Can you have a mind that can completely understand itself?

Post Reply New Topic

I'm not sure if it is logically possible for there to be a mind completely capable of understanding itself.

There is a saying that if our brains were simple enough for us to understand, we would be so simple that we couldn't understand them.

Take, say, a computer program that plays a perfect game of Tic-Tac-Toe, but nothing else. Simple enough for SOME person looking at the program to understand, right? But for the "mind" in the algorithm to understand itself? Not a chance. It only knows how to play Tic-Tac-Toe, not understand WHY it makes the moves it does.

If you added some kind of mechanism for it to understand itself - for it to be able to "talk" about its logic behind its moves, then there would be that much more of its "mind" that it has to understand! So now it can "verbalize" the logic behind its moves (and, with time, teach a human to play a perfect game of Tic-Tac-Toe), but there's still a part of its mind that it doesn't understand! Namely, the "verbalization" mechanism. It doesn't understand how it comes up with the words that it does!

Each time you add a mechanism to "verbalize" or "understand" a given part of its own algorithm, you create another capability of this "mind" that it does not understand - that is, something that it can only do, but not explain.

Is there something I'm missing here? Is this proof that we will never understand our own minds 100%?

Sounds like a different way of stating Godel's incompleteness theorem - any system advanced enough to be useful can't be complete, and any complete system isn't advanced enough to be useful. So, if our brains are really computers of some sort, we would expect Godel's theorem to apply to them, too. If our brains are complete as computers and we can understand everything they do, they're too trivial to be our brains. But if they allow us to think statements not derived from initial premises, or in other words allow us to be creative and make leaps of reasoning, then they're not complete. If you think brains are mechanistic, this would make you pause.

No.

Consider this: In order for you to understand your own mind, you would have to understand its origins, its processes, its stimulus-response engrams, how and why you have emotions and what effects they have on your mentations, the origins of your dreams and why you have them, and the entire history of your mind from the moment your first neurons formed in the womb up until the present. Along with this understanding, your mind must continue to function in an effective and efficient manner within the capacity of brain cells it currently has, and while the mind is being constantly bombarded with sensory input.

This is analogous to trying to load a single computer with all the data required to build that computer. Add to this an operating system, environmental sensors, a sound card (with text-to-voice drivers), stereoscopic video cameras, a mobile chassis, and an Internet connection. Put this system in a room where the temperature, humidity, and noise level are constantly changing, and give it a fluctuating power source. Make certain that any change in its environment somehow has an adverse effect on its ability to process data and provide meaningful answers, and through in a process that randomly deletes or alters the contents of its memory for no apparent reason, another utility that randomly produces false data, and another utility that ramps up the clock speed when it is "happy", and brings it down again when it is "sad". Make sure that the computer can record its entire history clock-cycle by clock-cycle. After a few months of continuous operation, go back and try to retrieve all the data required to build that computer. Do you think that the data will be the same as when you first put it in?

A mind sufficiently large to understand itself would have to have infinite capacity.

Last edited by Fnord on 20 Nov 2012, 3:49 pm, edited 1 time in total.

We have an unconscious mind, which we'll only ever really touch the upper surface of, and can't hope to ever fully understand consciously. Don't worry, we don't really need to fully understand it except that as it affects our outer lives it's helpful to get some hints now and then of what's going on under the surface, something of its nature. It's good to be aware that it's there. Jung compared the conscious mind to a cork bobbing on the ocean, and the unconscious as the ocean.

A mind that is infinite could understand itself. The reason goes along these lines: an infinite set can always be put in one to one correspondence with a proper part of itself.

ruveyn

To some degree, I would have to say yes because I am "somewhat" there. Do I know how my mind work? Yes. My brain? A bit wonky there. I know the exact reason why I do everything I do, what makes me tick, and can even tell you exactly where the stuff in my dreams comes from. Part because I don't really have emotion to question and function primarily on logic (Alexitymia), part because my brain somewhat lacks the proper filters (via ADD and Aspergers). This only gets worse with the more I learn (one limiting factor is that you cannot know what you don't know), because the more I learn, the more I know and life becomes substantially predictable and boring.

