RE: Language?

>In a Chomskian sense there is no "foreign grammar" - the very notion of
>there even being a "grammar" demands that its being "anything like
>ours" is
>not *assumption* but *fact* - what we generally agree on as "axiom". Also
>"never" is such a long time that there is the possibility that if you wait
>long enough even the Pico bus will come by and they'll put a tunnel under
>the Channel. Not only "may" we be talking about the same world, it is the
>only one we've got and symbols, although they *could* refer to *anything*,
>in fact *don't* - they refer to something.
>
>>One might have some success with very limited domains
>
>In the sense herein *all* domains have limits, else they wouldn't be
>domains, but universes.
>
>This is all very fun, but what we're about is (or should be, IMO) dancing
>the dance we sing of in these rather flighty songs.

William once more makes the key point in a picturesque fashion. There's
dancing to be done. One last little bit of fun, just to make the logicians
wince :

You've got two agents, with the same kind of interpretation mechanism
looking at the same world, but the agents speak different languages. Is it
possible for one agent to communicate a reference to an object in the world
to the other agent?
 Start by adding a couple of assumptions - there's a one-to-one
correspondence between objects and names, and each agent has a word for each
object, and there's nothing else in this kind of world but these objects.
Ok, a world with one object in it - can one agent refer to it in a way that
the other agent can understand - easy - just say the local name and as
there's only one object in the world, the other knows that must be the thing
that is being referred to. So the two agents can find each other's name for
object 1. Ok, now let's have n objects in the world, which both agents know
the other's name for - can they communicate the name of object n+1? Yep -
just by saying its name, which will be a name that isn't used of any of the
other n objects.
<sound of running feet...>

Received on Saturday, 12 May 2001 04:14:15 UTC