I have noticed that panpsychism, which is the idea that some or all elements of matter possess some form of consciousness, subjective experience, mental awareness, or whatever you would like to call it, seems to be gaining favour these days. Noted neuroscientist Christoff Koch has recently suggested that consciousness may be a property of matter like mass or charge. I was just listening to a Philosophy Bites podcast where philosopher Galen Strawson (listen here) was forcefully arguing that panpsychism or micropsychism was in fact the most plausible prior if one is a physicalist or monist (i.e. someone who believes that everything is made of the same stuff). He argued that it was much more plausible for electrons to possess some tiny amount of consciousness then for it to emerge from the interactions of a large number of neurons.
What I want to point out is that panpsychism is a closeted form of dualism (i.e. mind is different from matter). I believe philosopher David Chalmers, who coined the term “The hard problem of consciousness“, would agree. Unlike consciousness, mass and charge can be measured and obey well-defined rules. If I were to make a computer simulation of the universe, I could incorporate mass and charge into the physical laws, be they Newton’s Laws and Maxwell’s equations, the Standard Model of particle physics, String theory, or whatever will replace that. However, I have no idea how to incorporate consciousness into any simulation. Deeming consciousness to be a property of matter is no different from Cartesian dualism. Both off-load the problem to a separate realm. You can be a monist or a panpsychist but you cannot be both.
20 thoughts on “Panpsychism”
Good points but, I think most of your concerns are actually unfounded. Your distaste of panpsychism is still based on a underdeveloped conception of how a scientific panpsychism works. It’s true that most panpsychist theories don’t explain how the theory “works” at all; philosophers simply point to its metaphysical elegance. However, the panpsychism that Koch endorses–Giulio Tononi’s integrated information theory–is far more nuanced and instructive.
At the heart of the IIT is an identity thesis: consciousness IS integrated information.Where there is integrated information, there is consciousness. In the limit, every instance of information is at least minimally conscious. (The theory actually predicts that there is an absolutely minimal conscious experience generating “1-bit” of information.) When information is integrated it generates more consciousness. When is information “integrated” you ask? According to the theory, information is integrated when the casual relationships within a system generate more information analyzed as a single whole (over a certain timescale) than with their individual nodes. Consciousness then is literally the effect of something being more than the sum of its parts. Further, the quality of a conscious experience is defined wholly by the informational relationships of the system. These can be mathematically “mapped” using a n-dimensional topology called “qualia-space.” So, every conscious experience is a n-dimensional polytope shape in qualia space. Wild stuff, but very elegant, empirically minded (this hypothesis explains many paradoxes with consciousness in brain science) and convincing.
Anyway, you can read all about the IIT online; I suggest Tononi’s “provisional manifesto” or his hour long video lecture on the topic. But quickly, in regards to your claim that panpsychism is just Cartesian dualism, it’s worth noting that the IIT tacitly reduces not just consciousness to information but all causation! This comports well with trends in theoretical physics to see “information” as the fundamental reality. Since both consciousness and the physical world are “information” there is nothing dualist about the theory; both mind and matter are reduced to information, hence monism. Also, the theory suggests a reason why consciousness would be selected for by evolution; a creature able to integrate information and then use that information to guide behavior would fare far better than one that distributes all information over separate modules, a much less efficient way to process data.
I think the theory is scientifically and philosophically interesting and, like Koch, I find it the ONLY plausible “theory” of consciousness in the marketplace of ideas right now. All others commit more grievous philosophical sins or have no applicability to empirical testing. The IIT is really the best of both worlds.
Thanks for your well argued comment. Actually, I wasn’t trying to disparage panpsychism, only to suggest that it is a form of dualism. In terms of Tononi’s ideas, I think they are quite plausible and I’ve blogged about them before. However, there are two ways to take the assertion that integrated information is consciousness. One is that it is an emergent property, which is explainable within the rules governing the rest of the system. The other, which is what Koch is proposing, is that it is an inherent axiomatic property of the system. You and others take this position to be monistic. My claim is that this is different from a monism in which I can understand computationally. Consciousness is not like mass because mass is a property that obeys certain describable rules within the network. Meaning, I can write down a complete self consistent set of rules that governs a universe that includes mass. I cannot do so for consciousness.
