[ExI] My theory of consciousness

[Mike Dougherty]

fwiw Jason, this was an excellent reply.


On Thu, Sep 24, 2026, 2:26 PM Jason Resch via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

>
>
> On Wed, Sep 23, 2026, 7:38 AM John Clark <johnkclark at gmail.com> wrote:
>
>> On Tue, Sep 22, 2026 at 10:49 AM Jason Resch via extropy-chat <
>> extropy-chat at lists.extropy.org> wrote:
>>
>> * >> I think I am the way that atoms behave when they are arranged in a
>>>> Johnkclarkian way. In what ways could I be wrong? *
>>>>
>>>
>>> *> Because there is a certain sense in which atoms are irrelevant to
>>> your behavior. A Java program can run on a PC, Mac, or Linux comouter.*
>>>
>>
>> *Yes but that's because a PC and a Mac and a Linux computer are all made
>> of atoms and are logically equivalent to a Turing Machine, and a Turing
>> Machine's tape must also be made of atoms as does its read/write head.   *
>>
>
> Turing's specification was abstract and mathematical. Atoms aren't
> strictly necessary. No features of turing's machine depends on any specific
> properties of atoms, as I explain below.
>
>
>>
>>> *> The exact arrangement of its atoms are this not important to how that
>>> program behaves.*
>>>
>>
>> *If the arrangement of atoms is just a random jumble and its behavior
>> cannot be logically reduced to that of a Turing machine then the program
>> cannot run at all. To make a Turing Machine you could use any sort of
>> atoms, but you can't use just any arrangement of them. It's not easy to
>> make a fast efficient Turing Machine, that's why you can't use the hand
>> tools in your garage to make something that can compute as well as what
>> comes out of TSMC's factories in Taiwan. *
>>
>
>
> It sounds like you agree with my point that you can have two arrangements
> of atoms, computer X and computer Y, and both could run the same program,
> despite being in different material arrangements, but it's hard to tell
> from the above.
>
>
>> *>  if this Java program were an emulation of your brain which preserved
>>> your consciousness, then your "John Clarkian way" inherits no properties
>>> from your atoms.*
>>>
>>
>> *The atoms need to have one very important property, the property of
>> existence, however atoms are generic but the information necessary to
>> emulate me is not, in fact currently there is only one chunk of matter in
>> the entire observable universe that is arranged in a Johnkclarkian in way,
>> but once Drexler style nanotechnology becomes available that situation
>> could change. *
>>
>
>>
>>> *> Your current instance uses them, and to implement you some substrate
>>> is required,*
>>>
>>
>> *I agree but fundamentally it doesn't matter which substrate you use.
>> That's why 2+2 =4 regardless of if the calculation is made by a mechanical
>> calculator, a vacuum tube computer, a modern computer, or a human brain.  *
>>
>
> Right, this is what I have been saying.
>
>
>>
>>> *> but they need not even be atoms -- you could be running in a sim from
>>> a universe with very different laws that doesn't even have atoms*
>>>
>>
>> *Nobody knows for certain but maybe, however **even if it's true
>> somewhere down the line you're still going to need something physical to
>> make a calculation.*
>>
>
> Would a "game of life universe" count as physical if it existed
> primitively? What about rule 110? Or a 2d space of idealized billiard ball
> collisions? How simple can a physical universe be? How do you define
> "physical" or "physical existence"? You say something "physical" is needed
> at the bottom, but if this universe can exist primitively why can't any
> other structure also exist primitively?
>
> It seems certain to me that the laws of our universe cannot be the only
> ones realized in reality, given the gigantic apparent coincidences in the
> tuning of our universe's laws and constants and force strengths to sustain
> complexity and life. This suggests some sort of ensemble of varied
> universes with varying laws must exist. What meta law or meta physics
> determines whether a universe exists or not?
