Lojban In General

Posted by adamgarrigus on Mon 23 of Nov, 2009 11:26 GMT posts: 92

On Sat, May 30, 2009 at 2:57 PM, Luke Bergen <lukeabergen@gmail.com> wrote:

> I know I've read something about a "great rafsi relocation" and that not
> wanting to cause another one of these is the main reason behind not changing
> "tel" for stela (lock) to something else is out of the question. But that
> was a while ago. Is it still completely out of the question to change just
> one little rafsi so we can have 100% consistency instead of 99.9%
> consistency?
>
> I just picture teaching my kids lojban and them saying "but dad, you said
> that there are no exceptions toe the rules in lojban, why is the rafsi of
> 'te' 'ter' when all the others are just themselves plus 'l'?".

Are there any rules about rafsi besides their forms & the fact that every
gismu has at least one? Certainly there is no rule saying that the rafsi for
the SE-cmavo have to rhyme. Moreover, {ter-} is so commonly used that the
anomaly doesn't seem to cause much difficulty. And I would argue that
mistakenly using {tel-} in its place is unlikely to confound, due to the
semantic gap between {te} and {stela}. Lastly, a weighty argument against
drastically changing word meanings is retention of the intelligibility of
existing texts.

I agree that it's annoying, though.

mu'o mi'e komfo,amonan

Posted by adamgarrigus on Mon 23 of Nov, 2009 11:26 GMT posts: 92

On Sat, May 30, 2009 at 3:51 PM, komfo,amonan <komfoamonan@gmail.com> wrote:

> On Sat, May 30, 2009 at 2:57 PM, Luke Bergen <lukeabergen@gmail.com>wrote:
>
>> I know I've read something about a "great rafsi relocation" and that not
>> wanting to cause another one of these is the main reason behind not changing
>> "tel" for stela (lock) to something else is out of the question. But that
>> was a while ago. Is it still completely out of the question to change just
>> one little rafsi so we can have 100% consistency instead of 99.9%
>> consistency?
>>
>> I just picture teaching my kids lojban and them saying "but dad, you said
>> that there are no exceptions toe the rules in lojban, why is the rafsi of
>> 'te' 'ter' when all the others are just themselves plus 'l'?".
>
>
> Are there any rules about rafsi besides their forms & the fact that every
> gismu has at least one?
>

Well, I guess there are at least three gismu without rafsi — {brodi},
{brodo}, {brodu} — so scrap that rule. mu'o mi'e komfn

Posted by stevo on Mon 23 of Nov, 2009 11:26 GMT posts: 381

In a message dated 5/30/2009 17:09:25 Eastern Daylight Time,
komfoamonan@gmail.com writes:

> Well, I guess there are at least three gismu without rafsi — {brodi},
> {brodo}, {brodu} — so scrap that rule. mu'o mi'e komfn
>

Don't those three gismu each have a four-letter rafsi {brody-} and a
five-letter rafsi that matches the gismu itself?

mu'o mi'e stevon

On Sat, May 30, 2009 at 19:57, Luke Bergen <lukeabergen@gmail.com> wrote:

> I know I've read something about a "great rafsi relocation" and that not
> wanting to cause another one of these is the main reason behind not changing
> "tel" for stela (lock) to something else is out of the question. But that
> was a while ago. Is it still completely out of the question to change just
> one little rafsi so we can have 100% consistency instead of 99.9%
> consistency?
>

The entire rafsi/lujvo system itself suffers from systemic inconsistency.

> I just picture teaching my kids lojban and them saying "but dad, you said
> that there are no exceptions toe the rules in lojban, why is the rafsi of
> 'te' 'ter' when all the others are just themselves plus 'l'?".
>

I can't imagine that your kids are going to be miffed about the rafsi for
'te' as opposed to the fact that the rafsi list itself increases the memory
load of lojban by about 150% for no apparent benefit whatsoever.

