[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >
Mike S., On 01/10/2012 18:43:
On Mon, Oct 1, 2012 at 1:15 PM, And Rosta <and.rosta@hidden.email <mailto:and.rosta@hidden.email>> wrote:
Mike S., On 01/10/2012 17:32:
> The official grammar doesn't bar "sa bbbe ccci", nor does it bar "sa
> bbba la ccca ra ddda ggga", so any rule that (arguably reasonably)
> excludes these formulas would not be a mere reformulation of Xorban,
> but rather a novel formulation representing a subset of Xorban as we
> have it.
If we equate the official grammar with the codified grammar, then it is clearly incomplete, because it says nothing about binding relations. Once it does say something about binding relations, the core grammar should bar "sa bbbe ccci". As I recall, we agreed there should be some sort of appendix that specifies a way of rendering "sa bbbe ccci" grammatical, but we have no agreement about what rules that appendix should consist of, and sorting out this appendix is not a priority.
In "sa bbbe ccci", we have two distinct problems: unbound "e" and "i", and "a" bound uselessly. The former problem we have discussed at some length with solutions:
1) Free variables (generally) are ungrammatical
This is core Xorban without appendices.
2a) Explicitly free variables are implicitly bound by "lV smV" 2b) Explicitly free variables are implicitly bound by "lV RV" where is R is the most recently occurring restriction of "V".
There have been others. John WC has suggested that explicitly free variables are bound by an unspecified quantifier. I have suggested that formulas with unbound variables attach as extra complements to a quantifier that could bind them and/or as complement to a following word.
We all know who supports what on that.
I don't think we do, because John WC's suggestion appeared only today, and I have formulated only vague proposals, too vague to be considered properly yet.
Of the other problem, it's tempting to me to want to declare it ungrammatical, but if we were determined to try to make sense of it, it might be something like, sa bbbe ccci sa [le sme] [je pseka] bbbe [li smi] [je psika] ccci "There exists an A pertaining to bbb E such that A pertains to ccc I. ... which is not at all necessarily good, but may be the best thing we might make of a bad sentence.
It is ungrammatical in the core grammar. Some of us think it wasteful for any well-formed phonological form to not correspond to a complete or incomplete sentence, so the appendix could have some extra rules for sa bbe ccci. I think the rule I'd go for would be that the phrase is incomplete and sa must have a complement containing what sa binds. --And.