> [mailto:[log in to unmask]] On Behalf Of Jens

> --- [log in to unmask] wrote:
> > Semantically the
> > words may
> > have meanings that are wider, narrower, or shifted
> > one way or
> > another.  This is something that does need to be
> > addressed when
> > working out a semantic schema for a new language.  
> This is something that I think about a lot. I think
> Kjell has suggested that "if a difference is in the
> source languages, it should be in the IAL." I guess
> it's a good place for a start to discussion. And I
> find it easy to agree with a modified version of that:
> "if a difference is in ALL the source languages, it
> should be in the IAL."
> The problem of course is when a difference exists in
> some of the source languages but not in others. Then
> what do you do? My own method, as might be expected,
> is to (to a reasonable degree) take the simplist
> forms. So I would do just the opposite of what Kjell
> suggests, in a sense. "If a difference does not exist
> in at least one of the source languages, don't include
> it." So "hoof" and "fingernail" should be the same,
> "arm" and "hand" should be the same, and "maternal
> uncle" and "paternal uncle" should be the same. 

It's not just "arm" because you could be even more general and
have just a term for "limb", and maybe even just use "leg" for
"arm" which isn't that farfetched when you consider that most
other mammals have no "arms" though its possible that the front
"legs" could be distinguisted as such lexically.  "finger" and
"toe" can both be eliminated by just having "digit", and
compounding with "hand" or "foot" if more precision is needed.

These too are not really good examples.  I think colors may be a
better example of what I was getting at because they can show
where there are not always hard lines that divide meanings.
"orange" for example could be "yellow" or "red" in another
language depending on the shade, and the dividing point may or
may not be the same in each language, where yet another language
could have two or three distinct words for shades of "orange".
So how light does grey have to be before we call it "white", and
how dark before we start calling is "black"?

> Of course, I have to add the caveat "within reason",
> because suppose, hypothetically, that one language
> doesn't distinguish between aunts and uncles, another
> doesn't distinguish between uncles and fathers,
> another doesn't distinguish between fathers and
> brothers, then you might conceivably find yourself
> having a single word for everything in the world.
> Also, there are certain simplifications that seem to
> create difficulties. For example, there are, it is
> said, languages that only have numbers to three, and
> after that everything is "many". But from the
> standpoint of an industrial society, that would seem
> too limiting. 

I've done some looking around and found different systems for
kinship.  It's not just the distinction of maternal v. paternal,
but sometimes older and younger are an important difference
where in some cultures the degrees of separation play a part.