusion is that they’ll soon cross the uncanny valley in every respect. There are already processors that can reliably pass the Turing test.
And there will be a ton of money in sexbots both for the sex industry and also for the spying/government industry. To answer your specific question, yes, I think the uncanny valley would make a huge difference in our perception of consciousness - look how we can't even credit crows or dogs with self-perception. I think this is really just a human bias masquerading as science.
Interesting! If you haven't read it already, I think you'll find Gunnar's article on sex robots interesting. He raises some similar points. https://subtlesparks.substack.com/p/sex-robots-and-the-future-of-physical
On crediting crows or dogs with self-perception. It wasn't that long ago that only half the crowd would raise their hands to the question of who thinks animals are conscious. But, in a recent public seminar I attended, almost all the hands went up (even for insects). So, I wonder whether the perception is changing.
That would be a radical change, long overdue imo.
I think I don't understand what is meant by *conceivable*. At least, I am confused.
You posit an s-zombie that can run twice as fast as me. Is it conceivable? It certainly is to me.
But then you go on to say that it's not conceivable because if it's physically identical to me, it's not physically possible to be twice as fast as me (muscles, sinews etc). This seems to me to sneak in the meaning of *possible* to say that it's not *conceivable*. That's cheating, no?
I can conceive of flying elephants even if they are not possible.
Separate point: if it's possible to conceive of a mysterious non-physical substance that gives me consciousness (but which there is no evidence of), surely it's possible to conceive of a go-faster-ness substance (which there is no evidence of) too?
I get ya Ragged Clown. “This seems to me to sneak in the meaning of *possible* to say that it's not *conceivable*. That's cheating, no?” Suzi knows about the ambiguity that traps the meaning of “imagine” in a loop with “conceive.” I’m with you. I can “imagine” an s-zombie but I can’t “conceive” one. The distinction I think is “I can believe it could be real.” This doesn’t apply to s-zombie. But I can imagine s-zombies like flying elephants
Great point!
Conceivability and possibility are a minefield. Thinkers disagree about what is conceivable and they even disagree about what conceivable means.
Terry is right, I was (trying) to use 'conceivable' in a stricter philosophical sense, meaning "logically possible without contradiction," rather than simply "imaginable." So, I take it to mean that, to conceive of something without contradiction means we can think it through in detail without running into logical inconsistencies -- so it's more rigorous than just imagining something in my mind.
Your flying elephants example is a good one -- I can imagine them, but they might not be logically conceivable in the stricter sense.
But this brings up a broader discussion about the idea of 'a priori' knowledge. The conceivability argument supposes that we can know things through reason alone, without empirical evidence.
But we might want to question what this means. What does it mean to think about something logically, without needing empirical evidence? I might want to say that a square circle is illogical. But can I really make this claim without empirical evidence? When we are asked to conceive of something -- to think about whether it is logically consistent -- how much information about the world are we willing to add to premises before it becomes 'a posteriori'? And is it even possible to think about something logically, without needing empirical evidence?
I'm working on an article that digs into these sorts of issues.
Thank you, Suzi! I think, I've got it.
Back in my younger days, I used to play a game called robocode where you'd write some computer code to make a robot fight a battle against other robots. Robocode was 100% software, but you could certainly build a similar game with physical robots made of aluminium and rubber but controlled by the same software. Each robot would be identical except for a few lines of software code.
Some folks would build a complex robot that would take all of the inputs (phenomena?) from its radar and build a model of the world (consciousness?) that it would use to make decisions about where to fire its gun and how to best fight its battle.
Other folks would write simpler code that would make the robot react directly to its inputs. For example, a robot might detect a robot to its right and it would turn to its right and fire its gun. No complex model (no consciousness) would be required. The behaviour of the two types of robot would be similar. Someone who was not experienced with the workings of the game would not be able to tell them apart.
Could we build a zombie that simply reacts to its inputs like the second set of robots with no complex model of the world? Is the difference between me and a p-zombie analogous to the difference in a few lines of software code in our robots? Is the computer code physical? Are the rules stored and executed in 85 billion neurons different from computer code?
Another commenter suggested that Chalmers and Co seem to be assuming a world where consciousness is not physical and his thought experiments rely on that assumption rather than proving it. Take away the assumption and the experiment proves nothing.
Oh I love this! This is a fantastic thought experiment.
I'm particularly fascinated in the way you have defined consciousness here. Is consciousness simply a complex model of the world, or is there something more to it?
I'm working on an upcoming article that touches on this issue, so I'd be very interested in your thoughts once it's done.
