[ExI] How Robots discover the hard problem of consciousness

[BillK]

On Thu, 5 Nov 2026 at 21:53, Jason Resch via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> I recently gave this talk on how robots trained with a singular goal (in
> my example, how to drive safely) inevitably must ascend through increasing
> levels of consciousness in order to become more competent in that goal:
>
> https://docs.google.com/presentation/d/1B9tcSNf5UdYkO78Hazwj2Gs_zRyZ0Fo2YMUW7Xusv4o/edit?usp=drivesdk
> Jason
> _______________________________________________



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

Synthesis and Executive Summary

In What Is It Like to Be a Process?, Jason K. Resch outlines a rigorous
computational-functionalist theory of consciousness called the
Centered-Knowledge Thesis. The presentation addresses fundamental open
questions
in the philosophy of mind, computational theory (specifically referencing
Matthias Scheutz’s work on functional realization), and cognitive robotics.

1. The Core Thesis: P_C \equiv K_C

Resch posits an identity theory: Phenomenal Character (P_C) is identical to
Centered Knowledge (K_C).

  - Consciousness is not an epiphenomenal "glow" or a secondary byproduct
    produced by physical matter.
  - Rather, subjective experience is what it is like from the inside to
occupy
    an organized system of causally effective distinctions that steer an
agent's
    future transitions.
  - This is constrained by Causal Closure and the Closure Constraint: Any
    phenomenal difference that can be noticed, compared, remembered, or
reported
    must correspond to an actual difference in the causal-computational
    organization producing that judgment.

2. Computation as Unfolded Causal Event Graphs

Addressing Scheutz's problem of physical implementation:

  - Computation cannot be defined by arbitrary state mapping
(Putnam/Searle) nor
    by microphysical identity. It is defined by bisimulation over causal
paths.
  - Resch represents processes as acyclic causal event graphs G = (V, E)
    (drawing on Lamport's happened-before relations and spacetime causal
sets).
  - Crucially, recurrence is not essential for consciousness; it is an
    engineering economy. Recurrent loops across time unfold into
feed-forward
    causal event graphs. A feed-forward network with sufficient circuit
depth
    realizes the exact same event graph as a recurrent circuit with temporal
    depth.

3. The Epistemic Firewall and the "Center Rule"

  - The Center Rule: To determine the boundaries and grain of a mind, one
traces
    intervention-sensitive dependencies backward from behavioral "anchors"
    (decision/selection points) to the sensory inputs. The adequate
interface is
    the coarsest level that preserves every dependency.
  - The Epistemic Firewall: Thanks to computational virtualization and
multiple
    realizability, a process cannot inspect the underlying physical
substrate
    realizing its interface (carbon, silicon, vacuum tubes). The substrate
is
    epistemically opaque, explaining why physics appears to observers
purely as
    structure and dynamics.

4. Evolution, Valence, and the Spectrum of Mind

  - Consciousness is an evolutionary prerequisite for intelligent,
non-random
    action: agents must discriminate environmental states to select
appropriate
    continuations.
  - Valence (Pain/Pleasure): Pain is not an arbitrary tag; it is an
organization
    that forces avoidance, error-correction, and resource reallocation.
  - Continuum of Consciousness: Analogous to biological life (which spans
    seamlessly from single-celled bacteria to blue whales without a "magic
    moment"), consciousness spans from minimal systems (a single used
    distinction in a thermostat or nematode) up to human and metacognitive
    architectures.

5. Derivation of the Meta-Problem and Argonov's Test

Through a thought experiment of an autonomous vehicle evolving across 10
levels
of cognitive complexity (ConsScale) solely to avoid traffic accidents, Resch
shows that:

  - The machine must eventually build models of other agents' perspectives
    (Level 5), self-models (Level 6), and inspect its own internal latency
logs
    (Level 8).
  - In doing so, the machine naturally encounters the explanatory gap:
reading
    its own code/weights (a third-person description) does not place it in
the
    state of executing/occupying that code (first-person occupation).
  - The vehicle independently discovers Nagel’s Bat problem, Mary’s Room,
and
    the Hard Problem of Consciousness. Because it was never trained on
    philosophy, it satisfies Argonov’s (2014) criteria for validating
machine
    consciousness via spontaneous phenomenal judgment.

6. Defeating Rival Metaphysics

  - Philosophical Zombies: Refuted via a causal dilemma. If qualia (Q)
affect
    behavior/reports, zombies are physically non-identical; if Q has no
causal
    impact, conscious beings could never know or report that they have it.
  - Mental Causation (Kim’s Exclusion): Dissolved by interventionist
causality
    (Woodward). Higher-level computational organization is the true
invariant
    difference-maker across time, not microphysics acting across vertical
    levels.
  - Panpsychism & Russellian Monism: Dismissed because non-relational
"intrinsic
    natures" cannot account for the structured, relational, differential
    character of phenomenal experience.
---------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20261105/79c4e856/attachment.htm>