Mini-rant: gismu place consistency

[Robin Lee Powell]

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.

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.

-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/

[vitci'i]

On 12/16/2011 01:17 PM, Robin Lee 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.
>
> 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.

Perhaps there should be a central repository of things we'd like to fix
when the next major version of lojban occurs. (There almost certainly
will be one eventually if the language lasts long enough. Maybe it'll be
in ten years, maybe a hundred, but it *will* need a major update
*eventually*.)

[Robin Lee Powell]

I'm really hoping to avoid that. But I think there's at least one
page on the wiki; search for loccan maybe?

[vitci'i]

On 12/16/2011 01:25 PM, Robin Lee Powell wrote:
> On Fri, Dec 16, 2011 at 01:21:40PM -0600, vitci'i wrote:
>> On 12/16/2011 01:17 PM, Robin Lee 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.
>>>
>>> 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.
>>
>> Perhaps there should be a central repository of things we'd like
>> to fix when the next major version of lojban occurs. (There almost
>> certainly will be one eventually if the language lasts long
>> enough. Maybe it'll be in ten years, maybe a hundred, but it
>> *will* need a major update *eventually*.)
>
> I'm really hoping to avoid that. But I think there's at least one
> page on the wiki; search for loccan maybe?

Really? You want -tel- to still belong to {stela} instead of {te} in the
year 2200? When we finally clone a velociraptor, do you want to have to
ask {na'e gledji cinse fi ma lo xadni} because we don't have a gismu for
biological sex? When the Sun burns out and you're visiting your
grandkids-xiso'i in Andromeda, do you still want *two different short
rafsi* allocated to *Antarctica*?

[Robin Lee Powell]

LOL.

All those things can be fixed incrementally, was my point.

[vitci'i]

I'm not convinced that *all* bugs can be fixed by minor updates, but okay.

In that case, perhaps there should exist some sort of bug tracker for
Lojban.

[Luke Bergen]

Hmmm, what about words like {vecnu} where the audience seems less important than the direct object?  Or is that my lingual bias coming through?

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.
To unsubscribe from this group, send email to lojban+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.

[Robin Lee Powell]

This isn't about importance; that's what got us into this mess.
It's about ease of memorization. :)

-Robin

[John E Clifford]

The last one, which may have died years ago was called "LoCCan3", I think. Many
things -- not merely vocabulary -- have been added since then.

-Robin

--

[Ian Johnson]

The nice thing about {vecnu} if you aren't going to have two words for "buy" and "sell" is that both {vecnu} and {te vecnu} have the purchased-thing in the x2.

mu'o mi'e latros

[Luke Bergen]

Right, but my question is that with Robin's initial email he said that the audience or second agent would be in x2 which feels awkward for things like {vecnu} where it seems like the direct object is more important than the secondary agent (the te vecnu)

[Robin Lee Powell]

It was a rant; it wasn't intended to be well thought out. :)

-Robin

[la gleki]

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.

[MorphemeAddict]

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. 

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.
 
 

[Jonathan Jones]

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. Any volunteers?
 

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 )

[la gleki]

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

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".

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

 

Any volunteers?

Well, i wouldn't recommend changing gismu definitions. Gismu space is undepletable. We can create alternative gismu for that. 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.

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

[Jonathan Jones]

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?)
 

[Jonathan Jones]

As a side note, I believe that the grouping of gismu into families has merits of its own, in the same way and for nearly the same reason that grouping the cmavo into selma'o is.
 

[la gleki]

Why not using {ka'a} cmavo instead of all those places?

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.

Oh, that's another case. As lojbab mentioned this wasn't intentional. Fixing this is a must. E.g. I can do that now in Russian translation. This fix won't change the corpus much.

 

 

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.

Well, .ui I'm already working on categorisation of gismu.

Your comments might instigate me to revise this categorisation.

[Jonathan Jones]

I don't understand the question.
 

[la gleki]

ko'a vofli seka'a ko'e  ...

ko'a bevri seka'a ko'e ...

why bother changing the order? it's better to ignore those inconsistent places or use jvajvo based on {klama}.

[Jonathan Jones]

Because no it isn't, and what reason could you possibly have to think so?
 

[Pierre Abbat]

On Saturday, March 09, 2013 07:13:55 Jonathan Jones wrote:
> 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.

li'o

> 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?

I think we should drop the direction place of "mitre" and the subunit places.
I have no idea what to put in a subunit place, and the direction place could
be expressed with a preposition if appropriate. "gucti" has just as much
reason to have a direction place as "mitre" does, but "caltre" has none, even
though by jvojva it has to have one if "mitre" does.

