Why Can We Only Think One Thing at a Time? Four Perspectives from Jellyfish to 86 Billion Neurons

> A single-threaded speculation based on *The Unbearable Slowness of Being*, cross-validated through four independent lenses — evolution, information theory, complex systems, and philosophy.

1. Evolutionary Perspective: Not a Flaw, but a Legacy

Imagine a jellyfish: no brain, just a ring of nerve net drifting like a transparent necklace in the sea. It senses chemical gradients, flexes its body, chases food. In that primitive world, there was no need to consider multiple options simultaneously — you are in one place, sense one gradient, make one action.

From this starting point, evolution did not "add processor cores." It stacked layers on the original chassis: brainstem for breathing, cerebellum for movement, limbic system for emotion, neocortex for abstract reasoning. But the deepest logic — "one input, one action," executed serially — never disappeared. 86 billion neurons do not mean 86 billion parallel trains of thought; the vast majority maintain your life: regulating blood sugar, parsing retinal signals, keeping balance.

[Evidence strength: medium]

The octopus is an interesting contrast. With 500 million neurons distributed across eight arms, each arm has its own ganglia that can sense and grasp autonomously — like eight parallel processors. Yet when it comes to decisions — where to flee, whether to attack — they are still made by the central brain. The arms' "parallelism" is local reflex; high-level integration remains serial. The octopus doesn't have eight streams of consciousness; it has one consciousness commanding eight semi-autonomous arms.

Seriality may not be a defect, but the necessary cost of a "unified agent." An organism that cannot compress perception into a coherent "now" and converge decisions onto a single "next step" is merely a collection of reflexes, not a "self."

[Evidence strength: medium]

2. Information-Theory Perspective: 10 bits/s Is Verification Bandwidth, Not Search Bandwidth

Sensory organs have enormous input bandwidth — the retina around 10⁸ bits/s, the cochlea around 10⁵ bits/s. That torrent of data doesn't vanish; it is simply never "reported."

Consciousness is not the brain's entire output. It is a summary, a "reportable layer." Like a corporation where tens of thousands of employees work in parallel but the CEO can say only one thing at a time to shareholders. The 10 bits/s figure measures what you can tell someone in the next second, not the brain's actual processing volume.

[Evidence strength: weak — depends on the assumption of a functional divide between conscious and unconscious processing]

A computational insight: P vs NP tells us verifying a solution can be fast, while finding one may require exhaustive search. Unconscious processes may be doing parallel "searching" — billions of neurons exploring possibilities simultaneously. But consciousness is a "verifier": it can hold only one hypothesis at a time, check its consistency, then accept or reject it.

If consciousness could simultaneously verify two contradictory hypotheses, action would collapse — being "certain" of both turning left and turning right results in standing still, or worse. Seriality is not a limitation of search; it is a hard requirement of decision-making.

[Evidence strength: speculative — the P/NP analogy applied to consciousness is a metaphor, not a theorem]

3. Complex-Systems Perspective: Seriality as a Product of Phase Transition

The brain is an ocean of 86 billion coupled oscillators. Each neuron fires on its own rhythm, like countless small drums beating randomly. Most of the time this is chaos. But when certain stimuli appear, scattered rhythms begin to lock phase — they synchronize, like a stadium crowd spontaneously falling into a shared clapping rhythm. Independent units suddenly become one whole.

Dehaene's Global Workspace Theory holds that the brain contains a "broadcast system": information that enters gets globally broadcast. But this system can broadcast only one thing at a time — not because of insufficient hardware, but because two simultaneous broadcasts drown each other out, like two people shouting into the same microphone.

[Evidence strength: medium — GWT has experimental support, but the inference that "broadcasting must be serial" awaits verification]

Here is an elegant phase-transition picture: below threshold, the system is decentralized chaos. Above threshold, one coherent pattern "condenses" out and suppresses competitors. Consciousness is not a lightbulb switch but a superconductor transition — the instant of going from disorder to order, when an "order parameter" dominates global dynamics. Seriality is an inherent property of this order parameter.