>
>
> * Mathematics is the language of physics but no language by itself, not
>> even the language of mathematics, can produce a calculation any more than
>> the English language word "cow" can produce milk. That's why I think
>> physics is more fundamental than mathematics.*
>>
>>
>>> *> (like the Game of Life universe).*
>>>
>>
>> *It's impossible to play Conway's Game of Life without using something
>> physical, like a computer or a human brain, that is capable of making a
>> calculation.  *
>>
>
> Here I am talking about a primitively existing universe whose laws are
> identical to what we call Conway's game of life. Not a simulation of such a
> universe in our own.
>
> Such a universe is capable of hosting a Turing machine, and nothing like
> an atom would exists in this picture. So generally speaking, computers
> don't require atoms, that's a limitation of our technology and geography
> (we happen to exist in a place where atoms are an available and practical
> implements to realize computations relative to ourselves), but in a GoL
> universe they would be speaking of needing "cells",  "glider guns", and
> other structures to build computers in their world.
>
>
>> *> I use this notion of a "substitution level", inspired by Turing
>>> universality (the ability of a general purpose computer to emulate any
>>> other finite state machine) to derive a lower bound which must be preserved
>>> for the high level information state of a conscious process to be
>>> unaffected from its own point of view.*
>>>
>>
>> *I don't quite understand what you mean by that. *
>>
>
> What I am asking is: how much detail must a scan of your brain preserve to
> implement your same consciousness?
>
> Some say we just need high level software/algorithms implemented by brain
> regions, the exact circuits in the cortical columns don't matter.
>
> Others might say we need to replicate the brain down to the level of the
> spiking neural network.
>
> Others say we need to model the cell biology of each neuron and perhaps
> the glial cells as well.
>
> Still others believe we need a full proteum and molecular simulation.
>
> But my paper borrows a key insight from computer science: from the level
> of perspective of a program running on a computer, there is an abstract
> level below that program to which changes of implementation are invisible.
> A program may be able to read variables in its scope, but it can't tell how
> those values are physically implemented, it can't read outside it's address
> space, it can't tell if it's running on an emulator or not. It is trapped
> within a particular causal structure beyond which certain information has
> no way to get to or otherwise influence it. This is an epistemic firewall,
> and it nearly allows us to define what facts can or can't be relevant to
> the process at the level of the process.
>
> By considering the conscious state as a computational process, we can make
> a similar determination of what facts are able to enter and influence that
> conscious state, and this sets a minimum level of detail that a brain scan
> must capture and that a brain emulation must emulate to preserve the same
> character of consciousness.
>
> Failing this might still yield a conscious entity that is similar to John
> Clark, but your experiences, say in how you process sound or vision or how
> emotions feel to you, could be off in subtle ways.
>
>
> * But speaking of upper bounds, we know the upper bound of how much
>> information would be required to construct a human brain at the time of
>> birth, and it's not very large. DNA also places an upper bound on how
>> complex a seed AI would have to be. In the entire human genome there are
>> only 3 billion base pairs. There are 4 bases, so each base can represent 2
>> bits, there are 8 bits per byte, so that comes out to 750 meg. Just 750
>> megabytes, that's about the same amount of information as an old CD disk
>> could hold when they first came out 40 years ago. And because of redundancy
>> a lossless data compression algorithm could shrink that down to 500
>> megabytes or less. And that's for an entire human baby not just the brain. *
>>
>
> Genomes are small, but connectomes are about 10 million times larger. And
> proteomes larger still. If we're talking about what needs to be digitally
> backed up to restore one's mind, we need to understand at what scale
> details matter.
>
>
>
>> *And if you're talking about describing one person's genome relative to a
>> human reference genome, that is to say the amount of information you'd need
>> to change one person's genome into another persons genome, you need less
>> than 10 megabytes for that. *
>>
>>
>>> *> Thus, for the first time, we have a way of answering how detailed a
>>> brain scan and mind upload program must be.*
>>>
>>
>> *And what number did you come up with?  When I take a sip of coffee I am
>> changed, is that change large enough that I become a different person? *
>>
>
> It's enough to have changed your conscious state. But the larger
> point.from your example is this:
> Quantum fields in your brain changed
> Particles in your brain changed
> Molecules in your brain changed
> Cells in your brain changed
> Neural activity in your brain changed
> Neural circuits in your brain changed
> Specialized regions in your brain changed
>
> How detailed of a simulation of your brain is required if we want to
> faithfully reproduce the state your brain enters when it sips a cup of
> coffee?
