Lesson 8 — What's Contested? What Don't We Know?
How Does the Brain Actually Work?
Learning Material
1 pagesLesson 8 — What's Contested? What Don't We Know?
Understanding the Complex: How Does the Brain Actually Work?
Neuroscience is a young field working on the hardest problem in biology. That means genuine scientific disagreement is the norm, not the exception. And beyond the science, there are policy and ethics questions — about brain-computer interfaces, cognitive enhancement, mental privacy — where values differ legitimately.
This lesson distinguishes between four open areas: two that are primarily scientific disputes, and two that involve real value disagreements where reasonable people take different positions.
Dispute 1: Is the brain really a computer?
This question sounds philosophical. It's also deeply practical, because the answer determines research strategy.
The "brain-as-computer" camp — broadly aligned with classical cognitive science and parts of AI research — holds that cognition is computation: symbol manipulation or pattern recognition that is, in principle, substrate-independent. The brain is the hardware; the mind is the software. This view underlies much of AI research: if intelligence is computational, then silicon can replicate it.
The opposing tradition — associated with philosophers like Hubert Dreyfus and Merleau-Ponty, and with some neuroscientists — argues that cognition is fundamentally embodied and enactive. The brain doesn't represent the world and compute responses to representations. It is dynamically coupled to the body and environment. Thinking is not just in the head; it's distributed across a body moving through a world. The brain isn't a computer any more than a hurricane is a fan.
The empirical question: Is neural computation best understood as symbol manipulation (connectionism, predictive coding) or as dynamic systems coupled to body and world? There is no consensus. Predictive coding — the idea that the brain primarily generates predictions and only processes errors — has attracted enormous interest and significant criticism. Embodied cognition theories generate provocative experiments but struggle with formalization.
What's clear: the brain-as-computer metaphor, useful in many ways, is not literally true. What to replace it with is not agreed.
Dispute 2: Will we ever fully simulate a human brain?
Henry Markram's Human Brain Project was premised on the assumption that if you have enough data about the brain's structure and physiology, you can simulate it. Critics disagreed — not just about strategy, but about the underlying premise.
The disagreement has two dimensions. First, empirical: do we have enough data yet? Almost certainly not. Mapping synaptic connectivity at the resolution needed for a faithful simulation requires petabytes of imaging data and computational infrastructure that doesn't yet exist.
Second, conceptual: even if we mapped every synapse, would a simulation running that map actually think? This touches on deep questions in philosophy of mind — questions about whether simulation is sufficient for cognition, or whether the specific physical substrate (wet, warm, embodied biology) matters.
Most neuroscientists are skeptical of whole-brain simulation in the near or medium term — not because the brain is magical, but because it's more complex than current models can capture and because understanding is a prerequisite for accurate simulation.
Value question 3: Can BCIs read thoughts — and should they be allowed to?
Current BCIs cannot read thoughts in any meaningful sense. They can decode motor intentions (the signal to move a limb), simple emotional states from EEG, and reconstructed fragments of speech from neural signals. They cannot access the rich, private content of what someone is thinking.
But technology doesn't stay still. Researchers at the University of Texas Austin published a non-invasive "mental decoder" in 2023 that could reconstruct the rough semantic content of speech someone was hearing or imagining. The resolution was low — the decoder captured general meaning, not specific words. But the direction of travel was clear.
This raises questions that are not yet technically urgent but are politically important to address in advance. Should employers be allowed to use neural monitoring in workplaces? Should courts be able to compel neural data as evidence? Should insurers have access to risk profiles inferred from brain activity?
These are not sci-fi scenarios. Weak forms of neural monitoring (EEG-based alertness monitoring for truck drivers, for instance) already exist. The regulatory framework hasn't caught up.
Value question 4: Where is the line between therapy and enhancement? (Beutelsbach)
This is where values diverge most clearly, and where this course presents both positions without taking sides.
The case for permitting cognitive enhancement: Humans have always used tools — writing, caffeine, education — to extend cognitive capacity. Brain-computer interfaces, pharmacological enhancement, or neurofeedback that improves memory or attention would simply be more powerful versions of the same. Restricting access to enhancement technologies replicates existing inequalities (the wealthy already have better nutrition, education, and healthcare). Free societies should allow individuals to decide what to do with their own brains.
The case for caution or restriction: Enhancement technologies could dramatically amplify cognitive inequality — those who can afford them become more productive, more competitive, more cognitively powerful than those who can't. Unlike education, which is hard to reverse, neural enhancement might create irreversible divergences. The pressure to enhance, once available, may become coercive in professional environments — you don't "have" to get the implant, but your colleagues who do outperform you. And the long-term safety of implants that modify neural tissue is unknown.
Neither camp is wrong about its primary concern. This is a genuine value question: how to weigh individual autonomy against collective equity, and how much we trust markets and regulators to navigate the difference. Democratic societies will need to decide — and it's better to start the conversation before the technology arrives.
Next lesson: What Comes Next? — Brain organoids, optogenetics as therapy, the near-term future of BCIs.
Reading time: approx. 10–12 minutes