[Evidence strength: speculative — the phase-transition analogy is beautiful, but a neural "order parameter" has not been empirically established]

4. Philosophical/Cognitive Perspective: Seriality Defines the "Self"

William James's "stream of consciousness" is not a metaphor but a phenomenological fact. What we experience is not "multiple processes competing" but a continuous, indivisible "now." Even as the brain processes countless streams in parallel, experience is always linear — like a book that can only be read one line at a time.

Split-brain research offers stark evidence. After the corpus callosum is severed, the two hemispheres process information independently. The left hemisphere sees "spoon," the right sees "knife." Ask the left hemisphere and it says "spoon"; test the right hemisphere and it picks "knife." But when pressed on why it chose the knife, the left hemisphere confabulates: "I saw a knife-shaped object."

It has no access to the right hemisphere's input, yet it must produce a unified narrative. This suggests that conscious seriality is not just a feature of information processing but a basic requirement of the "self." The self is not an observer; it is a narrator — it must weave scattered neural events into the story of "what I experienced." If the story loses coherence, the self dissolves.

[Evidence strength: strong — split-brain research is classic and replicable]

Dissociative identity disorder seems to challenge this: one person appears to have multiple "selves." But is this parallel consciousness? Clinical observation shows these personalities appear alternately, never simultaneously. When one emerges, the other recedes. This is not parallel computing but time-division multiplexing — the same hardware running different programs, one at a time.

Schizophrenic patients' experience of "thought insertion" is also revealing: they report that "someone else's thoughts entered my mind." Note the wording — not "I coexist peacefully with another idea," but "my single stream of consciousness has been violated." Even in the most disordered states, the basic structure of subjective experience remains "one line" — merely twisted and interrupted.

[Evidence strength: medium]

The Integrated Picture: Four Perspectives Converge

| Perspective | Core Claim | Evidence | |---|---|---| | Evolution | Seriality is a legacy of primitive nervous systems | Medium | | Information theory | Seriality is the bottleneck of the conscious verifier | Speculative | | Complex systems | Seriality is an emergent property of global phase-locking | Speculative | | Philosophy/cognition | Seriality defines the "self" as narrator | Strong |

The final picture looks like this: the brain's 86 billion neurons are indeed parallel, and the vast majority of their activity forever remains outside consciousness. Consciousness is not the "sum" of these neurons but the product of some special relationship among them — a globally coherent pattern, a self-referential narrative.

Seriality is not a computational defect but a defining feature of "one agent." A system that must integrate massive parallel information, produce coherent behavior, and maintain a unified self-model must converge to seriality at some level.

But there is a blind spot none of the four perspectives fully resolves: if seriality is "necessary," why must it be *this one*? Why not multiple consciousnesses taking turns at the helm? The information-theory view says "verification must be serial" but doesn't explain why the verifier can't time-share multiple tasks. The complex-systems view says "phase transition produces seriality," but in principle different physical conditions might allow different emergent modes.

The most honest answer: we don't know. As the authors of the 10 bits/s paper put it — naming is not understanding. Saying "attentional bottleneck" just gives ignorance a name. Saying "stream of consciousness" just gives another ignorance a name.

The jellyfish's simple ring and the human neocortex share the same deep logic: one organism, in one world, one step at a time. Why it must be "one," we still don't know. But this "not knowing" is itself the most valuable discovery — it tells us that beneath the clamor of 86 billion neurons lies a most basic fact that remains unexplained.

References

• Zheng, J., & Meister, M. (2024). The unbearable slowness of being: Why do we live at 10 bits/s? *Neuron*, 113(2), 192-204. https://doi.org/10.1016/j.neuron.2024.11.008
• Hasselmo, M. E., et al. (2002). SPEAR model. *Neural Computation*, 14(4), 793-817.
• Dehaene, S., et al. (2006). Global workspace theory of consciousness.
• Gazzaniga, M. S. (2000). The split-brain revisited. *Scientific American*.
• James, W. (1890). *The Principles of Psychology*.