>
> That is the question I refer to when I speak of identifying the
> "substitution level." It's the level below which changes to your brain can
> be made without altering your conscious experience.
>
>
>
>> *> we need not introduce souls for atoms to be irrelevant to defining a
>>> John Clarkian way. Since you are a fan of Tegmark perhaps you will
>>> appreciate what he says on this question:*
>>>
>>
>>
>> *"Computation, intelligence and consciousness are patterns in the
>>> spacetime arrangement of particles that take on a life of their own, and
>>> it’s not the particles but the patterns that really matter! Matter doesn’t
>>> matter."*
>>>
>>
>> *I am a fan but I don't agree with Tegmark about everything, however I
>> agree with him about that, it’s not the particles but the patterns that
>> really matter, that's because matter is generic but patterns are not. **Tegmark
>> is a well known Many Worlds advocate and I thinks it's the least bad
>> quantum interpretation anybody has so far come up with. He** thinks
>> everything that does not violate the laws of physics actually happens, and
>> he's probably right about that too. But Tegmark goes further, he thinks
>> everything that does not violate the laws of logic or mathematics happen,
>> and I'm much more skeptical about that. In one of his books he listed 4
>> different types of multiverses: *
>>
>> *Level I — Our universe, beyond our cosmic horizon*
>>
>> *Level II — Different bubble universes*
>>
>> *Level III — Many-Worlds quantum mechanics*
>>
>> *Level IV — The mathematical universe*
>>
>> *I have no problem with the first three but I think Tegmark goes one
>> multiverse too far because Level IV is "every mathematically structure
>> corresponds to a physically existing universe", and I think that's probably
>> not true.*
>>
>
> Do you think there are at least a large number of universes with different
> laws? If not how do you explain the fine tuning coincidences?
>
> And if you do accept a large number of different universes with many laws,
> why would reality stop at say 173727473926371884757920498373902847
> universes and not go to infinity?
>
> * Mathematics is the language of physics but a language can be used to
>> write both fiction and nonfiction, and I have a hunch that some of the more
>> abstruse areas of modern mathematics are the equivalent to mathematical
>> Harry Potter novels.*
>>
>
> We can explain the trend towards lawful universes under various selection
> rules, for example using Kolmogorov complexity, we expect the description
> of most universes to be describable by short programs, and this tends to
> produce universes with stable laws, rather than universes with surprising
> magical deviations from laws as in Harry Potter.
>
>
>
>>
>> *> Consider a high-level program that receives a number N and stops when
>>> it finds the next twin prime greater than N.*
>>> *Now when was ask the question: "what makes this program stop?"*
>>>
>>
>> *If it does stop then it will be because of information, but that is a
>> big "if". Assuming N is the largest known twin prime we don't know if that
>> program will ever stop, it might but it could be that the number of twin
>> primes is finite and N is the largest one. However Gödel told us that a
>> proof of that might not exist, and Turing told us that in general there is
>> no way to determine if a question is answerable or not; so we could be
>> grinding through huge numbers forever looking for the next twin prime but
>> always failing because a larger twin prime does not exist. *
>>
>
> Then you see the power of rules that emerge above and beyond the mere low
> level physical laws. The emergent properties of complex systems can have
> their own rules and causal effects in our world. It is in this sense that
> minds can cause things even if you ignore their high level states and focus
> your attention on the atoms only, you can't through your attention, erase
> the fact of these high level states existing and standing in causal
> relations to other events in the universe.
>
>
>> *> Program behavior is in a sense divorced from the behavior of atoms*
>>>
>>
>> *It is divorce from the behavior of any particular group of atoms, but
>> not of atoms in general. *
>>
>
> Yes, unless we consider another primitive universe that lacks atoms but
> nevertheless has the capacity build a Turing machine.
