[net.nlang] the hardest natural language

budd@arizona.UUCP (06/13/83)

With regards to Dan Berry's article on variations on "it's greek to me",
there was a technical report a few years back by Arnie Rosenberg at
IBM Yorktown Heights on "The Hardest Natural Language"  (a take-off, of
course, on the popular theme of that time that saw such classics as
"the hardest regular language", "the hardest context free language", "the
hardest language accepted by a push down automata with three pebbles", etc.)

Anyway, the idea was that language X was said to be "harder" than language Y
if in language Y the phrase for something unintelligible was "It's X to me".
He gave many examples of these, and then took the TRANSITIVE CLOSURE in order
to come up with the "hardest" language.  Interesting facts were

1. almost every language had such a phrase, and

2. in taking the transitive closure there were no cycles.

Unfortunately, i believe when i moved a few years back i discarded my copy
of the paper.  If i remember right, however, Chinese was at the top, since
the appropriate phrase in Chinese translated roughly to "it's divine speech
to me", and he was unable to discover if THIS language had a next higher
counterpart.

Does anybody have a copy of this paper to verify my memory?
--tim budd  ( kpno!arizona!budd )