BPFK: Old General Negators

From Lojban
Jump to: navigation, search

 forum_name: General Negators
topic_title: A Shepherd Volunteer
    subject: A Shepherd Volunteer
   username: kd
  post_time: 2003-04-06 20:53:12
  post_text:

Looking over various bits of the corpus, I have decided to take on the responsibility of being a catherd. I don't want too big a category, but I also don't want something boring like "mathematical constants (PA5)". Therefore, I volunteer to be the shepherd of the NAI discussions.


 forum_name: General Negators
topic_title: Shepherding Philosophy
    subject: Shepherding Philosophy
   username: kd
  post_time: 2003-04-12 23:06:37
  post_text:

In general, I am anti-prescription. The BPFK exists because total lack of prescription simply cannot work right now. Consequently, I wish to minimize the prescriptivism necessary in the paradigm that I have volunteered to shepherd. I will operate under the idea that it is best to Let Usage Decide. Therefore, I intend to write a descriptive record of this paradigm, and post a draft of it for comments. Once the descriptive record is in its final form, it will contain an account of what usage thinks the meaning and grammar is of the relevent cmavo. It will not be based on the current baseline. I will then post a proposal that we officialize the previous usage, which will of course be open for amendment so that the final form may differ from what currently exists. Hopefully, the result will be an official version which is an optimized variant of the version currently in use.


 forum_name: General Negators
topic_title: Descriptive Record
    subject: Descriptive Record
   username: kd
  post_time: 2003-04-14 21:34:23
  post_text:

This is nowhere near done, but as promised in my Shepherding Philosophy I am posting a descriptive record of past usage of NAI. Please comment on this, to tell me of any usage I am missing or any that I have misinterprteted. Thank you.

Currently, this document does not reflect "standard Lojban" but rather Lojban as it is actually used. The difference is small, but I feel that it is important to have a reference for BOTH. The reference for "standard" NAI usage is, of course, the Codex Woldemar.

NAI contains only one cmavo, "nai". Nai negates the previous word, and is usually written as part of a compound with that word. Rather than brivla (which use other selma'o), nai negates cmavo. It is a contradictory negator in almost all cases. However, it can only negate certain cmavo and not others.

First and foremost, NAI is used to negate attitudinals. "mi klama ui" would express happiness about going, while "mi klama uinai" would indicate unhappiness about going. It is used as a contradictory negator of attitudinals. (This parenthetical note is an edit. The foregoing sentence is left in for the historical archive purposes, but it is a mistake and ought to be read as "It is used as a scalar negator of attitudinals." I apologize for the error. - kd) This seems to be one of the more common uses of NAI.

NB: The Woldy seems to think that such compounds as "uicainai" are permitted, with the nai indicating the opposite end of the spectrum. What the opposite end for ru'e, which is not an end of the spectrum at all, would be is unclear. It would appear that the Book intends for CAI NAI to be equivalent to NAI CAI. Usage seems to have decided otherwise; I have been unable to turn up any use of CAI NAI ever, probably because this feature is redundant.

The second use of NAI is to negate cmavo of selma'o BAI. This is fairly common, and is an example of contradictory negation. For instance, "mi klama ki'u lenu broda" means "I came because of the broding." On the other hand, "mi klama ki'unai lenu broda" indicates that "I came despite the broding." Now if only Webster's had an entry for "Broding"...

NAI can also be used to negate jeks. The canonical example of this is the phrase "fange jenai cizra", which is translated as "Foreign but not strange." That is, this use would imply a contradictory negation of the part after the JA.

NB: The Woldy seems to think that JOI NAI is permissible. It would indicate that a different connective applies. Sorry, badyxu'e, but "joinai" has only been used in other languages. BIhI is in the same boat; the book thinks that BIhI+NAI compounds are valid but do not seem to be a feature of Lojban as it is spoken.

Another use is to negate tenses. This includes FAhA and PU. For example, "mi pu klama" indicates that I came in the past. "mi punai klama" indicates coming, but not previous coming. This is not a particularly common use of NAI, but it does occur.

Yet another use is in negating CAhA. This use of NAI is not permitted under baselined Lojban as it currently exists. However, the compound "ka'enai" is one of the twenty most common cmavo compounds. It appears in the translations of Le Petit Prince and of Alice in Wonderland. Even though this usage is currently not part of the baseline, it is undeniably part of the language, rather like "ain't" in certain English dialects. "ka'enai" seems to have the same meaning as the crunchier "na'eka'e". Surprisingly, this is an example of NAI being used for scalar negation; however, the difference in meaning between this and "na ka'e" is not obvious.

In logical connectives, we see such compounds as "ganai". When a logical connective - GA or GI - is negated with NAI, it indicates that if that part of the sentence is negated then the whole assertion becomes true. For instance, "ga mi klama gi broda" indicates that either I go or there is broding, or both. "ganai mi klama gi broda" indicates that either I fail to go or there is broding, or both - in other words, the broding occurs if I go. "ga mi klama ginai broda" would indicate that either I go or there is no broding, so if there is broding then I go.

The above summarizes all of the frequent uses of selma'o NAI that I have been able to uncover. If you have any comments about it, post them here or contact me at ragnarok@pobox.com. If there are no comments, I will assume that everybody is pretty much agreed that I have accurately described what NAI is and isn't; remember though that this a description of NAI in actual use rather than a suggestion for what NAI ought to be.

Once I have published a final version (the above is merely a draft), it is my intent to submit a proposal making it the new standard. This will, of course, be open to amendment should that be seen as necessary.


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject: Is CAhA boolean? This determines CAhA nai.
   username: xod
  post_time: 2003-04-18 19:17:19
  post_text:

If CAhA are boolean, then na'e CAhA = to'e CAhA = CAhA nai = na CAhA, and there is no ambiguity. Lacking ambiguity, I support the use of nai with CAhA.


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject: CAhA Status
   username: kd
  post_time: 2003-04-18 19:39:38
  post_text:

I am not an expert on CAhA. I hope that the Catherd of CAhA figures this out and answers all of our questions. Until then, I am tempted to say yes, I believe it is. I must confess a small amount of bias here, as I favor ka'enai (unlike the no joinai part of my description).

I encourage you to expand upon why CAhA NAI is good iff CAhA is boolean, as most if not all BPFK members have preconceived ideas on ka'enai and other issues and it is doubtful that the BPFK will accomplish anything unless we can put aside these notions and discuss the actual merits of such changes. CAhA NAI so far has about the following argument:

Most Pro-CAhA NAI People: CAhA NAI is more consistent. Most Anti-CAhA NAI People: No it isn't. MPCNP: Yes it is. MACNP: No it isn't. MPCNP: 'Tis. MACNP: No it isn't; it's just contradiction.

An argument is a connected series of statements to establish a definite proposition.


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject:
   username: xod
  post_time: 2003-04-18 20:25:01
  post_text:

For a bit {0, 1}, na'e ("other than 0") = to'e ("opposite of 0") = na ("not true that 0"). All possible negations collapse into one single state, and only one meaning. Which means that there is no ambiguity introduced by using ka'enai. Ambiguity is the only reason I would override the strong usage pattern which has violated what I think was merely a grammatical oversight.


 forum_name: General Negators
topic_title: Shepherding Philosophy
    subject:
   username: nitcion
  post_time: 2003-04-19 09:43:34
  post_text:

You can do your paradigm as you see fit. The rest of the community are free to object, and to call for votes to reject your proposals, if they feel it necessary. I think Let Usage Decide is a crock, but it's not up to me, but up to the votes. As for ignoring the baseline, that's your call. And it's my call to vote against any such proposal, if they feel the deviation from the existing baseline unwarranted. Make very sure you've broken your proposal up into lots of discrete pollable points, because I can see a lot of this coming up.

You all know what my thinking on prescription is (I've issued enough manifestos on it), and what the board has said. But it comes down to what the community wants, not what you or I want; in particular, it comes down to the Commission consensus, in the first instance. Do as you see right, and be prepared for the votes to go against you, as well as for you. (This applies to the most fanatical fundamentalist as well, of course.)


 forum_name: General Negators
topic_title: Discussion of Changes 1: CAI NAI
    subject: Discussion of Changes 1: CAI NAI
   username: kd
  post_time: 2003-04-19 09:45:21
  post_text:

This will be the first of three topics in NAI for the discussion of the changes that would be called for by officializing the descriptive record. Here we will discuss the removal of CAI NAI compounds from the grammar. I do not believe such compounds are harmful in any way, but usage seems to find that they are not necessary. It would also appear that they are redundant.


 forum_name: General Negators
topic_title: Discussion of Changes 2: JOI NAI
    subject: Discussion of Changes 2: JOI NAI
   username: kd
  post_time: 2003-04-19 09:47:25
  post_text:

The book believes that JOI NAI is permissible. I believe that JOI NAI compunds could be useful on occasion in Lojban text, though they are unlikely to be useful in translations. I have not been able to track down any usage of JOI NAI, so officializing the Descriptive Record would mean removing them.


 forum_name: General Negators
topic_title: Discussion of Changes 3: CAhA NAI
    subject: Discussion of Changes 3: CAhA NAI
   username: kd
  post_time: 2003-04-19 09:52:40
  post_text:

This will probably be a big sticking point for most of Lojbanistan. Many people have preconceived notions about whether CAhA NAI compounds should be valid. They are amazingly common in usage. I support allowing that fact to dictate their validity, as I see nothing harmful about them. However, I will keep an open mind about CAhA NAI, and I beg everyone else to also. We must listen to debate, and be willing to learn the Lojbab Lesson. Please, do not support or oppose CAhA NAI without a reason, and certainly not without truly listening to the other side's point of view.

That said, I find it highly unlikely that anyone will be able to point to anything harmful about CAhA NAI. Since usage has decided that it is acceptable, I favor CAhA NAI unless it is actually harmful, and encourage the rest of the community to do so as well. Since we are in a time for fixing what some see as broken, "The BNF doesn't let you do that" is no longer an argument.


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject:
   username: nitcion
  post_time: 2003-04-19 09:57:13
  post_text:

To this, I can only say that changes in the syntax of the language are to be undertaken with reluctance, since the syntax has been the stablest part of the language for a while now (for all its faults). John Cowan has made a case for the ambiguity of CAhA NAI, which you'd need to address. (And if you're not addressing the criticism, but merely pointing out that people do say ka'enai, then you're only doing part of the job.) In any case, you'd need to get at least one of John or Jordan to change their minds, given the consensus requirement.