--
Adam Raizen <adam.raizen@gmail.com>
Got sente?

Posted by stevo on Mon 23 of Nov, 2009 11:26 GMT posts: 381

In a message dated 5/31/2009 06:56:22 Eastern Daylight Time,
adam.raizen@gmail.com writes:

> I can't imagine that your kids are going to be miffed about the rafsi for
> 'te' as opposed to the fact that the rafsi list itself increases the
> memory load of lojban by about 150% for no apparent benefit whatsoever.
>

You don't see a benefit in being able to combine five-letter gismu and get
a word that's often only six letters? I think it's a great idea. I just
wish that it had been done better. The rafsi are definitely a pain to learn.

mu'o mi'e stevon

On Sun, May 31, 2009 at 21:52, <MorphemeAddict@wmconnect.com> wrote:

> In a message dated 5/31/2009 06:56:22 Eastern Daylight Time,
> adam.raizen@gmail.com writes:
>
>
> I can't imagine that your kids are going to be miffed about the rafsi for
> 'te' as opposed to the fact that the rafsi list itself increases the memory
> load of lojban by about 150% for no apparent benefit whatsoever.
>
>
>
> You don't see a benefit in being able to combine five-letter gismu and get
> a word that's often only six letters? I think it's a great idea. I just
> wish that it had been done better. The rafsi are definitely a pain to
> learn.
>

Okay, so there's some benefit, but it's not nearly enough to begin to
justify the huge memory load it requires. Why couldn't the rafsi themselves
be the gismu?

--
Adam Raizen <adam.raizen@gmail.com>
Got sente?

Posted by lukebergen on Mon 23 of Nov, 2009 11:26 GMT posts: 493

because sometimes rafsi by themselves are other words entirely. For
instance "ta'e" is the rafsi for tanxe "box". But it also means
"habitually". This is why rafsi are not true lojbo valsi in and of
themselves. I'm not sure it's possible, given the number of gismu, to have
consistency in how all the rafsi are are formed. But I do wish that the
cmavo would have some internal consistency, especially within individual
selma'o in how the rafsi are formed.

Oh well, probably just a pipe dream.

- Luke Bergen

On Sun, May 31, 2009 at 3:58 PM, Adam Raizen <adam.raizen@gmail.com> wrote:

> On Sun, May 31, 2009 at 21:52, <MorphemeAddict@wmconnect.com> wrote:
>
>> In a message dated 5/31/2009 06:56:22 Eastern Daylight Time,
>> adam.raizen@gmail.com writes:
>>
>>
>> I can't imagine that your kids are going to be miffed about the rafsi for
>> 'te' as opposed to the fact that the rafsi list itself increases the memory
>> load of lojban by about 150% for no apparent benefit whatsoever.
>>
>>
>>
>> You don't see a benefit in being able to combine five-letter gismu and get
>> a word that's often only six letters? I think it's a great idea. I just
>> wish that it had been done better. The rafsi are definitely a pain to
>> learn.
>>
>
> Okay, so there's some benefit, but it's not nearly enough to begin to
> justify the huge memory load it requires. Why couldn't the rafsi themselves
> be the gismu?
>
> --
> Adam Raizen <adam.raizen@gmail.com>
> Got sente?
>
>

On Mon, Jun 1, 2009 at 00:40, Luke Bergen <lukeabergen@gmail.com> wrote:

> because sometimes rafsi by themselves are other words entirely. For
> instance "ta'e" is the rafsi for tanxe "box". But it also means
> "habitually". This is why rafsi are not true lojbo valsi in and of
> themselves.

Of course it doesn't work with the current Lojban phonological system (given
the design goals), but there are a lot of very smart people here and I'm
sure that coming up with a system that fulfilled the same design goals as
Lojban but didn't require learners to memorize 0-3 extra combining forms for
every root word would not prove to be overly ambitious.

--
Adam Raizen <adam.raizen@gmail.com>
Got sente?