My guess is that the brain takes all these sensory inputs and past memories and different parts of the brain think about different aspects of the memories and the phenomena separately. There is some kind of voting scheme where each part of the brain shouts “my thoughts are the most important!” and the ‘winning thoughts’ spread across the brain until every part is thinking about the same thing. Consciousness is the illusion that this coherent thinking is something new and special.
Tying this back to my previous post about robocode, the consciousness has a model of the world that combines many elements from different segments of the brain and the whole brain has access to this model.
I can imagine a machine that acts like me but is not conscious. It might be beyond today's capabilities but I see nothing in theory that would prevent the construction of such a machine. Any motor action that I can do could be generated by a system of pulleys and levers. Any argument that a machine that acts like me must be conscious should be able to identify exactly the observable activity that is impossible for an unconscious machine to perform.
I find more interesting actual p-zombies - somnambulists and the walking concussed. They appear to be completely conscious but, in fact, are not. They show brain activity predominately in the cerebellum and brain stem, but not so much in the cortex.
I agree that reality is more interesting than hypothetical scenarios, but such people are not P-zombies. This is part of the problem; often when people try to imagine p-zombies, they actually imagine something completely different. If they are distinguishable from normal people by observation or brain scans or dissection **or literally any other possible method** they are not p zombies and completely miss the point of the thought experiment.
I think you’re right that something that superficially looks and acts like you, good enough to fool most people most of the time, could exist without having subjective experience. Ie a highly convincing android. I used to think such an android would have to be conscious; since the rise of LLMs I’m much less sure. It may be possible to fake the appearance by using a different kind of design.
But that also is not a p zombie because you could tell which is which with a metal detector or an autopsy etc. The p zombie thesis asserts not that a machine could pass the Turing test etc, but that it is literally impossible to tell if a person is a zombie or not BY ANY MEANS WHATSOEVER. If it is true, I could never have a good reason to believe that I am not the only non-zombie to ever exist.
Somnambulists and the walking concussed may not meet the strict definition of p-zombie but they effectively do lack phenomenal consciousness. Yet they appear to be conscious. Somnambulists can walk, talk, eat meals, and have sex. Yes, a brain scan would reveal a physical difference, but that difference shows us that consciousness really is physical. So, I don't see it as completely irrelevant to the thought experiment.
“Yes, a brain scan would reveal a physical difference, but that difference shows us that consciousness really is physical.”
Well, exactly.. But zombism asserts that that isn’t actually evidence of anything, at least not of consciousness. Because those brain scans are just the correlates of consciousness, or rather correlates of the sounds people make with their mouths when they claim they are conscious. And as we'll all heard, correlation doesn't necessarily establish causation. According to the thought experiment, all that could happen exactly the same without anyone actually being conscious. Or at least such a world is “conceivable” therefore it’s “metaphysically possible” therefore…ghosts? Or something.
If you can’t tell, I think it’s nonsense. I can conceive of being able to run twice as fast as I actually can with all the laws of physics of being the same as easily as I can conceive of p-zombies, but I don’t think imagining fantasy worlds is a useful method for learning the truth about reality. I know non-physicalists would take “ghosts” as an insulting straw man, but the way I see it, if you posit “something else” you need to specify what that something else is and provide evidence for it. The weight of evidence seems to suggest that consciousness is a property of brains; brains are arrangements of atoms that owe their powers to the functional organization they have due to their evolutionary history, and nothing else is necessary. It’s not enough to just say “I don’t understand how that could work, and you can’t explain it to my satisfaction, therefore magic.” Occam’s razor would seem to suggest that it’s just something we don’t understand yet. Yes as previously, it was inconceivable how combinations of “mere inert matter” could have the miraculous powers of living tissues, hence the theory of vitalism; the belief that some mysterious non-physical force was needed to animate inanimate matter. I think consciousness will turn out the same. I could be wrong. Maybe it will only make sense when we have breakthroughs that turn our understanding of physics inside out. Maybe such discoveries could vindicate some form of dualism. But that’s all just speculation. Regardless, I don’t think we’ll get any closer to the truth by imagining six impossible things before breakfast.
We don't fundamentally disagree. Chalmers argument fails because conceivability does not imply physical realizability.
Well said.
FWIW I find it harder to imagine that miraculous power of living than I do consciousness. Consciousness seems much more straightforward to me.
Language. Can’t live with it, can’t live without it. You say conception, I say concoction
Well said!
Excellent and thought provoking presentation. It does remind me of the trapped prior. I don’t think it proceeds to any new conclusion without an inclination to believe one way or the other at the outset. It is enjoyable though and refines our reasoning in showing us where the holes are. Thank you.
