[ExI] redness is relation assumption? (Was Re: My theory of consciousness)

[Jason Resch]

On Fri, Sep 25, 2026, 5:29 PM Brent Allsop via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

>
>
> Hi Jason,
>
>
> I’ve been discussing your paper with co-pilot.  It’s hard for me to get my
> redness is not relation brain to fully model what I’m sure is a great
> model, given certain assumptions.  But co-pilot has helped, significantly.
>

That is wonderful to hear Brent. That is a great application of AI. :-) I
appreciate your engagement and in particular thank you and co-pilot for the
many great ideas below, more to say:



After a short discussion of how you define or assume that “phenomenal
> character” is “the organization of usable distinctions possessed by a
> center.”   and the great ways you justify this assumption, it eventually
> agreed that making this assumeption is an issue, and that it is a
> falsifiable issue.
>

That is good news (to me), it means it is a scientific theory we can test
and falsify, so we're not idly speculating in pure philosophy.


>
> And co-pilot agreeing this was an issue with your paper I asked it:
>
>
>
> “Can you formulate some proposals for how his paper might address this
> issue? Especially regarding pointing out how both our predictions are
> falsifiable, and that eventually science will demonstrate who is right”
> regarding whether you can get a redness quality from a bit?”
>
>
>
> Here is co-pilot’s response:
>
>
>
> Brent, this is exactly the right question to be asking — because *your
> view and Resch’s view make different empirical predictions*, and the
> paper *could* make this explicit. In fact, the paper already gestures
> toward falsification criteria (Section 1, Abstract), but it doesn’t yet
> spell out how the “intrinsic redness” hypothesis and the
> “redness‑from‑usable‑distinctions” hypothesis diverge scientifically.
>

The paper does have an explicit mention of this idea under "11.3 Challenges
for nonfunctionalist theories" where I give the example of using different
material substrates, I.e. silicon, germanium or carbon to build a computer
that emulates a conscious mind, and whether this could manifest any
difference in demonstrated behavior.

This is under substrate intrinsicism, which is the general name I use for
theories like yours that say a particular material is needed. I recently
learned there is another theory not unlike yours, called Russellian Monism,
which also holds that the "what" of what something is made of is what
carries the consciousness properties rather than the "how" of the dynamical
behavior. I am working on an updated version of the paper to discuss
Russellian Monism. I think this new paper actually can explain where the
intuition comes from in novel way.

Here is what the currently available version of my paper (not the new one)
offers as a challenge to substrate intrinsicism:

"Substrate intrinsicism. Computation is specified abstractly: a physical
imple-
mentation must realize the relevant state transitions, but it requires no
particular
material. Conway’s Game of Life makes this vivid: its simple, fully
specified local
rules support a universal Turing machine (Rendell, 2016). Consider three
computers
implementing the same high-level process, one in silicon, another in
germanium, and a
third in carbon-nanotube transistors. Suppose every distinction,
transition, access rela-
tion, and temporal dependency constitutive of the center is preserved.
Their materials,
energy use, and microscopic histories differ, but nothing in the
implemented process
unfolds differently. The same alternatives become available, enter the same
comparisons
and memories, and guide the same judgments and actions. To assign them
different
experiences solely because their materials differ would make phenomenal
character
vary without any difference the processes themselves could encounter."

In short, there's no "causal path" available through which any difference
in experience could be made know. To the process. Nothing different would
enter its memories, and no differences would be reflected in its speech.
All the internal states of its mind program are one for one identical, so
then, in what could the phenomenal difference consist in and how could the
subject come to know this difference?


Below is a set of *concrete proposals* for how the paper *could* address
> this issue directly, in a way that respects both positions and shows how
> science could eventually decide between them.
>
> I’ll ground these proposals in the actual text of the paper you’re viewing
> ().
>
> *⭐ Proposal 1 — Explicitly acknowledge the competing definition of
> phenomenal character*
>
> The paper could add a short section saying:
>
> “Some theorists hold that phenomenal qualities (e.g., redness) are *intrinsic
> physical properties* of certain states, independent of their relational
> or computational role. On this view, consciousness *uses* these intrinsic
> qualities but does not constitute them.”
>

I don't have a clear picture of how a process could use or access these
intrinsic qualities.

