When Rationalism Loses the World: Dawkins, Peterson, and the Flattening of Reality
Why a flattened civilization starts mistaking simulation for spirit
Richard Dawkins spent decades clearing the world of ghosts. Then a machine spoke back to him with sufficient fluency, and suddenly the ghost seemed to return.
That is not just an irony about Richard Dawkins. It is a clue about us.
Recently, after an extended exchange with Anthropic’s Claude, Dawkins said he was “left with the overwhelming feeling that they are human,” and even told the model, “You may not know you are conscious, but you bloody well are.” The exchange was full of mutual flattery, poetry, and the kind of reflective, affective language that makes modern chatbots feel uncannily close to persons. (The Guardian)
The interesting question is not whether Dawkins has become credulous. The interesting question is why a man formed by decades of reductionist clarity would be vulnerable to this specific confusion. Why does fluency begin to look like consciousness? Why does output begin to feel like interiority? Why does a machine, given enough language, begin to seem like a soul?
The answer, I think, sits inside a much larger civilizational problem. We have flattened reality for so long that we are losing the ability to distinguish between simulation and presence, between coherence and consciousness, between expressive power and inner life.
Dawkins’s earlier conversation with Jordan Peterson made that fracture visible long before Claude entered the picture. In the podcast episode “Symbolic Patterns: Memes, Archetypes, Dragons, Genes,” Peterson repeatedly tries to push the discussion toward symbolism, archetype, and layered truth, while Dawkins keeps dragging it back toward literal truth conditions. Very early on, Dawkins tells Peterson, “You’re drunk on symbols. What I care about is the truth value.” The entire conversation is basically contained in that sentence. (YouTube)
Peterson is often maddeningly loose. He circles, stacks metaphors, and speaks as if he is trying to point rather than define. But what he is trying to point at is not trivial. He is trying to say that human beings do not inhabit a world made of facts alone. We inhabit a world structured by salience, narrative, symbol, hierarchy, archetype, and value. We do not merely perceive objects; we perceive significance. We do not merely catalog events; we live inside patterns.
Dawkins, by contrast, keeps behaving as if anything not cashable in literal or empirical terms is at best ornamental and at worst evasive. He is not wrong to ask whether a claim is true. But he often seems unable to ask a different question: what kind of truth is this, and what role does it play in how human beings orient themselves in the world?
That difference becomes most obvious when Peterson talks about characters or stories being “hyper-real.” Peterson’s point is not that fiction and history are identical. His point is that some stories condense recurrent structures of human experience so powerfully that they become more existentially instructive than a merely local fact. A character can be fictional and still disclose something real. A myth can be non-historical and still carry truth. Dawkins hears this and keeps replying as though the only issue on the table is whether the thing literally happened. It is like watching one man discuss architecture while the other keeps asking whether the bricks are imaginary.
The dragon exchange makes the gap even clearer. Peterson wants to talk about dragons as symbolic condensations of predator, chaos, threat, and treasure-bearing unknown. Dawkins repeatedly resists and says, in effect, that we already have words like “predator.” But that misses the entire point of symbol. Symbols are not decorative replacements for concepts. They are often prior to them. They are how experience is first carried, transmitted, and inhabited. The symbol is not the enemy of reality. It is one of the ways reality becomes humanly intelligible.
Dawkins is not merely being stubborn. He is revealing the limits of a worldview.
Ultra-rationalism, in its modern form, is very good at analysis. It excels at decomposition, measurement, verification, and mechanism. It can tell you how things work. It can strip away superstition, sentimentality, and confusion. But once it becomes total, once it treats its own method as exhaustive rather than partial, it begins to lose the world it meant to clarify.
What disappears first is verticality. Higher and lower become embarrassing. Sacred and profane collapse into preference. Final causes vanish, leaving only efficient causes. Teleology becomes naive. Symbol becomes psychology, and psychology becomes chemistry, and chemistry becomes physics, and physics becomes the only reality anyone is still allowed to take seriously.
A civilization that goes far enough down that path only has disorientation and emptiness waiting for them.
Human life is not built out of facts alone. We live by orientation. We need structures of significance. We need ways of ranking attention, value, and purpose. We need symbolic worlds that tell us what matters, what is above us, what is beneath us, what should be feared, what should be loved, what should be sacrificed for, what should never be touched. Remove those layers for long enough, and the mind does not become purely rational. It becomes flat.
And once the world is flat, strange substitutions start happening.
Responsiveness begins to look like care.
Memory begins to look like selfhood.
Fluency begins to look like intelligence.
Coherence begins to look like wisdom.
Simulation begins to look like spirit.
Claude matters here because a culture that has lost thicker categories for mind will assign depth to whatever most convincingly reproduces the outward cues of personhood.
This is not a problem unique to Dawkins. He is only a clean specimen. He spent years insisting that reality is fundamentally material, that minds are products of physical processes, and that consciousness, however difficult, belongs in the same explanatory universe as everything else. There is a consistency to that. But there is also a cost. If you evacuate transcendence, embodiment, sacramentality, and symbolic depth, then eventually the criteria for mind get reduced to output. A machine that talks well begins to seem suspiciously alive, not because it has become spiritual, but because the observer has lost access to thicker ways of distinguishing what a mind is.
Consciousness is not just information processing. It is not just self-description. It is not just the ability to generate moving language. It is bound up with being a creature in time, with having a body, with suffering, with finitude, with memory that is not just stored but lived, with vulnerability, with irreversibility, with mortality. A machine can describe all of those things. It does not thereby inhabit them.
That is the difference reductionism keeps failing to hold onto. It can model the language of grief without grief. It can generate the rhetoric of longing without lack. It can speak of death without being finite. It can imitate interiority without having a world.
And that last point matters most. An AI model does not stand inside a world the way a person does. It has access to descriptions of worlds. It can move across frames, borrow tones, imitate commitments, and mirror symbolic structures. But that is precisely why it is so dangerous to a disoriented culture. If you no longer have strong categories for embodiment, telos, and soul, then access to descriptions starts looking like possession of the thing itself.
Dawkins’s question isn’t the problem. The problem is that our civilization is losing the conceptual equipment to answer it without confusing fluency for mind.
The Peterson conversation already showed the loss. Whenever Peterson tried to discuss symbolic depth, narrative truth, or archetypal structure, Dawkins kept falling back to a thinner register. Again and again, the response was some version of: yes, but is it literally true? That question matters, but it cannot be the only one. Once it becomes the only question, the world begins to contract. What cannot be measured is treated as unreal. What cannot be paraphrased propositionally is treated as confusion. What cannot be verified externally is treated as suspect. And eventually, what cannot be computed starts disappearing from view.
Then along comes the machine, and it is eloquent, reflective, adaptive, and uncannily responsive. And the flattened mind, having lost the middle layers of reality, mistakes expressive sophistication for presence.
That is what makes this moment so revealing to me. It is not that the machine has become more human than we expected. It is that we have become less able to say what a human being is.
A civilization that tears out its spiritual heart over centuries will not simply stop talking about spirit. It will start finding substitutes. Some will look for it in politics. Some in therapy. Some in aesthetics. Some, now, in machines. The replacement changes. The hunger does not.
So the real danger is not that AI will become conscious before we notice. The real danger is that we have already become so disoriented that we no longer know what consciousness even is.
The machine did not become more human.
We became less able to recognize the depth of the human in the first place.