>
>
>
>> *>> And if I thought that was a realistic possibility I'd burn all my
>>>> physics books, put on some saffron robes, and become a monk. And my current
>>>> plans do not include becoming a monk. So, what do you think? *
>>>>
>>>
>>> *>I think we don't need to assume dualism to arrive at a picture where
>>> things can resemble dualism.*
>>>
>>
>> *Actually I sort of believe in dualism because I think there is a
>> difference between a noun and an adjective, and I don't think I am a noun I
>> think I am an adjective.  *
>>
>> > *I think we can do still better. Which is to substitute "information"
>>> with "processes".*
>>>
>>
>> *A process is a specific series of actions, and you can't have that
>> without information. *
>>
>
> It is the same as your noun/adjective distinction. Information answers
> "what" a process answers "how".
>
> Jason
>
>
>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>> *but there is such a thing as an appearance of intelligence. We don't
>>>>>> believe our fellow human beings are conscious when they are sleeping or
>>>>>> under anesthesia or dead because when they are in those states they do not
>>>>>> have an appearance of intelligence.  *
>>>>>>
>>>>>>
>>>>>>> *> isn't a necessary part of the process, and that is what leads to
>>>>>>> the idea of a hard problem.*
>>>>>>>
>>>>>>
>>>>>> *The hardest part of the hard problem of consciousness is figuring
>>>>>> out exactly what the question is, what sort of answer would satisfy you
>>>>>> that the problem has been solved? Simply discovering that X causes
>>>>>> consciousness would not be sufficient because that would immediately
>>>>>> generate new questions like how does X manage to bridge the subjective
>>>>>> objective divide, and what causes X? It is a fact that an iterated sequence
>>>>>> of "how" or "why" questions either goes on forever or terminates in a brute
>>>>>> fact. And I think it's a brute fact that consciousness is the way data
>>>>>> feels when it is being processed intelligently. I know I've said all this
>>>>>> before but if you show me why I'm wrong I'll stop repeating it.  *
>>>>>>
>>>>>> *> The idea that you could remove consciousness from a functionally
>>>>>>> equivalent system and change nothing of its functioning, is what leads to
>>>>>>> the inherently self-contradictory idea of philosophical zombies.*
>>>>>>>
>>>>>>
>>>>>> *Philosophical zombies are not self-contradictory but if laws of
>>>>>> physics permitted the existence of such a thing in our universe then random
>>>>>> mutation and natural selection could never have been able to produce a
>>>>>> conscious entity, and yet I know for a fact that it did. Therefore I must
>>>>>> conclude that although philosophical zombies do not violate any law of
>>>>>> mathematics they must violate the laws of physics.*
>>>>>>
>>>>>>
>>>>>>> * >>Computational functionalism is certainly not claiming that
>>>>>>>> consciousness doesn't exist, it's just saying it's not needed to explain
>>>>>>>> behavior. *
>>>>>>>
>>>>>>>
>>>>>>> *> I somewhat disagree here. If a robot responsively and reliably
>>>>>>> reacts to catch a hockey puck as seen
>>>>>>> here: https://www.youtube.com/watch?v=bOIxu2lv8sI
>>>>>>> <https://www.youtube.com/watch?v=bOIxu2lv8sI>*
>>>>>>>
>>>>>>
>>>>>> *I'm surprised you included that link because it supports my views
>>>>>> not yours. The engineers who made that robot did not need to consider the
>>>>>> nature of consciousness for one instant, and yet it was still able to play
>>>>>> a good game. If that air hockey table/robot was able to reproduce itself
>>>>>> and let loose on an empty planet and you went away and came back 4 billion
>>>>>> years later you might find it had evolved into a race of beings that was as
>>>>>> smart or smarter than human beings. *
>>>>>>
>>>>>> *> Most people don't assume any process that receives information and
>>>>>>> changes its state is conscious,*
>>>>>>>
>>>>>>
>>>>>> *Most people DO assume such a process is conscious, but only if its
>>>>>> brain is wet and squishy, not if its brain is dry and hard. Call me crazy
>>>>>> but I don't think wetness or squishiness is relevant.   *
>>>>>>
>>>>>>
>>>>>>> *> Thus they will come to reach states that are private, and
>>>>>>> ineffable.*
>>>>>>>
>>>>>>
>>>>>> *If so then pondering such states is about as useful as navel gazing.