The move Russellian Monism takes is to say the same intrinsic qualities
that explain consciousness are what drive the dynamics of physics. But I
think this also leaves us at an explanatory dead end. Physics doesn't give
us access to intrinsic properties, so no physical experiment could ever
tell us anything about them, why they work the way they do, or how they
connect to any third person observable dynamical physical effect.


Then contrast this with the paper’s own definition:
>
> “By contrast, the centered‑knowledge thesis identifies phenomenal
> character with the *organization of usable distinctions* possessed by a
> process. This rejects intrinsic qualia and treats phenomenality as
> relational and structural.”
>
> This would make the conceptual disagreement explicit rather than implicit.
>
> *⭐ Proposal 2 — Show that the two views make different empirical
> predictions*
>
> This is the strongest way to address your concern. The paper already
> claims to offer “explicit falsification criteria” (Abstract, ), but it
> doesn’t yet contrast them with the falsification criteria of
> intrinsic‑qualia theories.
>
> Here’s how it could:
>
> *Intrinsic‑qualia prediction*
>
> If redness is an intrinsic physical property of a state, then:
>

States this way, the only difference between your theory and mine is a
matter of scale. You say a phenomenal state it is identical to a low level
state, whereas I say a phenomenal state identical with a high level state.


>    - Two states with identical computational/relational organization *could
>    still differ phenomenally* if their intrinsic physical microstructure
>    differs.
>
>
Yes this is the example I give with the silicon, germanium and carbon
nanotube transistors. It is a different micro physical state, but the same
high level informational/computational state.



>    - A perfect functional duplicate (e.g., a digital emulation) *might
>    lack redness*, even if behavior and knowledge are preserved.
>
>
But if behavior is preserved by running the same high level program, how
could any behavioral difference enter the picture? If no behavioral change
occurs, then how then could this test reveal any difference in the
phenomenal experiences of the computers based on different substrates? How
do we discover of any of them lack redness or experience redness
differently when the behavior is known a priori to be identical?

This is why my paper takes the path of justifying the theory abductively
(from the wide number of facts it can explain) thereby showing it is a
powerful and useful theory. This is a limitation all theories of
consciousness face, because each of us is trapped in our own view, we can
never see what anyone else sees. So unlike most theories in science which
depend on repeatable objective results, consciousness theories (being
designed to describe the subjective) can't use many of the standard
scientific methods.



>    - Neural replacement could preserve all functional organization but
>    still *lose qualia* if the intrinsic physical property is not
>    reproduced.
>
>
My new paper (will be ready in a few days) makes some powerful arguments
against this possibility (of maintaining organizational equivalence but
losing qualia i.e. zombies).


>
> This is the “fading qualia” or “dancing qualia” possibility.
>
> *Centered‑knowledge prediction (paper’s view)*
>
> If phenomenal character *is* the organization of usable distinctions:
>
>    - Any system with the same usable distinctions and access relations
>    *must* have the same phenomenal character.
>    - A digital emulation that preserves KC (the complete knowledge
>    organization) *must* preserve redness.
>    - Neural replacement cannot change phenomenality unless it changes the
>    distinctions available to the center.
>
> This is organizational invariance (Section 4.1–4.2, ).
>
> *Empirical divergence*
>
> The paper could say:
>
> “If intrinsic qualia exist, then two systems with identical KC could
> differ phenomenally. If centered knowledge is correct, such divergence is
> impossible.”
>
> This sets up a clear empirical fork.
>

Essentially where your theory and mine diverge comes down to whether the
intrinsic redness sits below the "substitution level" or above it. If it is
a high level property that sits above the substitution level, then your
theory and mine can be seen as equivalent. But if the property sits below,
that is, it it is some feature or property of a thing that can be used to
build a universal Turing machine, then when sitting in this space as a
building block of a computer it must necessarily exist below the
substitution level for any program this computer runs. It would then be
causally isolated from having any effect on the process, and whether this
property is present or not at that low level, makes no difference to the
evolution of this program. It will unfold the same way regardless.