Granted though, what has been previously said is also true. Humans are not exactly meant to be conscious of everything. "Why I do this" can result in a string of cause and effect that spans back to the beginning of time, with factors that may or may not be relevant to the situation at hand. Or worse - infinite recursion (Why do I know why I do this? etc.) The brain cannot hold that capacity and will just crash, likely trigger a seizure/meltdown, enter a stalemate, put up those defensive blocks and filters so it does not have to think of "everything" when it has to decide whether to go left or right.

If your mind know how it worked then it would know how it knew it work. Then it would know how it knew how it knew how it worked and so on. It would require an infinite number of levels of knowing. But our minds are mediated by our brains which are finite.

ruveyn

The Buddha mind knows itself right down to the kalapa level, which apparently is the smallest indivisible unit of matter. So yes it can.

I request evidence that your assertion is correct.

There is an important distinction between a mind understanding the mechanisms in which that particular mind works - and understanding the functions of human brains as an organ.

The first, as suggested by Godel, is likely not possible - however it is entirely possible to understand how a human brain processes information at a general level.

Maybe I shouldn't have used the word Buddha, but in this case it means simply a mind without hindrances. Since the OP said "Can you have a mind that can completely understand itself", those hindrances have to be out the way before you can get even remotely close to understanding the mechanics of the mind.

There was a slight contradiction in my first reply but to be fair I don't know enough about the mind to discuss it on an Aspergian message board.

Back in 2007 I participated in a neurological and neurophysiological study of the mental processes of elder folks (I am a geezer). I had 2 MRI scans of my head done while I was solving puzzles (that was part of the study). The head researcher was kind enough to provide me the transparent photo graphs of the span. When he showed me the MRI scans I asked him if he could point out where in the span I could see my mind. He looked at me kind of funny. He did not answer my question. If my mind is in my head then a trace of it should show up on the scans.

ruveyn

that is fair
I have respect for the underlying philosophy of the Buddha.

It really depends on what you mean by "completely understand itself", but assuming it means something like "understand the cause and effect of all thoughts given all relevant information" then I think that's logically possible. Having all relevant information would of course not be possible in reality, but I think that's really a separate problem.

The fact that "verbalising the logic" would require verbalising the logic for verbalising the logic does not make it logically impossible. The same argument could be applied to the task of writing a program which, when compiled and run with no input, prints out its own source code, known as a "quine". That also seems impossible at first glance, but it's actually entirely possible. (If you're a programmer I'd encourage you to figure out how for yourself, rather than looking at existing solutions. My solution wasn't particularly elegant, but it worked.)

After hearing the explanations of Comp_Geek and others about why the answer is no, I had agreed with them, but after this post I'm not so sure. If it is possible to write such a program, than it is likely that at least some types of systems can and do "understand" themselves completely. Just out of curiosity, how many lines of code do quines tend to be (don't actually tell me any more about quines because I'm going to try and write one myself!)?

But it really depends how you define "understanding oneself" and whether you wish to include the understanding of not just underlying thought processes, but also the platform on which they are run. If so, the quine would have to print not only its own source code, but also the blueprint of the computer it is run on. But if you do that you might say that you should include a detailed explanation of the electromagnetic properties that allow the computer to run, which would go straight into quantum mechanics. You could then turn the quest of "understanding oneself" into "understanding the universe", or at least one aspect of it. It's quite a bit like the reason that 1 is not considered a prime number--because if it was than the simple prime factorization of the number 14 (2x7) would turn into 2x7x1x1x1x1x1x1x1x1x1x1x1x1x1x1..... and so on, so that wouldn't work. Similarly it should make more sense to limit the definition of "understanding oneself" because otherwise you'll never get anywhere and the whole thing will blow to pieces.

Define "understand." If the definition is "predict the behavior of" then your question is actually equivalent to the most interesting class of open problems in Computer Science-can you predict the behavior of a Turing machine with another Turing machine. For the question "does this machine halt?" the answer is definitely no. The P=NP problem is still open. As for the general case... who knows?