* Tuesday, 2011-11-08 at 09:09 -0300 - Jorge Llambías <jjllambias@gmail.com>:
> On Mon, Nov 7, 2011 at 8:13 PM, Martin Bays <mbays@sdf.org> wrote:
> >
> > Let's call this "naive kind theory".
> ...
> > That's what I meant by "everything which can broda" - everything which
> > you might refer to in any context by {lo broda}. So this domain is what
> > you get if you just glom together willy-nilly all the various domains
> > you find yourself working in.
>
> So you've shown that you cannot glom together willy-nilly all the
> various domains and end up with a reasonable domain. That sounds very
> reasonable.
>
> But this problem of domain glomming is not specific to kinds.
>
> A: ie pei no da nenri lo vi tanxe
> B: ie go'i
> A: ie nai pei su'o vacri cu nenri lo vi tanxe
> B: .y .y ie go'i
> A: ua .i do jinvi ge lo du'u no da nenri lo vi tanxe gi lo du'u su'o
> da nenri lo vi tanxe .i do jinvi lo nafsi'u
That isn't really the same phenomenon, though. If you want to look at it
in terms of the domains changing, then the point is that the second
domain is just a more accurate refinement of the first one. In
particular, if we collect up everything mentioned into one domain, we
just get the second domain, which is an entirely reasonable one.
That isn't what's happening when we pass from a domain which contains
hats but no kinds of hats to a domain which contains kinds of hats but
no hats. There, we're changing which aspect of the world we're looking
at. If we try to collect up everything mentioned into one domain, we get
a monstrous domain which must not be permitted.
At least from the naive point of view, this is unnatural.
Martin
Attachment:
pgpd1WMPYnjCg.pgp
Description: PGP signature