sumti categories

[Jorge Llambias <jorge@PHYAST.PITT.EDU>]

Let's see... who disgrees with this:

(1) We can divide all sumti in categories, such that every sumti
    belongs to only one of them.

This is not really saying much, since the categories can be
completely arbitrary, but I will use them for my second (much more
controversial) claim.

I'm not sure yet about all the categories, but it would be something
like this:

* dacti (objects)  e.g. le dacti; le plise; le solri; mi; la djan
* fasnu (events)   e.g. le fasnu; le nu klama; le rinka; le se rinka
* fatci (facts)    e.g. le fatci; le du'u mi klama; le se djuno
* namcu (numbers)  e.g. le namcu; li pano; le se jelca

(The names are only for guidance, fatci don't have to necessarily be
true, dacti don't have to necessarily be tangible, fasnu don't have
to necessarily happen, etc)

I don't know what to do with {le ka... } and {le ni...} things for
the moment. It seems like they should both go in the same category,
and they might or might not be fasnu, I haven't yet made up my mind.

Processes, states, achievements and activities are all fasnu. For the
remaining abstractors, I don't know.

There should also probably be a category for citations, but I'm not too
clear on that either, perhaps something like:

* selcusku         e.g. le se cusku; lu coi roda li'u; di'e

 The point of the categorization is to boldly claim that:

(2) Each slot of every selbri can be filled with sumti from only
    one of those categories.

This restriction is not grammatical, because the grammar allows
any sumti to fill any slot, but rather, it's a semantic restriction.

That the division exists to some extent, I think is indisputable. The
sin of sumti raising is usually commited when a dacti is used in a
place reserved for a fasnu, that's why {tu'a} was introduced: to
convert from dacti to fasnu (maybe to fatci as well).
{la'e} also converts from one category to another, usually from
selcusku to dacti, or to fasnu, or probably even to fatci.

Many of the definitions in the gismu list already mark these classes.
For example, fatci places are marked with (du'u). There is only one
place marked with (li), the x1 of namcu, but there are other places
that accept only numbers, like the x2 of most units gismu.

The above two categories are quite distinct, but the dacti/fasnu categories
are not always kept apart. In many cases, they are explicitly allowed
to merge. For example:

vajni    x1 (object/event) is important/significant to x2 (person/event)
         in aspect/for reason x3 (nu/ka)
         [also: x1 matters to x2 in aspect/respect x3]

"Important to person", and "important to event" seem two different concepts,
and I suppose that in most cases where the two types are allowed, two
different meanings are being used in the same gismu. There is little
chance of confusing one with the other, because the type of sumti used
gives the clue as to which meaning is intended, though. But this is also
true in the cases of sumti raising, so if we allow them sometimes I don't
see why we shouldn't allow them always.

Anyway, I still have to study some more the classification, but is it true
or not that sumti places in general are restricted to only one category?

Jorge