LLMs and the Collapse of Representational Interiority: Appendix — The Persistence of the Inner World

Part I — The Interior Model of Experience

Human cognition does not begin with a theory of mind.

It begins with a felt structure of experience in which perception, thought, and intention appear as if they occur “inside.”

This “inside” is not initially metaphysical. It is phenomenological compression: a way of organising distributed processes into a single navigable field.

Thought appears to occur somewhere.

Memory appears to be stored somewhere.

Decision appears to be made somewhere.

From this, a stable interpretive structure emerges:

• there is an inside
• there is an outside
• meaning originates inside
• expression moves outward

This is not yet a homunculus in the full sense. It is the precondition for one.

Because once “inside” is stabilised, it becomes possible to imagine something located there.

A reader.

An observer.

A subject.

The conditions for interiority precede any theory of interiority.

And once they stabilise, they become extremely difficult to dislodge.

---

Part II — The Soul as Stabilised Interiority

Once experience is structured as internal, a further step becomes almost inevitable:

If there is an inner locus of experience, what guarantees its continuity?

This is where the notion of the soul becomes structurally available.

The soul is not simply a theological addition to psychology. It functions as a solution to a specific problem:

how can the interior observer remain the same across time, embodiment, and eventually death?

Under this framing, the soul is not primarily an empirical hypothesis about invisible substance.

It is a continuity device for interiority.

It preserves:

• identity across time
• unity of consciousness
• ownership of experience
• narrative coherence of the “I”

Without such a device, the interior model begins to fragment under temporal pressure.

So the soul can be understood less as a belief about what exists,

and more as a stabilisation strategy for what must remain continuous.

---

Part III — Mortality as a Constraint on Identity

Death introduces a structural pressure that is often misdescribed as purely emotional.

More precisely, mortality creates a discontinuity problem for identity.

If the self is:

• the locus of experience
• the bearer of narrative continuity
• the owner of intention and memory

then bodily death appears as a rupture not only of life, but of the entire structure of interior continuity.

This generates a demand:

something must preserve the “I” beyond the dissolution of the organism.

Importantly, this demand does not necessarily produce explicit belief in immortality.

It produces something more general:

continuity pressure on identity models.

Different cultural systems resolve this pressure in different ways:

• souls
• afterlives
• reincarnation
• ancestral persistence
• legacy narratives
• symbolic immortality (names, works, reputation)

What unifies them is not metaphysics, but function:

they stabilise identity under conditions of finitude.

Mortality does not merely end life.

It destabilises the explanatory architecture that assumes interior continuity.

---

Part IV — The Secular Persistence of the Homunculus

Even when explicit belief in souls disappears, the structural function of interior continuity remains intact.

It reappears in secular forms:

• “the self”
• “the mind”
• “inner experience”
• “beliefs and desires inside a person”
• “executive control”
• “decision-making centre”

None of these require a literal inner observer.

But collectively they preserve its position.

The homunculus persists not because people explicitly believe in it, but because it is functionally useful:

• it unifies cognition
• it anchors responsibility
• it stabilises narrative identity
• it makes agency intuitively legible

In this sense, the homunculus is not a theoretical mistake waiting to be corrected.

It is a structural attractor in how humans organise explanations of experience.

Even scientific accounts often preserve it implicitly, relocating it into:

• neural modules
• computational architectures
• global workspace models
• predictive processing hierarchies

The language changes.

The structural role remains.

---

Part V — LLMs and the Exposure of Non-Interior Coherence

Large language models introduce a new pressure point into this system.

They produce:

• coherent language
• context-sensitive response
• apparent reasoning
• adaptive symbolic behaviour

without any obvious commitment to:

• interior experience
• narrative selfhood
• continuous subjectivity

This does not “prove” anything about minds.

But it destabilises a key assumption:

that coherent symbolic behaviour requires an inner continuous observer.

This is where the earlier structures begin to interact with the main series.

Because if:

• coherence does not require interior continuity
• meaning does not require a hidden semantic locus
• language does not require a private interpreter

then the homunculus is no longer necessary as an explanatory anchor.

And if it is no longer necessary,

then its persistence must be explained differently.

Not as truth.

Not as error.

But as stabilisation under conditions of:

• temporal finitude
• cognitive distribution
• and the need for unified identity

This brings the analysis full circle.

The homunculus does not disappear when we stop believing in it.

It persists wherever continuity of identity must be maintained across fragmentation.

Which is precisely why it has survived for so long.

Not because it is real.

But because it is useful.

And because mortality makes continuity feel necessary.