But over and beyond the semantic arguments for the acceptability or not of CAhA NAI, the case for syntactic conservatism is a reasonable one. The question to decide is whether adding CAhA NAI to the syntax is too much of a change to countenance, and there are opinions on both sides. (I myself have not made up my mind yet; I think there are much worse violations of continuity than this one, but I am more sympathetic to conservatism than not.) This too will be one for the polls, in the first instance.


 forum_name: General Negators
topic_title: Shepherding Philosophy
    subject:
   username: kd
  post_time: 2003-04-19 10:01:26
  post_text:

I do not believe in "ignoring the baseline", as you put it. I do believe that since total lack of prescription won't work right now, we should base our prescription on description. However, that means we need a descriptive record that isn't based on the baseline, and a record of what is OK in the baseline. The latter, of course, is the CLL.

Just because I intend to first propose making the descriptive record official doesn't mean I support that proposal. I don't really care either way about CAI NAI, and I would rather keep JOI NAI. But that isn't my call, it's the community's. And if the community votes to keep doing what it has been doing, either in the sense of making no changes or in the sense of not using JOI NAI, that's up to the Commission consensus.

Lastly, I am not prepared to have votes go against me, because my priorities are different from what I think you think they are. My first priority is for goodness to occur, and if anyone votes against what they think is goodness, I will be shocked. I think fifteen people know goodness better than one, so I will be content with whatever the vote is provided it is discussed thoroughly beforehand, so that nobody votes against something they later decide really would have been goodness. Just because I think JOI and CAhA should both be negatable with NAI doesn't guarantee that those are goodness. If the BPFK doesn't think they are, then the BPFK is probably right.


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject:
   username: kd
  post_time: 2003-04-19 10:08:54
  post_text:

As xod points out, John's argument that ka'enai is ambiguous is absolutely correct iff CAhA is non-boolean. I think CAhA is boolean, because I can't imagine how it would work otherwise - either you can or you can't. Nick, you are right, merely pointing out that people do say ka'enai isn't enough. This is why the question is here of the booleanity of CAhA - it determines whether or not there is an answer to the most common objection other than the non-baseline status of CAhA NAI (which is subject to change) and therefore whether CAhA NAI is harmful. I believe that CAhA NAI is a harmless addition, but I cannot claim to be an expert on CAhA.


 forum_name: General Negators
topic_title: Descriptive Record
    subject:
   username: nitcion
  post_time: 2003-04-19 10:11:04
  post_text:

Well.

I'll respond more to the topic as you're currently breaking it up, but:

(a) That UI CAI NAI has never been used is not enough reason to make it ungrammatical. We change the grammar because it is broken, not because Lojbanists to date have underused it. (They've underused everything.) UI CAI NAI is not strictly speaking the same as UI NAI CAI (even if the distinction is hairsplitting), and I see no justification for nuking it from the grammar. That constitutes the kind of change I do not believe the BPFK is authorised to make.

uinai seems to me pretty overtly scalar, btw.

And oh, there is a documented instance of CAI NAI, from '95. Granted, it was generated by a random sentence generator :-) , but you should still have mentioned it. :-1/2

(b) Same goes for JOI NAI and BIhI NAI; they may or may not be useless, but the fact that they haven't been used yet doesn't tell me much, given that we are only starting to use Lojban. It is a prescribed feature of Lojban, and I see no reason to undo it unless it leaves to manifest confusion.

(c) There is a case for CAhA NAI (analogy), there is a case against it. Let the battle begin.

(d) If there are no comments, I will assume that everyone is agreed? Uh uh. Only the poll gets to say.

Craig, we will have nasty fights about this, but I do appreciate you doing this nonetheless.


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject:
   username: nitcion
  post_time: 2003-04-19 10:20:07
  post_text:

Agreed. If the issue of CAhA NAI *depends* on the issue of whether CAhA admits of scalar negation (is scalar or categorical), then it's appropriate, as you've hinted, to defer the decision to what gets decided about CAhA.

Now there is something I think you can do within the ambit of NAI, and which is the kind of thing I'd expect the shepherd to either do themselves, or get someone else to. That is,

  • get all the cmavo NAI can combine with (yes, that does mean you have to be able to read BNF --- and Craig, I think that *is* a fair requirement);
  • find what prescription and usage say aboput whether NAI in each instance is scalar or contradictory;
  • make rulings for the rest (where there is no prescription, or no usage, or their testimony is inconclusive), based on the patterns established.

I know the latter in particular is not what you want to do. That doesn't mean it won't happen. :-) And whether or not you want to prescribe, telling us what has already happened --- not just with the four contexts you pointed out --- is a valuable service to Lojban. (Though I do believe NAI is scalar with attitudinals, not contradictory.)


 forum_name: General Negators
topic_title: Descriptive Record
    subject:
   username: kd
  post_time: 2003-04-19 10:42:05
  post_text:

a) I agree that ungrammaticalizing things is a bad idea. I had not, when I wrote my shepherding philosophy, expected that things would be not just underused but used only by random sentence generators. As for uinai being scalar, it was a typo. I'll fix it in a moment.

b) Again, I agree. I also think JOI NAI could be useful, whereas CAI NAI is reasonably redundant.

c) Bring it on. ;)

d) Only the poll gets to say what will be done. Lack of differing opinion gets to say that the descriptive record accurately reflects usage, which was what I meant.


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject:
   username: kd
  post_time: 2003-04-19 10:44:41
  post_text:

I absolutely agree that it is fair to require me to read BNF. I am actually just starting to learn. It is not fair to require me to already be able to read it, but it would not be reasonable for a Shepherd to continue to be unable to understand the formal grammar.


 forum_name: General Negators
topic_title: Discussion of Changes 1: CAI NAI
    subject: Discussion of Changes 1: CAI NAI
   username: Anonymous
  post_time: 2003-04-19 12:52:22
  post_text:

Craig:

Here we will discuss the removal of CAI NAI compounds from the grammar. I do not believe such compounds are harmful in any way, but usage seems to find that they are not necessary. It would also appear that they are redundant.

Neither redundancy, or lack of necessity is justification to change the grammar. Lack of usage means relatively little at this stage, because large areas of the language have seen little usage, and some of them may be areas which would open up distinctions.


 forum_name: General Negators
topic_title: Discussion of Changes 3: CAhA NAI
    subject: Post subject: Discussion of Changes 3: CAhA NAI
   username: Anonymous
  post_time: 2003-04-19 13:14:51
  post_text:

Craig:

Since we are in a time for fixing what some see as broken, "The BNF doesn't let you do that" is no longer an argument.

We are in a time for defining the language, wherein fixing what is definitely broken is something which can be considered, and in which "The YACC doesn't let you do that" (the BNF is secondary to the YACC grammar) is a perfectly valid argument, although it is one which can be countered by near-unanimous consensus (but you'd better make sure that the proposed alternative validly YACCs).

I'm not going to vote against all proposed changes, but the existing prescription remains the presumed correct one, insofar as it is understood.


 forum_name: General Negators
topic_title: Descriptive Record
    subject: Descriptive Record
   username: Anonymous
  post_time: 2003-04-19 13:46:22
  post_text:

Could I request that you break up this descriptive record into separate chunks for each of the categories you envision. I see 3 or 4 topics of X+NAI treated sequentially, but I'm reluctant to try to address all of them in one topic on the forum, and it seems like people are opening multiple new topics which is likely even more confusing.

My suggestion: There be a summary topic wherein we collect a list (with minimal annotation) of all of the usages of NAI that need to be defined and/or decided, and then discuss each of these usages as a separate topic before assembling a definition on the skeleton of the summary topic.

If there were a summary topic, then IMHO, an early step would be to go thorough and list all YACC grammar rules that include NAI. And then, since NAI also has a freeform grammar, systematically consider combinations of NAI with each of the other places wherein NAI might possibly occur. Thus we would hit upon the Y+NAI that showed up on Lojban List, which is not currently included in your discussion.

Since we do not have a complete corpus of Lojban text, the descriptive record is inherently incomplete if it is only based on the corpus on file. We need to specifically look for other usages besides the ones that are common in usage, and in some cases they may emerge in text not yet part of any official corpus (or of course may start to emerge simply because the question is asked).

"Let usage decide" works well only when there is a LOT of usage that comprehensively explores the possibilities of the language. Something merely not-yet being used does NOT decide or even indicate that it is useless, because a lot of the language hasn't been explored, and some parts of the language were inserted because they were thought to possibly be more likely useful for non-English users who look at language a different way (and we hardly have a heavy amount of non-English native usage, except for Jorge).

Something that is used in a certain way, but then usage BY THE SAME PERSON evolves away from it towards an alternative in that kind of context DOES indicate that "usage has decided".


 forum_name: General Negators
topic_title: Discussion of Changes 1: CAI NAI
    subject:
   username: kd
  post_time: 2003-04-19 14:43:29
  post_text:

Absolutely. I had posted my Shepherding Philosophy before discovering certain things (JOI NAI, CAI NAI) were missing from all usage. I still intend to submit a proposal which includes removing them, but I intend to vote against said proposal.


 forum_name: General Negators
topic_title: YACC Rules With NAI In Them
    subject: YACC Rules With NAI In Them
   username: kd
  post_time: 2003-04-19 14:59:33
  post_text:

I admit that I still don't understand all of what this means, but it will be useful both to me once I do and to others who already do.

