Re: Never violate a universal unless it seems like a good idea at the time

Quoting Roger Mills <romilly@...>:

> Andreas Johansson/Greenberg wrote:
> > I note with a certain dismay that #40, which as I said appears to be the
> only
> > one that Tairezazh defies, appears to belong to the "absolute" category:
> >
> > Universal 40. When the adjective follows the noun, the adjective expresses
> all
> > the inflectional categories of the noun. In such cases the noun may lack
> overt
> > expression of one or all of these categories.
> >
> I need to go back and re-read that one in context, if there is any--- does
> the 2nd sentence mean (1) even if the noun lacks inflectional categories,
> the adjective must have them; or (2) if the noun lacks inflectional
> categories, then the adjective will also lack them, by definition.
No context. I read it as saying that any inflectional categories expressed on
the noun must also be expressed on the following adjective, but that
inflectional categories expressed on the adjective need not necessarily be
expressed on the noun.

Eg, if we had a language with _bomp_ "apple", _chek_ "red" and _-a_ "PLURAL",
this universal would allow _bomp chek_, _bompa cheka_ and _bomp cheka_
for "red apples", but forbid _bompa chek_.

> Kash seems to violate this-- nouns are marked for case/number, adjectives
> are not, e.g. standard _puna-ç-i virik_ house-pl-gen. pretty 'of (the)
> pretty houses'  (although in archaizing/formal/poetic work you can say _puna
> virik-(i)ç-i_, translating the pl and gen to the adjective).  But total
> concord is flat out wrong: *punaçi virikiçi.
Tairezazh is similarish - the only way to say that is _daivei dair_ "of a
beautiful house", where _-i_ is the genitival marker.

> If (1) is the case, Indonesian languages violate it, though if (2) is the
> case, then they squeak in under the wire-- IN nouns have no case markings,
> and neither do adjectives. OTOH if you _do_ pluralize the noun, you don't
> pluralize the adj.-- rumah-rumah besar 'big houses'.
If reduplication counts as inflection, I think this is a clear violation.

> Perhaps #40 is not as absolute as claimed.  Incidentally, there is a huge
> website (German, IIRC) with 100s of proposed "universals", some of them,
> suspiciously, based on evidence from just a handful of languages.
Yeah, I've seen that one. I can't tell how reliable those universals are, but
have you checked out the Raritätenkabinett? It's a collection of rare
typological traits in various languages. Pretty cool.

                                                    Andreas