Talk:BPFK Section: Intervals
Archived Tiki forum discussion for BPFK Section: Intervals. Original authors, UTC times, and post IDs are retained below.
BPFK Section: Intervals (Tiki thread #5795)
BPFK Section: Intervals
Anonymous · 2006-01-19 19:34 UTC · Tiki post #5796 (reply to #5795)
interval from A to B, not a point on that interval. A point on the interval could be denoted by {lu'a .abu bi'i by}, but {bi'i} can connect brivla and even whole bridi, so I don't see how to denote a point on an interval between them. Any suggestions?
What does {mi klama le zarci .ibi'ibo do cliva le tcadu} mean, anyway?
phmaBPFK Section: Intervals
rlpowell · 2006-01-19 19:44 UTC · Tiki post #5797 (reply to #5795)
> It seems to me, based on the majority of usage, that abu bi'i by > is the interval from A to B, not a point on that interval.
Agreed.
> A point on the interval could be denoted by {lu'a .abu bi'i by}, > but {bi'i} can connect brivla and even whole bridi, so I don't see > how to denote a point on an interval between them. Any > suggestions?
jbini
> What does {mi klama le zarci .ibi'ibo do cliva le tcadu} mean, > anyway?
I haven't the *slightest* idea.
-Robin
-- http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ Reason #237 To Learn Lojban: "Homonyms: Their Grate!"
Proud Supporter of the Singularity Institute - http://singinst.org/BPFK Section: Intervals
lojbab · 2006-01-19 19:59 UTC · Tiki post #5798 (reply to #5795)
>>What does {mi klama le zarci .ibi'ibo do cliva le tcadu} mean, >>anyway? > > I haven't the *slightest* idea.
My interpretation: It asserts to two events that occurred or will potentially occur, not necessarily in any particular order, and draws focus to the time-space interval between those two events, for purposes that must be inferred from context since nothing is asserted about said interval.
lojbabBPFK Section: Intervals
Anonymous · 2006-01-20 02:39 UTC · Tiki post #5799 (reply to #5795)
> On Thu, Jan 19, 2006 at 02:31:39PM -0500, Pierre Abbat wrote: > > A point on the interval could be denoted by {lu'a .abu bi'i by}, > > but {bi'i} can connect brivla and even whole bridi, so I don't see > > how to denote a point on an interval between them. Any > > suggestions? > > jbini
How would you use {jbini} to connect two brivla? Taking your plea as an example, {lo milxe bi'o banli certu} means someone whose expertise ranges from slight to great, which doesn't make sense, unless you're talking about his expertise increasing. {loi milxe bi'o banli certu} does make sense, as a mass of people can have all levels of expertise, but how to say that a single person has expertise somewhere between two limits is not clear. Maybe we could use some experimental cmavo for this.
phmaBPFK Section: Intervals
Anonymous · 2006-01-20 13:39 UTC · Tiki post #5800 (reply to #5795)
> It seems to me, based on the majority of usage, that abu bi'i by is the > interval from A to B, not a point on that interval. A point on the interval > could be denoted by {lu'a .abu bi'i by}, but {bi'i} can connect brivla and > even whole bridi, so I don't see how to denote a point on an interval between > them. Any suggestions?
I posit that the meaning should be "value or values in the interval". This covers both senses, since in extreme cases it could refer to all values in the interval, requiring context to decide which one applies in a given case, but it does favour the presumably minority usage.
With brivla connection, I can think of many examples for the "value within the interval" sense:
ko'a milxe bi'i banli lo ka certu She is somewhere between middling and great in expertise.
lo xamsi cu crino bi'i blanu The sea is between green and blue
ta tirxu bi'i cinfo That's something between a tiger and a lion (a tigron?).
Maybe there are examples where a whole interval sense for {broda bi'i brode} does make sense, but I can't think of any right now.