> 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.

There's no big problem with species. There are a few species words with a
third place (tirxu, mledi, lanme, mlafi'e) and one missing x2 (remna), which
should have it, as paleontologist talk about several species of Homo.

Pierre
--
When a barnacle settles down, its brain disintegrates.
Já não percebe nada, já não percebe nada.

[Jonathan Jones]

On Sat, Mar 9, 2013 at 2:26 PM, Pierre Abbat <ph...@bezitopo.org> wrote:

I think we should drop the direction place of "mitre" and the subunit places.
I have no idea what to put in a subunit place,

I'm pretty sure it's there to be an identifier: {gutci fo la.intc.} (feet) vs. {gutci fo lo degji} (cubit).

It's seems a circumspect way to do so to me, though, and I'm just guessing.

If that is indeed the intention, I agree that that place is pointless and should be dropped, possibly replacing it with a me'e place. (In the x2, pushing the rest one down?)
 

and the direction place could be expressed with a preposition if appropriate.  "gucti" has just as much reason to have a direction place as "mitre" does, but "caltre" has none, even though by jvojva it has to have one if "mitre" does.

Well, sure, I see your argument. I don't disagree. I said minimal, for reasons of past-usage breaking, whereas as what you're suggesting, I would call "optimal".
 

> 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.

There's no big problem with species. There are a few species words with a
third place (tirxu, mledi, lanme, mlafi'e) and one missing x2 (remna), which
should have it, as paleontologist talk about several species of Homo.

I was referring to the grouping, not the consistency, in this. Obviously the size of a gismu family isn't really relevant to the consistency issue.

[John E Clifford]

taDA!

---

From: Jonathan Jones <eye...@gmail.com>
To: loj...@googlegroups.com
Sent: Saturday, March 9, 2013 2:26 PM
Subject: Re: [lojban] Re: Mini-rant: gismu place consistency

[Jonathan Jones]

On Sat, Mar 9, 2013 at 3:17 PM, John E Clifford <kali9...@yahoo.com> wrote:

taDA!

????
 

[John E Clifford]

See earlier request from Robin et al.

---

From: Jonathan Jones <eye...@gmail.com>
To: loj...@googlegroups.com

Sent: Saturday, March 9, 2013 4:40 PM

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

[Michael Turniansky]

On Sat, Mar 9, 2013 at 4:47 PM, Jonathan Jones <eye...@gmail.com> wrote:

On Sat, Mar 9, 2013 at 2:26 PM, Pierre Abbat <ph...@bezitopo.org> wrote:

I think we should drop the direction place of "mitre" and the subunit places.
I have no idea what to put in a subunit place,

I'm pretty sure it's there to be an identifier: {gutci fo la.intc.} (feet) vs. {gutci fo lo degji} (cubit).

It's seems a circumspect way to do so to me, though, and I'm just guessing.

If that is indeed the intention, I agree that that place is pointless and should be dropped, possibly replacing it with a me'e place. (In the x2, pushing the rest one down?)

  No, it's a numeric, such as mi li mu lo merko  li pano --> I am 5'10"  It appears at the end to not break the consistency (such as it is) with other measurment gismu.

          --gejyspa

[Ian Johnson]

I think we should drop the direction place of "mitre" and the subunit places.

I have no idea what to put in a subunit place, and the direction place could
be expressed with a preposition if appropriate.  "gucti" has just as much
reason to have a direction place as "mitre" does, but "caltre" has none, even
though by jvojva it has to have one if "mitre" does.

I actually made good use of the mitre direction place in coining {clatre}, {gantre}, and {tsutre}. Please leave it, because it is helpful, even if difficult to use by itself.

mi'e la latro'a mu'o

[selpa'i]

la latro'a cu cusku di'e

Who is this a response to? Please do not remove the line from your quote
that indicates who said it. (I had to go look through the older mails to
find out that it was Pierre who said that one month ago.) Thank you.

mu'o mi'e la selpa'i

[Ian Johnson]

On Fri, Apr 5, 2013 at 8:21 PM, selpa'i <sel...@gmx.de> wrote:

Who is this a response to? Please do not remove the line from your quote that indicates who said it. (I had to go look through the older mails to find out that it was Pierre who said that one month ago.) Thank you.

mu'o mi'e la selpa'i

Sorry about that, sloppy selection. In my sort-of defense it was a response more to Michael's post; I hadn't read Pierre's post before, noticed it quoted in Michael's post, and wanted to comment on it.