Great point, John! How do we break free from our inclinations pushing us one way or the other? It's a challenge!
I had to think about it for a while, but I finally decided p-zombies are *not* conceivable. I don't know that it comes from a claim to understand the physical world, but from the observation that much in this world exists *because* we have experience. Why would zombie world ever have music, sports, movies, and theatre? Why would they even have aspirin? For me, a key question with thought experiments is how did they come to be? I cannot conceive of any evolution in any physical world that could lead to zombie world. Thus p-zombies are not conceivable to me.
I don't accept that one can will an entire world into existence by fiat. Human language is too facile at generating logical gibberish. The various forms of the Liar Paradox show how possible it is with language to make inherently impossible statements.
The second challenge seems to have legs, too. An epistemic gap *now* doesn't imply we won't close it later (as the H₂O example shows). I don't think we know enough now to make any claims about an epistemic gap. We're only in the foothills Mind Mountain. The gap speaks only to our ignorance now, not what we'll discover.
Even if the gap remains — if there proves something ineffable about consciousness — that doesn't deny physicalism. Science already understands there are areas inaccessible in principle. A full account of consciousness might be included, but that doesn't deny physicalism.
Even the third challenge works for me. Exactly so. Chalmers *assumes* zombie world lacks magical nonphysical Q, but physicalism insists that consciousness is physical, so what can zombie world lack?
As an aside, SF author Robert Sawyer wrote about p-zombies in his novel "Quantum Night". Turns out 4/7 of the population are p-zombies. Sawyer's character asserts, though, that Chalmers's zombies should be called philosopher's zombies because they are anything but philosophical (which is another count against their conceivability). The characters are Canadian, so they call them "p-zeds", which I liked.
Well now you’re talking! I am going to read this book. Thank you, John.
It was the first Sawyer I read, and it made me an instant fan.
I think there are big chunks of philosophy that should pay more attention to evolution. Ethics, for example, and aesthetics.
PS. I think you might be missing a ‘not’ in your last paragraph… “should NOT be called”.
I *am* prone to forgetting a "not" now and then, but in this case I didn't. Sawyer's character felt they should be called philosophER'S zombies rather than philosophICAL zombies. Because they *are* the former, but not the latter.
Yes. Got it. Sorry!
I still have no idea what the difference is between logical possibility versus metaphysical possibility. If Chalmers is talking about logical possibility then everything turns on the definition of consciousness…a sticky issue.
Water is still something like “life liquid” to me. H2O is simply the scientific designation. Nothing about the prescientific understanding of water has been negated by it.
I agree with Mike that consciousness should make a causal difference, but I don’t think that necessarily means phenomenal experience as such is open to third person observation. But I think in terms of a more classical understanding of consciousness as including desires, motivations, qualities like color and scent, emotions, ideas, morality, beauty, will, a stream of experience with both a background and intentionality that we experience and can’t help but assume about each other in our everyday lives. To call these things acausal seems absurd. I am commenting on this post because I’m interested in the topic, not because my particles are creating the illusion of my interest. If you believe nothing else exists outside of what physics currently allows, if you believe in a deterministic universe, this phenomenal “inner” world is going to be a very hard problem indeed.
Yep, good point! The physicalist faces a triple challenge here. They have to explain how subjective experiences could arise from physical processes. They need to grapple with the consequences such a view has on free will. And, as you pointed out, they must consider the implications of 'no free will' on the very necessity of conscious experiences. If our actions are determined, why do we need consciousness at all?
EXIST, CONSCIOUSNESS & DEGREE OF REALITY - Zombie Gedanken Experiment
Next Chapter
https://substack.com/profile/78789807-seremonia/note/c-69444344?r=1awqlr
"Conceivable implies possibility"—if something can be imagined, it means there is a possibility for it to exist, even though it may not necessarily exist in our reality, but only in a possible world that might differ from ours.
This can be understood as follows
{ ... It is all about the question, "🧩 What is exist?" Exist is "what limit our perception"
📍"Exist" limiting in such a way that it changes our state of consciousness ... }
Thus, as long as something can alter or limit our perception, then it exists for us, albeit in different degrees of reality (more realistic, less realistic, and so on).
However, this knowledge does not provide a solution to the zombie thought experiment in terms of proving the existence of metaphysical (non-physical) consciousness. Why?
REBUTTAL. Because "physical & consciousness" can be considered two different things, as we can think of the things always have to be paired as "if physical exists, then non-physical exists", in the sense that one accepts the principle of pairing, "every empirical reality is paired," But it does not necessarily follow that if there is something physical, that the non-physical side doesn't have to be a consciousness. Why? Because we can't even imagine "consciousness in the imagined zombie," except merely imagining expressions of consciousness.
