[oz] Use of elidable {cu}
24 views
Skip to first unread message
Felipe Gonçalves Assis
Jan 6, 2014, 6:06:12 PM1/6/14
to loj...@googlegroups.com
coi selpa'i
I know about the "cmevla as brivla" rule in the translation grammar. But still there remains a great amount of elidable {cu} occurrences throughout the text. E.g.,
(1) {lo zdàni cu pu se cintypu'i}
(2) {la nakfàmti .xènris. cu no roi cmìla}
(3) {xy ji'a cu grùsi}
(4) {ca lo càbdei cu na kèlci}
Granted, these are perfectly correct. It just made me curious to hear any thoughts you have on the usage of {cu}.
In any case, the common usage exemplified above seems inconsistent with more rare examples such as
(5) {sei ny spùda}
(6) {sei ny spùsku} (3 times)
(7) {pa la màxpre pu pònse}
(8) {la .oz. na dùnda lo bèsna do}
(9) {la .tòtos. ca lo nùncfa na nèlci lo vi ziljmìna be lo bènde}
(10) {la .dòrotis. na sai ka'e vìska}
(11) {la .tòtos. ja'a ka'e go'i}
mi'e .asiz.
selpa'i
Jan 6, 2014, 6:22:42 PM1/6/14
to loj...@googlegroups.com
la .asiz. cu cusku di'e
most of the {cu} that seem elidible actually are not. I will explain why
below.
> E.g.,
>
> (1) {lo zd�ni cu pu se cintypu'i}
> (2) {la nakf�mti .x�nris. cu no roi cm�la}
These two are truly stylistic choices, but there is an advantage to
using {cu} before TENSE+selbri. It makes it less likely to accidently
say something like the following (at least I assume that's true):
lo nu broda pu cinri
! "The event of brodaing in the past is interesting."
Using {cu} before {pu} is necessary there to prevent the accidental
slipping of the {pu} into the abstraction. So maybe that's one reason to
use {cu} even when it's not necessary: to be less likely to forget it
where it must be used.
> (3) {xy ji'a cu gr�si}
> (4) {ca lo c�bdei cu na k�lci}
Probably the same habit as in (1) and (2).
> Granted, these are perfectly correct. It just made me curious to hear
> any thoughts you have on the usage of {cu}.
>
> In any case, the common usage exemplified above seems inconsistent with
> more rare examples such as
>
> (5) {sei ny sp�da}
> (6) {sei ny sp�sku} (3 times)
> (7) {pa la m�xpre pu p�nse}
> (8) {la .oz. na d�nda lo b�sna do}
> (9) {la .t�tos. ca lo n�ncfa na n�lci lo vi ziljm�na be lo b�nde}
> (10) {la .d�rotis. na sai ka'e v�ska}
> (11) {la .t�tos. ja'a ka'e go'i}
The actual answer to your question is this: What happens when you erase
all the spaces from the text? The answer to this question will reveal
why there are {cu} in seemingly unnecessary places. To name just one
general example ({nu} is by far not the only place where it occurs):
{nu + BY + broda} = lujvo (nuBYbroda}
{nu + BY + cu + broda} = abstraction containing a bridi
In other words, leaving out {cu} will make the text parse differently!
I'm very careful with that; one needs to keep in mind that the whole
point of my orthography is to have full audio-visual isomorphism - the
speech stream should fully correspond to the string of letters. There
are no spaces in the speech stream. Everything depends on clusters and
stress. Therefore, spaces are to be meaningless in writing, too. And my
orthography enables exactly that: to omit all the spaces. You can either
use {cu} to prevent the accidental lujvo as in the example above, or you
can use a glottal stop, like so:
{nu + BY. + broda}
Then, it also cannot parse as a single word.
You will find both methods used in the text.
mi'e la selpa'i mu'o
> I know about the "cmevla as brivla" rule in the translation grammar. But
> still there remains a great amount of elidable {cu} occurrences
> throughout the text.
There may be some that really are elidible, but I'm fairly sure that
> still there remains a great amount of elidable {cu} occurrences
> throughout the text.
most of the {cu} that seem elidible actually are not. I will explain why
below.
> E.g.,
>
> (1) {lo zd�ni cu pu se cintypu'i}
> (2) {la nakf�mti .x�nris. cu no roi cm�la}
These two are truly stylistic choices, but there is an advantage to
using {cu} before TENSE+selbri. It makes it less likely to accidently
say something like the following (at least I assume that's true):
lo nu broda pu cinri
! "The event of brodaing in the past is interesting."