Posted by stevo on Mon 23 of Nov, 2009 11:26 GMT posts: 381

In a message dated 5/31/2009 23:57:24 Eastern Daylight Time,
adam.raizen@gmail.com writes:

> I'm sure that coming up with a system that fulfilled the same design
> goals as Lojban but didn't require learners to memorize 0-3 extra combining
> forms for every root word would not prove to be overly ambitious.
>

One possibility would be to create the rafsi first, perhaps using all the
CVC and CCV forms, neither of which can be a cmavo, then adding more or less
arbitrary -CV to each one to make a gismu. Thus the first three letters of
each gismu would provide a unique rafsi.

mu'o mi'e stevon

Posted by lukebergen on Mon 23 of Nov, 2009 11:26 GMT posts: 493

That would be quite an undertaking. with only CVC and CCV forms that would
leave...... 17*5*17 = 1445 forms for CVC and (due to initial consonant
pairs restriction) 48*5 = 240 for CCV form. Added together we have 1685....
that would be just about enough to cover the gismu that we have now with a
little left over to spare.

I doubt that it would happen in lojban, but an interesting thought
experiment morpheme

- Luke Bergen

On Mon, Jun 1, 2009 at 3:31 AM, <MorphemeAddict@wmconnect.com> wrote:

> In a message dated 5/31/2009 23:57:24 Eastern Daylight Time,
> adam.raizen@gmail.com writes:
>
>
> I'm
>
> sure that coming up with a system that fulfilled the same design goals as
> Lojban but didn't require learners to memorize 0-3 extra combining forms for
> every root word would not prove to be overly ambitious.
>
>
> One possibility would be to create the rafsi first, perhaps using all the
> CVC and CCV forms, neither of which can be a cmavo, then adding more or less
> arbitrary -CV to each one to make a gismu. Thus the first three letters of
> each gismu would provide a unique rafsi.
>
> mu'o mi'e stevon
>

OK, I'll put this in the file for LoCCan3. It will pretty much kill the 'easy to learn' 'feature' of the present system, but — absent a test — no one seems to have believed that anyhow. I think this is the first suggestion about phonolgy — and immediately suggests the need to revise the lujvo rules; any suggestions?

From: Luke Bergen <lukeabergen@gmail.com>
To: lojban-list@lojban.org
Sent: Monday, June 1, 2009 8:28:14 AM
Subject: lojban Re: sel ter vel xel

That would be quite an undertaking. with only CVC and CCV forms that would leave...... 17*5*17 = 1445 forms for CVC and (due to initial consonant pairs restriction) 48*5 = 240 for CCV form. Added together we have 1685.... that would be just about enough to cover the gismu that we have now with a little left over to spare.

I doubt that it would happen in lojban, but an interesting thought experiment morpheme

- Luke Bergen

On Mon, Jun 1, 2009 at 3:31 AM, <MorphemeAddict@wmconnect.com> wrote:

In a message dated 5/31/2009 23:57:24 Eastern Daylight Time, adam.raizen@gmail.com writes:

I'msure that coming up with a system that fulfilled the same design goals as Lojban but didn't require learners to memorize 0-3 extra combining forms for every root word would not prove to be overly ambitious.

One possibility would be to create the rafsi first, perhaps using all the CVC and CCV forms, neither of which can be a cmavo, then adding more or less arbitrary -CV to each one to make a gismu. Thus the first three letters of each gismu would provide a unique rafsi.

mu'o mi'e stevon

Posted by lagejyspa on Mon 23 of Nov, 2009 11:27 GMT posts: 350

On Mon, Jun 1, 2009 at 11:22 AM, John E Clifford <kali9putra@yahoo.com> wrote:
> OK, I'll put this in the file for LoCCan3.  It will pretty much kill the
> 'easy to learn' 'feature' of the present system, but — absent a test — no
> one seems to have believed that anyhow.  I think this is the first
> suggestion about phonolgy — and immediately suggests the need to revise the
> lujvo rules; any suggestions?
>

