On Sat, Feb 09, 2013 at 11:17:59PM -0500, Jacob Errington wrote:
> On 9 February 2013 22:14, v4hn <
me@v4hn.de> wrote:
> > On Sat, Feb 09, 2013 at 09:31:12PM -0500, Jacob Errington wrote:
> > > With this definition, we can easily create a predicate meaning "to look
> > for
> > > properties that make you happy", e.g. {.i mi sisku lo ka mi gleki ce'u}.
> >
> > Didn't you mean to say events/states here instead of properties?
> >
>
> No, I did intend to say properties, due to my general philosophy about
> Lojban predicates: if an intrinsic connection between a sumti and an
> abstraction exists in a given selbri, then that abstraction is a property
> of that sumti.
Ok, that seems to be a sane perspective. Although, I'm rather sure,
it overrides quite some learning material, so you have to deal with
alternative views as well..
> > That's what gleki2 is supposed to be. Mixing up terms here is confusing.
>
> It's been said in at least a few other posts, [...] that the type
> restrictions in brackets in the gismu list are not prescriptive.
> That being said, the gismu list simply tells us that the x2 must be an
> abstraction, with the *suggestion* that it should be an event or state. I
> disagree with that suggestion, and due to its non-prescriptive nature, am
> entitled to use a ka-abstraction there.
Yes, you are. but in {.i mi sisku lo ka mi gleki ce'u} you didn't say that the
{ce'u} place is to be a ka-abstraction. Therefore, this can't just be
translated as "to look for properties that make you happy", because "to look
for events that make you happy" is at least an equally good translation.
"to look for abstractions that make you happy" would be more fitting
for all possible interpretations, I suppose.
I'm sorry about that confusion then. You're right, I should have made it more clear. I also agree that "abstractions" would have been better overall.
Also, at least in my philosophy, you can become happy about an event
you're not involved in. {mi gleki lonu do citka lo plise} is a perfectly
valid sentence, so you're argument from above doesn't really restrict
the type of abstraction here, necessarily.
Right. That's the downside to this system: it winds up requiring some extra verbosity if you want to use an event that doesn't involve the formal argument. The solution that I made up for this when I first considered a new system for abstractions involved introducing a small exception: lifri2 is a {li'i}, rather than a {ka}, and the li'i-bridi doesn't need to contain ce'u. When a li'i-abstraction is used inside a ka-abstraction, the ce'u-place typically finds its way into li'i2, and then all is well.
{.i mi gleki lo ka [se] li'i do citka lo plise}.
The major advantage, however, of my abstractions system is that is makes producing jvajvo simpler. If we consider any lujvo of the type -dji, the jvajvo become a bit annoying, because djica2 is a {nu} (something I have yet to believe should be a {ka}).
e.g. ctidji = x1 djica lo nu *x2* citka x3 kei x4
Saying that there's a place merger is pretty wrong, because the Lojban definition then becomes slightly ridiculous. Place mergers should only occur on the same abstraction-level.
e.g. pampe'o = x1 boi x2 prami gi'e pendo
Because of this inconvenience with {djica} and other nu-type selbri, many lujvo makers simply drop the annoying x2 place. When speaking the full structures, leaving out the x1 is simple due to the bridi-tail counting rule, e.g. {.i mi djica lo nu citka lo plise}, but if we use the jvajvo, FA cmavo or repetition become inevitable, e.g. {.i mi ctidji fi lo plise}.
Indeed, ka-selbri are nicer in jvajvo: {.i mi ctika'e lo plise} -> {.i mi kakne lo ka [ce'u] citka lo plise}.
.i mi'e la tsani mu'o