%token NAI_581 /* attached to words to negate them */

text_0  : text_A_1

                       |  indicators_411  text_A_1
                       |  free_modifier_32  text_A_1
                       |  cmene_404  text_A_1
                       |  indicators_411  free_modifier_32  text_A_1
                       |  NAI_581  text_0
                       ;

indicator_413  : UI_612

                       |  CAI_515
                       |  UI_612  NAI_581
                       |  CAI_515  NAI_581
                       |  Y_619
                       |  DAhO_524
                       |  FUhO_536
                       ;

COI_A_417  : COI_520

                       |  COI_520  NAI_581
                       ;

NU_A_426  : NU_586

                       |  NU_586  NAI_581
                       |  NU_586  free_modifier_32
                       |  NU_586  NAI_581  free_modifier_32
                       ;

EK_root_911  : A_501

                       |  SE_596  A_501
                       |  NA_578  A_501
                       |  A_501  NAI_581
                       |  SE_596  A_501  NAI_581
                       |  NA_578  A_501  NAI_581
                       |  NA_578  SE_596  A_501
                       |  NA_578  SE_596  A_501  NAI_581
                       ;

JEK_root_926  : JA_546

                       |  JA_546  NAI_581
                       |  NA_578  JA_546
                       |  NA_578  JA_546  NAI_581
                       |  SE_596  JA_546
                       |  SE_596  JA_546  NAI_581
                       |  NA_578  SE_596  JA_546
                       |  NA_578  SE_596  JA_546  NAI_581
                       ;

JOIK_root_931  : JOI_548

                       |  JOI_548  NAI_581
                       |  SE_596  JOI_548
                       |  SE_596  JOI_548  NAI_581
                       |  interval_932
                       |  GAhO_656  interval_932  GAhO_656
                       ;

lexer_G_935  : lexer_G_707 GA_537

                       |  lexer_G_707  SE_596  GA_537
                       |  lexer_G_707  GA_537  NAI_581
                       |  lexer_G_707  SE_596  GA_537  NAI_581
                       |  lexer_G_707  simple_tag_971  GIK_root_981
                       |  lexer_G_707  JOIK_root_931  GI_539
                       ;

lexer_H_940  : lexer_H_708 GUhA_544

                       |  lexer_H_708  SE_596  GUhA_544
                       |  lexer_H_708  GUhA_544  NAI_581
                       |  lexer_H_708  SE_596  GUhA_544  NAI_581
                       ;

modal_974  : modal_A_975

                       |  modal_A_975  NAI_581
                       ;

GIK_root_981  : GI_539

                       |  GI_539  NAI_581
                       ;

GIhEK_root_991  : GIhA_541

                       |  SE_596  GIhA_541
                       |  NA_578  GIhA_541
                       |  GIhA_541  NAI_581
                       |  SE_596  GIhA_541  NAI_581
                       |  NA_578  GIhA_541  NAI_581
                       |  NA_578  SE_596  GIhA_541
                       |  NA_578  SE_596  GIhA_541  NAI_581
                       ;

time_direction_1035  : PU_592

                       |  PU_592  NAI_581
                       ;

space_direction_1048  : FAhA_528

                       |  FAhA_528  NAI_581
                       ;

interval_property_1051  : number_root_961 ROI_594

                       |  number_root_961  ROI_594  NAI_581
                       |  TAhE_604
                       |  TAhE_604  NAI_581
                       |  ZAhO_621
                       |  ZAhO_621  NAI_581
                       ;

token_1100  : any_word_698

                       |  BAhE_503  any_word_698
                       |  anything_699
                       |  any_word_698  BU_511
                       |  any_word_698  DAhO_524
                       |  any_word_698  FUhO_536
                       |  any_word_698  FUhE_535
                       |  any_word_698  UI_612
                       |  any_word_698  UI_612  NAI_581
                       |  any_word_698  Y_619
                       |  any_word_698  CAI_515
                       |  any_word_698  CAI_515  NAI_581
                       |  UI_612  NAI_581
                       |  CAI_515  NAI_581
                       ;

 forum_name: General Negators
topic_title: YACC Rules With NAI In Them
    subject: YACC Rules With NAI In Them
   username: lojbab
  post_time: 2003-04-20 13:39:09
  post_text:
I admit that I still don't understand all of what this means, but it will be useful both to me once I do and to others who already do.

%token NAI_581 /* attached to words to negate them */

Just defining that NAI exists.

text_0 : text_A_1

      | indicators_411 text_A_1
      | free_modifier_32 text_A_1
      | cmene_404 text_A_1
      | indicators_411 free_modifier_32 text_A_1
      | NAI_581 text_0
      ; 

Presuming that the problem for you is not knowing the format:

Translating to EBNF without simplifying (I take out the numbers because it makes it easier to read, but they are just ordering to find them easier):

text : [text_a] | [indicators text] | [free_modifier text_a]

    | [cmene text_a] | [indicators free_modifier text_a]
    | [NAI text]

In other words, YACC is pretty much the same thing as EBNF, but not simplified to remove duplicate appearances (such as the multiple text_a rules), and with each alternative written on a separate line. (I can't read EBNF easily because the alternatives are crammed together.)

In the above rule, it merely says that NAI can stand by itself at the beginning of any text string.

indicator_413 : UI_612

             | CAI_515
             | UI_612 NAI_581
             | CAI_515 NAI_581
             | Y_619
             | DAhO_524
             | FUhO_536
             ; 

Says that NAI can appear after UI or CAI. (Not clear whether defining these combinations is part of UI/CAI paradigm or NAI paradigm.)

COI_A_417 : COI_520

         | COI_520 NAI_581
         ; 

Says that NAI can appear after any of the members of COI (not clear whether defining these is part of COI paradigm or NAI paradigm)

NU_A_426 : NU_586

        | NU_586 NAI_581
        | NU_586 free_modifier_32
        | NU_586 NAI_581 free_modifier_32
        ; 

Says that NAI can appear after NU (with or without a free-modifier following it). Probably hasn't seen much usage. (Again, not clear whether defining the compounds is part of NU or part of NAI.)

EK_root_911 : A_501

           | SE_596 A_501
           | NA_578 A_501
           | A_501 NAI_581
           | SE_596 A_501 NAI_581
           | NA_578 A_501 NAI_581
           | NA_578 SE_596 A_501
           | NA_578 SE_596 A_501 NAI_581
           ;

JEK_root_926 : JA_546

            | JA_546 NAI_581
            | NA_578 JA_546
            | NA_578 JA_546 NAI_581
            | SE_596 JA_546
            | SE_596 JA_546 NAI_581
            | NA_578 SE_596 JA_546
            | NA_578 SE_596 JA_546 NAI_581
            ;

lexer_H_940 : lexer_H_708 GUhA_544

           | lexer_H_708 SE_596 GUhA_544
           | lexer_H_708 GUhA_544 NAI_581
           | lexer_H_708 SE_596 GUhA_544 NAI_581
           ;

GIK_root_981 : GI_539

            | GI_539 NAI_581
            ;

GIhEK_root_991 : GIhA_541

              | SE_596 GIhA_541
              | NA_578 GIhA_541
              | GIhA_541 NAI_581
              | SE_596 GIhA_541 NAI_581
              | NA_578 GIhA_541 NAI_581
              | NA_578 SE_596 GIhA_541
              | NA_578 SE_596 GIhA_541 NAI_581
              ;

lexer_G_935 : lexer_G_707 GA_537

           | lexer_G_707 SE_596 GA_537
           | lexer_G_707 GA_537 NAI_581
           | lexer_G_707 SE_596 GA_537 NAI_581
           | lexer_G_707 simple_tag_971 GIK_root_981
           | lexer_G_707 JOIK_root_931 GI_539
           ;

JOIK_root_931 : JOI_548

             | JOI_548 NAI_581
             | SE_596 JOI_548
             | SE_596 JOI_548 NAI_581
             | interval_932
             | GAhO_656 interval_932 GAhO_656
             ;

Says that NAI can appear after any of the logical and non logical connectives A, JA, JOI, GA, GIhA, GUhA, as well as GI. In the non-logical connectives, it does not appear after any of the "interval" constructs, only after JOI.

modal_974 : modal_A_975

         | modal_A_975 NAI_581
         ; 

Says that NAI can appear after any modal.

time_direction_1035 : PU_592

                   | PU_592 NAI_581
                   ;

space_direction_1048 : FAhA_528

                    | FAhA_528 NAI_581
                    ; 


Says that NAI can appear after any simple tense (PU or FAhA). Again, the compounds may need to be defined as part of those paradigms.

interval_property_1051 : number_root_961 ROI_594

                      | number_root_961 ROI_594 NAI_581
                      | TAhE_604
                      | TAhE_604 NAI_581
                      | ZAhO_621
                      | ZAhO_621 NAI_581
                      ; 

Says that NAI can appear after ZAhO, or TAhE, or a number+ROI construct. Again, compounds may need to be defined as part of those paradigms.


token_1100 : any_word_698

          | BAhE_503 any_word_698
          | anything_699
          | any_word_698 BU_511
          | any_word_698 DAhO_524
          | any_word_698 FUhO_536
          | any_word_698 FUhE_535
          | any_word_698 UI_612
          | any_word_698 UI_612 NAI_581
          | any_word_698 Y_619
          | any_word_698 CAI_515
          | any_word_698 CAI_515 NAI_581
          | UI_612 NAI_581
          | CAI_515 NAI_581
          ; 

This merely defines the stuff that supposedly is handled before the parser operates. It repeats the "Indicator" definition, because NAI in combination with CAI or UI as an indicator can occur standing alone, or after any other Lojban word, and not merely in those places where indicators are explicitly referenced in the rest of the grammar.

Bottom line: It matches most of what you said in the descriptive summary, but there are several constructs that you have not seen used. (Whether they have been used is unclear.)

Of these, the only one I'm not sure I can explain is the NAI at the beginning of text, but I suspect that is a fossil of when NAI as a standalone was a valid attitudinal. Most of the others which you haven't discussed are using NAI like NAhE as opposed to NA.


 forum_name: General Negators
topic_title: YACC Rules With NAI In Them
    subject: Standalone Nai
   username: kd
  post_time: 2003-04-20 14:55:44
  post_text:

Since standalone NAI is not even mentioned in the book, and I have not seen any usage of it, it did not make my descriptive record. I will therefore suggest removing it in my first proposal, and unlike other things that simply "haven't been explored" and should be left in, I actually think that if nobody knows what standalone NAI means and there isn't any documentation to tell us, it truly ought to be removed from the grammar.


 forum_name: General Negators
topic_title: YACC Rules With NAI In Them
    subject:
   username: xorxes
  post_time: 2003-04-21 12:07:53
  post_text:

Standalone nai can be used as a response to a pei question.

Would this be the correct place to discuss the experimental cmavo ja'ai, of selma'o NAI?

How about the proposal to include nai in selma'o UI, thus extending (and simplifying) rather than restricting its grammar?


 forum_name: General Negators
topic_title: YACC Rules With NAI In Them
    subject:
   username: kd
  post_time: 2003-04-21 16:18:26
  post_text:

I should have thought of pei -> standalone nai. Duh. Thanks.