I'm sure I haven't a clue what "LoCCan3" is, and I have REALLY no
desire to throw away all the aging brain cells I've already invested
in learning lojban, but one of the phonological problems I've always
had learning lojban is the of all the theoretical vastness of the
gismu space, there is severe "clumping" in some areas, and other areas
are completely unused. For example, there is ganra/grana/garna/gerna,
xrani/xanri, cernba/cinba, tsina/tsani/tinsa, tinsa/rinsa/risna,
racli/ralci, etc. etc. but not a single gismu with a pattern of *f*j*

• p*v* *v*v* *x*v* *z*x* or *z*z*, nor any gismu using any of more than

470 theoretically possible consonantal triples (in ANY order) such as
{d,f,z} {d,s,z} {d,p,t}, {l,l,m}, {s,t,v} I know that this is driven
by the gismu derivation paradigm, but I think any "reimagining" of
lojban should give reasonable weight to making the phonologic
separation between any two gismu as far apart as possible. This will
help in keeping the two words separate in one's mind.

(The other problem I've had, in learning the deep gismu list, is
the inconsistency. For example while some things have
source/location, some have composition (and I can see that, and don't
necessarily have a problem with that), and some have both source and
composition, those latter all have source first, then composition,
EXCEPT rokci. While most things that have a to and from argument have
it in the order "to" and "from", there are a few exceptions (ex. rinci
and sputu). While most body parts that have three places have the
body in x3, some don't, such as nazbi (possibly forgiveable, as the
subpart is there), kerfa, rango (but compare cigla). Some
measurements have subunits before the "by standard" piece, some after,
etc. I know this has all been discussed elsewhere.)

--gejyspa

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.

LoCCan3 is the shadowy third version of "the logical language" (the CC means "two consonants to be named later"). It exists, if at all, as a list of "corrections" to "mistakes" in the first two verisons (the list for Loglan, and probably Lojban, are out of date). Feel free to add your own.
The clump problem has been noted since at least 1960 on intelligibility grounds (which probably embraces learnabilty as well). Your idea wouldn't help much except to guarantee that all the possible similarities got in (with differentiation only in the second syllable, if at all).
The inconsistent assignment of similar relations to different places — and indeed, any rationale for what relations go where — goes back to the original prospectus for Loglan, where JCB proposed a frequency test. Which was, of course, never actually carried out, since it is often hard to see what would count as evidence and where to get it.

--- Original Message --
From: Michael Turniansky <mturniansky@gmail.com>
To: lojban-list@lojban.org
Sent: Thursday, June 4, 2009 10:29:02 AM
Subject: lojban Re: sel ter vel xel

On Mon, Jun 1, 2009 at 11:22 AM, John E Clifford <kali9putra@yahoo.com> wrote:
> OK, I'll put this in the file for LoCCan3. It will pretty much kill the
> 'easy to learn' 'feature' of the present system, but — absent a test — no
> one seems to have believed that anyhow. I think this is the first
> suggestion about phonolgy — and immediately suggests the need to revise the
> lujvo rules; any suggestions?
>

I'm sure I haven't a clue what "LoCCan3" is, and I have REALLY no
desire to throw away all the aging brain cells I've already invested
in learning lojban, but one of the phonological problems I've always
had learning lojban is the of all the theoretical vastness of the
gismu space, there is severe "clumping" in some areas, and other areas
are completely unused. For example, there is ganra/grana/garna/gerna,
xrani/xanri, cernba/cinba, tsina/tsani/tinsa, tinsa/rinsa/risna,
racli/ralci, etc. etc. but not a single gismu with a pattern of *f*j*

• p*v* *v*v* *x*v* *z*x* or *z*z*, nor any gismu using any of more than

470 theoretically possible consonantal triples (in ANY order) such as
{d,f,z} {d,s,z} {d,p,t}, {l,l,m}, {s,t,v} I know that this is driven
by the gismu derivation paradigm, but I think any "reimagining" of
lojban should give reasonable weight to making the phonologic
separation between any two gismu as far apart as possible. This will
help in keeping the two words separate in one's mind.