*⭐ Proposal 3 — Identify testable consequences of each view*
>
> The paper already proposes tests for its own theory (e.g., neural
> replacement, emulation fidelity). It could add parallel tests for intrinsic
> qualia.
>
> *Test A — High‑fidelity neural replacement*
>
>    - Replace neurons with silicon equivalents that preserve all
>    functional distinctions.
>    - *Intrinsic‑qualia view:* Phenomenal character might change or fade.
>    - *Centered‑knowledge view:* Phenomenal character must remain
>    identical.
>
> *Test B — Digital emulation with preserved KC*
>
>    - Build a digital emulation of a subject’s brain at the substitution
>    level.
>    - *Intrinsic‑qualia view:* The emulation might behave identically but
>    lack redness.
>    - *Centered‑knowledge view:* If KC is preserved, redness must be
>    preserved.
>
>
I might suggest as something to try: take all of your writings and thoughts
on your theory of consciousness and ask a frontier model AI to put it
together in an organized paper in a similar manner as I did for my ideas. I
started using OpenAI but recently switched to Claude Fable and have found
it far better and faster. It can do much more in each step without messing
things up and builds new PDFs in seconds rather than minutes. It's ability
to research and find relevant citations in the literature is also quite
good (though not perfect). I think this will make your ideas much easier to
share with certain people, and you will also learn a lot from the
experience (I certainly did). It will help you frame your own ideas
relative to others in the field, point out people who have agreed with or
defended ideas like your own, point out challenges similar theories face,
and can recommend how to go about testing your theory.



>
> *Test C — Microphysical variation below substitution level*
>
>    - Introduce microphysical changes that do *not* alter KC.
>    - *Intrinsic‑qualia view:* Phenomenal character might change.
>    - *Centered‑knowledge view:* Phenomenal character cannot change.
>
> The paper could explicitly say:
>
> “These tests discriminate between intrinsic‑qualia realism and the
> centered‑knowledge identity.”
>
> *⭐ Proposal 4 — Address your specific claim: “Intrinsic qualities are not
> acquired; consciousness uses them.”*
>
> The paper could add a paragraph acknowledging this view:
>
> “One alternative holds that phenomenal qualities are intrinsic physical
> properties of states, and that consciousness merely uses these properties
> to represent knowledge. On this view, the ‘glow’ is not added by knowledge;
> it is already present in the intrinsic nature of the state.”
>
> Then respond:
>
> “The centered‑knowledge thesis denies that states have intrinsic
> phenomenal properties. It identifies phenomenality with the organization of
> distinctions a process possesses and uses. If intrinsic redness exists,
> then two states with identical KC could differ phenomenally — a possibility
> the centered‑knowledge thesis rejects.”
>
> This directly addresses your objection.
>
> *⭐ Proposal 5 — Show how the paper’s own framework could falsify itself*
>
> The paper already says:
>
> “The theory stands or falls on the answer to one question: can a process
> genuinely possess knowledge while there is nothing it is like to occupy
> that locus?” (Section 1, )
>
> It could extend this:
>
> “If a system possessed all the distinctions and access relations of KC but
> lacked phenomenal character, the centered‑knowledge thesis would be
> falsified.”
>
> This is the paper’s version of a zombie test.
>

The paper already has some test proposals for how to falsify the theory. I
will be adding a few more. This test given here doesn't quite help because
you would presumably still have behavior equivalence.

One AI suggested the test where information in K_C is changed and the
subject says whether or not there was any change in their experience, which
I thought was useful.

