Hans Peter Willems - Mar 8, 2011:

As for NLP, I still fail to see the importance of that towards AI consciousness or even inteligence; we know for a fact that people who do not possess language traits (for various reasons) are still conscious for every definition of the concept. To me language are patterns, assigned representations to concepts. Nothing more, just that. Humanity hasn’t even settled on one uniform vocabulary and still we can translate from one language to another without loosing (much of) the meaning of the message. However when we translate into some computer representation we suddenly insist that it is no longer understanding and when we translate back to natural language, the message somehow magically regains it’s properties to feed ‘understanding’.

To me, natural language is our best model of reality. It is so powerful that people can list its faults not necessarily realizing they’re using natural language to do so. Natural language’s very ambiguity is its strength, providing it the flexibility to adapt to changing conditions (such as the words “web”, “hit”, “mouse”, “code”, etc. taking on new meanings as technology advanced). Natural language admits contradictions, paradoxes, etc. without going into infinite loops (“this statement is a lie”). Puns and double meanings create humor and creative reinterpretations (“time flies like an arrow; fruit flies like a banana”). Natural language is amazingly fault tolerant, allowing us to understand misspellings (“inteligence”, “loosing”) and syntax errors (“it’s” instead of “its”) where a formal computer language would stop cold.

In figuring out how to make a computer deal with natural language, I think we develop the tools to address other problems. For example, if computer languages allowed spaces within function names, it would force us to think in terms of how to recognize linguistic delimiters rather than explicit symbolic ones, and hence to get our programs to abstract out the concept of a group and force us to backtrack and reinterpret previous parses when the first guess at where the group ends doesn’t make sense as we continue to parse (“the horse raced past the barn fell”, “the government plans to raise taxes were defeated”).

—-

From the first post in this thread:

The brain operating system is not constantly being rewritten, just like a computer OS is fairly static. Of course there is an ‘upgrade’ sometimes but that can be seen as ‘evolution’ (and we do actually see software versions as the evolution of a computer program). Our ‘brain OS’ mainly implements the ‘plumbing’ (similar to a computer OS), all learning, experience and therefor consciousness is held in the data-model.

The point is that the evolution of a computer program is directed, “intelligently designed” if you will, but biological evolution (monoliths notwithstanding) is not the result of any one human or group of humans working for Microsoft or Apple. (Yet; it may be that we will use genetic engineering soon to “evolve” ourselves - but that would be like a program rewriting itself…) In my opinion, the software needs to evolve independently of us. So, for me, code is part of the “data-model”.

Jan Bogaerts - Mar 9, 2011:

In software design, this concept is generally described as grammar definitions.

Indeed (I had my share of that and wrote several parsers in my lifetime). Same as different programming languages are mapping different words to similar concepts (e.g. WHILE, FOR-NEXT, REPEAT UNTIL are all implementing countable loops), real world languages are doing the same.

Now, this link you make to programming languages can actually illustrate my point about words being mere symbolic representations of a concept: take a look at BrainFuck and you will see a clear implementation of this idea of symbolic patterns:

http://en.wikipedia.org/wiki/Brainfuck

Note: pay special attention to the ‘hello world’ example

Robert, thanks for the elaborate reply.

Robert Mitchell - Mar 9, 2011:

To me, natural language is our best model of reality.

I agree, but natural language is not, in itself, reality. It is a representation of reality, and indeed very capable of describing it. Nevertheless, this does imply that it ‘should’ be possible to describe reality in other ways, as long as there is some construct to keep things coherent.

Dave Morton - Mar 9, 2011:

In order to prevent any possible unpleasantness, I’d like to point out that, no matter what we feel the probability of Han Peter’s assertions are, it is still possible for him to be right.

I think the 2 fundamental approaches to intelligent agents are bottom up, or top down.  One approach involves building many specific modules and then building on top of them like a pyramid so the system can pick and chose, or perhaps figure out how to chain these together.

Then there is the top down approach, or starting with a kind of general purpose solution.

The former is sort of the path I’ve chosen.  Hans has chosen the top down approach.  Chatbots are such a new technology and the science behind them is in its infancy, thus from this point in time I think both options are equally valid paths to pursue.

Erwin—I really like the “auto hyperlink maker” functionality

For those interested, some more food for thought:

http://rjwagner49.com/Robotics/Philosophy/philo.html

http://consc.net/papers/qualia.html

Robert Mitchell - Mar 9, 2011:

Natural language’s very ambiguity is its strength, providing it the flexibility to adapt to changing conditions (such as the words “web”, “hit”, “mouse”, “code”, etc. taking on new meanings as technology advanced).

Well said.  NL’s ambiguity exists because that seems to be all there is in the real world.

http://www.timothyjpmason.com/WebPages/LangTeach/CounterChomsky.htm