I am not familiar with ja'ai. If you can provide documentation of a) what it is and preferably also b) where it has seen usage, that would be very helpful. But yes, I think this is the place for it.

As for other proposals, let's discuss ideas on a different topic. Basically, I plan to first submit a proposal making the baseline match usage, just as a starting point, and then amend that. I would be against adding any restrictions to the grammar that don't eliminate actual sources of confusion. I am currently againt moving nai to UI, but I could probably be convinced. Give it its own forum topic, as this is just a place for discussing the YACC rules.


 forum_name: General Negators
topic_title: YACC Rules With NAI In Them
    subject:
   username: xorxes
  post_time: 2003-04-21 17:03:56
  post_text:

ja'ai:nai::ja'a:na

It was proposed by Mark Shoulson, see for example:

http://groups.yahoo.com/group/lojban/message/1144 http://groups.yahoo.com/group/lojban/message/1164 http://groups.yahoo.com/group/lojban/message/3994

And here is another example of use:

http://groups.yahoo.com/group/lojban/message/4091


 forum_name: General Negators
topic_title: ja'ai
    subject: ja'ai
   username: kd
  post_time: 2003-04-21 18:45:01
  post_text:

Xorxes had wanted to use the NAI forum to discuss ja'ai, an experimental cmavo in selma'o NAI. It is affirmative rather than negative, rather like ja'a. This obviously needs its own topic, so this would be the space for that.

BTW, if ja'ai were to become an official part of the baseline, it should probably be given a non-experimental form. I would vote for ja'u, since it has the same first two vowels as ja'a. This doesn't mean I support adding it at all; I'm still undecided on that one, having yet to see a compelling argument either way.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: Making nai a UI?
   username: kd
  post_time: 2003-04-21 18:47:30
  post_text:
Xorxes:
How about the proposal to include nai in selma'o UI, thus extending (and simplifying) rather than restricting its grammar?
This clearly shouldn't be discussed in the list of YACC rules topic. Xorxes, please explain why this would be enough of a Good Thing to merit a substantial baseline change.
 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject: What John said
   username: xod
  post_time: 2003-04-22 10:50:33
  post_text:

In email:

xod: I hear that you argued somewhere that CAhA is not boolean. Would you care to join the thread and restate your argument, or point us to the place where yoy already posted it? Thanks!

John: I did not argue that, but I agree that CAhA+NAI is only truly ambiguous if it is so. My argument was rather against the supposed intuitiveness of CAhA+NAI, given that NAI is sometimes scalar, sometimes contradictory.

xod: Do you support CAhA NAI if we determine that the 4 CAhA are boolean?

John: No. I think that NAI is a hack, and I don't support extending it in any way. Indeed, I would like (were it possible) to get rid of some uses

like NU+NAI.

 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-22 11:54:50
  post_text:

In a sense this _is_ related to the list of YACC rules topic. If {nai} is moved to UI, the list vanishes. You just have to learn that {nai} changes the meaning of the previous word to some kind of opposite (which opposite will depend on the word, but that is already the case now). If some words don't have an obvious opposite meaning then the use of {nai} with them won't be obvious, that's all. That doesn't mean that they won't have a useful meaning in some context. The existing situation, where you have to learn which words are nai-able and which aren't, seems to be unnecessarily complicated. I don't know how "substantial" this change would be, but it seems to me that allowing or disallowing {nai} with particular selma'o is extremely arbitrary, so if we are going to make a change why not go for a simpler and less arbitrary grammar instead of adding more patches?


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: kd
  post_time: 2003-04-22 21:29:18
  post_text:

It is related, but at the same it belongs in its own topic. Anyway, if nai became a UI then the following statements might occur. What do they mean?

1 nai mi broda 2 mi nai broda 3 mi broda nai 4 zo'e nai broda 5 nai 6 se nai broda 7 mi fe nai do broda 8 lo nai broda 9 no nai broda

None of these make any sense now but all of them become meaningful under the "nai joins UI" proposal. So what meaning do they carry?


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-23 12:27:24
  post_text:
1 nai mi broda

It is already grammatical. I would interpret it as {na ku mi broda}.

2 mi nai broda

In some context, {mi nai} could refer to my double, the "me" that is not me. In general, perhaps just {na'e bo mi}.

3 mi broda nai

Probably {mi na'e broda}.

4 zo'e nai broda

{zo'e nai} could indicate a non-obvious referent.

5 nai

Already grammatical.

6 se nai broda

Perhaps an emphasis that places are not being rearranged.

7 mi fe nai do broda

"Other than the second place"?

8 lo nai broda

Nothing useful I can think of at the moment, which is not to say that some useful meaning cannot be found.

9 no nai broda

"Other than zero".


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: kd
  post_time: 2003-04-23 15:20:55
  post_text:

1 is already grammatical, but except if asked a question like (do brode xu pei) it is unlikely to carry any meaning under the present baseline.

2 Ok, I suppose that works.

3 Yeah, that makes sense too. Why na'e as oppose to na, by the way? Not that I care which it is, just playing devil's advocate.

4 But it is KOhA, so it ought to be like "minai". Thus na'ebo zo'e. Which I can't think of a meaning for.

5 Already grammatical, but meaningless without a prior pei. What meaning should it carry?

6 Ok, but that seems like a bit of a stretch. Like, "let's make nai a UI and I'll think of a meaning for that."

7 As I say, it suddenly becomes something people say. I've said "fe ui" before; fenai seems rather useless.

8 And then someone uses it because they've found a use, so you have to stop, revert to your native Spanish, and ask them. The need to do this now is why there is a BPFK.

9 Ok, that makes sense. How about:

10 da nai broda


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-23 16:28:33
  post_text:
1 is already grammatical, but except if asked a question like (do brode xu pei) it is unlikely to carry any meaning under the present baseline.

But that's not a problem we need to address here, since it is not related to the moving of {nai} to UI.

3 Yeah, that makes sense too. Why na'e as oppose to na, by the way? Not that I care which it is, just playing devil's advocate.

I see {nai} as modifying the word it attaches to, so it makes more sense to say that it changes the brivla rather than the whole bridi of which the brivla is a selbri. {i nai}, on the other hand, would negate the bridi.

4 But it is KOhA, so it ought to be like "minai". Thus na'ebo zo'e. Which I can't think of a meaning for.

If you can't think of a meaning for it, would you make it ungrammatical? But {na'e bo zo'e} is simply "other than the obvious value", which is pretty much what I suggested.

5 Already grammatical, but meaningless without a prior pei. What meaning should it carry?

Not affected by this proposal.

6 Ok, but that seems like a bit of a stretch. Like, "let's make nai a UI and I'll think of a meaning for that."

Lots of things in Lojban were "discovered" this way.

7 As I say, it suddenly becomes something people say. I've said "fe ui" before; fenai seems rather useless.

Yes, so it won't be used much, but that does not mean it has to be ungrammatical, which only serves to make the grammar more difficult to learn.

8 And then someone uses it because they've found a use, so you have to stop, revert to your native Spanish, and ask them. The need to do this now is why there is a BPFK.

Why couldn't I ask in Lojban? Anyway, if the goal of the BPFK is to find a meaning for every conceivable grammatical string, we will finish sometime in the year 3503, give or take.

10 da nai broda

na'e bo da broda


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: kd
  post_time: 2003-04-23 16:44:27
  post_text:

6 Yes, lots of things have been "discovered". The BPFK is here to stop that from being necessary.

8 I do not claim our job is to find a meaning for all grammatical strings. But if nobody is able to find one (we don't have to do it, but it should be doable) then that string is broken. The BPFK is here to make sure nothing is broken.

As for whether we need to address the meaning of the ones that are already grammatical, certainly. As you pointed out, standalone nai is grammatical in response to peis. The example I used in 1, btw, is a typo - the pei there should have read uipei. Anyway, if they would suddenly be expected to have meaning without that, then absolutely we have to say what it is.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-23 17:51:33
  post_text:

{lo nai broda} can mean {na'e bo lo broda}, because attaching {nai} to {lo} is the way to attach to the whole sumti. Another way would be {lo broda ku nai}. That's how it works for other UIs: {lo da'i broda}, for example.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: We should maintain the status quo
   username: arj
  post_time: 2003-04-24 09:22:11
  post_text:

I do not think that NAI should not be moved to selma'o UI.

These are some of the reasons why:

According to the Fregean principle of compositionality, doing this would require us to define some kind of semantic content that is independent of context, AND a rule that can combine nai with the word or phrase that is modified, UI-style. I don't think anyone here knows what kind of semantic content that would be, in any more detail than "the modified word/phrase is negated somehow".

The selma'o CAhA deals with modalities, hence with the truth-value of the predication, and it is therefore important that negation works in a well-defined way, which it apparently does in the current language. The selma'o UI, on the other hand, is much more loosely connected to truth-values.

According to the draft version[1] of the CLL, there are two kinds of UI, one which consists of "pure emotion indicators" (UI1), and one which establishes"an internal, hypothetical world which the speaker is reacting to, distinct from the world as it really is." (UI1) One of the arguments for the proposal to put NAI into UI, was to enable cmavo compounds like "ka'enai". However, since this is supposed to equal "na'e ka'e", it fails both semantic criteria for the selma'o UI: it does affect the truth value of the bridi, and it does not establish any hypothetical world.

The argument that the speakers of Lojban will have to remember which words are nai-able, and which are not, I consider a red herring. Lojban speakers already have to remember which words can take le, xi, or ni'u.

Finally, the reason we are considering this at all, is to save one syllable ("ka'enai" over "na'eka'e"). I don't think it is worth our efforth to shorten something that is already sayable.

--- [1] I haven't received my print version yet.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-24 10:37:14
  post_text:

The current rule for nai is: "the modified word is negated somehow". This proposal does not change that. The only change is that instead of having an arbitrarily selected set of nai-able words, every word becomes nai-able.

The classification of UIs is irrelevant here. I don't think {kau} or {xu} for example fit in any of the given categories, but it doesn't matter. We could even maintain nai in its own selma'o NAI, the important point is not the name of the selma'o but its distribution: it should be allowed to modify every word instead of an arbitrarily selected set of words. UI and CAI have different names but are the same thing in terms of their distribution, so we can propose to move {nai} to CAI instead, if that makes people more comfortable. Y, DAhO and FUhO have three different names but are the same thing in terms of distribution (and would also merge with UI/CAI if this proposal is adopted).