With sumti, when we need to be precise which sense we mean it can be done fairly easily:
pa me ko'a bi'i ko'e Exactly one value between ko'a and ko'e
su'o me ko'a bi'i ko'e At least one value between ko'a and ko'e
ro me ko'a bi'i ko'e Each value between ko'a and ko'e
loi ro me ko'a bi'i ko'e All values between ko'a and ko'e together
etc.
> What does {mi klama le zarci .ibi'ibo do cliva le tcadu} mean, anyway?
I can't make much sense of it, because the two assertions don't seem to naturally conform the end points of any group of assertions, but what about something like:
ro da poi prenu zo'u ro de prami da i ke'i bi'i ke'i bo ro di xebni da For every person, the truth lies somewhere between everybody loving them and everybody hating them.
Not something very useful to say perhaps, but at least it may be meaningful. (I think {i bi'i bo} should produce an interval of assertions, not of events.)
mu'o mi'e xorxesBPFK Section: Intervals
Anonymous · 2006-01-20 15:24 UTC · Tiki post #5801 (reply to #5795)
> I posit that the meaning should be "value or values in the interval". > This covers both senses, since in extreme cases it could refer to all > values in the interval, requiring context to decide which one > applies in a given case, but it does favour the presumably minority > usage. li'o > With sumti, when we need to be precise which sense we mean > it can be done fairly easily:
As long as there is a way to be precise, I can go along with that.
> Not something very useful to say perhaps, but at least it may be > meaningful. (I think {i bi'i bo} should produce an interval of assertions, > not of events.)
I agree.
phmaBPFK Section: Intervals
rlpowell · 2006-01-20 18:44 UTC · Tiki post #5802 (reply to #5795)
> On Thursday 19 January 2006 14:39, Robin Lee Powell wrote: > > On Thu, Jan 19, 2006 at 02:31:39PM -0500, Pierre Abbat wrote: > > > A point on the interval could be denoted by {lu'a .abu bi'i > > > by}, but {bi'i} can connect brivla and even whole bridi, so I > > > don't see how to denote a point on an interval between them. > > > Any suggestions? > > > > jbini > > How would you use {jbini} to connect two brivla?
lo jbini be lo broda ce lo brodi
Perhaps I'm mis-understanding your question?
> Taking your plea as an example, {lo milxe bi'o banli certu}
lo jbini be lo milxe ce lo banli be'o certu
> means someone whose expertise ranges from slight to great, which > doesn't make sense, unless you're talking about his expertise > increasing.
Oh, that's a good point. I guess I did want a point on the range, didn't I? Didn't notice.
> {loi milxe bi'o banli certu} does make sense, as a mass of people > can have all levels of expertise, but how to say that a single > person has expertise somewhere between two limits is not clear. > Maybe we could use some experimental cmavo for this.
- nod*
-Robin
-- http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ Reason #237 To Learn Lojban: "Homonyms: Their Grate!"
Proud Supporter of the Singularity Institute - http://singinst.org/BPFK Section: Intervals
rlpowell · 2006-01-20 18:45 UTC · Tiki post #5803 (reply to #5795)
> On 1/19/06, Pierre Abbat <[email protected]> wrote: > > It seems to me, based on the majority of usage, that abu bi'i by > > is the interval from A to B, not a point on that interval. A > > point on the interval could be denoted by {lu'a .abu bi'i by}, > > but {bi'i} can connect brivla and even whole bridi, so I don't > > see how to denote a point on an interval between them. Any > > suggestions? > > I posit that the meaning should be "value or values in the > interval". This covers both senses, since in extreme cases it > could refer to all values in the interval, requiring context to > decide which one applies in a given case, but it does favour the > presumably minority usage.
Agreed, once I realized that talking about the interval == talking about all the values on it.
> > What does {mi klama le zarci .ibi'ibo do cliva le tcadu} mean, > > anyway? > > I can't make much sense of it, because the two assertions don't > seem to naturally conform the end points of any group of > assertions, but what about something like: > > ro da poi prenu zo'u ro de prami da i ke'i bi'i ke'i bo ro di > xebni da For every person, the truth lies somewhere between > everybody loving them and everybody hating them. > > Not something very useful to say perhaps, but at least it may be > meaningful. (I think {i bi'i bo} should produce an interval of > assertions, not of events.)
- Boy* does this look like something usage should decide.
Re: BPFK Section: Intervals
xorxes · 2007-05-25 17:03 UTC · Tiki post #6168 (reply to #5795)
ko'a joi joi gi ko'e gi ko'i
My preferred grammar change would be as follows:
Eliminate selma'o BIhI and move its members to JOI. Eliminate selma'o GAhO and move its members to UI.
This has the advantage that ga'o and ke'i could be used with the more general senses of "including" and "except", and not just for the ends of intervals. I have often felt the need for UIs with those meanings.
(The other advantage is the reduction in the number of superfluous selma'o.)
mi'e xorxesBPFK Section: Intervals
PierreAbbat · 2007-05-25 22:14 UTC · Tiki post #6169 (reply to #5795)
> Actually, there is a place where this could happen: > > ko'a joi joi gi ko'e gi ko'i
Okay, so that means it is possible to have a phrase such as {ko'a bi'i ga'o bi'i gi ko'e gi ko'i}. Is that an interval whose closed right end is another interval, or is it an interval whose right end is an interval closed on the left? And what does it mean that an interval's end is another interval? If ko'a, ko'e, and ko'i are points, this could be a triangle with them as vertices; it could also mean an interval whose end falls somewhere in another interval.
> My preferred grammar change would be as follows: > > Eliminate selma'o BIhI and move its members to JOI. > Eliminate selma'o GAhO and move its members to UI. > > This has the advantage that ga'o and ke'i could be used with > the more general senses of "including" and "except", and not > just for the ends of intervals. I have often felt the need for > UIs with those meanings.
Can you give some examples?
Whether GAhO modifies the following or preceding word depends on whether it precedes or follows BIhI. UI always modifies the preceding word. I don't think moving GAhO to UI is right.
phmaBPFK Section: Intervals
JohnCowan · 2007-05-25 23:55 UTC · Tiki post #6170 (reply to #5795)
> Whether GAhO modifies the following or preceding word depends on whether it > precedes or follows BIhI. UI always modifies the preceding word. I don't > think moving GAhO to UI is right.
+1
-- In politics, obedience and support John Cowan <[email protected]>
are the same thing. --Hannah Arendt http://www.ccil.org/~cowanBPFK Section: Intervals
Anonymous · 2007-05-26 00:22 UTC · Tiki post #6171 (reply to #5795)
> Okay, so that means it is possible to have a phrase such as {ko'a bi'i ga'o > bi'i gi ko'e gi ko'i}. Is that an interval whose closed right end is another > interval, or is it an interval whose right end is an interval closed on the > left?
Currently it is not grammatical. With your proposed change it would have to be the former, because the parser will process the first BIhI first. The second BIhI will not be modifiable on the left unless the first one had been modified on the right.
An obvious place for the GAhOs on a BIhI GI ... GI ... construction however is after the GI's: BIhI GI GAhO ... GI GAhO ...
If GAhO were in UI that would be grammatical. As it is now, the GAhO BIhI GAhO GI ... GI ... structure is confusing.
> > My preferred grammar change would be as follows: > > > > Eliminate selma'o BIhI and move its members to JOI. > > Eliminate selma'o GAhO and move its members to UI. > > > > This has the advantage that ga'o and ke'i could be used with > > the more general senses of "including" and "except", and not > > just for the ends of intervals. I have often felt the need for > > UIs with those meanings. > > Can you give some examples?
mi pu vitke ro lo do pendo pe ga'o la djan I visited all your friends, including John.
mi pu vitke ro lo do pendo pe ke'i la djan I visited all your friends, except John.
> Whether GAhO modifies the following or preceding word depends on whether it > precedes or follows BIhI. UI always modifies the preceding word. I don't > think moving GAhO to UI is right.
KOhA GAhO BIhI KOhA would be parsed as (KOhA GAhO) BIhI KOhA instead of as KOhA (GAhO BIhI) KOhA, but it would retain the same meaning.
KOhA BIhI GAhO KOhA would remain as KOhA (BIhI GAhO) KOhA and could by convention retain the same meaning too. KOhA BIhI (KOhA GAhO) would be allowed as well, of course.
So the change would be basically backwards compatible.
mu'o mi'e xorxesBPFK Section: Intervals
PierreAbbat · 2007-05-28 00:08 UTC · Tiki post #6172 (reply to #5795)
> Currently it is not grammatical. With your proposed change it would have > to be the former, because the parser will process the first BIhI first. The > second BIhI will not be modifiable on the left unless the first one had > been modified on the right. > > An obvious place for the GAhOs on a BIhI GI ... GI ... construction however > is after the GI's: BIhI GI GAhO ... GI GAhO ... > > If GAhO were in UI that would be grammatical. As it is now, the > GAhO BIhI GAhO GI ... GI ... structure is confusing.
I think that's a good idea. I propose the following:
- If BIhI is used in afterthought (between two sumti), it can have GAhO before
or after it, and can have zero, one, or two GAhOs.
- If, however, BIhI is used in forethought, it can have zero or two GAhOs next
to it. To use one GAhO, put it after the GI. You can also use two GAhOs and put them both after the GIs.
- If both afterthought and forethought are used together, in the sequence ...
[GAhO] BIhI [GAhO] [GAhO] BIhI [GAhO] GI [GAhO] ... GI [GAhO] ..., proceed as follows:
- If there is a GAhO right after the second BIhI, the GAhO right before the
second BIhI also belongs to it, and if there isn't one, it's an error or nonsense.
- If there isn't a GAhO right after the second BIhI, any GAhO between the two
BIhIs belongs to the first.
- If the second BIhI has GAhOs, and there is also a GAhO after either GI, it's
an error or nonsense.
If we move the GAhOs to UI, sequences such as {.abu bi'i bi'i ke'i gi by gi ga'o dy} will be grammatical, but nonsensical. This is nothing new, e.g. {mi du ra'o do te .u ko'a}.
I think we should also note that it's preferred, in forethought, to put GAhO after GI.
PierreBPFK Section: Intervals
arj · 2007-05-28 19:05 UTC · Tiki post #6173 (reply to #5795)
> My preferred grammar change would be as follows: > > Eliminate selma'o BIhI and move its members to JOI. > Eliminate selma'o GAhO and move its members to UI.
This is ill-advised, because it breaks down the semantic unity of selma'o UI. Can the same effect be achieved by including GAhO in one of the alternations in free_32?
-- Arnt Richard Johansen http://arj.nvg.org/ Q: How many Prolog programmers does it take to replace a light bulb?
A: no.BPFK Section: Intervals
Anonymous · 2007-05-28 20:15 UTC · Tiki post #6174 (reply to #5795)
> On Fri, May 25, 2007 at 10:03:43AM -0700, xorxes wrote: > > > Eliminate selma'o GAhO and move its members to UI. > > This is ill-advised, because it breaks down the semantic unity of > selma'o UI. Can the same effect be achieved by including GAhO > in one of the alternations in free_32?
I think "including" and "except" would fit quite comfortaby semantically in UI3b (with ji'a, ku'i, mi'u, po'o, si'a).
Putting them in CAI would be equivalent, since as far as I'm concerned CAI is just a variant name for UI.
Putting them in FUhE, FUhO, or DAhO (three variant names for the same selmaho) would be almost equivalent. The only drawback is that now they couldn't be followed by NAI. Of course NAI should be merged with all of these too, in which case the distinction disappears.
Selma'o should be purely syntactic groups of words, not semantic. If we want semantic groups, we can use sub-selma'o, but not clutter the grammar with them.
mu'o mi'e xorxesBPFK Section: Intervals
arj · 2007-05-31 16:57 UTC · Tiki post #6175 (reply to #5795)
> On 5/28/07, Arnt Richard Johansen <[email protected]> wrote: > >On Fri, May 25, 2007 at 10:03:43AM -0700, xorxes wrote: > > > >> Eliminate selma'o GAhO and move its members to UI. > > > >This is ill-advised, because it breaks down the semantic unity of > >selma'o UI. Can the same effect be achieved by including GAhO > >in one of the alternations in free_32? > > I think "including" and "except" would fit quite comfortaby > semantically in UI3b (with ji'a, ku'i, mi'u, po'o, si'a).
Only if you take a detour through English.
I think it is fair to say right away that allowing GAhO anywhere seems too innovative, and I need a *very* good reason to accept it.
-- Arnt Richard Johansen http://arj.nvg.org/ There is a great deal of drinking in Japan, unbridled by licensing hours. It forms an important part of semi-official end of work or business negotiations [...], but is also rampant without any such
excuse. -- Ballhatchet, Kaiser: Teach Yourself JapaneseBPFK Section: Intervals
Anonymous · 2007-05-31 19:11 UTC · Tiki post #6176 (reply to #5795)
> On Mon, May 28, 2007 at 05:05:55PM -0300, Jorge Llambas wrote: > > > > I think "including" and "except" would fit quite comfortaby > > semantically in UI3b (with ji'a, ku'i, mi'u, po'o, si'a). > > Only if you take a detour through English.
ie nai .i zo ga'o sinxa lo nu da noi fanmo lo girzu cu ckaji lo kampu be gy. .i zo ke'i sinxa lo nu da noi fanmo lo girzu cu na'e ckaji lo kampu be gy. vu'u da .i zo ji'a sinxa lo nu da noi cmima lo girzu cu ckaji lo kampu be gy. .i zo po'o sinxa lo nu da noi cmima lo girzu cu ckaji lo na'e se ckaji be gy. vu'u da .i le glico bangu na srana
> I think it is fair to say right away that allowing GAhO anywhere > seems too innovative, and I need a *very* good reason to accept it.
I don't know what would count as a very good reason for you. Probably not the same things that would count as very good reasons for me.
I suspect the current grammar of Lojban is unlearnable. I'm almost certain nobody at this point has learned it in full, and even if it is not unlearnable, it is way too complicated for no good reason. Moving GAhO to UI won't solve this problem, but it is s step in the direction of making it more learnable.
(BTW, in the Lojban above I made a conscious grammar error, which I think should not be an error, also in the spirit of making the grammar more learnable.)
mu'o mi'e xorxesBPFK Section: Intervals
arj · 2007-05-31 19:56 UTC · Tiki post #6179 (reply to #5795)
On Thu, May 31, 2007 at 04:13:25PM -0300, Jorge Llambías wrote: > On 5/31/07, Arnt Richard Johansen <[email protected]> wrote: > >On Mon, May 28, 2007 at 05:05:55PM -0300, Jorge Llambías wrote: > >> > >> I think "including" and "except" would fit quite comfortaby > >> semantically in UI3b (with ji'a, ku'i, mi'u, po'o, si'a). > > > >Only if you take a detour through English. > > ie nai .i zo ga'o sinxa lo nu da noi fanmo lo girzu cu ckaji lo kampu > be gy. .i zo ke'i sinxa lo nu da noi fanmo lo girzu cu na'e ckaji lo > kampu be gy. vu'u da
That sentence is ungrammatical. I can sort of guess what you mean, but I'm not entirely sure.
In both of these sentences, I suspect that "girzu" should be replaced by "kuspe".
> .i zo ji'a sinxa lo nu da noi cmima lo girzu cu > ckaji lo kampu be gy. .i zo po'o sinxa lo nu da noi cmima lo girzu > cu ckaji lo na'e se ckaji be gy. vu'u da .i le glico bangu na srana > > >I think it is fair to say right away that allowing GAhO anywhere > >seems too innovative, and I need a *very* good reason to accept it. > > I don't know what would count as a very good reason for you. Probably > not the same things that would count as very good reasons for me.
Would it be helpful if I made a list of everything that I consider a sufficient condition for a grammar change?
> I suspect the current grammar of Lojban is unlearnable. I'm almost > certain nobody at this point has learned it in full, and even if it is not > unlearnable, it is way too complicated for no good reason. Moving > GAhO to UI won't solve this problem, but it is s step in the direction > of making it more learnable.
Lojban is bloated, I give you that. It has nooks and crannies bells and whistles which are arguably never really necessary, such as this one. But is that a good reason to remove it altogether, or relegate it to a sort of pseudo-deprecated status, as happened, against my vote, with zai:
> (BTW, in the Lojban above I made a conscious grammar error, which > I think should not be an error, also in the spirit of making the grammar > more learnable.)
I suspect it wouldn't be stringent enough, and a semantics-aware parser would have to have very ad-hoc rules to extract a semantic structure out of it.
-- Arnt Richard Johansen http://arj.nvg.org/ The names of a species, empire, language, homeworld, homestar and so on will all be self-evidently related; Ogrons come from Ogros, Arisians come from Arisia, Arcturans come from Arcturus, and Humans no
doubt come from Humus. --Justin B. Rye in A Primer In SF XenolinguisticsBPFK Section: Intervals
Anonymous · 2007-05-31 20:43 UTC · Tiki post #6181 (reply to #5795)
> (Replying in English for the benefit of the nalkboka'e.)
OK, the only point of my switch to Lojban was to show that the semantic fit of GAhO in UI3b had nothing to do with any detour through English.
> > ie nai .i zo ga'o sinxa lo nu da noi fanmo lo girzu cu ckaji lo kampu > > be gy. .i zo ke'i sinxa lo nu da noi fanmo lo girzu cu na'e ckaji lo > > kampu be gy. vu'u da > > That sentence is ungrammatical. I can sort of guess what you mean, > but I'm not entirely sure.
How would you say in Lojban what you guess I meant?
> In both of these sentences, I suspect that "girzu" should be replaced > by "kuspe".
Or perhaps {gunma}. If you look at the examples collected by Pierre, some of them are about a continuum and some of them about a sequence of discrete things.
In any case, whatever you use, I think {ji'a} is closer to {ga'o} than to {.oi}, for example, so I don't think you can argue that merging GAhO with UI "breaks down the semantic unity of selma'o UI".
> > I don't know what would count as a very good reason for you. Probably > > not the same things that would count as very good reasons for me. > > Would it be helpful if I made a list of everything that I consider a > sufficient condition for a grammar change?
Don't spend time on that for my benefit, but if you think it's worth discussing I will read it with interest. (I think I already know where our main differences lie: in the different weigh we assign to conservatism for its own sake.)
> Lojban is bloated, I give you that. It has nooks and crannies bells and
> whistles which are arguably never really necessary, such as this one.
> But is that a good reason to remove it altogether, or relegate it to a
> sort of pseudo-deprecated status, as happened, against my vote, with
> zai:
>
> http://www.lojban.org/tiki/tiki-index.php?page=BPFK+Section%3A+lerfu+Forming+cmavo+as+of++17+March+2004&bl
In this case, I don't really want to deprecate GAhO, just make it more useful.
> > (BTW, in the Lojban above I made a conscious grammar error, which > > I think should not be an error, also in the spirit of making the grammar > > more learnable.) > > I suspect it wouldn't be stringent enough, and a semantics-aware parser > would have to have very ad-hoc rules to extract a semantic structure out > of it.
The change I had in mind here was:
Eliminate selma'o VUhU, and move any of its cmavo worth saving into JOI.
mu'o mi'e xorxes
