On 22 July 2020 at 17:43 Mike Bennett <mbennett@hypercube.co.uk> wrote:

So I think if you are going to try and identify language primitives at this kind of level, you are right into Chomsky territory, trying to find a universal set of abstractions from which to specialise for different language groups.

Here you will find the bones of philosophers who have gone before.

My impression is that Denny wants a different approach. Something like: standardise on a thousand-page phrasebook that should exist for each of the target languages. Crank up its practical expressive power by attempting broader kinds of encyclopedic content. Leverage previous work in a piecemeal way.

It's going to be complex, as a software pipeline.

Charles