On Sunday, November 9, 2014 9:42:50 AM UTC-3, And Rosta wrote:
1. What does ro mean, and does it have EI? (A question settled a dozen years ago.)
2. Should there be a non-EI universal quantifier?
3. Should there be an EI universal quantifier? This is the question John seems to be addressing.
4. In any bpfk revision of the CLL specification, which meaning should be paired with the phonological form /ro/?
I believe that #1 and #4 are the question I'm trying to ask -- which hopefully have the same answer -- and I'm sorry if I invited the detour into other questions.
CLL 16.8 says:
sumti of the type “ro da poi klama” requires that there are things which “klama”
It's not entirely explicit in the section what the consequences when the requirements of the sumti are not met. I have assumed that according to this requirement, {ro da poi klama cu pavyseljirna} is then considered to be false in the case of {no da klama}. (If it has another truth value which is neither true nor false, then I'm barking up the wrong tree!)
However, if such sentences are considered to be false, then the definition of {ro} is incompatible with the description of how negation boundaries work. Supposing, for old time's sake, a universe without unicorns:
{su'o pavyseljirna na ku cu blabi} => There is at least one unicorn, such that it is not white. => FALSE.
Now we move the boundary, and "invert" the quantifier, while preserving the truth value of the statement:
{na ku ro pavyseljirna cu blabi} => It is not true that all unicorns are white. => FALSE.
But then we negate that and get:
{ro pavyseljirna cu blabi} => All unicorns are white. => TRUE.
The {ro} derived from these transformation is not one that "impl[ies] the corresponding existential claims". Either that definition of {ro} is invalid, or the derivation of {ro} from {su'o} by moving negation boundaries is. It's not just faulty examples at stake. The rules don't work together.
When I started this thread, I was under the impression that
the BPFK section on "Inexact Numbers" took no clear position on this problem. I now see that there is indeed a commitment to preserving the negation boundaries formula, although it is buried in the formal definitions and obscured by bad formatting. The part I'm looking at is this:
ro da = da'ano da = no da naku = naku su'o da naku
If this definition holds, then {ro pavyseljirna cu blabi} has the same truth value as {na ku su'o pavyseljirna na ku cu blabi}. If {ro} were held to import, then both sentences would be false. Wouldn't that commit us to hold the negation of the second sentence to be true?
{su'o pavyseljirna na ku cu blabi} => "There is at least one unicorn, such that it is not white."
If I have made a mistake in my reasoning, please point it out. Otherwise, I will assume that BPFK has settled questions #1 and #4 per the equivalence for {ro da} == {naku su'o da naku} on the "Inexact Numbers" page.
mi'e la mukti mu'o