On Mon, Aug 11, 2008 at 9:05 AM, John Vertical <[log in to unmask]> wrote:
>>Is there any linguistic theory dealing with language typology on a semantic
>>level, corresponding to the morphosyntactic typologies based on primary

> I've seen some that concern fairly closed word classes, especially colors: the
> most basic division is "white" vs. "black" (some binary systems I've seen
> described as "bright" vs "dull"), after that comes the concept of "red", after
> that "green" and "yellow", then "blue", then various other colors. As usual,
> there are exceptions - IIRC there was a language that failed to
> distinguish "blue", but did have words for "purple" and "brown".

Now that you mention it I remember heard about some such studies
a while ago.   gzb might already violate said conditional universal;
its root terms are for white, red, blue and green (the light-primaries
rather than the pigment-primaries).  Black and secondary colors
are derived with opposite-derivation from the basic terms, and
other terms can be derived from any word for a physical thing
with the suffix -kwa "color of"; potato > light brown and chocolate
> dark brown are the most common uses of that suffix, I think.
But for very light-brown/yellow I tend to use the basic term for
"red" rather than the "red + green" compound.

> Compass directions, maybe? I would assume any language distinguishing roots
> for any of the intercardinals should also distinguish some for the cardinals.

That's a sensible hypothesis.  Do you know of any languages that
have roots for the intercardinals rather than deriving them from
cardinals?

> I'm rather skeptical about the existence of semantic universals linking non-
> related concepts.

So would I be.  I was thinking along the lines of the color theory
you mentioned, like "if a lang distinguishes brown from orange
it will also distinguish blue from green" or, re kinship maybe,
"if a lang has separate roots for male and female cousins it will also
have such for male and female siblings".

>>If there are such hypothesized universals I want to violate some of them
>>in my next engelang and see what happens.

> Basic color terminology of "polka", "tweed", "chrome", "iridescent"
> and "matte"? :)

Good for a Dadalang!  And its basic kinship terms would be
"male first cousin", "female first cousin once removed",
"great aunt on the father's side" and "father-in-law"; complex
affixation would derive "father", "mother", "son", "daughter" etc.
from those.

The basic directional roots would all be 22.5 degrees clockwise
off from the cardinals.

-- 
Jim Henry
http://www.pobox.com/~jimhenry/conlang/fluency-survey.html
Conlang fluency survey -- there's still time to participate before
I analyze the results and write the article