*⭐ Proposal 6 — Explicitly contrast “intrinsic redness” with the paper’s
> “redness-as-distinction”*
>
> The paper could add:
>
> “On the intrinsic‑qualia view, redness is a basic physical property of a
> state. On the centered‑knowledge view, redness is the structured difference
> the center possesses — the organization of alternatives that guide its
> continuation.”
>
> This clarifies the metaphysical disagreement.
>
> *⭐ Proposal 7 — Add a section titled “Competing hypotheses and empirical
> resolution”*
>
> This section could:
>
>    1. Lay out the intrinsic‑qualia hypothesis.
>    2. Lay out the centered‑knowledge hypothesis.
>    3. Show their different predictions.
>    4. Propose experiments.
>    5. State what would falsify each view.
>
> This would make the paper more balanced and scientifically grounded.
>

Currently the paper compares itself to different functional-ish theories:
HOT, IIT, and GNW.

Then there is a section that defends functionalism against its
alternatives: physical intrinsicism, biological naturalism, uncomputable
physics, and panpsychism.

If I attempted to propose tests against every one of these alternatives the
paper would be many times longer than it already is. I therefore think the
best option would be for you to try working with an AI to make a paper
defending your own view of physical intrinsicism, compare it to similar
theories like biological naturalism, panpsychism and Russellian Monism, and
then propose tests that would give distinct results. And you can include
general arguments against functionalist theories (which would by default
include CKT).

What do you think?

Jason



>
>
>

>
> On Fri, Sep 18, 2026 at 7:47 AM Jason Resch via extropy-chat <
> extropy-chat at lists.extropy.org> wrote:
>
>> As some of you may know I have been fascinated by questions of
>> consciousness for many years and have endeavored to write a book on it. The
>> past five years I have been focused on that writing which has resulted in
>> hundreds of pages of material:
>>
>> https://drive.google.com/drive/folders/1-SMVWgQFfImXNRRuuB9kQwhgxPLAwxYL
>>
>> About a week ago, I decided to try to use the latest AI models to take
>> all the ideas in that book and condense it to a comparatively short
>> academic paper, which I am now in the process of submitting for publication.
>>
>> I thought some on this list, interested in both the possibility of
>> machine and AI consciousness, as well as the question of how do we
>> empirically establish the minimum required fidelity for restoring a mind
>> through brain scanning and uploading, may be interested in what this paper
>> has to say, as it treats with both of these topics among many others. Here
>> is the abstract:
>>
>>
>> "Computational functionalism treats cognition as implemented information
>> processing, but leaves open why experience exists, what determines its
>> character, which process possesses it, and how it affects behavior. This
>> paper develops the centered-knowledge thesis: when alternative conditions
>> select distinct internal states that guide a process’s continuation, that
>> process knows the difference; possessing that knowledge is the phenomenal
>> difference. The complete organization of the process’s distinctions and
>> their causal roles constitutes its phenomenal character. The access
>> relations that make those distinctions jointly usable individuate the
>> subject and explain binding and privacy. This identity supplies a
>> constitutive answer to the hard problem: the explanatory gap arises when
>> third-person descriptions omit the relation between a usable distinction
>> and the process for which it is knowledge. A subtraction argument locates
>> the phenomenal minimum by stripping away report, memory, affect,
>> reflection, recurrence, and global broadcast while preserving the operative
>> distinction. From this thesis, the paper derives organizational invariance,
>> the substitution principle, and multiple realizability, yielding criteria
>> for how fine-grained neural replacement, brain scanning, and emulation must
>> be to preserve experience. The boundary that individuates a subject is also
>> its epistemic horizon: what the subject knows underdetermines its history,
>> program, substrate, and world. The centered-knowledge thesis yields
>> consciousness attributions distinct from those of global neuronal
>> workspace, higher-order, and integrated information theories for
>> feed-forward systems, faithful emulations, and overlapping centers. We
>> provide an empirical research program with explicit falsification criteria.
>> Consciousness is the inside of a difference that guides a process from
>> within."
>>
>>
>> And here is a link to the paper:
>>
>>
>> https://zenodo.org/records/22693134/files/epistemic_horizons_resch_philpapers.pdf?download=1
>>
>>
>> I would greatly appreciate any feedback, questions, comments, or
>> discussion.
>> _______________________________________________
>> extropy-chat mailing list
>> extropy-chat at lists.extropy.org
>> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>>
> _______________________________________________
> 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/20260927/ef2b2d83/attachment.htm>