[CL] Quantified sentences
phayes at ihmc.us
Wed Dec 2 13:41:40 CST 2009
On Dec 2, 2009, at 8:56 AM, Randall R Schulz wrote:
> On Wednesday December 2 2009, Cameron Ross wrote:
>> Yes, according to the CLIF grammar it is allowed, but does it make
>> sense for this to be allowed?
> It's pointless, but that doesn't mean a special case should be made of
> it. Another legal but pointless use of a quantifier is to bind a name
> and then make no reference to it in the matrix.
> As far as why not disallow these things, consider the situation in
> you're generating CLIF programmatically rather than writing it by
> forbidding an empty binding list is just another special case to
> accommodate in your code generator.
Quite. Also, there was a general methodological principle in the CL
design, which was to always permit rather than forbid, unless there
was a good reason to forbid. So 'meaningless' but harmless cases, like
these, were deliberately NOT excluded, on principle. This keeps the
language simpler and removes a special-case burden in syntax checking,
etc.. Also, we have found a number of cases where apparently 'silly'
cases in fact have quite valid uses. The (and) and (or) were two, but
also for example we decided not to prohibit numerals and quoted
strings in relation/function positions, even though that seemed
insane. However, it turns out to be very useful, in the string case,
as a natural way to encode RDF datatyped literals. We did not forsee
this and would never had noted it if the construction had been
arbitrarily ruled illegal just because nobody could think of what it
could be used for at the time.
> Randall Schulz
> CL mailing list
> CL at philebus.tamu.edu
IHMC (850)434 8903 or (650)494 3973
40 South Alcaniz St. (850)202 4416 office
Pensacola (850)202 4440 fax
FL 32502 (850)291 0667 mobile
More information about the CL