(The other problem I've had, in learning the deep gismu list, is
the inconsistency. For example while some things have
source/location, some have composition (and I can see that, and don't
necessarily have a problem with that), and some have both source and
composition, those latter all have source first, then composition,
EXCEPT rokci. While most things that have a to and from argument have
it in the order "to" and "from", there are a few exceptions (ex. rinci
and sputu). While most body parts that have three places have the
body in x3, some don't, such as nazbi (possibly forgiveable, as the
subpart is there), kerfa, rango (but compare cigla). Some
measurements have subunits before the "by standard" piece, some after,
etc. I know this has all been discussed elsewhere.)

--gejyspa

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.

Posted by Minimiscience on Mon 23 of Nov, 2009 11:27 GMT posts: 3588

de'i li 04 pi'e 06 pi'e 2009 la'o fy. John E Clifford .fy. cusku zoi skamyxatra.
> LoCCan3 is the shadowy third version of "the logical language" (the CC means
> "two consonants to be named later"). It exists, if at all, as a list of
> "corrections" to "mistakes" in the first two verisons (the list for Loglan,
> and probably Lojban, are out of date). Feel free to add your own.
.skamyxatra

I would very much like to see this list, preferably as a website or page on
the tiki. If you're taking suggestions right now, my two most well-formed
complaints are that all of the cultural {gismu} should be removed (with the
exception of "{lojbo}," the continental {gismu}, and possibly the {gismu} for
the six source languages), and the metric/non-metric measurement words should
be merged into generic {gismu} for each type of measurement with a single
{gismu} meaning "metric" ("{metro}"?) added to the language.

mu'omi'e .kamymecraijun.

--
li'a .e'i ca vondei .i mi na'e pu'i kufra loi vondei

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.

The last time I checked it was not even all in one place (several people have made contributions over the years). I'll ask around and try to find what the overall status is.

--- Original Message --
From: Minimiscience <minimiscience@gmail.com>
To: lojban-list@lojban.org
Sent: Thursday, June 4, 2009 4:16:58 PM
Subject: lojban Re: sel ter vel xel

de'i li 04 pi'e 06 pi'e 2009 la'o fy. John E Clifford .fy. cusku zoi skamyxatra.
> LoCCan3 is the shadowy third version of "the logical language" (the CC means
> "two consonants to be named later"). It exists, if at all, as a list of
> "corrections" to "mistakes" in the first two verisons (the list for Loglan,
> and probably Lojban, are out of date). Feel free to add your own.
.skamyxatra

I would very much like to see this list, preferably as a website or page on
the tiki. If you're taking suggestions right now, my two most well-formed
complaints are that all of the cultural {gismu} should be removed (with the
exception of "{lojbo}," the continental {gismu}, and possibly the {gismu} for
the six source languages), and the metric/non-metric measurement words should
be merged into generic {gismu} for each type of measurement with a single
{gismu} meaning "metric" ("{metro}"?) added to the language.

mu'omi'e .kamymecraijun.

--
li'a .e'i ca vondei .i mi na'e pu'i kufra loi vondei

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.

