Re: [lojban] Re: Mini-rant: gismu place consistency

On Sat, Mar 9, 2013 at 5:32 AM, la gleki <gleki.is.my.name@gmail.com> wrote:

On Saturday, March 9, 2013 4:00:13 PM UTC+4, aionys wrote:
On Sat, Mar 9, 2013 at 4:33 AM, MorphemeAddict <lyt...@gmail.com> wrote:

On Sat, Mar 9, 2013 at 6:06 AM, la gleki <gleki.is...@gmail.com> wrote:

On Friday, December 16, 2011 11:17:26 PM UTC+4, Robin Powell wrote:

If I had it to do over again, every gismu with an agent place would
have it as the first place. Every gismu with an audience or second
agent of any type would have it as the second place. There would
probably be several other regularities as well.

Let's assume that
xau = BAI tag denoting agent or reference frame.
.iei = BAI tag denoting object.
.uai = BAI tag denoting property of the object.

Then we can rephrase many gismu as

ko'a djuno lo du'u ko'e brode kei ko'e = xau mi djuno .uai lo ka brode .iei ko'i
(the same for {morji} and company)

ko'a kakne lo ka brode = xau [je .iei] mi kakne .uai lo ka brode

ko'a te mukti lo nu brode = xau ko'a .iei ko'a .uai lo ka brode kei te mukti
(May be xau = gau or xau = ma'i)
These tags can replace FA (a separate set of SE is not needed here if we use {jai}).

For {vecnu, dunda, cpacu} and {benji} we can use {be'i} tags (if {benji} is about sending objects, not information) as the core semantic prime for those gismu seems to be the same.
For sending information another set of BAI might be needed.

I've been saying {manci} and {melbi} to my babies *a lot*, and I
*still* have trouble remembering that it's {do melbi mi} and {mi
manci do}. GRAAAAAAAAAAAAAAAAAAAAAAH.

This is not the problem. We can just use {selmanci} and forget about {manci}.
{xabju = se zdani} is a better example showing redundancy.

The problem isn't redundancy, it's consistency.
Place structures should be as consistent as possible.

Agreed. I will be very happy when we're allowed to review this and actually do something about it.

Speaking of, I do know that one thing that will be needed is to know which gismu have inconsistent structure, what would need to be changed to make it consistent with its "family", and how that change affects the corpus.

How is that possible ? I showed how to "fix" that using BAI, SE and (semiseriously) experimental BAI.

That's not a fix, it's a workaround. If you don't know the difference, look it up.

Are there other methods?
Even if all agentive gismu had agent in x1 then newbies would still complain that non-agentive gismu break this "rule".

Then they would be wrong, because the whole thing is about consistency of place structure among similar gismu, not about about making every gismu have the exact same place structure.

Or e.g. what's common in djuno3 and klama3, vecnu2 and lakne2? Nothing.

Well, obviously. They're not in the same family.

However, klama, litru, muvdu, pluta, bevri, and farlu are in the same family: transportation. Note that they also have similar place structures. In fact, {litru} = {klama zi'o zi'o}, {muvdu} = {klama fu zi'o}, {pluta} = {ve klama veka'a loi mokca fo zi'o zi'o}, {bevri} = {zo'e poi lo jgari ku'o klagau faxixa zi'o} and {farlu} = {klani'a fo zi'o zi'o ci'e zo'e}. As you can see, while there are some variances as to what places there are, all of the /shared/ places are in the same exact order, and usually even in the "x" place.

As another example, all of the gismu of the measurement family have the exact same place structure: "x1 is x2 [units of measurement] by standard x3", and for non-metric: "x1 is x2 [units of measurement] in standard x3 with sub-units x4", /except/ for mitre: "x1 is x2 meters in direction x3 by standard x4", which moves the typical x3 place to the x4 and places an additional "in direction" place in its stead, and minli: "x1 is x2 long local distance units with sub-units x3 in standard x4", which has the exact same places as it's brethren, except that the final two are in reverse order.

Still newbies will have to remember that both for djuno3 and klama3 in order to reach this place you have to use either {fi} or {te}. So what problems of learnability of the language would be solved? None.

Incorrect. Consistency of place structure amongst gismu families solves the problem of having to remember exceptions to the typical place structure, for the simple reason that there would /be none/.

Any volunteers?

Well, i wouldn't recommend changing gismu definitions. Gismu space is undepletable.

Which is not the point. The reason for the desire to /fix/ the place structure of the inconsistent gismu is because they are, as is, broken. And before you ask how they are broken, the answer is: because they are inconsistent.

We can create alternative gismu for that.

Which only makes things worse in this scenario. Say we make an "alternative gismu" for mitre. First off, what shape should it have? {mitre} is arguably the best shape for the Lojban word for "meter". Secondly, how would convince people to stop using mitre and instead use this new word, when they've already learned {mitre}="meter"? Would it not be simpler to merely say "From now on, the x3 place of mitre is the standard place, and the direction place is x4.", especially since mitre's x3 and x4 are so seldom used it wouldn't really affect the corpus, and would make it consistent with its brethren, thus making it easier to learn fully?

Besides, given that many hate the sounding of gismu ("selsiclu dukse" syndrom and abhorrence of "la kafkylerfu") and complain that 6-language mixture made gismu unrecognisable for everyone this can be even advisable.

Personal feelings as to the choice of the shapes of the gismu is irrelevant to this issue. It is an issue all its own and should be discussed separately.

Anyway you can add me to your team of volunteers. Just show several examples of how it might look like.

I gave two examples above, but this is basically the process in its entirety:

1) Take the whole of the gismu, possibly including the experimental ones, and group them into families. Species {badna}, Transportation {klama}, Mode of Travel {cadzu}, Measurement {mitre}, and Comparatives {barda} are a few families I can think of off the top of my head. Species might need to be split into the kingdoms (animal, plant, etc.) just because of size.

2) Look at each family, and determine the typical place structure of that family: what are the places, in what order do they appear, etc.

3) Find the gismu that do not have that place structure, and determine what the differences are, and if they are justified (for example, {pluta} has a slightly atypical structure, but it is obviously justified, whereas {mitre}'s inconsistency is much harder to argue for)

4) Determine what the smallest change would be required to bring the inconsistent gismu in line with the typical one for its family (such as reversing the x3 and x4 places of mitre and minli).

5) Determine how this change would affect the corpus (i.e. instances of usage thus far), and document this for each gismu. (For example, does reversing the x3 and x4 places affect the meaning of any currently existing bridi which have {mitre} in them? If so, how many are affected, and what would the new meaning be?)

stevo

-Robin
--
http://singinst.org/ : Our last, best hope for a fantastic future.
Lojban (http://www.lojban.org/): The language in which "this parrot
is dead" is "ti poi spitaki cu morsi", but "this sentence is false"
is "na nei". My personal page: http://www.digitalkingdom.org/rlp/

--
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+un...@googlegroups.com.

To post to this group, send email to loj...@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.

--
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+un...@googlegroups.com.
To post to this group, send email to loj...@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.

--
mu'o mi'e .aionys.

.i.e'ucai ko cmima lo pilno be denpa bu .i doi.luk. mi patfu do zo'o
(Come to the Dot Side! Luke, I am your father. :D )

--
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.