Re: [lojban] Beyond Whorf: "things," "qualities," and the origin of nouns and adjectives

[Ivan A Derzhanski <iad@MATH.BAS.BG>]

Jorge Llambias wrote:
> la tipitr park cusku di'e
> >               All natural human languages that I know of [...]
> >have (1) lots and lots of words of the type "stick, stone, [...]
> >star", (2) lots and lots of words of the type "big, small, long,
> >[...] dry," and (3) and lots and lots of words of the type "eat,
> >drink, bite, [...] hear." This is true, as an empirical fact, of
> >English, German, [...] and Lojban.
> 
> It is true of Lojban, although if we take place structures
> seriously in Lojban classes (1) and (2) are much smaller
> than in other languages, while most words fall in class (3).
> 
> Only one-place predicates can be of classes (1) and (2). All
> others must be of class (3), because they don't refer to a
> property or bundle of properties but always to relationships.

But so do many of the natlangs' nouns and adjectives.  What about
such relationships as `father', `friend', `part/piece'?  They are
nouns (under whatever definition might be applicable) in every lg
in the world, but they denote two-place relations.

Why is _fire_ a noun and _burn_ a verb, btw?  Don't they denote
the same part of reality (a process, as it happens)?

> Thus the Lojban word {botpi} is not class (2) like the English
> word "bottle". It does not refer to the bundle of properties
> that make up a bottle, it refers to the relationship that
> exists between bundles of bottle properties and bundles of
> bottle contents properties.

Which is just why it provoked so much debate here.  We want to name
entities -- entities that are more than bundles of properties -- by
stating (the) categories to which they belong.  Thus a bottle is not
a bottle because it actually or potentially bottles something; it is
a bottle because it is a vessel with a narrow neck.

--Ivan