4 ms·
> The problem with that line of reasoning is you're assuming the brain is a computer, or that it merely computes. The brain can compute. That's extraordinary.
by tialaramex 8y ago
> The problem with that line of reasoning is you're assuming the brain is a computer, or that it merely computes.
The brain can compute. That's extraordinary. I say one type of thing does that, computers. You say no, two things, computers and then also brains. But when pressed to explain what is a brain if not a computer you'll just sputter (probably at length) without offering any substance.
In a sense that's the wrong way up to explain it. Church-Turing intuitively defines computation (the things computers can do) in terms of what our brains can do, so the match is not a coincidence but it also isn't there for the reason you probably expect. Because it's an intuition Church-Turing isn't provable, but you may notice that we subsequently built an _entire world-changing industry_ upon it in a lifetime.
You pointed to a review, others have written entire books, always they can be summarised as simply arguments from incredulity. "What? Nonsense, the brain can't be a computer, I simply won't believe that". It's unfortunate that we have woken such people from their daydreaming, I have no doubt that if similarly aroused they'd give the mathematicians what for too, "What? Nonsense, how can there be numbers which aren't ratios of whole numbers, I simply won't believe it".
- vinceguidry 8y agoNot so extraordinary. What's extraordinary isn't that brains can compute, it's that anything else can. Brains computing is ordinary. What's extraordinary about the brain isn't that it can compute. What's extraordinary about the brain is that it can be rational and self-aware, things that computers cannot do. Computers can only be deterministic. Brains can be deterministic, but they can also be non-deterministic
- TeMPOraL 8y agoComputers can become non-deterministic in practice as soon as you botch your random number handling, or hook your input up to environmental noise. Is there anything suggesting the brain is non-deterministic in a theoretical way, not just the way computers are?
- vinceguidry 8y agoDeterministic means that given the same input, the system gives the same output. Computers would be useless if they were not deterministic. Hooking up random input to a deterministic process will give random output. Garbage in, garbage out. What you're asking is for computers to be rational. That given garbage input, it will produce intelligible output. Computers cannot do this unless you program them to. Human minds are non-deterministic in many, many, many ways. Hand the same input to the same mind and you'll get a different output every single time, unless the mind willed itself to act rationally. But they are deterministic enough that you can study their behavior. Other brains are not as non-deterministic, so their behavior is easier to study. Look at it from a thermodynamic standpoint. Biological systems arose to conserve order against entropy. A fully-deterministic system will shed order, it's only through non-deterministic means that biological systems can conserve order. The mind is the most complex system nature has devised that can not only slow the aggregation of entropy, but also create order! It's not breaking the laws of physics, but yet it can create all kinds of order. Conway's Game of Life is an excellent illustration of the concept. You have to work hard and study the domain in order to find stable systems. Otherwise they just drop to equilibrium fast.
- TeMPOraL 8y agoThermodynamics is defined in a deterministic universe. It's only through sheer amount of states any interesting system could be in that entropy arises. Chaotic systems (as studied in mathematics) are deterministic too. Conway's Game of Life is indeed an excellent illustration of the concept - of how chaotic and "surprising" behavior can arise in a fully deterministic system. > Human minds are non-deterministic in many, many, many ways. Hand the same input to the same mind and you'll get a different output every single time That's not non-deterministic. That's simply stateful. Most human-made systems you interact with daily are stateful, so it's not exactly surprising. When I say "deterministic in practice" vs. "deterministic in theory" I mean this: a system is deterministic in practice if you can actually predict its outputs based on its inputs with reasonable amount of effort. A LED hooked up to a switch and a battery is deterministic in practice. So is a program computing GCD on 32-bit integers. A system is deterministic in theory if it's deterministic, but actually predicting its outputs requires absurd amount of computation. Lorenz system and weather are two examples. So is protein folding and turbulent flow. I see no reason why a fly brain, a mouse brain, or a human brain wouldn't be such systems either. The entire universe could be one, if you subscribe to many-worlds interpretation of quantum mechanics.
- Zaphods 8y ago> The brain can compute. You'll see in my comment and your quote that I don't say the brain can't compute. I agree, the brain can compute. But that doesn't mean it is a computer, because computing is an ability. People can do many other things aside from computing, none of which rely on computation, for instance they can imagine, which is the ability to think new thoughts. Computers can't imagine because all they do is compute: that's their programming. No amount of programming can produce imagination. Computation and imagination are categorically distinct as different intellectual powers and abilities. You are conflating an ability with ontology. We know what a brain is. It's a collection of fatty material with neurons that do not explicitly fire exactly like a computer. Key word there is like. Church-Turing built a model of computational logic off of intuitions about the brain and formal mathematical logic. That's it's not provable doesn't prove your point; it removes any distinction between it being right or wrong: because it is a model (lets make something like the brain). That an industry was built on computation doesn't prove anything. We know computation is an ability. For instance it's also something we can do with abacuses. We could have built an enormous industry on building elaborate abacuses. We built computers do be extremely fast at computation. We didn't build computers to be brains. You'll notice, if you read the review, that the author of the review repeatedly cites cognitive neuroscientists, even evangelists of the singularity, philosophers, psychologists, and zoologists, who have published at length on this topic and repeatedly critcise and disrupt the simple idea that the brain is a computer or an algorithm or even a machine. An entire branch of philosophy developed off of Ludwig Wittgenstein to counter the computational model of consciousness. Numerous books in the Philosophy of Mind argue that the assumption that the brain is a computer is not just unsupported, it is logically nonsensical.
- parksy 8y ago"No amount of programming can produce imagination" is a very bold statement to make. The brain exists in a physical universe, made out of matter/energy, and its behaviours are entirely dictated by the laws of physics; that's a fairly accepted truth unless you have solid evidence otherwise. The laws of physics are mathematical and can be computed by their very nature, and we are already pretty good at simulating physical interactions to a quantum level, and this ability improves over time. At some point in time, unless there is "magic" or missing physics, a sufficiently powerful computer with a physically accurate simulation of a brain would produce virtually identical results to a real brain. So either there must be new physics involved, or, the notion that a sufficiently advanced computer simulation can't produce imagination must be abandoned. A team of scientists able to sufficiently model the physics of the brain (and presumably the entire central nervous system, I imagine a disembodied brain simulation would experience a horrific form of locked-in syndrome) would not need to be concerned about emergent properties of the simulation such as a sense of consciousness, or thought, or imagination. Those things will just happen once the simulation is perfected. Indeed the cognitive neuroscience folk, etc, would be invaluable to actually understanding, training, interpreting and caring for the brain simulation, and figuring out if its behaviours and interactions constitute consciousness etc, so I do not think this has to even be framed as programmers pretending to know about brain stuff vs brain people who dismiss any notion of computationally recreating consciousness. It would be a team effort that works both ways, but is already doomed to fail if half the team thinks it's impossible from the get-go.