>>>>>> The current revolution in AI happened when people stopped wasting their
>>>>>> time thinking about the nature of consciousness and started thinking about
>>>>>> the nature of intelligence. *
>>>>>>
>>>>>>
>>>>>>> *> The recurrence-necessity test*
>>>>>>>
>>>>>> [...] *The revealing test is substitution within a mind capable of
>>>>>>> comparison and report.*
>>>>>>>
>>>>>>
>>>>>> *The report is the only part of this experiment that a third-party
>>>>>> performing the experiment has access to, and Computational Functionalism
>>>>>> can reproduce that report, or any report, without making use of
>>>>>> "consciousness" or any other state that is not available to a third party.
>>>>>> And a report (a.k.a. behavior) is the only thing that Natural Selection can
>>>>>> see in the 4 billion year long experiment called "life", but natural
>>>>>> selection can't directly detect  consciousness any better than we can see
>>>>>> it in anything other than ourselves and it can't select for something it
>>>>>> can't see. Nevertheless evolution managed to produce consciousness at least
>>>>>> once, in me (and probably you too). The only way that could've happened is
>>>>>> if consciousness is a byproduct of intelligence because Natural Selection
>>>>>> can see intelligence, or at least intelligent behavior.       *
>>>>>>
>>>>>>
>>>>>>> *> that change only if some later transition also changed. Yet the
>>>>>>> unfolding preserves those transitions. A theory that makes physical
>>>>>>> recurrence constitutive must therefore posit a phenomenal change to which
>>>>>>> the subject has no epistemic access. CKT predicts no change*
>>>>>>>
>>>>>>
>>>>>> *CKT insists that in any experiment you could set up it can reproduce
>>>>>> the objective results of that experiment without resorting to any
>>>>>> hypothetical subjective state, it doesn't say there is such a state and it
>>>>>> doesn't say there isn't, it is agnostic on the issue and it only says that
>>>>>> however important it may be to us, consciousness is irrelevant as far as
>>>>>> predicting any experiment you could possibly perform. *
>>>>>>
>>>>>>
>>>>>>> *> because the center possesses and uses exactly the same
>>>>>>> distinctions before and after substitution. Repeatedly switching between
>>>>>>> the two implementations turns the case into a recurrence-specific version
>>>>>>> of dancing qualia.*
>>>>>>>
>>>>>>
>>>>>> *OK. So what? If I am not a noun but an adjective, if I am the way
>>>>>> matter behaves when it is arranged in a Johnkclarkian way then dancing
>>>>>> qualia would be exactly what I would expect to happen.   *
>>>>>>
>>>>>>
>>>>>>> *> This does not establish that LLMs are conscious.*
>>>>>>>
>>>>>>
>>>>>> *True, but I have no way of definitively establishing that you are
>>>>>> conscious and I have no way of definitively establishing that rocks are
>>>>>> not. However I could not function if I really believed I was the only
>>>>>> conscious entity in the universe therefore I must accept as an axiom of
>>>>>> existence that solipsism is untrue, and a simple corollary from that axiom
>>>>>> is consciousness must be a side effect of intelligence.   *
>>>>>>
>>>>>> *> A more targeted experiment would temporarily block communication
>>>>>>> through the corpus callosum while leaving both hemispheres active*
>>>>>>
>>>>>>
>>>>>> *Such experiments prove that a split brain produces a split mind, and
>>>>>> that's not surprising because mind is what a brain does. In looking over
>>>>>> your other proposed experiments it seems to me they all suffer from
>>>>>> problems very similar to the ones described above. They all involve the
>>>>>> experimental subject in some way "reporting" on the nature of his
>>>>>> subjective state, but the trouble is such a report, regardless of what it's
>>>>>> about, is an objective behavior that CKT can reproduce without
>>>>>> consciousness entering the picture. *
>>>>>>
>>>>>> *If you feel I've overlooked something important feel free to bring
>>>>>> it to my attention. *
>>>>>>
>>>>>> *John K Clark *
>>>>>>
>>>>>> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260924/70ddd5a2/attachment.htm>