[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
xod: > On Mon, 28 Oct 2002, And Rosta wrote: > > > > If you take that a step further, you'll see the logical error > > > > > > If ni uses ce'u, then it can't express "the degree to which", because > > > that's an abstraction of a filled bridi. If ni doesn't need ce'u, then it > > > makes sense, but loses its symmetry with ka, and becomes completely > > > identical to jei > > > > Okay. As I see it, ni doesn't have ce'u, it has no symmetry with ka, > > and the reason I had asked you to expand your ideas is that I'm > > interested to see how feasible it is to maintain that it becomes > > identical to jei > > It seems we're down to two uses of ni: ni + ce'u, used for counting the > valid sumti in a tergi'u, and ni without any ce'u, which is like jei, but > not restricted to [0, 1] Yes. I am in favour of the latter. > > > ni ko'a xunre: the degree to which A is red > > > ni ce'u xunre: the degree to which anything is red <-- makes no sense > > > > > > If ni and jei are redundant, ni should be the one to go: it is roundly > > > abused, it seems to expect a ce'u but shouldn't have one, and its values > > > are not defined to be in [0, 1] like jei is, and it doesn't have the usage > > > history of working like jei > > > > I have always used and understood ni in a way parallel to jei. Admittedly > > I am more than averagely free from semantic solecisms (my errors are > > largely syntactic) > > > > But the difference between ni & jei is analogous to (tho not synonymous > > with) that between "the extent to which" and "whether". That's a > > useful distinction in English, though I can see how it might also be > > useful to be able to conflate them > > Yes, indeed. The conflation is the essential insight of fuzzy logic Yes, but not all conflations give what I associate with "fuzzy logic", viz the idea that the boundary between True & False is gradient. --And.