It is no red herring to say that moving nai to UI (or equivalently extending the distribution of NAI to that of UI) simplifies the grammar. It certainly reduces the number of rules to learn. If le, xi or ni'u are arbitrarily restricted, then those arbitrary restrictions should also be removed. Are they?

I don't think ka'enai is the reason I'm considering this at all. It just happens to be the case where extending nai was found most obviously useful, but my main concern is the simplification of an overcomplicated grammar.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: kd
  post_time: 2003-04-24 16:33:50
  post_text:

I'm generally with you on this, tsali, though my mind is far from made up. However, I must take objection to your claim that the purpose of allowing ka'enai is to shorten na'eka'e by a syllable. This would be specially clear if you considered that ka'e is probably boolean, and if so na'eka'e = na ka'e, and yet some people argue against ka'enai iff ka'e is nonboolean. No, there are two reasons cited. Some people like ka'enai out of a perception that it is more consistent. I'm not sure what it is supposed to be more consistent with, but maybe if I knew I would agree. I am in favor of ka'enai for two reasons. One is that I think the language should follow the philosophy followed in other areas - if something does not cause confusion or interfere with logic, allow it. Ka'enai fits this criterion. The big reason, though, is that usage has already decided that ka'enai is acceptable. The BPFK declaring that it still isn't won't change that; why not make the official grammar reflect the actual grammar on areas where it causes no harm? If ka'enai were harmful, all of these arguments would go away instantly. But it isn't, and although there has been a lot of debate about the 'consistancy' claim, nobody has yet answered my reasons for supporting ka'enai except in a nonsubstantive way (eg calling ka'enai an 'abomination' or calling all of selma'o NAI - which is in the official grammar already, dammit - a 'hack'). So if you have an opinion on those that leads you to oppose ka'enai, then you ought to tell the community. And if you didn't see any reason why my arguments are worng, then you would support ka'enai (because any argument that adding it is harmful would be an argument against my arguments); so I conclude that you have some reason not to want greater flexibility and a realistic parser. Please share that with us, since our understanding your point of view will lead to more productive debate. </rant>


 forum_name: General Negators
topic_title: Discussion of Changes 2: JOI NAI
    subject:
   username: arj
  post_time: 2003-04-25 17:13:12
  post_text:

I must agree with Nick's comment in the Descriptive Record topic that underusage is not a good enough reason to make something ungrammatical. Our mandate is only to fix what is broken or underdescribed, not to make The Perfect Loglan.

And, I hope that your comment that officializing the descriptive record would mean removing JOI+NAI, is a misunderstanding. Since it is a descriptive record, it only says what the current Lojban definition is, and how usage differs from it. Prescription comes later, and then presumably in the form of which YACC rules should be removed/added/changed, and a fairly formal description of any changed semantics.


 forum_name: General Negators
topic_title: Discussion of Changes 2: JOI NAI
    subject:
   username: kd
  post_time: 2003-04-25 19:03:57
  post_text:

First of all, I've been rethinking the bit about actually submitting something we know nobody will like. That's just silly.

Second of all, the "officializing the descriptive record" phrase is shorthand for "making an official prescription that usage continue to be what it already is."


 forum_name: General Negators
topic_title: Discussion of Changes 2: JOI NAI
    subject: Re: Discussion of Changes 2: JOI NAI
   username: phma
  post_time: 2003-04-25 19:35:00
  post_text:
The book believes that JOI NAI is permissible. I believe that JOI NAI compunds could be useful on occasion in Lojban text, though they are unlikely to be useful in translations. I have not been able to track down any usage of JOI NAI, so officializing the Descriptive Record would mean removing them.

I think {jo'enai} and {ku'anai} could be useful. For instance, la muremoi klaji jo'u la somoi dargu is the intersection of 52nd Street and 9th Road; la muremoi klaji jo'unai la somoi dargu is all of 52nd Street except the intersection with 9th Road.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: kau is already in UI
   username: phma
  post_time: 2003-04-25 19:49:59
  post_text:

{kau} is in UI, even though it has a fairly narrow set of usage: it is meaningless in a simple sentence and is used mostly following a question word in a subordinate clause. So I don't see why putting {nai} in UI is going to cause problems.

I see {ka'enai} as equivalent to {na ka'e}, not {na'e ka'e}, which means some other member of CAhA applies. The usefulness I see for {ka'enai} is in conjunction with another member of CAhA, e.g. {pu'ijeca'anai}, "can and has but isn't". I think the reason people say {ka'enai} is that in English it's a contraction, and we're used to that order.


 forum_name: General Negators
topic_title: ka'enai survey
    subject: ka'enai survey
   username: kd
  post_time: 2003-04-25 20:48:08
  post_text:

I am creating this survey - NOT an official vote; those happen on the Twiki and certainly not for a while - to guage current opinions on CAhA NAI. Please express your view. I am assuming that everyone's views are subject to change if anything new is brought up in discussion.

If the poll is broken, disregard this message. (But, he edits in, it isn't.)


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-26 12:26:18
  post_text:

I think the idea is that the other members of KAhE (ca'a, pu'i and nu'o) all require ka'e to apply. You can't have one of them applying if ka'e doesn't. The only one that could apply if ka'e doesn't is... ka'enai.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: nitcion
  post_time: 2003-04-27 09:53:23
  post_text:

At this point, I'd like to remind you (or state, if I haven't done so already) that we're not really meant to state new arguments here if there's a chance they're repeating old arguments. (This is following the djez dictum: if you have an argument, relive it through the archives.) As part of his record, Craig will be posting links to the previous iterations of this argument in the archives.

It is true that NAI as a UI would not break much more than {kau} in UI currently does (though {kau} *is* on record as a hack), that it would simplify the grammar tremendously, and that meaning for the new NAI combinations could be found. It is also true that this is a significant change, and while I won't make a ruling as to whether it is within bounds, I don't think I need to: there's enough conservatism voiced here.

(I say this, even though I fully agree that the grammar is often dog droppings, and is dog droppings in this case too; I nonetheless don't think we should be messing with it any more, except in case of manifest brokenness. I argue this vehemently for NAI=UI; I argue this lukewarmly for CAhA NAI, but I nonetheless think this a valid stance. John did say in the board meeting that we should not be conservative for conservativeness' sake; I say it depends, and I want to be more conservative about the grammar than the semantics.)

(The consistency I've plead for though, for the record, as supporting CAhA NAI, is allowing CAhA to be negated like the other tenses. I believe that is the underlying analogy.)

One thing that will help tremendously, and which I think is part of our job here, is to establish whether there is a discernible rule as to when NAI is contradictory, and when it is scalar, in its existing occurrences. Craig, that would mean surveying all usage for the more unusual combinations --- other than UI and the logical connectives. (This would include ka'enai, in fact, qua usage. After all, we've just seen a debate here on whether ka'enai would be scalar or contradictory --- and this was John's stumbling block.) If we knew NAI tended one way or the other by default, or that there were identifiable factors at work, we could make a more informed decision.


 forum_name: General Negators
topic_title: Discussion of Changes 3: CAhA NAI
    subject:
   username: nitcion
  post_time: 2003-04-27 10:00:35
  post_text:

I agree with Bob, more or less. This is a weighing game: usage, logic, and the existing prescription all carry different weights for each of us, although the official priority is prescription first, then usage, then logic (but "design principles" above all of them -- with unambiguity, but not elegance, a design principle.)

But some stats on ka'enai would sure help. Craig, please provide us with relative counts for ka'enai and na'e ka'e/na ka'e, in existing Lojban text, and by period and author. Then we'll know how widespread it truly is.

If we establish a well defined semantics for CAhA NAI, and the relative frequency of ka'enai justifies it, then CAhA NAI certainly cheeses me off a lot less than would UI=NAI: I could live with that. But I need more data for an informed decision.


 forum_name: General Negators
topic_title: Discussion of Changes 2: JOI NAI
    subject:
   username: nitcion
  post_time: 2003-04-27 10:02:45
  post_text:

Supplication call. If CLL provides no justification for JOI NAI, I ask it of John, Nora or Bob.

If CLL *does* provide a justification, then of course it must be included in the record under preparation...


 forum_name: General Negators
topic_title: Is CAhA boolean? This determines CAhA nai.
    subject:
   username: nitcion
  post_time: 2003-04-27 10:09:51
  post_text:

ki'e xod.

If we established that NAI is only scalar with UI (or something of the sort), this would become a non-issue. And such a verdict is certainly within our ambit.

NU NAI is bizarre, and I certainly think Lojban Central has some 'splainin' to do about what they intended with it. (Note that John implies we can't get rid of it. In theory we can --- and that is the line Craig would presumably take --- and I will not issue a ruling on this for now. Further discussion of whether we can intervene with the grammar, if it is really necessary, goes to the Meta BPFK forum. But the BPFK has a voluminously stated policy, and I'd rather not hash it out again; let's rather leave this for the vote.)

A Founder is entitled to consider a core feature of the language a hack; that doesn't decide what we do with it, of course, but it is in the interest of the language to un-hack it as much as possible. How drastically we do so it to be determined.


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: nitcion
  post_time: 2003-04-27 10:17:23
  post_text:

Though I haven't mentioned experimental cmavo much lately, proposals such as ja'ai should indeed be subject to polling --- and thus, linked to (though not necessarily approved of) from the description page the shepherd prepares


 forum_name: General Negators
topic_title: Discussion of Changes 1: CAI NAI
    subject:
   username: nitcion
  post_time: 2003-04-27 10:19:34
  post_text:

But it is still a legitimate question (and one that may be posed as Supplication): what on earth is CAI NAI meant to be? If ui cainai != uinaicai, how does that work? Stats on any instances whatsoever of cainai actually seeing use will help.

(Craig, if you want people to help you out in the various searches I'm suggesting, go ahead and ask...)


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: fracture
  post_time: 2003-04-27 10:42:49
  post_text:
At this point, I'd like to remind you (or state, if I haven't done so already) that we're not really meant to state new arguments here if there's a chance they're repeating old arguments. (This is following the djez dictum: if you have an argument, relive it through the archives.) As part of his record, Craig will be posting links to the previous iterations of this argument in the archives.