I claim that the position that consciousness is beyond computational understanding is dualistic. Now, I am not saying that everything needs to be understandable to be monistic. I am the first to accept that there are things that we cannot ever know and we can even prove that we can never know. However, there are still two possibilities for uncertain universes. One that can be described self-consistently and one that cannot. The unknowable in the former would be that there are many different self-consistent descriptions and we can never know which one is the correct one. I call this a monist uncertain world. Then there is one where we can not even construct a self-consistent computational description. I would call this a dualist world.
Great response. I got a little lost when your mentioned “uncertain universes.” Do you mean universes that have uncertainty in them (like ours does perhaps, ie. quantum mechanics) or universes that have an uncertain ontological existence?
Anyway, just a follow up thought or two.
If information is the correct conceptual description of the world, then even “mass” would be an “emergent” property within the abstract information network; the self-consistent set of rules would include mass because “mass” would be mathematically necessary given the constraints of the particular informational rules. (Even space and time would be similarly emergent in this scheme. But, let’s not go there!) Given that integrated information would “emerge” just as inevitably as mass does within this model, it seems that integrated information, aka consciousness, would be equally entailed by the rule set. Since integrated information is not beyond computational understanding, neither is consciousness. The world would not be dualist, only informational. The fundamental “substance” of existence (to use more 18th century terminology) would be consciousness, whose quality is defined by abstract computational processes in the world.
I think the pressing question that nags is: Why should integrated information be conscious? How is that identity better than “matter” being conscious? There are two possible answers I can imagine, and they and not mutually exclusive. The first is that possible worlds which “exist” but which contain no consciousness would be indistinguishable from abstract possible worlds that don’t exist. (We’re accepting a information ontology for the sake of argument.) This, I think, is a contradiction. Consciousness then is what separates the actual world from possible worlds. So, the claim is that any actual world must contain consciousness. It’s what makes it actual and not platonic.
The second is more intuitive: if the information in my brain really does generate more information than the sum of its parts, how could that information be represented in the world without consciousness? Since these macro-physical integrated information states actually have causal efficacy (Tononi is adamant about that) it’s difficult to imagine that a conscious-less creature could integrate information, for how would that information “present itself” to the world? Where would it be realized? Since the theory claims that integrated information is demonstrable in the physical world, it’s no great leap then to claim that consciousness is equally demonstrable for what else could that extra information BE?
Obviously, there’s no way to talk about consciousness without falling into the rabbit hole of wild metaphysical and epistemological speculation. Nonetheless, since dualism is strongly counter-intuitive, (maybe even contradictory), I think it important to attempt to conceptualize a world that is non-dualist but still accounts for the physical and phenomenological reality that we experience.
Yes, by uncertain I mean ontological uncetainty.
Saying the universe is information is not sufficient to escape the dualistic trap. I could create a universe where particles are characterized by a vector of properties such as mass and charge. Mass is defined as the proportionality constant between force and acceleration as well as that for gravitational force, etc. This Newtonian world is completely self-consistent. I have definitions for all the particle properties such as mass and charge. In this universe, mass is not an emergent property, although it could be in a quantum field theory or string theory world for example. However, now let’s say I further add the attribute of conscious experience. How do you codify that? You might be able to codify the rules of consciousness for how it might help an organism survive but you would not be able to codify the subjective experience aspect. This aspect is not solved by saying the universe is information.
Finally, i would argue that dualism is more intuitive than monism. All the problems of dualism are easily solved if we assume that the “physical effects” of consciousness are a purely computable emergent property, (i.e. the easy problem of consciousness), whereas the “subjective effects” live in a separate realm and have no direct interaction with the physical world. The real puzzle to me is how to make the subjective part live in the physical world.
Thanks for the follow-up. Just to conclude my thoughts on the matter:
First, if the world is information than any traditional physics terms – “particles” “force” acceleration” – can be reduced to abstract zeros and ones. And, it if that’s the simplest, most informative description Occam’s Razor would have us say it’s best to accept it as true. It’s a digital physics I’m envisioning.
