Re: [lojban] 5 classes of sumti in Lojban and about a new word for "grammatical abstraction"

[Jacob Errington <nictytan@gmail.com>]

On 19 April 2013 23:37, la gleki <gleki.is.my.name@gmail.com> wrote: 

Well, I wanted to ask why should we have a separate fundamental class of sumti if {ni} is derived and therefore probably can be rephrased from another NU type of sumti.

Another question is isn't fasnu1 and djuno2 have something in common? We call both of them abstractions but acc. to your scheme they are in separate classes.

{ni} as it stands in current usage, has a handful of uses. For one, it can be used as a function (taking ce'u) for use in comparative selbri. {ka} (at least under strict fancylojban) is binary, either true or false, and is inadequate for useful comparisons. Typically, it doesn't make sense for a {ni}-abstraction to contain indirect questions, but if a decent example can be thrown together, I'm prepared to go with it.

That being said, as {ni} can be a function, both of these are allowed: 

{.i mi do zmadu lo ka ce'u se pampe'o} 

{.i mi do zmadu lo ni ce'u se pampe'o}

The former simply states {.i mi se pampe'o .ijenai do se pampe'o} due to the binary nature of {ka} abstractions. However, the statement with zmadu is a bit more connotative than the truth-functionally equivalent statement with connectives. On the other hand, the latter one states that the extent to which the function holds for me is greater than that for you. (Perhaps my girlfriend and I see each other more frequently, or some such.) It doesn't deny that you have a girlfriend, which is what the former does.

Non-function {ni} (i.e. {ni}-abstractions containing no {ce'u}) are a bit troubling, on the other hand. They have some kind of pseudo-numeric type that seems to act like a quantity. The problem with quantities is that Lojban doesn't really handle them in most selbri. Examples:

{.i lo ni do mamta cu zmadu lo ni mi djeca lo nu sipna} is nonsensical.

{.i lo ni mi djica lo nu sipna cu barda} is redundant to the functional form {.i mi barda lo ni djica lo nu sipna}

I'm willing to say that most cases of non-functional ni appearing are either nonsense or beta-reductions, i.e. produced by applying a functional {ni} to its arguments.

{ni} then isn't a separate class on its own. It also isn't entirely redundant to a ka-kau function, -- that is -- a ka-function with an indirect question. If we want to compare event frequencies, for instance, it is better done with ka-kau than {ni}:

.i do mi zmadu lo ka xokauroi pampe'o penmi

Assuming a constant interval is being tagged by xokauroi, then the statement means "You are greater than me in the property of how many times in some interval does one meet their significant other." In that case, the ka is no longer simply boolean; using indirect questions allows us to produce functions returning any type.

I wager that in other places in which non-functional {ni} abstractions might appear have a (complex) abstraction in some other place of the selbri handling it.

{.i lo ni do tcidu cu cinri} is perhaps sumti raising, with {.i lo ni do tcidu cu jai cinri fai lo ka ce'u zmadu lo ni lo'e prenu cu tcidu}.

In sum, {ni} is not a class of its own, but is convenient to have because it allows us to make simplifications with {ka}-abstractions. In my original abstractions paper, I suggested that {ni} is {ka la'u makau}, although this opinion has fallen out of favour with me, as I have begun to consider {ni}-abstractions as both a subset of ka-abstractions, but separate enough such that we don't have the tools necessary to properly convent between the two.

.i mi'e la tsani mu'o

-- 
You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.