It is true that NAI as a UI would not break much more than {kau} in UI currently does (though {kau} *is* on record as a hack), that it would simplify the grammar tremendously, and that meaning for the new NAI combinations could be found. It is also true that this is a significant change, and while I won't make a ruling as to whether it is within bounds, I don't think I need to: there's enough

conservatism voiced here.

kau probably should've been implemented with a special place in the sumti rules with its own selma'o. But *UI* itself is a hack. It shouldn't be used for anything more than things which actually fit that grammar class (i.e. words you can say anywhere). .inai makes no sense; neither does .ikau.

Maybe we should move kau into its own selma'o.

(The consistency I've plead for though, for the record, as supporting CAhA NAI, is allowing CAhA to be negated like the other tenses. I

believe that is the underlying analogy.)

Yeah, but as has been discussed, CAhA are *not* like other tenses. They don't have the same grammar as PU/FAhA inside of the tense.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: Re: kau is already in UI
   username: fracture
  post_time: 2003-04-27 10:45:58
  post_text:
{kau} is in UI, even though it has a fairly narrow set of usage: it is meaningless in a simple sentence and is used mostly following a question word in a subordinate clause. So I don't see why putting {nai} in UI is going to cause problems.

I see {ka'enai} as equivalent to {na ka'e}, not {na'e ka'e}, which means some other member of CAhA applies. The usefulness I see for {ka'enai} is in conjunction with another member of CAhA, e.g. {pu'ijeca'anai}, "can and has but isn't". I think the reason people say {ka'enai} is that in English it's a contraction, and we're used

to that order.

That's no reason to consider CAhA+NAI useful. You can already do that with "pu'ijenaica'a". (Which is better style anyway: pujecanai should be avoided in favor of pujenaica).


 forum_name: General Negators
topic_title: ka'enai survey
    subject: Re: ka'enai survey
   username: fracture
  post_time: 2003-04-27 10:48:22
  post_text:

The harmful thing with it is that it creates more places where NAI is used instead of less.

I would probably support removing PU+NAI and FAhA+NAI.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: kd
  post_time: 2003-04-27 12:29:01
  post_text:
The harmful thing with it is that it creates more places where NAI is used instead of less.

I fail to see the harm.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: xod
  post_time: 2003-04-27 16:10:09
  post_text:

Right now, I favor Jorge's idea, to move NAI to UI, or something equivalent. It will break almost no usage if any, and result in a Lojban that's more elegant.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: fracture
  post_time: 2003-04-27 18:19:02
  post_text:
Right now, I favor Jorge's idea, to move NAI to UI, or something equivalent. It will break almost no usage if any, and result in a Lojban that's more elegant.

How is it more elegant to add grammar rules which allow sentences which are not valid semantically?

We have this problem with kau currently; we should not create more similar problems with nai.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: xod
  post_time: 2003-04-27 18:32:27
  post_text:

More elegant, because we don't have to memorize the inelegant and miserable rules for NAI that stand now. When the CLL author declares those rules hackish, it is hard to defend them on their own strengths.

Semantic validity should be left up to the speakers and listeners to determine. That is, I think, a fairly well established principle of Lojban history.

Inasmuch as Lojban relates to Sapir-Whorf, we are behooven to maintain as much semantic freedom as possible. Inasmuch as other Lojbanists agree with you, that freedom will not be abused.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: fracture
  post_time: 2003-04-27 18:40:06
  post_text:
More elegant, because we don't have to memorize the

inelegant and miserable rules for NAI that stand now. When the CLL author declares those rules hackish, it is hard to defend them on

their own strengths.

Right. And I would support removing the hackish rules (PU+NAI, FAhA+NAI, NU+NAI, JOI+NAI).

Semantic validity should be left up to the speakers and listeners

to determine. That is, I think, a fairly well established principle

of Lojban history.

To an extent. We *could* implement the whole language grammar by declaring that every word is a UI and some sentences make sense semantically, and some don't.

More sane is to try to only make words UI when they absolutely make sense to be UI. NAI does not make sense to be UI, since it is not a word that can just be said anywhere with an obvious meaning (like an attitudinal); rather than declaring half of its usages to be semantically incorrect, let's just define the valid usages.

Inasmuch as Lojban relates to Sapir-Whorf, we are

behooven to maintain as much semantic freedom as possible. Inasmuch as other Lojbanists agree with you, that freedom will not be

abused.

SW is dead.

But no one has defined a decent way to test it with Lojban anyway (since it could only prove it true, not prove it false (and my understanding is that it is believed more likely false than true in modern linguistics)).


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: feasability?
   username: jkominek
  post_time: 2003-04-27 18:52:10
  post_text:

Has anyone even taken the time to, say, modify the YACC formatted grammar to account for this change, and see what it does to the structure of the parses?

If it can't be done, or dramatically restructures the parse tree of utterances, that would sort of add on to the "hey, its a bad idea" school of thought. And if you want the change made, it seems as though you ought to demonstrate that its even feasible before everyone gets worked up over whether or not to do it. (Has anyone even set out to prove that you can currently stick UI everywhere you can stick NAI? It'd be a bit embarassing to make the change and then find that somewhere you could stick NAI, you no longer can.)


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: kd
  post_time: 2003-04-27 18:55:21
  post_text:

Seeing as JOI NAI compounds have some potential usefulness, and both PU NAI and FAhA NAI have enough usefulness to see actual use, I doubt you'll get the necessary support. However, seeing as there is an actual forum topic for the moving nai to UI idea, you shouldn't be discussing it here. I do not want to see the forum turn into utter chaos, and I'm sure the BPFKJ will agree with me.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: fracture
  post_time: 2003-04-27 19:04:40
  post_text:
Seeing as JOI NAI compounds have some potential usefulness,

and both PU NAI and FAhA NAI have enough usefulness to see actual use, I doubt you'll get the necessary support. However, seeing as there is an actual forum topic for the moving nai to UI idea, you shouldn't be discussing it here. I do not want to see the forum turn into utter chaos, and I'm sure the BPFKJ will agree with

me.

Seeing use doesn't mean they are useful. PU+NAI exists simply because that was how it was done in loglan, but it can always be replaced with use of na (or nai/na on connectives). FAhA+NAI was added simply because FAhA are like PU in lojban grammar.

So they aren't useful. They don't even save syllables. It's simply a redundant feature that adds rules to the grammar, and therefore should not have been added.

Anyway though I didn't bring up UI. But I don't particularly think we should be Nazi about the organization of this stuff either.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: kd
  post_time: 2003-04-27 19:55:28
  post_text:

Every post on this topic except my first one and your first one has been about the UI issue. And while I agree that we shouldn't be fascists about the organization, there ought to be limits. If you want to find the UI discussion, you should know you can find it in the UI topic.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-28 11:02:41
  post_text:
Has anyone even taken the time to, say, modify the YACC formatted grammar to account for this change, and see what it does to the structure of the parses?

There is no need to modify anything. All you need to do is move nai from NAI to CAI or UI and leave the YACC as it is. (The rules involving NAI can be dropped or left as dead relics, whatever you prefer.) Since NAI always attaches to the previous word, just like UI, no change in the structure of parses results.

(Has anyone

even set out to prove that you can currently stick UI everywhere you can stick NAI? It'd be a bit embarassing to make the change and then find that somewhere

you could stick NAI, you no longer can.)

UI/CAI can attach to any word, therefore it can attach in particular to the restricted set of words that NAI can attach to. QED.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: jkominek
  post_time: 2003-04-28 11:20:00
  post_text:
There is no need to modify anything. All you need to do is move nai from NAI to CAI or UI and leave the YACC as it is.

Nope.

(The rules involving NAI can be dropped or left as dead relics, whatever you prefer.)

Exactly.

Since NAI always attaches to the previous word, just like UI, no change in the structure of parses results.

Uhh. No. We'll be going through the free_modifier rule, rather than directly to a terminal token. How can you even think that isn't a change in the structure?

UI/CAI can attach to any word, therefore it can attach in particular to the restricted set of words that NAI can attach to. QED.

UI/CAI can attach anywhere the grammar says it can. Having just spent all weekend retyping the machine grammar, I can assure you it does not say it can go everywhere in the parse. However, what I don't know is whether or not it can go between every input token.

I can tell you right now, I will never, ever vote for any change to the grammar until I see the proof that the before and after recognize the same set of strings. I can't imagine how anyone could settle for less, either.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-28 11:38:41
  post_text:

From the point of view of a human speaker of the language, what difference does it make if nai is attached through the free-modifier rule or through a terminal token? All a human speaker has to do is know that nai attaches to the immediately preceding word, it seems to me. I don't claim any expertise in machine grammars, so please correct me if I'm missing something here. Is the {nai} in {uinai} different from the {cai} in {uicai}?

What kind of proof are you looking for?


 forum_name: General Negators
topic_title: jboske, please.
    subject: jboske, please.
   username: rlpowell
  post_time: 2003-04-28 14:37:46
  post_text:

The sort of discussion that is currently going on here is not appropriate to the BPFK forums, as should be obvious if you read the usage guidelines.

If you're going to have a vast, nit-picky discussion like this, take it to jboske. It is the sheperd's job to sift jboske for the useful parts of the argument.

-Robin


 forum_name: General Negators
topic_title: jboske, please.
    subject:
   username: xod
  post_time: 2003-04-28 14:51:59
  post_text:

Let me be even more helpful by posting a link to the usage guidelines [1], and direct attention to 7.3, which specifies the ideal shape of these discussions.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: You're out of your mind.
   username: rlpowell
  post_time: 2003-04-28 14:54:09
  post_text:

Sorry, posted this in the wrong place the first time.

Legal if nai is in UI:

nai nai nai mi nai nelci nai nai nai

If you can't see why that's a bad idea, I can't help you.

Note further that if you actual read the grammar, you'll note that UI followed by NAI is handled specially. Without that special handling, is "ko'a .ui nai" equivalent to "ko'a .ui" and "ko'a nai", which is what one would expect out of a string of UI, or "ko'a .uinai", which is what we have now?

I don't know, and as no-one here has produced a grammar fix, you obviously don't know either.

Until someone produces a detailed grammar change report, not only are you talking out your ganxo, but you're doing it in the wrong place. You know where jboske is, go use it.

doi le jatna po byfy pe'u do vimcu piro le dei se cusku porsi

-Robin


 forum_name: General Negators
topic_title: Making nai a UI?
    subject:
   username: xorxes
  post_time: 2003-04-28 16:24:11
  post_text:

{nai nai nai mi nai nelci nai nai nai} would be equivalent to {nai nai nai na'e bo mi na'e na'e na'e nelci}, which is currently grammatical.