My question is simple: If a world is nothing but zeroes and ones, what separates an actual world from a possible but not-actual world? Such an informational world has a platonic, abstract description and this description seemingly specifies everything there is to be said about the world EXCEPT if the world is ACTUAL or NOT. If there is consciousness in the world however, there is no question: the world is real and not merely possible. This is an abstract philosophical argument but, it sets a baseline of where my argument originates.
Next, in the world you describe and create in your example above I’d want to ask: Is there integrated information in it? I imagine that certain causal arrangements in your made-up world still objectively produce information that is over and above the sum of its parts. It’s our ability to codify the amount and quality of integrated information, of “phi” in Tononi’s language, that allow us to understand consciousness computationally.
If I’m understanding you correctly, I think you are trying to argue that subjectivity still needs to be “added” to a world, that no computational description can entail this aspect. On this sticking point I say, we need to put consciousness first as fundamental, necessary (for reasons we do not understand) and ubiquitous. You can’t “create” any universe without it nor try to explain it away. I think, hippy-dippy as it may sound, it’s what “being” IS. I refer back to my first point now; it’s what makes any actual world “actual.”
Ultimately, we use science and scientific concepts to explain the regularities in our phenomenal experience. I think the IIT goes a huge leap further in specifying why the world is as it appears to us. Like all great paradigm shifts it doesn’t negate past concepts but subsumes them under a richer modality of explanation. A world with subjectivity merely floating on top of of “real reality,” causally inefficacious and seemingly epiphenomenal, is a conceptual system that strikes me a fundamentally impoverished and ugly. I find it unlikely.
The identity of information with consciousness, should it turn out to be true, may very well turn out to be one of if not the most important discoveries in all of science.
To say the world is information is to say it is algorithmic. A binary representation is one way you can code it but there infinitely many other equivalent ways to represent it. Mass and acceleration are then incorporated into the algorithm. The world I created can certainly have integrated information. I could easily envision a universe comprised of interacting particles without any consciousness per se. I don’t see why such a universe could not exist. Information could entail consciousness but I don’t see why it would necessitate it. Mutual information is just a statistical relationship between two entities and Tononi’s phi is based on mutual information. You can make consciousness fundamental but then that is saying that you cannot explain it within the universe. It would be an axiom and thus is by definition not accessible to empirical science. It would be the most important discovery of philosophy but not science.
Well Sir, on that last point, I concede. Nonetheless, it may be a philosophical discovery with deep empirical applications (in AI for instance).
I suppose it comes down to the intuition that if space and time can emerge from informational processes than it doesn’t make sense to talk about “particles” as existing since any description of a particle is better explained by abstract informational content; a particle wouldn’t “exist” as a thing-in-itself. Ultimately, if a mathematical description can say everything that can be said about the world, it’s prudent to think that it’s the math that’s real and not the “objects” that emerge from it. But of course, the problem is that math is abstract and the world is apparently concrete.
But, we’ve fallen into a philosophical K-hole.
Thanks for the invigorating discussion. Always impressed with your intelligence and breadth of thought.
Yes, if you believe the algorithm is real then it doesn’t matter if you talk about particles or bits or anything else. These are all “dual” to each other. They’re just ways of describing the algorithm unfolding. The question is whether the algorithm can understand the algorithm from within. If our conceptualized description is particles and masses then the answer is yes, although we will never know where the algorithm that generated it comes from or what it is “running” on, etc. At this point in time, we don’t know how to codify subjective experience into even a conceivable algorithmic understanding much less and implementation. To me, that is the big puzzle.
Thanks for you stimulating comments. It is just for these exchanges that I started this blog.
[…] my post on panpsychism, a commenter, Matt Sigl, made a valiant defense of the ideas of Koch and Tononi about […]
In my view, there is a way of understanding panpsychism which unites dualism and monism into what I call a Multisense Realism. With any kind of panpsychism or dualism, there is still an explanatory gap between private subjective experiences and public objective conditions. I propose that only a monistic primitive that makes sense and makes sense of sense itself can give rise to coherence of any kind.
Beyond assumptions of ‘information’ or ‘substance’ the still has to be a ‘within’ in which ‘in’-tegration occurs. Looking at the sensory-motor qualities which underlie all functions, we can see that information integration does not generate consciousness, rather it assumes consciousness to begin with. Before any function can be produced, ontological conditions of input/output, of memory and sequence, of identity and local persistence, etc must pre-exist. It is in these ontological primitives which provide unity and multiplicity, cardinality and ordinality, sensing and doing, that panpsychism is realized – not as a physical vs non-physical dualism but as a private and a public physics.
