[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >

[Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: [engelang] discourse-scoped existental quantifier


On Sat, Sep 8, 2012 at 8:36 PM, And Rosta <and.rosta@hidden.email> wrote:
 

As I wrote in the other thread:

> There's a slightly different quantifier that I'd been meaning to
> mention, but have kept on forgetting. It's an existential quantifer
> with scope over the whole discourse:
>
> "Once upon a time, there was a poor woodcutter. He lived in a hut
> with his dutiful daughter."
>
> In Livagian I treat it as a separate quantifier in its own right.
> (Livagian doesn't have Xorban's lV.)

How to do this in Xorban? Just "la [poor]a [woodcutter]a"? -- No, tthat's "The poor are woodcutters"? Maybe "sa [hereby-newly-introduced-into-discourse]a je [poor]a [woodcutter]a"?

--And.

I think we need something like this too.  To save on consonants, one idea I had was to apply some unary operator "NV" to "s-" i.e. "NV s-".  "NV s-" would announce that there will be an implicit binding of its variable in subsequent sentences, but the restriction would be stronger than those in implicit "l-" bindings (the ones you are on record as not liking).

By "stronger" I mean that whereas "Ba Ra Pa" yields the weak implicit binding "[la Ra]" in subsequent sentences with free "a", in contrast "NV sa Ra Pa" would yield the strong implicit binding "[sa je Ra Pa]".  For example, in:

sa xrma lu fu bjra vska'aku. [la xrma] xkra.
"I see some horse run.  It [horse] is black.

... the free variable "a" in the second sentence gets the weak implicit binding "la Ra" from original "sa Ra Pa". In contrast, in:

NV sa xrma lu fu bjra vska'aku. [sa je xrma lu fu bjra vska'aku] xkra.
"I see some horse run.  It [some horse that I see run] is black.

... the free variable "a" in the second sentence gets the strong implicit binding "sa je Ra Pa" from original "NV sa Ra Pa".

I am not sure whether it's always safe or not to use "s-" rather than "l-" on strong restrictions, but that seems preferable, as we are without question trying to bind to individuals here, not the myopic singularization.  I have already sketched out the rules for deriving restrictions on another thread and I think this would work and be a nice formalization.  In fact I would have the implicit restriction on "a" keep getting stronger throughout the discourse every time it appears in a predication: first [sa je Ra Pa], then [sa je Ra je Pa P1a], then [sa je Ra je Pa je P1a P2a], and so on.  The specific keeps getting more specific.

There is a known issue with strong implicit restrictions, namely that they sometimes cause unwelcome effects when "na" is involved.  For example:

na sa xrja vfla. na [la xjra] se sme nlveka
"It's not so that some pigs fly.  It's not so that there is something that is a wing of them [pig]"
"Pigs don't fly.  They don't have wings."

... gets clobbered with the stronger restriction:

na NV sa xrja vfla. *na [la je xjra vfla] se sme nlveka
"It's not so that some pigs fly.  *It's not so that there is something that is a wing of them [pigs that fly]"
"Pigs don't fly.  *They [pigs that fly] don't have wings."

However, I don't think this problem is insurmountable.  The solution is in noticing that you can't apply NV willy-nilly to "sa".  NV would be a +specific marker, and it's simply infelicitous to use it to mark inherently non-specific arguments as in "pigs don't fly".  I suspect that the exact rules could be spelled out without a great deal of trouble.