Using {cu} before {pu} is necessary there to prevent the accidental
slipping of the {pu} into the abstraction. So maybe that's one reason to
use {cu} even when it's not necessary: to be less likely to forget it
where it must be used.
> (3) {xy ji'a cu gr�si}
> (4) {ca lo c�bdei cu na k�lci}
Probably the same habit as in (1) and (2).
> Granted, these are perfectly correct. It just made me curious to hear
> any thoughts you have on the usage of {cu}.
>
> In any case, the common usage exemplified above seems inconsistent with
> more rare examples such as
>
> (6) {sei ny sp�sku} (3 times)
> (7) {pa la m�xpre pu p�nse}
> (8) {la .oz. na d�nda lo b�sna do}
> (9) {la .t�tos. ca lo n�ncfa na n�lci lo vi ziljm�na be lo b�nde}
> (10) {la .d�rotis. na sai ka'e v�ska}
> (11) {la .t�tos. ja'a ka'e go'i}
The actual answer to your question is this: What happens when you erase
all the spaces from the text? The answer to this question will reveal
why there are {cu} in seemingly unnecessary places. To name just one
general example ({nu} is by far not the only place where it occurs):
{nu + BY + broda} = lujvo (nuBYbroda}
{nu + BY + cu + broda} = abstraction containing a bridi
In other words, leaving out {cu} will make the text parse differently!
I'm very careful with that; one needs to keep in mind that the whole
point of my orthography is to have full audio-visual isomorphism - the
speech stream should fully correspond to the string of letters. There
are no spaces in the speech stream. Everything depends on clusters and
stress. Therefore, spaces are to be meaningless in writing, too. And my
orthography enables exactly that: to omit all the spaces. You can either
use {cu} to prevent the accidental lujvo as in the example above, or you
can use a glottal stop, like so:
{nu + BY. + broda}
Then, it also cannot parse as a single word.
You will find both methods used in the text.
mi'e la selpa'i mu'o
Jorge Llambías
Jan 6, 2014, 6:36:01 PM1/6/14
to loj...@googlegroups.com
On Mon, Jan 6, 2014 at 8:22 PM, selpa'i <sel...@gmx.de> wrote:
lo nu broda pu cinri
! "The event of brodaing in the past is interesting."
Actually it's even worse than that, it's just a sumti: "the brodaing in the past type of interesting thing."
Using {cu} before {pu} is necessary there to prevent the accidental slipping of the {pu} into the abstraction. So maybe that's one reason to use {cu} even when it's not necessary: to be less likely to forget it where it must be used.
Yes, it's a good idea not to rely on tenses to do the job of "cu", as they can't always do it.
mu'o mi'e xorxes
Felipe Gonçalves Assis
Jan 6, 2014, 7:16:52 PM1/6/14
to loj...@googlegroups.com
On 6 January 2014 20:22, selpa'i <sel...@gmx.de> wrote:
la .asiz. cu cusku di'e
E.g.,
(1) {lo zdŕni cu pu se cintypu'i}
(2) {la nakfŕmti .xčnris. cu no roi cměla}
These two are truly stylistic choices, but there is an advantage to using {cu} before TENSE+selbri. It makes it less likely to accidently say something like the following (at least I assume that's true):
lo nu broda pu cinri
! "The event of brodaing in the past is interesting."
Using {cu} before {pu} is necessary there to prevent the accidental slipping of the {pu} into the abstraction. So maybe that's one reason to use {cu} even when it's not necessary: to be less likely to forget it where it must be used.
On 6 January 2014 20:36, Jorge Llambías <jjlla...@gmail.com> wrote:
On Mon, Jan 6, 2014 at 8:22 PM, selpa'i <sel...@gmx.de> wrote:
lo nu broda pu cinri
! "The event of brodaing in the past is interesting."
Actually it's even worse than that, it's just a sumti: "the brodaing in the past type of interesting thing."
I just checked the parse. This completely breaks my mental grammar! I feel that the natural place of tenses is just before the selbri, and anywhere else they ought to be followed by either a sumti or {ku}. It is so much simpler to add a {ku} whenever we want an out of place tense. I have little hope that the grammar be reviewed, but don't you feel the same?
It only baffles me more that it works differently for negation. Am I safe to omit {cu} whenever there is a {na}?
