G''day. For the past few months I have been developing concepts and conclusions for a synthetic human project. The device will be written in C/C++ using some OpenGL graphics just so it looks nice. Anyway, I have some basics written down, and am about to incorporate some thoughts: When working with AI, it''s important to realize that you are creating a clone of the human thinking paths (in a way). Therefore, I have done this:

  
typedef struct {
char *subject;
char *averb;
char *lverb;
char *object;
char *compliment;
bool question;
} sentence;

  

This of course is very basic, but it gets to the point. This, I believe is how we think when we are about to say a sentence. For example, I am trying to get my point across to you, so I am thinking of ways. Showing examples is one, and I could go on from there. Anyway, the sentence structure above is all I can think of at the moment. Eventually all of the memory of Estill will be in a tree-form, much like ours. We see an apple, and we immediately declare the following: apple.color = red; apple.size = fairly_small; apple.location = x,y,z; apple.contents... and so on. I was thinking of a way to implement this... Okay, off the subject again... Can anyone think of any other way to implement the sentence structure and input/output of the human brain? I have done so minutely using a sample of the above code, but it''s very small.. ~Dwarf Thanks Why am I the only one on this away team wearing a red shirt?

Why don''t you just have a list of words that compose a sentance. For example:

Why Don''t you have list words compose sentance

I don''t know, just thinking aload. Possibly have properties of the words like verb, noun, etc. to help the comp.

BTW, what are you trying to do here?

sup,

have you read The Language Instinct by Steve Pinker -
http://www.amazon.com/exec/obidos/ASIN/0060958332/qid=1010181356/sr=2-1/ref=sr_2_11_1/104-3891240-1123135

that''s a really good summary of nearly contemporary linguistic theory.

then a good thing to do would be to chase through his bibliography and read as much as you can of the journal papers by the people he cites. (apologies if you''re already deep into that).

i''d be wary about throwing out terms like "clone" and "we see an apple and we immediately declare the following". you can''t call it a clone of human processes if you''re not using a neural network to do your storage and processing. and you can say this is how the brain works, cause well the brightest minds in linguistic theory are still debating over the fundimental structure of even the simple lexicon of human words let alone the pathways (both neural and conceptual) of the processing of those words into meaning.

anwyay, that aside, i''m always down with ambition. you can''t really achieve anything in this world if you limit yourself to what other people think of as possible. I''m working on a similarly complex project with AI as well

-me

Natural language processing is a huge topic, but it really only covers one aspect of the brain. Pinker''s book mentioned above is great (in fact all of his books are pretty good), and there''s plenty of other literature about. Some stuff by Jerry Fodor and Chomsky is probably relevant but it gets pretty technical from all accounts.

I don''t know whether the brain gets much of its problem-solving power from language, but good luck anyway.

Just to let you know: for your programs knowledge you should implement a semantic net. I''m not giving free code, so look them up on google nad you''ll should find something. OK, if you really need help, post again and i''ll give you the structs.

Also will you be simulating heart rate, breathing, etc. so tht they could go up when adrenaline is released and you could have a simulated endocrine system to control other emotions and... I''m getting ahead of myself...