Introduction
According the general consensus, natural language is too fluid and too complex to be processed by software into depth. Therefore, natural language is processed top-down – “as good as it gets” – lacking a deeper level of understanding, unable to touch the deeper meaning.
In order to fix the lack of understanding, semantics is added to information systems artificially, in the hope those systems will grasp the deeper meaning. Unfortunately, adding artificial semantics might not solve the problem, if the root of the problem lies deeper than that: in understanding intelligence itself.
A correct procedure to include Artificial Intelligence to information systems:
• Analysis: Intelligence must be defined in an unambiguous, fundamental (= natural1) and deterministic (=implementable) way, because without such a fundamental relationship, any attempt to add AI to information systems will surely fail;
• Design: Methods must be developed to implement that unambiguous, fundamental and deterministic definition of intelligence in software.
Hints for fundamental research:
• Grammar gives structure to natural language, in a similar way as Leonardo da Vinci gave structure to his Mona Lisa; (Click on the picture)
• Children have no problem in deriving the logic behind grammar, because it is hard-wired in our brains;
• Find that logic contained within grammar, and it will provide the foundation for AI through natural language.
1 Natural rather than artificial. The system itself might be artificial, but not its foundation.