{ko'a ui nai} remains as now, because nai modifies ui, just as in {ko'a ui cai}.


 forum_name: General Negators
topic_title: ja'ai
    subject: forms of officialized experimental cmavo
   username: And
  post_time: 2003-04-29 19:02:30
  post_text:
BTW, if ja'ai were to become an official part of the baseline, it should probably be given a non-experimental form. I would vote for ja'u, since it has the same first two vowels as ja'a. This doesn't mean I support adding it at all; I'm still undecided on that one, having yet to see a compelling argument either way.

I think this should be part of a different discussion. Personally I favour new cmavo being maximally mnemonic -- so prefer ja'ai to nai. But that's part of a general policy issue on the forms that any new cmavo in general should take.

--And.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: what it takes to be a UI: {kau}
   username: And
  post_time: 2003-04-29 19:06:29
  post_text:
I do not think that NAI should not be moved to selma'o UI.

These are some of the reasons why:

According to the Fregean principle of compositionality, doing this would require us to define some kind of semantic content that is independent of context, AND a rule that can combine nai with the word or phrase that is modified, UI-style. I don't think anyone here knows what kind of semantic content that would be, in any more detail than "the modified word/phrase is negated somehow".

It's a good principle, but are we going to apply it to existing members of UI and remove them from UI if they fail the test? {kau} would be a prime candidate.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: Summary: nai, ui, meaningless sentences
   username: xod
  post_time: 2003-04-29 20:10:53
  post_text:

Should this be done by the Shepherd? Should it go in its own thread?

Objection: nai should not go in UI, because it doesn't make sense everywhere, and would lead to meaningless sentences.

Answer: Jorge has shown that it makes sense in most places.

Answer: kau and possibly other cmavo already in UI already don't make sense everywhere, and lead to perhaps more meaningless sentences than nai.


 forum_name: General Negators
topic_title: Making nai a UI?
    subject: simplifying the grammar
   username: And
  post_time: 2003-04-29 20:55:25
  post_text:
I don't think ka'enai is the reason I'm considering this at all. It just happens to be the case where extending nai was found most obviously useful, but my main concern is the simplification of an overcomplicated grammar.

I opine that the appropriate degree of simplicity/complexity is the one that makes all meaningful strings licit, if other principles (e.g. unambiguity) are not compromised, and (though this is not essential) all meaningless strings illicit. (It doesn't really matter if a licit string is meaningless, because it is a vacuous licitness.)

If we do not impose this sort of criterion, then we might just as well simplify the grammar out of existence by allowing every possible string -- putting everything in UI, say.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: And
  post_time: 2003-04-29 21:09:12
  post_text:
We *could* implement the whole language grammar by

declaring that every word is a UI and some sentences make sense semantically, and some don't.

More sane is to try to only make words UI when they absolutely make sense to be UI. NAI does not make sense to be UI, since it is not a word that can just be said anywhere with an obvious meaning (like an attitudinal); rather than declaring half of its usages to be

semantically incorrect, let's just define the valid usages.

Hey, I pretty agree with you, believe it or not. But I'd suggest the following guiding principles:

1. No meaningful string should be prohibited by the Formal Grammar unless it is for other reasons detrimental to allow it. 2. A truly kosher Lojban sentence is one that is (a) allowed by the Formal Grammar and (b) meaningful.

From those principles it follows that so long as we do our job in defining which strings are meaningful, we don't really have to worry about meaningless sentences being allowed by the Formal Grammar.


 forum_name: General Negators
topic_title: ka'enai survey
    subject: Poll Addition/New Poll
   username: rlpowell
  post_time: 2003-05-01 04:26:30
  post_text:

My answer to the poll is "Yes, as long as it doesn't involve like moving nai to UI.".

And option for such an answer would be nice. 8)

-Robin


 forum_name: General Negators
topic_title: ja'ai
    subject: Why?
   username: rlpowell
  post_time: 2003-05-01 04:27:59
  post_text:

Can someone explain to me how this would be used? What would the semantic effects be? What problems does it solve?

-Robin


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: kd
  post_time: 2003-05-01 06:48:40
  post_text:

Then a bit of clarification is in order. The poll is only about making a rule that NAI can follow CAhA. Making nai a UI is a seperate issue, and I doubt it will ever have enough support to pass. This, well, I'm hoping the "I don't like to change things" people can be persuaded, because it's not IMO a very good reason.


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: xorxes
  post_time: 2003-05-01 13:30:24
  post_text:

It serves to emphasize the positive. For example, if you say {na.e}, I can counter with {ja'a.e}, and emphasize the {ja'a}, but what if you say {.enai}? Then I can counter with {.eja'ai} and emphasize the {ja'ai}.


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: nitcion
  post_time: 2003-05-02 13:59:40
  post_text:

Craig, you know what it will take for me (and, more or less, John) to be persuaded: a non-adhoc rule on when NAI is scalar and when contradictory, which will only emerge from studying the behaviour of NAI other than with UI and A. I can see myself yielding; but not until the semantics is resolved to my satisfaction.

Nick, repeating himself...


 forum_name: General Negators
topic_title: ka'enai survey
    subject:
   username: xorxes
  post_time: 2003-05-02 14:36:44
  post_text:
a non-adhoc rule on when NAI is scalar and when contradictory, which will only emerge from studying the behaviour of NAI other than with UI and A

Contradictory negation can only apply to a bridi, therefore NAI is contradictory when it applies to a bridi (e.g. after gi'e, .e, ge, .i, etc.) It negates not the word it attaches to but the bridi introduced by that word.

Scalar "negation" applies to the meaning of a word, not to a bridi. Therefore, NAI is scalar when it modifies a word's meaning directly and not a bridi (e.g. with UI, BAI, and almost all selma'o with semantic content).

In some cases, for purely structural words, neither contradictory nor scalar negation may apply. In this cases NAI changes the _function_ of a word to an opposite function. Examples could be {xinai} for superindex, {kinai} to cancel stickiness, {da'onai} to retain the referent of a pro-sumti, and others.


 forum_name: General Negators
topic_title: Compromise?
    subject: Compromise?
   username: kd
  post_time: 2003-05-28 22:23:04
  post_text:

It seems clear to me that there is just no way that those who like ka'enai and those who dislike it can agree on any proposal. There are too many in either camp for this to work. Somehow, we have to compromise.

I suggest adding ka'enai to the YACC, but issuing an official statement along the lines of:

CAhA NAI compounds are in the new YACC grammar. This way, official parsers will not break when they encounter these forms. However, CAhA NAI is a nonstandard use and will not be included in any baseline materials other than the grammar.

Hopefully, this will at least declaw the cats.


 forum_name: General Negators
topic_title: Compromise?
    subject: Re: Compromise?
   username: jkominek
  post_time: 2003-05-28 22:35:08
  post_text:
CAhA NAI compounds are in the new YACC grammar. This way, official parsers will not break when they encounter these forms. However, CAhA NAI is a nonstandard use and will not be included in any baseline materials other than the grammar.

I feel like you were laughing at the anti-ka'enai people while you wrote that.

Something like:

Muahaha! Maybe I can trick them into accepting it if I put a slight admonishment against its use in there!

How long are we to believe that this gentle chastisement will keep people from using it? Five minutes? Ten? Would you never again use ka'enai if that were the sort of statement put out?


 forum_name: General Negators
topic_title: Compromise?
    subject:
   username: kd
  post_time: 2003-05-28 22:51:19
  post_text:

No, djez, it was serious. I'm sorry it came across as it did.

I freely admit that after such a thing I would still use ka'enai in idle conversation. Basically, I think people would still say it, but in anything formal it would still be considered worng. Therefore, in large translation projects, it wouln't appear. Ideally after such a declaration even Alice would give it up.

The change would be that although ka'enai would keep its current official no-no status, the parser would accept it.


 forum_name: General Negators
topic_title: Compromise?
    subject: my subject disappeared!
   username: jkominek
  post_time: 2003-05-28 23:09:47
  post_text:
No, djez, it was serious. I'm sorry it came across as it did.

Just letting you know how bizarre it seems from here.

I freely admit that after such a thing I would still use ka'enai in idle conversation. Basically, I think people would still say it, but in anything formal it would still be considered worng. Therefore, in large translation projects, it wouln't appear. Ideally after such a declaration even Alice would give it up. The change would be that although ka'enai would keep its current official no-no status, the parser would accept it.

The other problem with this is that it introduces a distinction between the grammar, and "correct" Lojban, which is a hideous idea.

hmm, ok, to decide if this sentence is valid i need to check it against the grammar... ok, grammar likes it. now to check it against the list of exceptions!

that'd be fun!


 forum_name: General Negators
topic_title: Compromise?
    subject: I wonder where it could have gone.
   username: kd
  post_time: 2003-05-29 07:31:25
  post_text:

My thinking is this: There are some people who will use ka'enai regardless of what the bpfk says and does. There are some people who will not use ka'enai, even if the bpfk were to officialize it. If we can get everyone to understand and accept everyone else's usage by adding ka'enai to the grammar, we should. In fact, ka'enai is already commonplace. However, there are many people who don't want to do that. So the compromise is to keep it out of anything official, and even declare by fiat that it is to'e lobykai. Then, people will still use it in informal chitchat - English grammars saying you can't end a sentence with a preposition haven't stopped people from trying to - but hopefully not in formal situations. Meanwhile, it isn't in any LLG materials other than the YACC grammar itself, so no newbies learn it and it either proves its goodness or dies out.


 forum_name: General Negators
topic_title: Compromise?
    subject: re: Compromise?
   username: greg
  post_time: 2003-05-29 07:41:02
  post_text:

Those who dislike ka'enai are often those who worship the YACC.

How do you expect your proposal to seem to them to be a compromise? It effectively outlaws anyone who claims that ka'enai is wrong...

The way you put your argument, it seems that ka'enai should be accepted on grounds that otherwise whole utterances will be stumped when computer-parsed. It is then up to the discretion of the parser-programmer to decide what options he wants to turn on.

