On Sat, Jan 20, 2024 at 05:09:58PM -0000, Curt wrote: > On 2024-01-19, David Wright <deb...@lionunicorn.co.uk> wrote: > > On Fri 19 Jan 2024 at 17:25:10 (+0000), debian-u...@howorth.org.uk wrote: > >> Greg Wooledge <g...@wooledge.org> wrote: > >> > >> > I won, and you lost > >> > >> There shouldn't be a comma in that sentence, in English. There is in > >> the closely related expression "I won, you lost." > > > > That's rather proscriptive. "I won and you lost." and > > "I won, and you lost." are two different sentences. > > > > AFAIK, "you lost" is an independent clause and should be separated from > the independent clause that precedes it with a comma before the > coordinating conjunction.
Regardless of which grammar rules are right, wrong, or optional, the point of this is that parsing natural language text is *stupidly difficult*. A person who has to ask why "grep -c" doesn't count the number of commas in a single line of text probably isn't able to take on this quest. Any serious inquiries about natural language parsing should be directed to an appropriate artificial intelligence mailing list instead of this one.
