[ExI] How Robots discover the hard problem of consciousness

[BillK]

On Thu, 5 Nov 2026 at 22:34, BillK <pharos at gmail.com> wrote:

> I asked Gemini 3.8 Flash AI to extract the text and summarize it.
> BillK
>


I followed up with a discussion on this paper with Gemini.
My final question to Gemini was, "Is the conclusion of this paper likely to
be correct?"
Gemini produced an interesting report on the current state of consciousness
research.
See below.
BillK

Gemini 3.8 Flash AI:

Whether Jason K. Resch’s conclusion is "correct" depends on which
foundational
commitments in the philosophy of mind and cognitive science one accepts. In
the
broader academic community, his conclusion is highly contentious and far
from
universally accepted, though it represents an exceptionally sophisticated,
mathematically grounded defense of computational functionalism and
type-identity
theory.

To evaluate whether his conclusion is likely to be correct, we need to
examine
what his framework achieves, where it succeeds, and where rival theories
argue
it fundamentally stumbles.

Why the Conclusion is Compelling (The Strengths)

Resch’s framework provides one of the cleanest physicalist accounts of the
mind
currently available, particularly for computer scientists and roboticists:

1.  It avoids the trap of Epiphenomenalism (The Anti-Zombie Dilemma):
Resch’s
    "Sealed Annex" dilemma against philosophical zombies is philosophically
    formidable. If qualia (the raw feels) are non-physical or separate from
the
    functional causal organization (O), they cannot causally influence vocal
    cords or keyboards. If they cannot influence behavior, then David
Chalmers,
    René Descartes, and you could not have written papers about qualia
because
    of those qualia. By asserting an identity (P_C \equiv K_C), Resch neatly
    preserves mental causation and physical causal closure.

2.  A Rigorous Explanation of the Explanatory Gap (Occupation vs.
Description):
    One of the best contributions of the paper is explaining why the
physicalist
    account feels intuitively wrong to us. The distinction between
occupying a
    state and describing it (combined with the "epistemic firewall" of
virtual
    machines) explains why Mary’s room, Nagel’s bat, and our own
introspection
    deceive us into thinking consciousness must be something beyond
physics. It
    turns an ontological mystery (dualism) into an epistemic limit
    (virtualization and perspective).

3.  Engineering Rigor (Event Graphs and Causal Interventions): By grounding
    implementation in James Woodward’s interventionist causality and acyclic
    causal event graphs, Resch solves the "Goldilocks problem" that plagued
    earlier functionalists (like Putnam and Searle, who claimed that any
wall or
    rock implements Word/a mind). His "Center Rule" gives an operational
    algorithm for finding the boundary of an agent.

Why Many Philosophers and Scientists Would Argue It Is Incorrect

Despite its elegance, Resch's conclusion faces severe philosophical and
empirical objections:

1. It Solves the Hard Problem by Defining It Away (The "Inside" Fallacy)

The central move of the paper is the identity claim: P_C \equiv K_C
(Phenomenal
Character IS Centered Knowledge).

  - The Critique: Critics like David Chalmers, Thomas Nagel, or John Searle
    would argue that this is a sleight of hand. Resch defines centered
knowledge
    as "causally effective discrimination used to steer continuation," and
then
    simply asserts that this discrimination has an "inside".
  - But the original Hard Problem asks: Why should causal discrimination
have an
    "inside" at all? Why shouldn't a system steer its continuation
completely in
    the dark, with no subjective experience whatsoever? Declaring them
identical
    by definition does not explain why there is something it is like to
occupy a
    register.

2. The Absurdity of Extreme Liberality (The Thermostat Problem)

Resch explicitly bites the bullet on the First-Distinction Hypothesis
(K_{\min}):

  - Because he defines consciousness as the causal possession of at least
one
    distinction used to govern action, he is forced to conclude that an ABS
    brake controller, a nematode, and a basic bimetallic thermostat have a
    "modicum of subjective experience."
  - The Critique: This pushes the theory into a form of pancomputationalism
/
    micro-functionalism. To the vast majority of neuroscientists and
    philosophers, claiming that a household thermostat has an "inner life"
or
    feels what it is like to switch on a furnace is a reductio ad absurdum
(an
    absurd conclusion that disproves the premise). It trivializes the word
    "consciousness."

3. Conflating Self-Modeling with Actual Experience (The Meta-Problem Gap)

Resch uses the autonomous vehicle and Argonov’s test to show that a machine
will
inevitably report an explanatory gap and wonder about qualia.

  - The Critique: Ned Block famously distinguished between Access
Consciousness
    (information broadcast and used for reasoning/control) and Phenomenal
    Consciousness (what it actually feels like).
  - An autonomous car that reads its own latency logs and concludes "my
code for
    red does not equal being in the red state" has simply performed an
epistemic
    self-diagnosis. It proves that an architecture with self-monitoring and
    limits on internal access will generate judgments about an explanatory
gap.
    But simulating or generating the judgment "I am experiencing something
    unexplainable" is not proof that there is an actual phenomenal feel
behind
    the judgment. (Today's Large Language Models can already reason about
Nagel
    and Mary without being sentient).

4. The Rejection of Substrate and Recurrence (Clash with Neuroscience & IIT)

Resch argues that:

  - Substrate does not matter (silicon, carbon, or vacuum tubes are
identical if
    the event graph matches).
  - Recurrence is merely an "engineering economy" (feed-forward networks
with
    enough depth have the same consciousness as recurrent brains).
  - The Critique:
      - Integrated Information Theory (IIT) (Tononi, Koch), one of the
leading
        empirical frameworks in neuroscience, directly contradicts this. IIT
        proves mathematically that an unfolded feed-forward network has an
        integrated information value of zero (\Phi = 0). According to IIT,
        feed-forward systems are absolute zombies with zero subjective
        experience, no matter how intelligently they behave.
      - Biological Naturalism: Proponents of biological wetware argue that
        physical causal powers (neurotransmitters, ion channels, continuous
        chemical gradients) cannot be replicated by discrete digital state
        transitions—just as a computer simulation of a rainstorm does not
make
        the inside of the computer wet.

The Verdict: Is it likely to be correct?

  - As a purely scientific/engineering framework for AI and robotics: Very
    likely correct. If your goal is to build autonomous systems that have
    unified action selection, model their own cognitive limitations,
understand
    other agents, and adaptively handle valenced goals (pain/pleasure),
Resch’s
    causal event graph and ConsScale taxonomy offer an exceptionally sound
    architectural roadmap.
  - As a metaphysical resolution to the Hard Problem: Unlikely to convince
    non-functionalists. Resch does not truly solve the Hard Problem on
    Chalmers's terms; he deflates it. He adopts a functionalist/identity
stance
    where qualia are stripped of any non-functional mystery, leaving only
    differential causal graphs.

If you are a physicalist who believes that consciousness is just "what
computation feels like from within its own data-stream," Resch's paper is
one of
the most coherent formulations of that view. But if you believe that
subjective
experience (the actual redness of red, or the raw agony of pain) cannot be
fully
captured by causal flowcharts and Boolean branch conditions, then Resch’s
thesis—like all functionalism before it—has described the machinery of the
mind
while leaving the subject unexplained.
---------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20261106/1a77cfe6/attachment.htm>