Re: grammar

[math.ucla.edu!jimc]

> Date:  Fri, 15 Dec 89 22:36:52 EST
> To:  "est@cs.nyu.edu"@mintaka.lcs.mit.edu
> From:  "Keith F. Lynch" <KFL@ai.ai.mit.edu>
> Subject:  Re: grammar
   
> > The actor of gerna is that which "grammars".  lojban takes this to be
> > the grammar itself.

> Once I understand why that is, I will be well on my way to fully
> understanding the language.  Currently, it's not clear to me why it
> is the grammar which "grammars", rather than a person or a language,
> or how to tell just what the X1 place is for all the other gismu.

I believe that this is still current Lojban doctrine (and note that when
I assert something I don't mean to deny that other points of view are
possible; I'm just trying to keep it short.)  A word is a symbol for a
relation, and the relation consists of a list of sets of thus-related
objects.  My favorite example in -gua!spi is "eat" as in "the rat eats
the cheese".  A pair (ordered) containing the example rat and cheese is
on the list for "eat", as are many other pairs of (same or other) rats
and cheeses, etc. etc.  The first, second... slots in these member sets
are the cases: X1 for the eater and X2 for the food.  

> I know I'm not supposed to categorize gismu into nouns, verbs, etc,
> but I find that in the ones that seem to be nouns, the X1 place is
> always (?) <X1> is a <gismu>.  With adjectives, <X1> is <gismu> seems
> to hold.  And with verbs, <X1> does <gismu>, i.e. <X1> is a <gismu>er.

In Old Loglan, JCB realized that the language would be much easier to 
learn if the cases were consistent and so he tried hard to get the
actor in X1 if there was one, etc.  Review the verbs of motion; the
cases are almost identical (and when creatively viewed, quite a number
of quasi- motions can be equally well interpreted as verbs of motion). 
However, none of Loglan, Lojban or -gua!spi have these case assignments
hardwired into the language.  Hence you may feel, correctly, that the
speaker is an "actor" in the relation of "gerna/grammar", but somebody
decided (wisely, I think) that the word was more useful with the
relation-system (the grammar itself) in the X1 case, following the
pattern of the noun-oids.  It's valuable to remember the general rules
you note, but also remember that they are only a guide, and exceptions
are made freely and without apology.  

The doctrine of words as symbols for relations is well known to
philosophers; perhaps someone could post a good readable reference.

		-- jimc