The actual answer to your question is this: What happens when you erase all the spaces from the text? The answer to this question will reveal why there are {cu} in seemingly unnecessary places. To name just one general example ({nu} is by far not the only place where it occurs):
{nu + BY + broda} = lujvo (nuBYbroda}
{nu + BY + cu + broda} = abstraction containing a bridi
In other words, leaving out {cu} will make the text parse differently! I'm very careful with that; one needs to keep in mind that the whole point of my orthography is to have full audio-visual isomorphism - the speech stream should fully correspond to the string of letters. There are no spaces in the speech stream. Everything depends on clusters and stress. Therefore, spaces are to be meaningless in writing, too. And my orthography enables exactly that: to omit all the spaces. You can either use {cu} to prevent the accidental lujvo as in the example above, or you can use a glottal stop, like so:
{nu + BY. + broda}
Then, it also cannot parse as a single word.
You will find both methods used in the text.
That is very reasonable. I still find it much easier to use a glottal stop after every Cy not followed by another one, since you need it sooner or later and I just don't have the mental power to evaluate, in practical time, when it is safe to skip it.
Jorge Llambías
Jan 6, 2014, 9:07:38 PM1/6/14
to loj...@googlegroups.com
On Mon, Jan 6, 2014 at 9:16 PM, Felipe Gonçalves Assis <felipe...@gmail.com> wrote:
On 6 January 2014 20:36, Jorge Llambías <jjlla...@gmail.com> wrote:
lo nu broda pu cinri
! "The event of brodaing in the past is interesting."Actually it's even worse than that, it's just a sumti: "the brodaing in the past type of interesting thing."I just checked the parse. This completely breaks my mental grammar!
It's weird, I agree.
I feel that the natural place of tenses is just before the selbri, and anywhere else they ought to be followed by either a sumti or {ku}. It is so much simpler to add a {ku} whenever we want an out of place tense. I have little hope that the grammar be reviewed, but don't you feel the same?
I did propose once that tags should have priority in their function as selbri tags over their function as sumti tags.
It only baffles me more that it works differently for negation.
That's because for some reason it was decided that the "ku" of "na ku" should never be elidable.
Am I safe to omit {cu} whenever there is a {na}?
If the selbri starts with "na", yes, but not if it's something like "pu na broda", then the "pu" might still get stolen by some subordinate bridi.
guskant
Jan 7, 2014, 2:15:06 AM1/7/14
to loj...@googlegroups.com
{lo broda pu brode} is parsed differently between jbofi'e (yacc) and camxes (peg):
jbofi'e
[(lo broda KU) CU <<pu brode>> VAU]
camxes
[(lo broda KU) (pu KU) CU <<brode>> VAU]
It seems that jbofi'e is in accordance with CLL, while camxes isn't.
(cf.
CLL9.10:
An alternative to using “ku” is to place the modal cmavo right before the selbri, following the “cu” which often appears there. When a modal is present, the “cu” is almost never necessary.
CLL10.1:
The placement of a tense construct within a Lojban bridi is easy: right before the selbri. It goes immediately after the “cu”, and can in fact always replace the “cu” (although in very complex sentences the rules for eliding terminators may be changed as a result).
)
Le mardi 7 janvier 2014 11:07:38 UTC+9, xorxes a écrit :
On Mon, Jan 6, 2014 at 9:16 PM, Felipe Gonçalves Assis <felipe...@gmail.com> wrote:I feel that the natural place of tenses is just before the selbri, and anywhere else they ought to be followed by either a sumti or {ku}. It is so much simpler to add a {ku} whenever we want an out of place tense. I have little hope that the grammar be reviewed, but don't you feel the same?
I did propose once that tags should have priority in their function as selbri tags over their function as sumti tags.
Then why are peg parsers against your proposition and jbofi'e and CLL?
Is it a bug, or an intended behavior?
Gleki Arxokuna
Jan 7, 2014, 2:38:11 AM1/7/14
to loj...@googlegroups.com
I can digest all that but ...
{nu broda pu brode} parses while {broda pu brode} does not, that's what I'm kucli about.
--
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.
For more options, visit https://groups.google.com/groups/opt_out.
Jorge Llambías
Jan 7, 2014, 8:05:39 AM1/7/14
to loj...@googlegroups.com
On Tue, Jan 7, 2014 at 4:15 AM, guskant <gusni...@gmail.com> wrote:
Le mardi 7 janvier 2014 11:07:38 UTC+9, xorxes a écrit :I did propose once that tags should have priority in their function as selbri tags over their function as sumti tags.
