Sorry, I didn't see that you'd replied to my comment until today.
> A car's horsepower is in the engine.
Where is it in the engine? The engine can go 180hp. But the engine does not contain 180 hp. That's what the concept of an ability or a power is. A broken engine cannot go 180 hp. But, if as you say, it is in the engine, then that distinction would be irrelevant. We would still say a broken engine can go 180 hp. But we don't.
> Similarly, if brain can compute, it is a computer. It's also an organ.
Right, you'll see I have never said the brain can't compute. But that doesn't mean it is simply a computer. If the assumption that the brain is a computer is to stand then the abilities of a computer should be compared to the abilities of a brain, or a person. There are those that match. We agree on that. But there are those that do not. And that means the assumption that the brain simply is a computer is flawed. It is an organ that can compute. But to extend from that that it is a computer is eliding the crucial difference between the two. That is your assumption.
> Evidence needed. Why would it [imagination] not be computational?
Let's ignore that the premise you are making: that imagination is computational, requires you to support it as well;
> imagination-like computations
> A fuzzy search on a graph
> A series of simulations with relaxed constraints and somewhat randomized initial states
All require you to posit things that are -like, or somewhat like imagination. But computers are programmed. They can't think new thoughts. They are closed deterministic systems. That their output seems imaginative or novel does not mean the computer has the ability to imagine, it means the computational output was unexpected to you or the people who wrote the code. The idea that imagination is computational is a category error.
> not a big leap to conclude that imagination is nothing but a more complex variant of such computations
This is actually an enormous leap. Can computers imagine? You will find zero agreement in that regard. That doesn't prove your point. You'll need to provide evidence that computers can actually violate their programming, cannot just compute and instead imagine. But that's not what computers do. Computers compute. That they can do things that seem like imagination to you does not mean they can imagine.
>> It means you can fuzzily model the brain on a computer, but that model will have glaring gaps.
> Models exist on a map, not in the territory. So do brains and computers. The territory is made of whatever sub-quark substrate the reality is made of. When you say "brain", what you're really referring to is a model, and a pretty black-boxy one. Viewing the brain as a computer is an attempt to apply a model that's little more transparent (and therefore more useful); as long as it matches observable evidence (and it does), it's the right thing to do.
Excuse my original words, I meant "fuzzily model the brain as a computer
Again, I don't think applying the computer as a model is completely invalid. But it has limitations. You can't just brush off those limitations when you talk about the brain as a computer. They fundamentally mean the comparison is less useful. Supposing it is 1 to 1, which you are doing leads you to build on assumptions that are unsupported. You have to accept that the assumption that the brain is a computer has serious criticisms brought against it. And you have to defend that assumption. You can't simply ignore them and argue that you are right.
For instance viewing the brain as computer frequently does not match the observable evidence. We can imagine. Computers cannot. That is observable. So how do you support the assumption that the brain is a computer in spite of that?