* Sunday, 2011-10-16 at 23:14 -0300 - Jorge Llambías <jjllambias@gmail.com>:
> On Sun, Oct 16, 2011 at 10:56 PM, Martin Bays <mbays@sdf.org> wrote:
> > * Sunday, 2011-10-16 at 20:49 -0300 - Jorge Llambías <jjllambias@gmail.com>:
> >
> >> On Sat, Oct 15, 2011 at 8:29 PM, Martin Bays <mbays@sdf.org> wrote:
> >> >> The drawback of that approach is that you cannot combine predications
> >> >> that "resolve" differently.
> >> [...]
> >> > In any case, this drawback seems a rather small one to me.
> >>
> >> It's impotant though. For example, compare:
> >>
> >> (1) ca lo prulamnicte mi tavla su'o da poi do nelci ke'a
> >> "Last night I talked to someone you like."
> >>
> >> (2) ca lo prulamnicte mi citka su'o da poi do nelci ke'a
> >> "Last night I ate something you like."
> >>
> >> You want to accept (1) but reject (2), even though to me they have the
> >> exact same logical structure.
> >
> > More precisely, I'd give the lojban in (2) the less likely (without
> > context) of the meanings of the english - the one for which a reasonable
> > response would be "you bastard!".
>
> But shouldn't "you bastard!" be a response to "last night I ate
> something you liked"? I think the one you call less likely is even
> less likely than you suggest, and the reasonable response would be "I
> most certainly don't anymore".
Yes, true.
> > You, meanwhile, would copy to lojban this ambiguity in the english.
> >
> > (You wouldn't call it an ambiguity, I know; but consider that in english
> > we can (mostly) disambiguate to the obvious option by making it "I ate
> > of something you like" (although this construction is rare in modern
> > english))
>
> There are ways to change the level of abstraction in Lojban too,
> usually by being more wordy. The inconvenience is being forced to do
> it even when you don't need to. You are essentially forbidding a
> certain level of abstraction.
>
> >> You would need to say something like "ca lo prulamnicte mi citka su'o
> >> da poi ckaji su'o de poi do nelci ke'a" instead of (2).
> >
> > That's abbreviable to
> > {ca lo prulamnicte mi citka su'o ckaji be su'o se nelci be do},
> > which isn't all that verbose.
>
> But I bet nobody will talk like that.
Most people would bet no-one would talk lojban ;)
But many verbosity problems could be resolved by introducing new
(meanings for) cmavo. e.g. it could be {mi citka su'o lu'ai lo se nelci
be do}, with {lu'ai} meaning "the exemplars/realisations/instances of".
Moreover, I don't see a problem with having {lu'a} have this meaning, in
addition to its "elements of" meaning when applied to sets.
> > Alternatively, how about
> > {ca lo prulamnicte mi ckaji citka su'o se nelci be do}
> > "I exemplar-ate a foodstuff you like" -> "I ate an exemplar of
> > a foodstuff you like" -> "I ate of a foodstuff you like"?
>
> Where presumably "ckaji citka" is not a kind of "citka"? i.e. "ko'a
> ckaji citka ko'e" does not entail "ko'a citka ko'e"?
Fair point. I suppose it would have to be a lujvo; eating up lujvo-space
like that might not be such a good idea.
> >> Or consider:
> >>
> >> (3) mi zukte lo se zukte be do
> >> "I'm doing what you are doing."
> >>
> >> You have to say: "mi zukte lo ckaji be lo se ckaji be lo se zukte be do."
> >
> > To copy the kinds approach, yes. It could be abbreviated to
> > {da se ckaji lo se zukte be mi .e do}, of course.
>
> I think you will need "da se ckaji lo se zukte be mi be'o .e lo se zukte be do"
>
> "lo se zukte be mi .e do" is "zo'e noi ke'a se zukte mi .e do", and
> presumably there's no such thing in your universe.
Right again. Make that {da zo'u mi .e do zukte lo ckaji be da}.
With {lu'a} as above, it could be just {lu'a da mi .e do se zukte}
(although then there's even less to indicate the intended value of da).
> > Neither this nor the kinds version really gets across the meaning of the
> > english, though, since there's no indication of the level of generality
> > of the property/kind involved. Maybe something like {mi zukte lo panra
> > be lo se zukte be do} is clearer.
>
> In any case, wouldn't you agree this is more than a small drawback?
Yes, I suppose I would. But I still prefer required verbosity to
undisambiguable ambiguity.
Martin
Attachment:
pgpmJop65FDf0.pgp
Description: PGP signature