Then why are peg parsers against your proposition and jbofi'e and CLL?Is it a bug, or an intended behavior?
You'd have to ask Robin that. He probably didn't bother with the "lo broda pu brode" case because the meaning is not really changed either way, The "lo nu broda pu brode" case is handled the same by both parsers, and it may be what CLL calls a "very complex sentence".
Jorge Llambías
Jan 7, 2014, 8:08:54 AM1/7/14
to loj...@googlegroups.com
On Tue, Jan 7, 2014 at 4:38 AM, Gleki Arxokuna <gleki.is...@gmail.com> wrote:
I can digest all that but ...{nu broda pu brode} parses while {broda pu brode} does not, that's what I'm kucli about.
In the first case you have a tanru "(nu broda pu KU KEI) brode", in the second case you start with a bridi "broda pu KU" and then you want to add a selbri to it, which you can't do.
Gleki Arxokuna
Jan 7, 2014, 9:54:25 AM1/7/14
to loj...@googlegroups.com
Yes, that's what I'm curious. {nu} turns a string of words into a tanru.
mu'o mi'e xorxes
--
Jorge Llambías
Jan 7, 2014, 10:12:09 AM1/7/14
to loj...@googlegroups.com
On Tue, Jan 7, 2014 at 11:54 AM, Gleki Arxokuna <gleki.is...@gmail.com> wrote:
On Tue, Jan 7, 2014 at 5:08 PM, Jorge Llambías <jjlla...@gmail.com> wrote:
On Tue, Jan 7, 2014 at 4:38 AM, Gleki Arxokuna <gleki.is...@gmail.com> wrote:
I can digest all that but ...{nu broda pu brode} parses while {broda pu brode} does not, that's what I'm kucli about.In the first case you have a tanru "(nu broda pu KU KEI) brode", in the second case you start with a bridi "broda pu KU" and then you want to add a selbri to it, which you can't do.Yes, that's what I'm curious. {nu} turns a string of words into a tanru.
Not just any string of words, it has to be a bridi, and NU ... KEI turns it into a tanru unit.
Felipe Gonçalves Assis
Jan 7, 2014, 10:27:06 AM1/7/14
to loj...@googlegroups.com
In
http://users.digitalkingdom.org/~rlpowell/hobbies/lojban/grammar/lojban.peg.txt
(where can I find the file used by camxes?)
There is this rule
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause? free*) ) / termset / NA-clause KU-clause free*
I don't have a firm grasp of formal grammars, so correct me if I am wrong.
1. This rule is relevant.
2. If we make the KU-clause mandatory, instead of optional, my mental grammar is realized.
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause free*) ) / termset / NA-clause KU-clause free*
3. It would contradict the CLL, though. http://dag.github.io/cll/9/9/http://users.digitalkingdom.org/~rlpowell/hobbies/lojban/grammar/lojban.peg.txt
(where can I find the file used by camxes?)
There is this rule
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause? free*) ) / termset / NA-clause KU-clause free*
I don't have a firm grasp of formal grammars, so correct me if I am wrong.
1. This rule is relevant.
2. If we make the KU-clause mandatory, instead of optional, my mental grammar is realized.
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause free*) ) / termset / NA-clause KU-clause free*
mu'o
mi'e .asiz.
Jorge Llambías
Jan 7, 2014, 10:50:24 AM1/7/14
to loj...@googlegroups.com
On Tue, Jan 7, 2014 at 12:27 PM, Felipe Gonçalves Assis <felipe...@gmail.com> wrote:
In
http://users.digitalkingdom.org/~rlpowell/hobbies/lojban/grammar/lojban.peg.txt
(where can I find the file used by camxes?)
There is this rule
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause? free*) ) / termset / NA-clause KU-clause free*
I don't have a firm grasp of formal grammars, so correct me if I am wrong.
1. This rule is relevant.
2. If we make the KU-clause mandatory, instead of optional, my mental grammar is realized.
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause free*) ) / termset / NA-clause KU-clause free*
That's one way of dealing with it, although for my taste it has the disadvantage of making a terminator non-elidable, which is exceptional. Also, because of the silly restrictions on compound tags, you would have to be very aware of which tags can be combined without ku and which strings of tags would require ku insertions to make them work.
