Thursday, April 11, 2013

The Human Mind is a Chinese Room

The following is an essay I wrote as part of an Introduction to Philosophy course I finished recently on Coursera that I felt like sharing more generally. 

Could a computer have a mind?

In the debate over can a computer have a mind, one of the most well known arguments for the negative is Searle's Chinese Room Argument. To make his case, Searle asks us to consider a man locked in a room. In the room with the man are some pens, some paper, and a large book. From time to time, the man will receive a piece of paper with some symbols on it which the man will look up in the book and then copy some other symbols onto a blank piece of paper and send that out of the room. Unknown to the man, the symbols are words in Chinese and the book is telling him what to say in Chinese. Searle claims that since the man is just following instructions based on the symbols and not their meaning he doesn't understand Chinese, and since any computer program would also just be following instructions based on the symbols, it too can not understand Chinese. That is, a computer can operate on a syntactic level, but it can not operate on a semantic level, and thus can't be said to have a mind.

I disagree with conclusion, and to explain why I wish to present a counter example, namely the human brain. The human brain is made up of around 86 billion neurons. Each individual neuron is functionally very simple. It receives electrical impulses from other neurons and from time to time based on the received impulses it will send out an electrical impulse to other neurons, who go on to do the same thing. Now for an individual neuron all of these electrical impulses only have syntactic information. Either there is an impulse or there isn't, that's all. It might have originated with a pain receptor in your foot after stepping on a nail, or in a photoreceptor in your eye when you look at a red rose, but to a neuron it's just an electrical impulse with no semantic information associated with it.At this functional level, a neuron is very similar to a circuit in a computer, various electrical inputs are received, and based on that, a certain output is made. From this we can see that a neuron operates on a syntactic level, and not on a semantic level.

So what happens when we start grouping together neurons? For a moment, let us go back to Searle's Room, and this time instead of just one person, we have two people, each with their own book, receiving and writing symbols and passing them between each other as their books instruct to achieve the same outcome. Again, everything is happening on a syntactic level so no semantic understanding occurs. If we continue to add more people to the room each with their own book of instructions, nothing changes to move us from the syntactic to semantic level. So if we remain operating on a syntactic level by adding more people to the room, we will also remain at the syntactic level by adding more neurons to a brain. So a human brain (and indeed any other creatures brain) must operate at a syntactic level and not a semantic level.
But the human brain does operate on a semantic level. We know a rose would smell as sweet by any other name. Semantics are who we are, it's how we think.

We can come to either one of two conclusions. We could look at Searle's claim that a system that operates on a syntactic level can not also operate on a semantic level and say "well, the human brain does that, why cant something else?", or if we concede that yes, the human brain only operates at a syntactic level, then the question is "If the human brain operates only on a syntactic level and we say people have minds, why can't something else that only operates syntactic level also have a mind?" In either case, Searle's argument does not provide a reason to believe that a computer can not have a mind.

No comments: