TECH: PROPOSED GRAMMAR CHANGE 2$i
John Cowan
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 36
CURRENT LANGUAGE:
There are three basic kinds of vocative phrases: "DOI name", "DOI selbri",
and "DOI sumti". (Here DOI stands for possible multiple COIs with or without
following DOI as well). The third case, "DOI sumti", is the general case
which can handle whatever is needed with some extra cmavo, since "DOI name"
really means "DOI la name", and "DOI selbri" really means "DOI le selbri".
Relative clauses are currently allowed after "DOI name", and either before
or after the selbri in "DOI selbri". However, if relative clauses precede
the selbri, then a full sumti-tail-1 (essentially a description without a
descriptor) is permitted.
PROPOSED CHANGE:
Only allow a selbri in the context "DOI relative-clauses ...".
This allows "DOI selbri" to have relative clauses before or after the selbri.
In addition, a new rule is added allowing relative clauses both before and
after the selbri. Quantifiers are disallowed altogether. Legal cases are:
doi pe mi pendo
do pendo poi melbi
doi pe mi pendo poi melbi
all of which are natural and easy to understand.
RATIONALE:
The current language allows vocative phrases of certain types only if
a preposed relative clause is present: "DOI relative-clauses quantifier
selbri", "DOI relative-clauses quantifier selbri relative-clauses",
"DOI relative-clauses quantifier sumti", and possibly other forms. All
of these are meaningful, but their existence makes vocative phrases hard
to teach.
Nothing is lost by making these forms ungrammatical, because if they are needed,
a full sumti can be used instead.
ADDITIONAL NOTE:
Jorge also proposed the form "DOI relative-clauses sumti", but I reject this,
because it would not be clear whether the relative-clauses were to be taken as
inside-the-ku or outside. There is no other place where relative clauses
can appear before a sumti as such.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-35 DATED 29 MARCH 1994
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-36 DATED 26 OCTOBER 1995
19c19
< grammar.235
---
> grammar.236
484c484,486
< sumti_tail_A_112 DOhU_gap_457
---
> selbri_130 DOhU_gap_457
> | DOI_415 relative_clauses_121
> selbri_130 relative_clauses_121
DOhU_gap_457
co...@digitalkingdom.org
John Cowan
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 38: ANNULLED
PRESENT LANGUAGE:
There is currently no defined way to do lambda quantification.
PROPOSED CHANGE:
Add "ce'u" of the new selma'o CEhU. This is a new kind of quantifier_300,
parallel to number-strings and parenthesized mathematical expressions.
RATIONALE:
Lambda quantification is needed to specify which place(s) of a "ka"
abstraction are being abstracted over.
In early versions of this change, "ce'u" was a PA digit, which would not
require a grammatical change, but would allow lots of new kinds of
garbage. Now "ce'u" is limited to quantifying sumti and forming indefinite
descriptions.
(This change involved renumbering selma'o CMENE, CO, COI, CU, CUhE.)
OBJECTION:
It is sufficient to make "ce'u" a KOhA, somewhat analogous to "ke'a",
the KOhA used within relative clauses. Both of these work only in
restricted contexts. "ce'u broda" becomes "ce'u poi broda" without
change in meaning.
Jorge and And argued for overloading "ke'a" in both senses, but Lojban
Central finds that too confusing.
RESOLUTION:
No grammar change: "ce'u" is a KOhA. The renumbering of selma'o stands.
John Cowan
Robin Lee Powell
John Cowan
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 40
CURRENT LANGUAGE:
geks can be used to join two independent levels, sentence-40 and bridi-tail-50.
The former corresponds roughly to afterthought ijeks, the latter exactly
to afterthought giheks.
There is a limitation in gekked sentences: the first sentence must have
at least one term before the selbri, or the construct will look like a
gekked bridi-tail.
PROPOSED CHANGE:
Unify the two types of geks as follows: unrestricted sentences (including
bare bridi-tails) can be gekked, and the result has the grammar of a
bridi-tail. This allows currently forbidden sentences like:
1) gonai catra la djonz. gi la djonz. catra
Either the-obvious-one killed Jones or Jones killed the-obvious-one.
which currently is excluded because it garden-paths the parser into expecting
a mere bridi-tail after "gi".
We also get a new sentence type consisting of a gekked sentence with joint
head-terms:
2) mi [cu] ge le zarci cu klama gi le rilti cu dansu
I both to-the office go and to-the rhythm dance.
Example 2 parses because the gekked sentence functions as the equivalent
of a bridi-tail and can be preceded by head-terms and an optional "cu".
RATIONALE:
The arbitrary restriction on gekked sentences is lifted. Since bridi-tail-50
has fundamentally the same semantics as sentence-40 (not all bare bridi-tails
are semantic observatives; some full sentences with head-terms are in fact
semantic observatives), logical connectives between the two are naturally
unifiable.
This change is a pure extension: everything sayable before is still sayable.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-39 DATED 6 JANUARY 1996
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-40 DATED 6 JANUARY 1996
19c19
< grammar.239
---
> grammar.240
504,512c504
< | sentence_A_41
< ;
<
< sentence_A_41 : GEK_807 sentence_A_41 GIK_816 sentence_40
< | prenex_30 sentence_40
< | statement_42
< ;
<
< statement_42 : terms_80 front_gap_451 bridi_tail_50
---
> | terms_80 front_gap_451 bridi_tail_50
513a506
> | prenex_30 sentence_40
532c525
< bridi_tail_C_53 : gek_bridi_tail_54
---
> bridi_tail_C_53 : gek_sentence_54
536c529
< gek_bridi_tail_54 : GEK_807 bridi_tail_50 GIK_816 bridi_tail_50
---
> gek_sentence_54 : GEK_807 sentence_40 GIK_816 sentence_40
538,539c531,532
< | tag_491 KE_551 gek_bridi_tail_54 KEhE_gap_466
< | NA_445 gek_bridi_tail_54
---
> | tag_491 KE_551 gek_sentence_54 KEhE_gap_466
> | NA_445 gek_sentence_54
Robin Lee Powell
co...@digitalkingdom.org
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 41: ANNULLED
PROPOSED CHANGE:
A new syntax is added for nested forethought relative clauses, using
the new cmavo "pe'e" of the new selma'o PEhE. The center-embedding of
le (poi le (poi le tcadu cu se klama ku'o) nanmu cu se viska ku'o) verba
which is the forethought relative clause version of
le verba poi le nanmu poi le tcadu cu se klama [ku'o] cu se viska
can be rewritten more clearly as
le poi le tcadu cu se klama pe'e nanmu cu se viska ku'o verba
le poi viska le poi klama le tcadu vau nanmu vau verba
OBJECTION:
This change is simply too hard to explain and too complex to work out
at this late date. Loglan historically didn't have preposed relative
clauses, and they still are not as well supported as postposed ones.
Complexities like this work better in languages where nouns and verbs
are visibly distinct, like German or Finnish.
Robin Lee Powell
John Cowan
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 42
CURRENT LANGUAGE:
Lojban uses the same words (TAhE/ROI and ZAhO) for space interval modifiers
as for time interval modifiers. However, they are distinguished with a
"fe'e" flag proceeding any such modifiers. The rules for multiple
modifiers are that two ZAhOs may be consecutive but two TAhE/ROIs may not.
This rule applies to both time and space. Once a modifier has been
given, an interval size cannot follow it. So the following is not grammatical:
mi ze'a reroi ze'i klama
I [medium interval] [twice] [short interval] go
Over a medium period I go twice for a short period each.
PROPOSED CHANGE:
Allow mixtures of ZEhA [PU], TAhE, ROI, and ZAhO for time; and VEhA [PU],
VIhA, FEhE+TAhE, FEhE+ROI, and FEhE+ZAhO for space. Note that each of
the space interval modifiers must be flagged if there is more than one.
As a side effect, space information may now precede time information
in a tense.
RATIONALE:
Simplicity and uniformity, plus the ability to specify the size of
subintervals.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-40 DATED 6 JANUARY 1996
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-42 DATED 6 JANUARY 1996
19c19
< grammar.240
---
> grammar.242
1451a1452
> | space_1040 time_1030
1534,1536c1535,1537
< | interval_modifier_1050
< | ZEhA_622 interval_modifier_1050
< | ZEhA_622 time_direction_1035
interval_modifier_1050
---
> | time_int_props_1036
> | ZEhA_622 time_int_props_1036
> | ZEhA_622 time_direction_1035 time_int_props_1036
1542a1544,1547
> time_int_props_1036 : interval_property_1051
> | time_int_props_1036 interval_property_1051
> ;
>
1575,1576c1580,1581
< | FEhE_530 interval_modifier_1050
< | space_intval_A_1047 FEhE_530
interval_modifier_1050
---
> | space_int_props_1049
> | space_intval_A_1047 space_int_props_1049
1578c1583
< FEhE_530 interval_modifier_1050
---
> space_int_props_1049
1590a1596,1599
> space_int_props_1049 : FEhE_530 interval_property_1051
> | space_int_props_1049 FEhE_530
interval_property_1051
> ;
>
1600,1604d1608
< interval_modifier_1050 : interval_property_1051
< | interval_property_1051 event_mod_1052
< | event_mod_1052
< ;
<
1608a1613,1614
> | ZAhO_621
> | ZAhO_621 NAI_581
1621,1630d1626
<
< event_mod_1052 : event_mod_A_1053
< | event_mod_1052 event_mod_A_1053
< ;
<
< event_mod_A_1053 : ZAhO_621
< | ZAhO_621 NAI_581
< | ZAhO_621 interval_property_1051
< | ZAhO_621 NAI_581 interval_property_1051
< ;
co...@digitalkingdom.org
Robin Lee Powell
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 39
CURRENT LANGUAGE:
A high-priority MEX operator is created by prefixing BO to it.
PROPOSED CHANGE:
Rather than using "bo", use "bi'e" in the new selma'o BIhE.
RATIONALE:
There is a hidden conflict between suffixed "bo" at the end of something
and prefixed "bo", which appears only here. This does not show up in
Yacc because all uses of suffixed "bo" are within the lexer rules.
Changing to a new flag eliminates the problem.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-37 DATED 26 OCTOBER 1995
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-39 DATED 6 JANUARY 1996
5c5
< COPYRIGHT 1989,1990,1991,1992,1993,1994 THE LOGICAL LANGUAGE GROUP, INC.
---
> COPYRIGHT 1989,1990,1991,1992,1993,1994,1995,1996 THE LOGICAL LANGUAGE GROUP,
INC.
19c19
< grammar.237
---
> grammar.239
198c198
< %token CMENE_517 /* DJAN = names; require consonant end, then pause no
---
> %token CMENE_518 /* DJAN = names; require consonant end, then pause no
201,202c201,202
< %token CO_518 /* GO = tanru inversion */
< %token COI_519 /* UI = vocative marker permitted inside names; must
---
> %token CO_519 /* GO = tanru inversion */
> %token COI_520 /* UI = vocative marker permitted inside names; must
204,205c204,205
< %token CU_520 /* GA = separator between head sumti and selbri */
< %token CUhE_521 /* tense/modal question */
---
> %token CU_521 /* GA = separator between head sumti and selbri */
> %token CUhE_522 /* tense/modal question */
293a294
> %token BIhE_650 /* prefix for high-priority MEX operator */
742c743
< /* Support for right-grouping (short scope) infix expressions with BO. */
---
> /* Support for right-grouping (short scope) infix expressions with BIhE. */
745c746
< | MEX_B_312 BO_508 operator_370 MEX_A_311
---
> | MEX_B_312 BIhE_650 operator_370 MEX_A_311
846,847c847,848
< cmene_A_405 : CMENE_517 /* pause */
< | cmene_A_405 CMENE_517 /* pause*/
---
> cmene_A_405 : CMENE_518 /* pause */
> | cmene_A_405 CMENE_518 /* pause*/
896,897c897,898
< COI_A_417 : COI_519
< | COI_519 NAI_581
---
> COI_A_417 : COI_520
> | COI_520 NAI_581
988,989c989,990
< CO_443 : CO_518
< | CO_518 free_modifier_32
---
> CO_443 : CO_519
> | CO_519 free_modifier_32
1015,1016c1016,1017
< front_gap_451 : CU_520
< | CU_520 free_modifier_32
---
> front_gap_451 : CU_521
> | CU_521 free_modifier_32
1413c1414
< | CUhE_521
---
> | CUhE_522
co...@digitalkingdom.org
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 44
CURRENT LANGUAGE:
A (nearly) full text can be encapsulated between TUhE and TUhU long-scope
parentheses if they stand alone. However, if a prenex or a tag comes first,
only paragraphs are permitted.
PROPOSED CHANGE:
Allow the same things between all TUhE-TUhU pairs.
RATIONALE:
This change is really part of Change 27 that was not correctly implemented
when that change was installed.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-43 DATED 6 JANUARY 1996
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-44 DATED 6 JANUARY 1996
19c19
< grammar.243
---
> grammar.244
438,439c438,439
< | prenex_30 TUhE_610 paragraphs_4 TUhU_gap_454
< | tag_491 TUhE_610 paragraphs_4 TUhU_gap_454
---
> | prenex_30 TUhE_610 text_B_2 TUhU_gap_454
> | tag_491 TUhE_610 text_B_2 TUhU_gap_454
co...@digitalkingdom.org
Robin Lee Powell
John Cowan
upload_19...@teddyb.org
Veijo Vilva
> Date: Mon, 11 Mar 1996 15:58:51 -0300
> From: "Jorge J. Llambias" <jo...@INTERMEDIA.COM.AR>
> Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
>>CHANGE 46
>...
>>The corresponding forethought syntax remains "NUhI GEK terms NUhU GIK
>>terms NUhU",
> Is NUhI really necessary? Couldn't it be just "GEK terms /NUhU/ GIK
> terms /NUhU/"? That would make it much more in tune with the
> general use of geks.
This would work at the yacc level
> (Even better if the first NUhU could be avoided.)
1 shift/reduce conflict
> I supppose that there must be some yacc problem with that, but
> I don't see why. If "ge sumti gi sumti" is ok, why would "ge sumti sumti
> gi sumti sumti" cause any problems?
15 shift/reduce + 15 reduce/reduce conflicts
The shift/reduce come from the "ge sumti sumti" part and the reduce/
reduce conflicts from the "gi sumti sumti" part. The last one is
quite obvious as there is no way to discriminate between, e.g.
[ge sumti sumti gi sumti sumti] [sumti] and
[ge sumti sumti gi sumti] [sumti] [sumti]
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
Jorge J. Llambias
>CHANGE 46
...
>The corresponding forethought syntax remains "NUhI GEK terms NUhU GIK
>terms NUhU",
Is NUhI really necessary? Couldn't it be just "GEK terms /NUhU/ GIK
terms /NUhU/"? That would make it much more in tune with the
general use of geks. (Even better if the first NUhU could be avoided.)
I supppose that there must be some yacc problem with that, but
I don't see why. If "ge sumti gi sumti" is ok, why would "ge sumti sumti
gi sumti sumti" cause any problems?
>and the syntax "NUhI terms NUhU", with no logical connective,
>is added as well.
Probably harmless. That means that "NUhI terms NUHU EK NUhI terms NUhU"
would be acceptable, right? Longer than the current grammar, but much
easier to understand. The current grammar, without the second NUhI, is
very counterintuitive for me.
>The other problem is that of indicating that two numerically quantified
>sumti have co-equal scope:
>
> ci nanmu re gerku cu batci
>
>says that three men bite two dogs each, for a possible total of six dogs,
>whereas
>
> ci nanmu ce'e re gerku cu batci
> nu'i ci nanmu re gerku nu'u cu batci
>
>says that three men bite two dogs each, the same two dogs.
I liked this at first, but now I don't know. I think termsets are orthogonal
to the scope problem. Suppose you say:
ci nanmu ce'e re gerku pe'eje ci ninmu ce'e re cinfo cu batci
Three men two dogs, and three women two lions bite.
If you use termsets for the scope distinction, then you can't use
them for their simple original function. If you want to solve the
scope problem with this, then perhaps there could be two members of
selmaho CEhE, one for normal embedded scope and one for equal scope?
Jorge
Jorge J. Llambias
la veion mi di'e spuda
>> Couldn't it be just "GEK terms /NUhU/ GIK
>> terms /NUhU/"? That would make it much more in tune with the
>> general use of geks.
>
> This would work at the yacc level
>
>> (Even better if the first NUhU could be avoided.)
>
> 1 shift/reduce conflict
Is it possible to identify when this happens? When a {gi} is found,
the only possibility is that it is closing the latest open gek,
isn't it? In other words, under what circumstances would the first
nu'u not be elidable?
Jorge
Veijo Vilva
> Date: Wed, 13 Mar 1996 14:26:10 -0300
> From: "Jorge J. Llambias" <jo...@INTERMEDIA.COM.AR>
> Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
>>> (Even better if the first NUhU could be avoided.)
>>
>> 1 shift/reduce conflict
>
> Is it possible to identify when this happens? When a {gi} is found,
> the only possibility is that it is closing the latest open gek,
> isn't it? In other words, under what circumstances would the first
> nu'u not be elidable?
After some juggling I found a solution at the YACC level:
--- in the rules for term_set_83 replace
GEK_807 terms_80 GIK_816 NUhU_gap_460
terms_80 NUhU_gap_460
with
GEK_807 terms_80 term_81 GIK_816
terms_80 term_81 NUhU_gap_460
Now, a termset must have at least two terms.
Veijo Vilva
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
Veijo Vilva
> Date: Wed, 13 Mar 1996 22:50:29 +0200
> From: Veijo Vilva <ve...@XIRON.PC.HELSINKI.FI>
> Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
> --- in the rules for term_set_83 replace
>
> GEK_807 terms_80 GIK_816 NUhU_gap_460
> terms_80 NUhU_gap_460
>
> with
>
> GEK_807 terms_80 term_81 GIK_816
> terms_80 term_81 NUhU_gap_460
>
> Now, a termset must have at least two terms.
This has the weakness compared to the present grammar that single
non-sumti terms cannot be connected. Allowing single terms here
caused a conflict with rule sumti_D_94 : GEK sumti GIK sumti.
The following modification will allow the connection of single
terms:
term_81 : sumti_90
| term_81_A
;
term_81_A : modifier_82
| term_set_83
| NA_KU_810
;
Add the following line to the rules for term_set_83:
GEK_807 term_81_A GIK_816 term_81_A NUhU_gap_460
The NUhU isn't absolutely necessary -- only it must either be always
required or always omitted.
Veijo Vilva
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
Robin Lee Powell
Veijo Vilva
> termsets work without the mniddle NUhU worthwhile. I think I am still
> skeptical that it is, and do not mind the middle NUhU as much as the
> plurality of rules, though it seems that what you are saying is that all
> you eliminate is the direct equivalent of sumti_94 GIks.
>
I am not proposing anything, I was just providing a technical answer
to Jorge - without comments, pro or con. I'm not too keen on termsets
without the middle NUhU. They are easy to make to work in YACC but
perhaps not at the human level (count the terms between the GEK and
the GIK in order to know where and how the terms after the GIK are
supposed to terminate, eh? I think I'd rather have the NUhU).
> I think that termsets and their power have never been well explored, and they
> will come to be more useful as people get more fluent in the language, and so
> I would rather have a more consistent set of rules, and a broader range of
> things called termsets, including single sumti_90, not merely modifier_82
> permitted (I can imagine there to be times when brevity might lead to the
> BAI put only on the first term in the termset, and not on the others.)
Agreed.
Veijo
Jorge J. Llambias
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
la veion mi di'e spuda
>> Couldn't it be just "GEK terms /NUhU/ GIK
>> terms /NUhU/"? That would make it much more in tune with the
>> general use of geks.
>
> This would work at the yacc level
>
>> (Even better if the first NUhU could be avoided.)
>
> 1 shift/reduce conflict
Is it possible to identify when this happens? When a {gi} is found,
the only possibility is that it is closing the latest open gek,
isn't it? In other words, under what circumstances would the first
nu'u not be elidable?
Jorge
Robin Lee Powell
Robin Lee Powell
Veijo Vilva
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
> Date: Mon, 11 Mar 1996 15:58:51 -0300
> From: "Jorge J. Llambias" <jo...@INTERMEDIA.COM.AR>
> Subject: Re: TECH: PROPOSED GRAMMAR CHANGE 2$i
>>CHANGE 46
>...
>>The corresponding forethought syntax remains "NUhI GEK terms NUhU GIK
>>terms NUhU",
> Is NUhI really necessary? Couldn't it be just "GEK terms /NUhU/ GIK
> terms /NUhU/"? That would make it much more in tune with the
> general use of geks.
This would work at the yacc level
> (Even better if the first NUhU could be avoided.)
1 shift/reduce conflict
> I supppose that there must be some yacc problem with that, but
> I don't see why. If "ge sumti gi sumti" is ok, why would "ge sumti sumti
> gi sumti sumti" cause any problems?
15 shift/reduce + 15 reduce/reduce conflicts
The shift/reduce come from the "ge sumti sumti" part and the reduce/
reduce conflicts from the "gi sumti sumti" part. The last one is
quite obvious as there is no way to discriminate between, e.g.
[ge sumti sumti gi sumti sumti] [sumti] and
[ge sumti sumti gi sumti] [sumti] [sumti]
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 43
CURRENT LANGUAGE
MEX operators, which are meant to be formally parallel to selbri,
can be logically connected by guheks, jeks, or joiks, and can be grouped
by KE/KEhE. It is not possible to use JEK+BO or JOIK+BO to connect them.
PROPOSED CHANGE:
Allow JEK+BO and JOIK+BO.
RATIONALE:
This should have been done as part of Change 2 above, but was overlooked.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-42 DATED 6 JANUARY 1996
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-43 DATED 6 JANUARY 1996
19c19
< grammar.242
---
> grammar.243
777a778,779
> | operator_B_372 JOIK_BO_822 operator_A_371
> | operator_B_372 JEK_BO_821 operator_A_371
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 45
CURRENT LANGUAGE:
Grammatically, I and ijek are treated identically, although in the
semantics, I constitutes a stronger boundary.
Prenexes can be attached only to sentences or to TUhE...TUhU groups, although
logically a prenex can persist across several sentences connected by
ijeks.
PROPOSED CHANGE:
Treat I as a higher-priority break than ijek (which is higher than
I+BO or ijek+BO; no distinction is made between I+BO and ijek+BO).
Shift all the sentence fragments (the forms of utterance_20 which are not
sentence_40) to a higher level; they can only be connected by I, not by any
lower-level form.
Attach prenexes to the new level "statement_11"; statements contain ijeks
and I+BOs, but not bare Is.
RATIONALE:
1) It has always been a rule that I and ijek have different semantic
implications: I is a pure separator, whereas ijek connects as well as
separating. In particular, logical variables persist across ijek
boundaries always, but (by default) not across I boundaries. This change
makes the grammar reflect the semantics.
2) Logically connecting sentence fragments never did make very much sense,
but was allowed because of the lack of distinction between I and ijek.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-44 DATED 6 JANUARY 1996
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-43 DATED 6 JANUARY 1996
19c19
< grammar.244
---
> grammar.245
164c164
< J, K, M ,N, G, O, V, W, F, P, R, S, Y, L, Q. This ensures that the longest
---
> J, K, M ,N, G, O, V, W, F, P, R, T, S, Y, L, Q. This ensures that the
longest
342,343c342,343
< %token lexer_S_719 /* flags an I, not BO */
< /* %token lexer_T_720 /* null */
---
> %token lexer_S_719 /* flags simple I */
> %token lexer_T_720 /* flags I_JEK */
369,370c369,370
< /*%token lexer_S_995 /* : lexer_S_719 I_root_996 */
< /*%token lexer_T_1000 /* null */
---
> /*%token lexer_S_995 /* : lexer_S_719 I_545 */
> /*%token lexer_T_1000 /* : lexer_T_720 I_545 simple_JOIK_JEK_957
*/
396a397
> | I_JEK_820 text_B_2
420,421c421,423
< paragraph_10 : paragraph_A_11
< | paragraph_10 I_819 paragraph_A_11
---
> paragraph_10 : statement_11
> | paragraph_10 I_819 statement_11
> | paragraph_10 I_819 fragment_20
428,433c430,431
< paragraph_A_11 : paragraph_B_12
< | paragraph_B_12 I_BO_811 paragraph_A_11
< | paragraph_B_12 I_BO_811
< /* this last fixes an erroneous start to a sentence,
< and permits incomplete JOIK_JEK after I, as well
< in answer to questions on those connectives */
---
> statement_11 : statement_A_12
> | prenex_30 statement_11
436c434,444
< paragraph_B_12 : utterance_20
---
> statement_A_12 : statement_B_13
> | statement_A_12 I_JEK_820 statement_B_13
> | statement_A_12 I_JEK_820
> ;
>
> statement_B_13 : statement_C_14
> | statement_C_14 I_BO_811 statement_B_13
> | statement_C_14 I_BO_811
> ;
>
> statement_C_14 : sentence_40
438d445
< | prenex_30 TUhE_610 text_B_2 TUhU_gap_454
443c450
< utterance_20 : EK_802
---
> fragment_20 : EK_802
453d459
< | sentence_40
459d464
< | terms_80 ZOhU_628 free_modifier_32
506d510
< | prenex_30 sentence_40
1240a1245,1248
> I_JEK_820 : lexer_T_1000
> | lexer_T_1000 free_modifier_32
> ;
>
1492c1500,1503
< lexer_S_995 : lexer_S_719 I_root_956
---
> lexer_S_995 : lexer_S_719 I_545
> ;
>
> lexer_T_1000 : lexer_T_720 I_545 simple_JOIK_JEK_957
Robin Lee Powell
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 37:
CURRENT LANGUAGE
Nothing can intervene between LA or DOI and a CMENE.
PROPOSED CHANGE
Allow relative clauses between LA or DOI and a CMENE.
RATIONALE
This will allow names with relative clauses that are part of the name, like
"la poi banli .karl." (Karl the Great, i.e. Charlemagne) and the like.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-36 DATED 26 OCTOBER 1995
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-37 DATED 26 OCTOBER 1995
19c19
< grammar.236
---
> grammar.237
489a490,493
> | DOI_415 relative_clauses_121 cmene_404
> DOhU_gap_457
> | DOI_415 relative_clauses_121 cmene_404
> relative_clauses_121 DOhU_gap_457
605a610
> | LA_558 relative_clauses_121 cmene_404
John Cowan
Sender: Lojban list <LOJ...@CUVMB.BITNET>
Poster: John Cowan <co...@LOCKE.CCIL.ORG>
Subject: TECH: PROPOSED GRAMMAR CHANGE 2$i
-------------------------------------------------------------------------------
CHANGE 46
CURRENT LANGUAGE:
Termsets are a pair of term sequences joined by a logical connective;
the start is NUhI and the elidable terminator is NUhU. The logical
connective can be afterthought EK, in which case NUhU must precede it,
or forethought GEK-GIK, in which case NUhU must precede the GIK.
PROPOSED CHANGE:
The syntax "NUhI terms NUhU EK terms /NUhU/" is REMOVED from Lojban.
It is a hybrid of forethought (the NUhI) and afterthought, and is
overly restrictive.
Instead, a "term group" construct is introduced, joining the terms
with the new cmavo "ce'e" of the new selma'o CEhE. A logical connective
is permitted but not required, thus: "term ce'e term ... pe'e EK term
ce'e term ..." "pe'e" belongs to the new selma'o PEhE.
