Sent from my iPad
On Oct 15, 2012, at 5:31 PM, Jorge Llambías <jjllambias@hidden.email> wrote:
On Sun, Oct 14, 2012 at 11:37 PM, John E Clifford <kali9putra@hidden.email> wrote:
> ________________________________
> From: Jorge Llambías <jjllambias@hidden.email>
>
>
> Now let's define the unary version of a quantifier such that it gives
>
> Where is the quantifier here? Is ni'u one?
This is terrible terminology even by Logjam standards. I see the point but hope there is a nonconfusing way of making it. Proportional predicate?
No, bbbake is a binary quantifier predicate (as defined in the first
paragraph) and bbba would be the unary version of that predicate.
> the same meaning as the binary version when applied to an _expression_
> with "ni'u" standing for the "ni/nu" of the x2 and x1 of the binary
> version. So for example, if "mstake" means "most A are E", we can
> have:
>
> la prna le li ckli nlceki mstake
> A/person(A): E/( I/chocolate(I):like(E,I) ): most(A,E)
> Most people like chocolate.
>
> = la je prna nu li ckli nlcaki le je prne ni li ckli nlceki mstake
> = la je prna ni'u li ckli nlcaki msta
> Peeople who DO like chocolate are most (people).
>
> So instead of "la prna le li ckli nlceki mstake" we can say "la je
> prna ni'u li ckli nlcaki msta", which uses one fewer variable, but
> also we can say "la je frmra je se xsle pnsake ni'u drxake msta",
> "farmers who own some donkey and DO beat it are most (of the farmers
> who own some donkey whether they beat it or not)", which doesn't have
> a "mstake" form without repetition.
>
> So this is short for la je frma je se xsle pnsake ni'u drkake li je frmi so xslo pnsiko ni drxiko mstaki, which is not of the form la ccca ni'u ddda le ccce ni ddde fffake because the break is across the quantifier boundary.
You can use the equivalent "la je je frmra se xsle pnsake nu drxake li
je je frmri so xslo pnsiko ni drxiko mstaki" if you prefer.
The form is "la je ccca nu ddda le je ccce ni ddde fffake"
OK, so we work with the weakest claim about farmers with donkeys. That helps, as does shifting the conjunction so that what is added in the first case at least looks. But, of course, the shift is illegitimate, since the original is not conjunction-conjunction, but conjunction-particular, and the addition has to go in the particular, not outside (to keep the meaning). This not quite to say you can't get what you want, just that your explanation needs to be different. Whether this internal fusion is going to work in general is another matter, which depends on what the whole array of fusion techniques is going to be.
>
>
> Unary quantifiers are not the only kind of predicates that can make
> use of a ni/nu comparison. Another example we considered is prfrake "A
> prefers E" perhaps reduced from something like prfrakeki "A prefers E
> from among I":
>
> lo je ckfa ni'u ldra prfra'aka
> "I prefer my coffee with milk."
> (From the choices of coffee whether with milk or not with milk, I
> prefer coffee WITH milk.)
>
> I think more generally "ni'u" can be thought of as a focus marker, for
> example tied to something like an implicit "I make assertion x rather
> than more general assertion y", where x is the one with ni and y the
> one with nu.
>
> I don't think I will ever understand this, since I don't see the point of the complication. Why not just say that I prefer coffee with milk or most people like chocolate or whatever it is that the donkey sentence is trying to say (since that is really very unclear: beats all his donkeys, beats some and not others, doesn't beat any?) Or why not contrast with the negative case rather than the general?
The difference that "ni'u" captures is the difference between, for
example, "I prefer coffee with milk (rather than, say, tea)" vs "I
prefer coffee with milk (rather than with nothing or with something
else)". "la ni'u je ckfa ldra prfra'aka" vs "la je ckfa ni'u ldra
prfra'aka".
I prefer coffee with milk to anything else - even with implicit salience restrictions,this seems a little extreme.
Contrasting with the negative case is an alternative way of thinking
about it. "The A&B's are most of the A(whether or not)B's" is
equivalent to "The A&B's are more than the A¬B's". But you need
more transformations, because "most" is defined as something with
respect to the total, rather than as a direct comparison of something
with its complement.
I would have thought this was a argument for having the basic word be "more", with most falling out as an easy special case. "more" is also more useful since it is not restricted to two classes. |