I don't mind ka'enai and, as such, I would make sure that any parser of mine would preparse ka'enai as na'eka'e. But I'm against putting CAhANAI in the grammar unless you do the research on NAI that John and Nick require.


 forum_name: General Negators
topic_title: Compromise?
    subject:
   username: kd
  post_time: 2003-05-29 14:36:27
  post_text:

In all things subject to change, the rule is to let usage decide. The bpfk exists to change things once again. Usage favors ka'enai. So the official parser - not just any that you might happen to write that preprocess ka'enai - should understand it. However, those that worship the grammar as is wouldn't want ka'enai to be part of the standard form of the language, so it seems like a compromise to me to have its official status be something that should be understood but not used. For all those who dislike this and threaten vetoes: Do you have a compromise that all but one (at most) of us can agree on?

As for the research on NAI that you allude to, I assume you mean figuring out whether it is contradictory or scalar in the circumstances given. I maintain that although only the CAhA shepherd can say so officially, CAhA is boolean and so na'eka'e is equivalent to naka'e - iow, it doesn't matter. And no amount of NAI research will tell you any other answer, as NAI is irregular as to which form of negation it means.


 forum_name: General Negators
topic_title: One change I think we can agree on
    subject: One change I think we can agree on
   username: kd
  post_time: 2003-05-29 14:48:15
  post_text:

mi pensi ui y nai OUGHT to be equivalent to mi pensi ui nai, with an y thrown in there. I suggest the following changes to the YACC grammar:

%token NAI_581 /* attached to words to negate them */

Should be replaced with

%token NAI_581_A /* attached to words to negate them */

Then, a rule should be created before all the other rules:

NAI_581 : something goes here, but I can't yet figure out what

| NAI_581_A | Y_619 NAI_581

;

The result, I think, is that the Y would get absorbed by the NAI such that ROD Y NAI compounds are grammatical iff ROD NAI compounds are grammatical.

If this fix is buggy or if you dislike it, let me know here or at my e-mail address, ragnarok at pobox dot com


 forum_name: General Negators
topic_title: One change I think we can agree on
    subject:
   username: xorxes
  post_time: 2003-05-29 15:31:32
  post_text:

How does {ui y cai} parse? {ui y nai} should parse the same way.

{y} should probably be in selma'o BAhE, so that it attaches to the following word without changing its syntax.


 forum_name: General Negators
topic_title: One change I think we can agree on
    subject:
   username: JohnCowan
  post_time: 2003-05-29 16:11:02
  post_text:

IMHO the Right Thing is to revert to an earlier version of the Lojban grammar, wherein Y was a selma'o by itself which could not accept NAI; this was changed between the 1990 and 1997 grammar baselines, moving almost every stand-alone free modifier into UI except for CAI. Formerly, DAhO, FUhE, FUhO, and POhA were separate selma'o that could not accept NAI.

In the old grammar, "UI Y NAI" would be ungrammatical because of the stray NAI, so we would have to add explicit rules for UI Y NAI and CAI Y NAI.


 forum_name: General Negators
topic_title: One change I think we can agree on
    subject:
   username: kd
  post_time: 2003-05-29 16:38:27
  post_text:

Xorxes: Good point on the cai. I agree about putting it in BAhE; this takes it out of my jurisdiction so whoever shepherds the Y and BAhE selma'o, please consider it.

John: It sure looks to me from the refgram like there is still a seperate selma'o Y, which cannot take a NAI. Or a CAI, for that matter.


 forum_name: General Negators
topic_title: One change I think we can agree on
    subject:
   username: JohnCowan
  post_time: 2003-05-29 17:22:23
  post_text:

Oops, yes. In that case I would favor adding explicit rules UI Y NAI and CAI Y NAI.


 forum_name: General Negators
topic_title: One change I think we can agree on
    subject:
   username: JohnCowan
  post_time: 2003-05-30 10:20:08
  post_text:

UI Y CAI parses as three separate entities; the association between UI and CAI is not enforced by the grammar.

The trouble with making ".y." part of BAhE is that that would forbid it with no cmavo following.

I still think the simplest solution is to add explicit rules (and preparser behavior) for UI Y NAI and CAI Y NAI, since NAI is what causes the problem.


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: clsn
  post_time: 2003-06-01 11:33:26
  post_text:

Just to be on the record, since this is originally my proposal anyway...

Yes, I support ja'ai. Arguments pro?

  • Except for NAI, so far as I know, every negator in Lojban comes with a corresponding affirmator: na/ja'a na'e/je'a na'i/jo'a (you can argue about no/su'o if you want). And for good reason, presumably.
  • To emphasize the positive. mi .enai do cu klama vau xu? .i na go'i .i mi .eja'ai do cu klama. Note that this is possible in the current language with alternating na.e and ja'a.e, but not the symmetric case of .enai/.e. This is important for A, and JA, and also for UI: .ianaipei? .iaja'ai (NOT the same as .iasai!) It isn't hard to come up with other examples.

Basically, I think the lack of ja'ai is essentially a typo, something that was "accidentally" left out of the grammar. After all, we have the other affirmators. (Actually, it could be even more important than just emphasis: what about overwriting a negation on a tense or something when using go'i?) It should have been there all along, let's fix it.

As to the form... That's a toughie. On the one hand, it should be a "full-fledged" cmavo, in cmavo space. On the other hand, ja'ai is so nicely mnemonic. It is, is it not, within the BPFK's power to elevate an experimental form to "official" status, right? Which would mean that ja'ai would no longer be considered "experimental"... ah, but then people somehow have to remember which CV'VV are experimental and which are "occupied"... I think at this point I favor retaining ja'ai as the form, but could likely be convinced.


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: nitcion
  post_time: 2003-06-01 11:49:43
  post_text:

Any cmavo the BPFK approves of is, of course, no longer experimental. I agree that the mnemonic function of ja'ai is nice, and am wondering whether we might not give official status to say CV'VV cmavo. But I feel the assignments should be decided at a separate stage, once we have enough experimentals approved that we know what sort of numbers we're talking. Bob has insisted --- and I agree --- that xVV needs to be exhausted before we go to CVVV. (And hey, four monosyllables there up for grabs.)

But it is true for all proposals -- even my pet {lau} -- that the form someone is eyeing should not influence the allocation or de-allocation of a cmavo; the allocations should happen all together, and be judged by the BPFK as a body. Later. If {ja'ai} is a good cmavo to have around, then it should be approved; what it ends up actually looking like, I cannot see being decided before another year passes...


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: And
  post_time: 2003-06-01 13:52:30
  post_text:

I support (a) the ja'ai function, (b) the ja'ai form, (c) Nick's suggestion of postponing decisions about forms until later.


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: noras
  post_time: 2003-06-01 15:29:35
  post_text:
... before we go to CVVV. (And hey, four monosyllables there up for grabs.)...

I beg to differ. There are no CVVV monosyllables. As defined in CLL, multiple vowels group in twos starting at the left. Thus, for example, "cuai" = "cua,i" = /shwa-ee/ (two syllables).


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: And
  post_time: 2003-06-01 15:40:00
  post_text:
... before we go to CVVV. (And hey, four monosyllables there up for grabs.)...
I beg to differ. There are no CVVV monosyllables. As defined in CLL, multiple vowels group in twos starting at the left. Thus, for example, "cuai" = "cua,i" = /shwa-ee/ (two syllables).

Wow. Are CVVV really allowed with no /'/? That'd make a hell of a difference.

I'd taken Nick to be talking about CV'VV/CVV'V, and about the unassigned xVV monosyllables.


 forum_name: General Negators
topic_title: ja'ai
    subject:
   username: nitcion
  post_time: 2003-06-01 21:05:20
  post_text:

And's right: I was referring to xVV.

As to monosyllabic CVVV and all: one of the tasks of the VTBPK will be to decide what cmavo space --- experimental and not --- is. There's a case to be made against making {tio} a legal cmavo, for example (confusion with {tco} through palatalisation); such a case hasn't been made yet, because (I feel) the morphology hasn't been defined exhaustively enough. So we may get some pleasant surprises down the road in cmavo space; or we may not. (Right now, I'm betting on not...)


 forum_name: General Negators
topic_title: ja'ai
    subject: CVVV cmavo
   username: And
  post_time: 2003-06-02 17:18:08
  post_text:
As to monosyllabic CVVV and all: one of the tasks of the VTBPK will be to decide what cmavo space --- experimental and not --- is. There's a case to be made against making {tio} a legal cmavo, for example (confusion with {tco} through palatalisation); such a case hasn't been made yet, because (I feel) the morphology hasn't been defined exhaustively enough. So we may get some pleasant surprises down the road in cmavo space; or we may not. (Right now, I'm betting on not...)

I don't know why you should bet on not. After all, it does involve undoing anything in the current baseline, does it?

Anyway, what I really want to say is that I would like to ask that the cmavo space issue be decided sooner rather than later, since, like it or not, it is bound up with other issues, such as whether to get rid of cmavo currently using up monosyllables.

(I don't know why I bother, really, because I'm sure that any position I advocate will be (*not* as a consequence) vetoed in favour of the status quo. But I have to pretend this is not the case, else why bother be here?)


 forum_name: General Negators
topic_title: NAI affirmator poll
    subject: NAI affirmator poll
   username: clsn
  post_time: 2003-06-08 14:24:43
  post_text:

Just to get an idea of this, let's try an informal poll. Should we have a ja'ai-like word in NAI? This is separate from any discussion about removing or otherwise changing the selma'o NAI. It is also separate from any question regarding the form of the putative cmavo.


 forum_name: General Negators
topic_title: NAI affirmator poll
    subject:
   username: kd
  post_time: 2003-06-08 14:47:59
  post_text:

How can you half-veto to make it so? And what's up with your pic?


 forum_name: General Negators
topic_title: NAI affirmator poll
    subject:
   username: clsn
  post_time: 2003-06-08 15:40:08
  post_text:
How can you half-veto to make it so? And what's up with your pic?

"Half-vetos" work both ways. In order for something to pass, it must be approved by at least everyone-1, and in order for it to be defeated, it must be opposed by at least everyone-1; anything else and we have to talk more. So you can veto in either direction (half-veto, really). Of course, this isn't a binding poll, so it's more "I *would* half-veto, if/when this comes to an official vote"

As to my pic... well, that's what I look like! Part of me, anyway.


 forum_name: General Negators
topic_title: Discussion of Changes 1: CAI NAI
    subject:
   username: JohnCowan
  post_time: 2003-08-11 01:40:04
  post_text:

AFAIK uicainai = uinaicai. I can't imagine any other meaning.