Have you been in an argument that degenerated into simply arguing semantics? No? How about “Canadians want [more healthcare|social programs|less taxes|more services|more freedom|more security]”? Most debaters share a common goals but see different ways of achieving those goals (“A better Canada”) But really, what does a “Better” anything mean?
The semantic web, and before it, ‘classic’ AI researchers from the 60’s onward struggled with semantics – the ambiguity of language and meaning that is so easily resolved by us but so difficult to pin down when making natural language interfaces and more generally, organizing information in a way that bridges the gap between humans and machines. This is to say nothing of language barriers and differences between individuals.
"The semantic web, and before it, ‘classic’ AI researchers from the 60’s onward struggled with semantics – the ambiguity of language and meaning that is so easily resolved by us but so difficult to pin down when making natural language interfaces."
An article in Science online is outlined at Ars Technica – research that has taken advantage of functional MRI and it’s ability to monitor real-time brain activity to find common brain patterns associated with word meanings. The implied goal would be to find a common brain pattern ‘print’ for every meaning, which would then lead to the ability to categorize the semantics based not on ontologies (which themselves are debated and subject to semantics) but on an ‘absolute’.
This may be doomed from the outset if the same word creates different patterns in the tangle of sentence context, or if there are variation between individuals, or languages or for that matter, moods and distraction. Research is just beginning, but this is a unique and brave attempt to pin down an aspect of human interaction and analyze it in a truly new way.