The notion that "existence" does not always refer to something physical, but rather to anything that affects the limits of our perception, remains true. However, the zombie thought experiment has not yet provided a definitive solution as to whether consciousness is non-physical.
Physicality and consciousness may not automatically be paired, and we cannot fully imagine consciousness in zombies, because we can only imagine its expressions, not the consciousness itself.
This shows that the debate on the relationship between physicality and consciousness remains open, and while the zombie argument is intriguing, it has not definitively proven that consciousness must be understood as something non-physical.
To properly assess this case, reasoning beyond mere cause-and-effect thinking is required—reasoning that transcends into absolutes.
I won’t delve too deeply here, but will merely offer an outline of this densely packed discussion.
Perhaps over time, I will gradually include more details. For now, this will suffice.
Good writeup Suzie, but I was confused by the last section.
You object that Chalmers begs the question by “defining Q as something that can be absent while all physical facts (P) remain the same”. But this isn’t the definition of Q, Q is just qualia (“an arbitrary phenomenal truth”).
If you agree we’re forced to say Q is non-physical, then you’ve encountered the force of the p-zombie argument. But that Q is non-physical isn't assumed anywhere in the argument. P&(not-Q) just describes a p-zombie. If you agree with P1 and P2, then yes, “we must conclude, therefore, that Q cannot be physical”.
Why is this such a big problem, why does it make physicalism false? Well, who is reading this, talking about this? Non-zombies. The existence of your qualia isn’t something you can plausibly deny.
And you ask who has the burden of proof? Anyone who wants to argue their position on the question. If we want to say zombies aren’t conceivable, then we need to demonstrate the contradiction.
Notice that for the s-zombie this is easy to do, we only need to appeal to the physiology of muscles etc. Running isn’t something over and above the movement of the physical parts. But that sort of explanation fails with qualia. Hence, the hard problem for the physicalist.
Thanks Prudence!
I might push back a little.
Chalmers seems to want us to consider whether consciousness is either physical or not physical. So, when he asks us to consider a world physically identical to ours in every respect but then asks us to imagine this world not having 'something' that is in our world, he seems to want us to consider whether that 'something' is physical or that 'something' is non-physical.
If the 'something' is non-physical, a physically identical world without the 'something' is conceivable (the zombie world).
But if the 'something' is physical, the concept of a physically identical world without that 'something' is self-contradictory.
By including the entire physical world in his thought experiment, Chalmers leaves no room for the 'something' to be physical. He removes that option.
I'm struggling to see this any other way. What am I missing?
There’s room in the first premise. If consciousness is physical, then yes, P&(not-Q) will be contradictory. Which is just to say, p-zombies will be inconceivable. So, the physicalist just needs to point out the contradiction.
There’s an excellent comment here from Alex Popescu. He frames the issue as one of structure and situates it in the ongoing work in philosophy. We can then think of physics as structural and consciousness as intrinsic. We have the same physical/non-physical contrast, but it’s informative exactly what we mean. (An added bonus, it doesn’t trigger anyone’s ideological commitments).
Chalmers does talk about physics explanations being limited to structure and function. If you see the issue this way, that the language of physics is limited to describing structure and function, and consciousness is intrinsic, then it at least shows how intractable the problem is.
Thanks Prudence, as always, you've given me a lot to think about.
Hey Suzi,
I think we can put forward a more charitable version of the conceivability arguments, which doesn’t require that we beg the question against physicalism. On this version, it’s not that Q inherently requires a non-physical truth, it’s that it requires an assumption that Q is a non-structural truth. You did briefly mention structure at one point, but I don’t see it anywhere in your actual discussion of the conceivability arguments.
If you read Chalmers original article from 1995 (https://consc.net/papers/facing.pdf) on the hard problem (and subsequent work) as well other phenomenal “realists” like Goff, it’s obvious that they take the issue to be one of structure.
Reformulated this way, we would see that the conclusion follows from the assumption that if physical states are purely structural, and phenomenal states intrinsic (aka non-structural), then it’s not possible that the latter could be reduced to the former.
Why think this? The thesis that physical truths have to be purely structural follows from the work done on epistemic structural realism. Basically, if we accept the microphysical thesis (that all physical truths are equivalent to or reducible to microphysical truths) then all physical truths have to be captured in the mathematical vocabulary of physics (so it will be structural).
