The Foundations of Cognitive Theory
- Thread starter barehl
- Start date
Reality Check
Penultimate Amazing
Is your version of the Chinese room now empty, barehl
!Turing Machine
There is nothing preventing someone from implementing a real Turing machine running the example of pattern matching you gave and putting it into a Chinese room. There will be limitations, e.g. an infinite tape becomes finite.
Maybe that "A Turing machine realisation in LEGO" could run your example
.
Last edited:
RussDill
Philosopher
I disagree. One of the central tenants of the Searle's Chinese room thought experiment is that a conversation can be carried out. This means that not only must things be looked up in books, but some internal state be modified as well. Which means you have an input, an internal state, an output, and a set of rules by which the current state in transitioned to a new state based on the current state and the input. So the Chinese room is just a FSM. The process of providing output and advancing the state of the FSM is computation.
The whole point of the Searle's Chinese room is that he believes the same thing. That a computer and the Chinese room have the same capability. And since he believes it absurd that somehow a Chinese room could have consciousness, so too would it be absurd for a computer to have it. More specifically, since the human performing the Chinese room actions would not understand Chinese, the computer performing the Chinese algorithm would not understand Chinese.
And again, since human brains are conscious, and can be represented by FSMs, a Chinese room can be can be conscious.
Others have explained this, but you seem to think that the combination of two countable sets provides an uncountable set. A very basic example that students learn early on is the proof that the Cartesian plane is countable. In your thinking, the topic the question is about would be the y axis, and the question would be the x axis. If you are looking for some background, use these search terms:
'Theorem: The Cartesian product of finitely many countable sets is countable.'
Have you not read anything about the theory of computability? The entire premise is based on what grammar a given machine can accept (pattern matching). So here, once again, you are declaring that the human brain is above the Turing machine in computability classes and therefore either some law of physics allows for hypercomputation, or we have souls.
Ah, now I see why you think your AI theory is "just around the corner". And things are going to start happening. I've read many a paper that describe the exact above sequence. Not only that, but there are entire programming languages based on the above type of thinking.
You are describing a semantic network. It's something that has been tried to death, and the over-optimism coupled with the disillusionment of it's failures caused led to the AI winter.
https://en.wikipedia.org/wiki/History_of_artificial_intelligence#The_golden_years_1956.E2.80.931974
RussDill
Philosopher
Such a thing is laughably simple with a 2 state FSM.
PixyMisa
Persnickety Insect
How big is the room?
W.D.Clinger
Philosopher
So you're assuming the room contains a computing device that's powerful enough to perform whatever you mean by simple pattern matching but not powerful enough to match patterns that appear on the left hand sides of type 0 productions?
Set theory is powerful enough to serve as foundation for all standard mathematics, which includes addition.
{ anbn+1 | n ≥ 0 } is not a regular language, so no finite state automaton can recognize it.
Dancing David
Penultimate Amazing
I wonder why the basis of neurology and brain function is not being discussed.
Beelzebuddy
Penultimate Amazing
- Joined
- Jun 10, 2010
- Messages
- 10,570
Yes, that's becoming increasingly difficult to deny, no matter how charitable you're willing to be.
Which means I can weigh in on the Chinese Room.
Once we discard the argument from incredulity, frankly, I don't see what the fuss is about. If you don't adopt some form of dualism, it should be evident that "understanding," and "consciousness," and all of the other properties we ascribe to intelligence must be somehow emergent from components which are individually lacking. In the Chinese Room we have what appears to be a understanding of Chinese, even though none of the constituent parts understand Chinese. So... ta da? Seems like a perfect description of emergence.
The trick, of course, is building a lookup table which is complete enough to convincingly answer in Chinese.
If we alter the hypothetical a bit, I think my point will be more clear. To avoid strawmanning, let's switch languages to Hungarian. Instead of a lookup table of inscrutable questions and answers, the Hungarian Room has a perfect Hungarian-English phrasebook. Questions written in Hungarian come in, are translated into English, answered by you, translated back into Hungarian and sent out. The situation is unchanged to outside observers - you appear to speak perfect Hungarian - but on the inside, not only do you not understand Hungarian, the phrasebook has no capacity to select responses. But the both of you together, taken as a unit, do. That's how understanding emerges. There's nothing special about it. Take the book away and you won't understand Hungarian, drive an icepick through your Broca's area and you won't understand speech. Same difference.
