On Thu, Jan 17, 2013 at 2:54 PM, Gary Shannon <[log in to unmask]> wrote:

> [snip]
> I'm not a linguist either, but it seems to me that part of the
> definition of a sentence is that it COMPLETES a thought. Isn't it a
> requirement, then, that the sentence have an end? And aren't infinite
> sentences forbidden by the rules of grammar? Is it "a sentence" before
> it has a full stop? Once it has a full stop, can it be infinite?

Since when does a sentence need to communicate a thought?  Again, I have no
linguistics training, but isn't the famous sentence "colorless green ideas
sleep furiously" still technically a "sentence"?

I would really call this whole discussion a question of definition, and
linguists should (and I assume have?) define "sentence" for themselves.
Personally I would define it in terms of grammar.  Something like:

S = NP + VP
NP = (M +) N (+ embedding word + S)

Unless one added a specific rule banning infinite sentences, allowing
recursion allows infinite sentences if a sentence is defined in terms of a
recursive grammar.


Speaking as an engineer, not as a mathematician, I would propose that
> a practical upper limit on sentence length should be a sentence than
> can be read and understood in a single 12-hour sitting. And don't
> forget that a sentence that takes 6 hours to read will probably take
> AT LEAST six MORE hours to disentangle and understand. At least!

Why not just require that they be finite?  It's certainly more elegant, and
doesn't require dealing with "but it should be 11 hours!  12 is much too