Re: Universal grammar

In article <1161104368.641935.149100@xxxxxxxxxxxxxxxxxxxxxxxxxxx>, "Peter
T. Daniels" <grammatim@xxxxxxxxxxx> wrote:

Thank you for the references. It does not as such make any difference if
it is the original book. But the story above goes back to some of the
years 1972-75, so unless there is some earlier edition of the book you
mention, it is not the one.

The study of linguistic universals was initiated by Greenberg at the
conference published as *Universals of Language* in 1963 by MIT Press,
but I can't think of anything in that volume with the sort of
specificity you describe. It really didn't take off until he got the
funding at Stanford more than a decade later. Maybe you saw a preprint
or preliminary version of something from the Stanford Universals
Project, and that's definitely where you should start your
investigation.

I have made a note of your comments. There are really two ends: One is
curiosity of what this stuff was back then. I was in high school, and
there was some physics days arranged in Sweden for talented students, one
from each school, and the guy I shared hotel room with was interested in
this stuff and had a book.

The other end, though, is that I am into computer language construction,
mainly using Bison/Flex, which are Yacc/Lex compatible programs. There is
one constructed human language with a Yaccable grammar, Lojban
<http://www.lojban.org/>. But other human languages are quite
nondeterministic and ambiguous. Parsers, like Bison, now slowly get the
capacity of handling ambiguous grammars. Bison does that by what is called
GLR, "Generalized LR", which splits the parser when an ambiguity occurs,
and merges it again when it is resolved. This is faster than parsing all
possibilities. Parsing a human language is even more complicated, because
the lexer needs to be split as well, which is currently not within scope,
even though we have discussed the possibility.

Anyway, so I thought it might be good to have some references handy in the
case of need.

--
Hans Aberg
.