The corresponding forethought syntax remains "NUhI GEK terms NUhU GIK
terms NUhU", and the syntax "NUhI terms NUhU", with no logical connective,
is added as well.
RATIONALE:
1) Afterthought termset logical connection becomes genuinely afterthought,
without the need for any special start cmavo.
2) Providing termsets without logical connectives (in both forethought
and afterthought forms) provides cmavo grouping to solve two unrelated
problems. By grouping a "se BAI" with a "te BAI" (or "ve BAI, or whatever)
we can require that both BAIs refer to the same underlying event:
se ka'a ko'a ce'e te ka'a ko'e
The presence of the "ce'e" makes sure that "ko'a" is the destination and
"ko'e" the destination of the same act of going.
The other problem is that of indicating that two numerically quantified
sumti have co-equal scope:
ci nanmu re gerku cu batci
says that three men bite two dogs each, for a possible total of six dogs,
whereas
ci nanmu ce'e re gerku cu batci
nu'i ci nanmu re gerku nu'u cu batci
says that three men bite two dogs each, the same two dogs.
3c3
< THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-45 DATED 6 JANUARY 1996
---
> THIS DRAFT ALSO INCORPORATES CHANGE PROPOSALS 1-46 DATED 21 FEBRUARY 1996
19c19
< grammar.245
---
> grammar.246
197a198
> %token CEhE_517 /* afterthought term list connective */
258a260
> %token PEhE_591 /* afterthought termset connective prefix */
538c540
< GIK_816 subsentence_41 tail_terms_71
---
> GIK_816 subsentence_41 tail_terms_71
548,549c550,551
< terms_80 : term_81
< | terms_80 term_81
---
> terms_80 : terms_A_81
> | terms_80 terms_A_81
552,554c554,564
< term_81 : sumti_90
< | modifier_82
< | term_set_83
---
> terms_A_81 : terms_B_82
> | terms_A_81 PEhE_591 JOIK_JEK_422 terms_B_82
> ;
>
> terms_B_82 : term_83
> | terms_B_82 CEhE_517 term_83
> ;
>
> term_83 : sumti_90
> | modifier_84
> | term_set_85
558c568
< modifier_82 : mod_head_490 gap_450
---
> modifier_84 : mod_head_490 gap_450
562,563c572,573
<
< term_set_83 : NUhI_587 GEK_807 terms_80 NUhU_gap_460
---
> term_set_85 : NUhI_587 terms_80 NUhU_gap_460
> | NUhI_587 GEK_807 terms_80 NUhU_gap_460
565,566d574
< | NUhI_587 terms_80 NUhU_gap_460
< JOIK_EK_421 terms_80 NUhU_gap_460
649c657
< relative_clause_122 : GOI_542 term_81 GEhU_gap_464
---
> relative_clause_122 : GOI_542 term_83 GEhU_gap_464
713,714c721,722
< linkargs_160 : BE_504 term_81 BEhO_gap_467
< | BE_504 term_81 links_161 BEhO_gap_467
---
> linkargs_160 : BE_504 term_83 BEhO_gap_467
> | BE_504 term_83 links_161 BEhO_gap_467
717,718c725,726
< links_161 : BEI_505 term_81
< | BEI_505 term_81 links_161
---
> links_161 : BEI_505 term_83
> | BEI_505 term_83 links_161
Logical Language Group
termsets work without the mniddle NUhU worthwhile. I think I am still
skeptical that it is, and do not mind the middle NUhU as much as the
plurality of rules, though it seems that what you are saying is that all
you eliminate is the direct equivalent of sumti_94 GIks.
I think that termsets and their power have never been well explored, and they
will come to be more useful as people get more fluent in the language, and so
I would rather have a more consistent set of rules, and a broader range of
things called termsets, including single sumti_90, not merely modifier_82
permitted (I can imagine there to be times when brevity might lead to the
BAI put only on the first term in the termset, and not on the others.)
lojbab