On Thu, Jun 4, 2009 at 18:29, Michael Turniansky <mturniansky@gmail.com>wrote:

> (The other problem I've had, in learning the deep gismu list, is
> the inconsistency. For example while some things have
> source/location, some have composition (and I can see that, and don't
> necessarily have a problem with that), and some have both source and
> composition, those latter all have source first, then composition,
> EXCEPT rokci. While most things that have a to and from argument have
> it in the order "to" and "from", there are a few exceptions (ex. rinci
> and sputu). While most body parts that have three places have the
> body in x3, some don't, such as nazbi (possibly forgiveable, as the
> subpart is there), kerfa, rango (but compare cigla). Some
> measurements have subunits before the "by standard" piece, some after,
> etc. I know this has all been discussed elsewhere.)
>

This is my other major pet-peeve with Lojban as it currently stands, making
essentially unlearnable (at least correctly). Gismu places should be
zealously cropped to contain only those places which are absolutely
essential in casual conversation, and then cropped some more, no matter how
necessary someone argues that they are logically or philosophically. No
gismu should have more than 3 places. "By standard" places are going to be
used approximately 1/1000 (wild guess) of the time that a gismu that has
them is used (not to mention 'by epistemology', which will only be used in
philosophical discussions conducted in Lojban). That is not anywhere near
frequent enough for it to be retained in a living language.

--
Adam Raizen <adam.raizen@gmail.com>
Got sente?

On Sat, May 30, 2009 at 2:57 PM, Luke Bergen<lukeabergen@gmail.com> wrote:
> Is it still completely out of the question to change just one little rafsi
> so we can have 100% consistency instead of 99.9% consistency?

If you think that changing any small number of things will make Lojban
100% consistent, you have a lot to learn about Lojban.

> I just picture teaching my kids lojban and them saying "but dad, you said
> that there are no exceptions toe the rules in lojban, why is the rafsi of
> 'te' 'ter' when all the others are just themselves plus 'l'?".

There's no rules you can learn to generate rafsi, I'm afraid. You
have to memorize the rafsi. Some of them are sort of like the gismu,
some are sort of like cmavo, some are pulled out of someone's ass.
That's about all the structure you're going to find to it.

> I know it's a cosmetic thing but it does kind of stick in my head like a
> thorn ya' know?  "sel, ter, vel, xel".  Blech.

OK look, I'm probably just rationalizing, but here's how I think about it.

First, leave "vel" and "xel" out of it. Quick, what's the first lujvo
that comes to mind with "vel" or "xel"? The first one that comes to
my mind is: A complete blank. None at all. If I think for a while I
can remember occasionally seeing "velsku" and "xelkla". But they're
rare enough that it doesn't matter much at all what they sound like.

It's the complete opposite with "sel". It's hard to find lujvo that
don't have "sel". People all the time refer to second place things as
"sel" + the shortest rafsi they can find-- selkla, selsku, selcme.
Those words are basically an accented "sel" with a few sounds tacked
on the end that tell you what you're even talking about.

In between we have "ter", which is much less common than "sel", but
much more common than "vel" and "xel". We've already got an excess of
"sel" all over the place, putting the language in danger of having
everything sound the same, so I find it a relief that there's a little
bit of a different sound to "ter" in there.

Obviously it's not perfect spartan efficiency that to say the x2 and
x3 of cmene we can say se cmene, selcmene, or selcme and te cmene,
tercmene, or tercme. But Lojban is not logical in the old dry way of
the philosophical languages, organizing everything into perfect rows.
It turns out that a little bit of messiness, a little bit of diversity
is necessary in order for a language to be practically useful. It
might seem more orderly to have all the conversion rafsi sound the
same, but I think it is more practical and aesthetic to hear and write
a diversity of sounds.

People complain all the time about the complexity and quirkiness of
Lojban's way of forming words, but there have been countless projects
over centuries which have tried to invent logical orderly
word-formation processes, and most of them are astoundingly ugly.
There's a reason, though perhaps an illogical one, why Lojban's quirky
lujvo succeeded where Ro's sterile efficiency failed. Of course it's
easier to think of a simpler way for rafsi to go together. It's
simple to just make everything even blocks, for instance. Then it's
easy to split everything apart, because it's all the same. In Lojban
everything is not all the same-- it's a quirky, weird little language.
That is why it's the first logical language which so many speakers
have come to love.

mu'o mi'e la stela se ckiku mi'e la .telselkik.

To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.