It's a top-down approach. Computer Science is generally cool with neurobiology, but the assumptions, ambiguities, and amount of stuff which is likely to be completely wrong coming out of neuroscience would take it too far from its mathematical, anything-not-proven-is-hearsay roots. Cognitive Science would have fared better, but despite barehl's attempts to combine them, Cog. and Comp. Sci. have the same kind of disconnect.
The problem for me is that "emergence" doesn't have any explanatory power. It's an observation that cries out for an underlying explanation. I want to know the rules. Otherwise, it sounds too much like, "...and here, in step #2, emergence happens and presto! we're done."
Without a recipe book, emergence feels a bit like dualism repackaged.
Beelzebuddy
Penultimate Amazing
- Joined
- Jun 10, 2010
- Messages
- 10,570
Nah, it just sounds magical because the things it's commonly used for have so much special pleading attached. All it means is something produced by interactions between simpler processes. The macrostructure of sand dunes and snowflakes are emergent properties too.
[ETA] To clarify: what does it mean when we say that the room understands a language? I would say it requires the appearance of taking in a string of that language, parsing it for a correct meaning, producing some meaningful response, and encoding that response back into the language. The Chinese Room has meaning and response built into its set of questions and answers, but the Hungarian Room performs each step explicitly, which is why I brought it up. The phrase book performs the Hungarian decoding and encoding, while the human creates the response to the query. Neither element can satisfy the entire definition alone, but together they can, thus it is an emergent property of the room as a whole.
Last edited:
RussDill
Philosopher
What I mean is that a 2 stage FSM can perform addition very easily assuming symbols are fed in from the right to left.
W.D.Clinger
Philosopher
Sure. If you thought I was contradicting you, I apologize.
It's been hard to guess what barehl is thinking. If he was thinking of finite state transducers, then your comment might be relevant. If he was thinking of recognizers that operate via pattern matching, then my comment might be relevant. It's entirely possible, perhaps even likely, that neither of our comments had anything to do with whatever barehl may be trying to say.
Beelzebuddy has done an admirable job of summarizing the main issues here, and I applaud his contributions.
barehl
Master Poster
- Joined
- Jul 8, 2013
- Messages
- 2,655
I'm tired of babysitting. I can't explain even the most basic concept because the replies consist of people arguing that their superhero Turing Machine can beat up my Chinese Room. I don't know what anyone wanted to get out of this thread but this isn't worth my time.
Brian-M
Daydreamer
- Joined
- Jul 22, 2008
- Messages
- 8,044
As I see it, this scenario is just a way of having a proxy conversation with the writer of the book, who is a conscious being.
I tentatively agree with this assertion on the technicality that if it's using set theory and logic to produce its responses, then it's not actually a Chinese Room. (ETA: In the sense of the Chinese Room as most people think of it, with a facilitator looking up fixed responses.)
However, I strongly disagree with the idea that you could not create an AI capable of solving problems and passing the Turing test based on these things. The reason for my disagreement is your inclusion of "logic" in the list.
I don't know why you included logic in the list when your examples only talked about set theory and didn't even address other forms of logic. With adequate information at its disposal, it could apply logic to work out the answer to the balloon question.
(I'm assuming that "information" is an unintended omission from the list, because your set-theory examples require it to have an adequate body of available information in order to work.)
Last edited:
Perhaps you should just accept that your genius puts you so far out of the reach of the minds of the other members of this forum that posting here is like trying to have a conversation with a group of baboons.
Perhaps you should just accept that your genius puts you so far out of the reach of the minds of the other members of this forum that posting here is like trying to have a conversation with a group of baboons.
That was my impression as well.
PixyMisa
Persnickety Insect
The Chinese Room is an inherently dualist argument. Dualists need no assistance; they beat themselves up.
You've spent a couple of years thinking about this and reading the literature. Good for you. Some of us have spent a couple of decades. Some of us actually work in relevant fields. You might do well to listen to what people are saying.
PixyMisa
Persnickety Insect
And just on the Chinese Room: I asked this before, but how big is it?
The answer is the first step on the road to understanding.
The answer is the first step on the road to understanding.
W.D.Clinger
Philosopher
The highlighted phrase is a bit of an exaggeration, but I respect the effort you've put into proving it.
Slings and Arrows
Graduate Poster