The way I would rather do it is:
term-1 <- sumti / ( !gek (tag !selbri-1 / FA-clause free*) (sumti / KU-clause? free*) ) / termset / NA-clause KU-clause free*
which means that the tag will not be absorbed as a term if it's directly followed by a selbri-1 (in which case it will be absorbed by the selbri rule),
Felipe Gonçalves Assis
Jan 7, 2014, 11:24:36 AM1/7/14
to loj...@googlegroups.com
On 7 January 2014 12:50, Jorge Llambías <jjlla...@gmail.com> wrote:
On Tue, Jan 7, 2014 at 12:27 PM, Felipe Gonçalves Assis <felipe...@gmail.com> wrote:
In
http://users.digitalkingdom.org/~rlpowell/hobbies/lojban/grammar/lojban.peg.txt
(where can I find the file used by camxes?)
There is this rule
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause? free*) ) / termset / NA-clause KU-clause free*
I don't have a firm grasp of formal grammars, so correct me if I am wrong.
1. This rule is relevant.
2. If we make the KU-clause mandatory, instead of optional, my mental grammar is realized.
term-1 <- sumti / ( !gek (tag / FA-clause free*) (sumti / KU-clause free*) ) / termset / NA-clause KU-clause free*
That's one way of dealing with it, although for my taste it has the disadvantage of making a terminator non-elidable, which is exceptional. Also, because of the silly restrictions on compound tags, you would have to be very aware of which tags can be combined without ku and which strings of tags would require ku insertions to make them work.
I don't see {ku} here as a terminator. To me, {pu ku} is a single entity, used whenever we want to move {pu} away from the selbri.
With regards to compound tags, don't you have to understand the rules anyway to get the correct meaning? Can you give an example?
With regards to compound tags, don't you have to understand the rules anyway to get the correct meaning? Can you give an example?
The way I would rather do it is:term-1 <- sumti / ( !gek (tag !selbri-1 / FA-clause free*) (sumti / KU-clause? free*) ) / termset / NA-clause KU-clause free*which means that the tag will not be absorbed as a term if it's directly followed by a selbri-1 (in which case it will be absorbed by the selbri rule),
That is a mid-ground more appropriate for an eventual proposal. I would support it.
In production, though, I still prefer to always use {ku}, because I find {gau se ka'a ko'a} and {bai gau se ka'a ko'a} confusing for the same reason: I keep expecting the selbri they refer to.
In production, though, I still prefer to always use {ku}, because I find {gau se ka'a ko'a} and {bai gau se ka'a ko'a} confusing for the same reason: I keep expecting the selbri they refer to.
Jorge Llambías
Jan 7, 2014, 12:05:28 PM1/7/14
to loj...@googlegroups.com
On Tue, Jan 7, 2014 at 1:24 PM, Felipe Gonçalves Assis <felipe...@gmail.com> wrote:
With regards to compound tags, don't you have to understand the rules anyway to get the correct meaning?
If you follow the official rules, yes.
Can you give an example?
Things like "ze'i co'a" being fine, but "co'a ze'i" oficially being really "co'a ku ze'i".
One problem with that is that "lo ze'i co'a broda" is fine, but "lo co'a ze'i broda" is not, although you can always cheat by using "lo co'a ja'a ze'i broda".
The way I would rather do it is:term-1 <- sumti / ( !gek (tag !selbri-1 / FA-clause free*) (sumti / KU-clause? free*) ) / termset / NA-clause KU-clause free*which means that the tag will not be absorbed as a term if it's directly followed by a selbri-1 (in which case it will be absorbed by the selbri rule),That is a mid-ground more appropriate for an eventual proposal. I would support it.
In production, though, I still prefer to always use {ku}, because I find {gau se ka'a ko'a} and {bai gau se ka'a ko'a} confusing for the same reason: I keep expecting the selbri they refer to.
Yes, it's more of a theoretical than a practical issue. Very few tags are actually used in the tag ku construct, and having more than one in the same sentence is rare.
guskant
Jan 7, 2014, 9:12:58 PM1/7/14
to loj...@googlegroups.com
I agree with xorxes.
As for zasni gerna: http://www.lojban.org/tiki/zasni+gerna
term <- FA#? sumti / tag? sumti / FA# KU#? / tag KU#? / NA# KU#? / gek term+ VAU#? GI# term+ VAU#?
should be replaced with
term <- FA#? sumti / tag? sumti / FA# KU#? / tag !selbri KU#? / NA# KU#? / gek term+ VAU#? GI# term+ VAU#?
, shouldn't it?
Reply all
Reply to author
Forward
0 new messages