Saturday, March 13, 2010

A Radio Show: Contextual Semantic Inference (or Something of the Sort)

This morning I heard a fragment of a conversation on the radio. A telephone caller was saying, "...and the guy said that for $700.." when the radio host broke in and said "he would fix the ducts" and the caller replied, "Exactly."

Interesting.

The host drew an inference.

Immediately.

And the host was correct.

The host drew the correct inference partly because he knew something about about the subject matter -- about some car, auto mechanics, whatever.

And the host's interjected conclusion was inference.

What shall we call this kind of inference without using ungainly language?

Contextual inference?

That's not pretty.

Domain-dependent inference?

More descriptive -- but a bit pretentious.

It was inference importantly based on knowledge of how some things work.

There's a lesson there for legal academics who theorize about inference.
The lesson has been known to AI people for quite a long time. Some of them like to talk about "ontologies."
They would make Aristotle proud. Sort of.
A theory of being perhaps should reach beyond and below things such as tables.
Still, the AI people are onto something. And they are wonderfully unabashed: they don't worry much about whether it's possible to have a theory of being.

&&&

The dynamic evidence page

It's here: the law of evidence on Spindle Law. See also this post and this post.

Post a Comment