Could we deny this? We could by, for example, denying the microphysical thesis (we can see now the reason why Chalmers adopts this premise). Chalmers and co. do of course consider the implications entailed by rejecting the microphysical thesis elsewhere but reject it on the grounds that a) either such intrinsic physical truths explain consciousness- in which case it seems that we are led to some kind of weird physicalist alternative like Russellian monism- which people like Goff tend to endorse, or b) they do not explain consciousness, in which case they are irrelevant, and the hard problem still goes through.
What about the assertion that phenomenal truths have to be non-structural? This, I take it, is the very reason for formulating the argument in terms of a conceivability-possibility divided. The assumption that phenomenal truths are intrinsic is supposed to follow from the fact that we conceive of them that way. As Chalmers puts it, the qualities that we find in experience, like redness, don’t appear to be structurally decomposable-they are presented to us as being simple properties of objects like apples, as though they existed in a non-physical edenic environment. This of course is contrary to the findings of modern science, which tells us that redness is a complex property, structurally decomposable into EM waves and so forth.
Of course this just pushes the argument one step back, namely that we have to further assume that our phenomenal insight is reliable. That when we have an experience of a phenomenal quality as intrinsic, we are therefore justified in thinking that it actually is intrinsic. To support this premise, Goff and co. have argued for a thesis of ‘revelation’ (see here: https://assets.zyrosite.com/Aq2ZN1NGQMILGL8l/real_acquaintance_and_physicalism-dWxM5kOPGWCzLaBG.pdf) wherein we are justified in assuming that our introspective capacities are reliable.
In short, if you grant the above, then the conceivability arguments are certainly not circular! There’s a reason that proponents of phenomenal realism have spent a lot of time justifying the reliability of our phenomenal insight (in opposition to the illusionists), and a reason that they have framed the argument as being about a structural-intrinsic divide from the very beginning.
Hence, the focus of many physicalists (who have engaged in the literature) on denying the reliability of our introspective powers, like the illusionists do. The alternative approach, to deny the purely structural nature of physical truths, has largely fallen out of favor as a workable approach I gather.
To sum up, the conceivability arguments are only circular if you assume that there is no justification for the premises of the arguments, but obviously that’s not the case, otherwise nobody would take them seriously to begin with!
Cheers,
Alex
Super-excellent comment.
Hi Alex!
Thank you for taking the time to write this amazing comment. You've given me a lot to think about.
It seems that you're saying the debate hinges on whether we accept that subjective experiences are non-structural (or irreducible)?
I do find myself on the side of the fence that believes our experiences can and do break down into simpler components.
Thanks for challenging my thinking!
Yes I think that’s the main component of the debate (whether phenomenal qualities are structural). That said, I’ve spoken with Chalmers via email in the past on this exact issue and he insisted that some version of the conceivability argument should still go through even if phenomenal states were purely structural. That is, he thinks there remains an explanatory gap between physical states and phenomenal states- even if both are conceived of as being purely structural. So make of that as you will.
Others in the literature do think that the explanatory gap, and all the associated arguments (e.g. conceivability, knowledge) just reduces to a structural-phenomenal gap though. I’m thinking of Torin Alter in particular. What I will say is that the main aim of these arguments is to point at something which proponents think that phenomenal states have but physical states lack. What that something is, of course, has to be spelled out in some detail in order to justify the premises of these arguments. In the absence of such justification, it’s probably best to just view such arguments as intuition pumps and nothing more.
So far the focus of most phenomenal realists (including Chalmers) has been on the structural-phenomenal gap, but it’s also possible to try to point to some other attribute. I would say that in the absence of the structural-phenomenal gap, defending such arguments becomes much much harder.
For what it’s worth, I’m a physicalist who too believes that phenomenal states are complex states that can be fully structurally decomposable. But I arrived at this position after spending most of my life on the other fence.
So Chalmers fell in the presupposition trap of only seeing the evidence that supported his initial theory.
The water-without-oxygen and philosophical zombie examples aren't quite analogous. We can imagine oxygen-free water only by conveniently ignoring water's chemical composition - a fundamental physical property. But Chalmers' zombie argument is more complex. It asks us to consider a world physically identical to ours in every respect, yet devoid of consciousness.
By definition, this zombie world includes all physical properties and physical phenomena. So we can't simply disregard inconvenient facts as we did with the water example. Instead, we're forced to rely on our preconceptions about consciousness to determine whether such a scenario is truly conceivable. This seems problematic to me.
The zombie argument's strength hinges on our ability to imagine a world that's physically indistinguishable from ours, yet lacks subjective experience. But can we really do this without sneaking in our assumptions about the nature of consciousness? I'm skeptical.
Great point Matt!