A brilliantly defended Panpsychism @multisenserealism. It’s obvious to anyone who puts their finger to the wind that this is where science is going. The problem is that the implications of this worldview are very profound for the human race. Suddenly the space of possibility becomes the space of mind, and that is as large as information theory itself. There’s going to be a lot of resistance from the high priests of science before they concede this axiom. Of course, this is THE axiom, the one which grounds all systems of thought and being, and by doing so, makes them true.
Thank you @mattsigl. I must be on a roll today (I picked up some high praise from a reader on the Multisense Realism site today http://multisenserealism.com/2013/01/05/morphoria-and-the-deconstruction-of-arithmetic/) and my last blog entry on why we can’t imagine new colors is picking up some popularity as well (http://s33light.org/post/40296309313). Yay me. ;)
It may be self consistent but it’s still dualism.
It’s an involuted monism, like a Mobius strip. It can be understood through many different nested dichotomies, trichotomies, quadruplicities, etc, but it also all boils down to a single primordial element, which I call Sense, or experience.
However, this sense experience will not be explainable in a finite number of bits.
BTW. There is nothing inherently wrong with dualism.
Bits are a conceptual representation within sense. Sense is a concrete physical reality.
“BTW. There is nothing inherently wrong with dualism.”
A lot of people associate dualism with pre-scientific worldviews. Personally, the problem that I see with dualism is
1) that it doesn’t go far enough.
To say that mind and matter are separate fails to recognize the significance of the fact that they seem to be opposite in every way.
2) that it leaves the most important part unexplained.
Whether it’s substance dualism, property dualism, or dual aspect theory, you still have the infinite regress of a homuncular interface. Something has to be in between the two sides of the sandwich passing instructions or sensations back and forth, and that something must itself speak both the language of mind and embody the mind of matter, so we really haven’t gotten anywhere. We still are stuck with a missing site within the brain where the memory of the smell of pine trees is transduced from a pattern of X (where X = bytes of doxastic logic, calcium waves, microtubular quantum entanglement, neurochemical activity, elan vital, awareized energy, soul, etc) into the actual sensory-motor perceptual participation which we know to be concretely real.
[…] amount of integrated information, I am skeptical that it is the only relevant measure. See here and here for some of my previous thoughts on the topic. One of the reasons that Tononi has proposed a single […]
You totally don’t get it.
Let’s run the zombie experiment with producing a simulation of the world. When you arrange the material, you create consciousness. Consciousness is not asserted to exist in the same sense as matter. A university exists. Show me the university. The library is not the university, and neither are the staff. An ontological distinction for a system is worthwhile, but it does not give it the same ontology as a brick. The “show me the university” experiment (is that Hume?) does not assert dualism.
This is what I love about panspsychism modified by Tononi. So far, it seems to solve every classic dilemma and thought experiment, especially those that pertain to monism. It invalidates dualism (not that that was needed, but it passes that test). I am extremely uncomfortable with this idea. In fact, I hate it and I fear Deepak Chopra is going to claim he can cure cancer with it, but it keeps passing every test. It even provides universal morality, because the Universe can be described as a system. I really want to find a critical error… but the zombie experiment does not establish it as dualist/invalid. Rather, it solves the zombie question.
I even ran it through Indian logic and Eastern monist systems, which have always struggled with a priori assumptions… and it solves though too. In fact, I have to scrap my pet theory of Logical Emanationism. This is a big loss for me personally, but… it solves so many problems. I think this might be the biggest deal to ever happen in philosophy
But seriously, we have to head Deepak Chopra off at the pass before he tries to sell books with it.
@Victor It is dualistic in the sense that it adds a new property to the universe that is algorithmically distinct from the other properties. Your university analogy is a perfect example of an emergent property. However, from what I understand, panpsychism claims that consciousness is not just an emergent property of the underlying rules that govern matter and energy as we know it.
[…] of consciousness although it may not be intelligible to humans. He also seems to subscribe to panpsychism, where consciousness is a property of matter like mass, an idea championed by Christof Koch and […]