response to Steven Belknap on language baselines and stability (long)

[Logical Language Group <lojbab@ACCESS.DIGEX.NET>]

Pardon for this length.  I put a priority on getting it out so Jorge
could see it before he goes off-line, which means I can't spend too long
editing or thinking.  Suffice it to say that I think Jorge is closer to
my way of thinking than Steven is.  Indeed, my experience is that the
longer someone is in the community and using the language, the more they
see the effects of changes proposed by people who come after them, and
the more conservative they get.  Thus Nora, pc, Cowan and I are even
more conservative than Jorge.  And there are a couple of old timers who
make ME seem like a profligate advocate of change (Oh, Tommy!)

>From: "Steven M. Belknap" <sbelknap@UIC.EDU>
>Subject:      la lojban zasni
>
>>>> We can imagine the impression
>>>> this gives outsiders, when all we can tyalk about is what changes we are
>>>> thinking about.
>
>At this point, who cares? Lets get some of these things worked out before
>we have several thousand speakers.

I think I dealt with this.  If we don't stop changing the language,
we'll never get to several thousand speakers, or even to several dozen.

As it is I am no longer effectively serving the non-net Lojban
community, since I can't keep up with the list as well as fill orders,
and put out newsletters (much less write books).  We have roughly 100
people on Lojban List - just about the same number as 2 years ago.  They
may know the language better, but the community is not growing.
Meanwhile, every once in a while we lose one of our top people.  Nick
and Colin Fine are inactive now and for the indefinite future.  Jorge is
shortly to join them.  Sylvia moved out of the DC area, and while she
still gets Lojban List, no longer actively works with the language.

In conlangs, failure to grow is death.

>The decisions made now are likely to affect many people. lojbab is the
>de facto leader of the lojban endeavor and he needs to carefully
>consider how he will use his authority to guide the birth and
>development of the language.

I most certaianly have done so, and will reevaluate as necessary.

>First, there are various improvements, extensions, and clarifications to
>the grammer, cmavo and gismu.  For example, there are the (related)
>issues of Guttman scales, standards, fuzziness, and cardinality, which
>are discussed in a long series of posts.  When lojbab gets to the end of
>these posts, I would be most interested in lojbabs opinion as to how
>lojban should deal with these issues.  I believe that some of these
>issues are best resolved with constructions already existing in lojban
>by the process Jorge designates "language exploration."

These need no intervention by Lojban Central.

>However, I believe *some* of these issues are best resolved by
>extensions to the language or changes to the language:  the inclusion of
>the selmaho <xoi> and the abolition of standard positions in gismu are
>the only two that I strongly feel require changes/extensions.

These almost certainly will not even be considered before the 5 year
baseline.

>Second, and more importantly, there is the issue of how change is to be
>managed in the language once it is baselined.  JCB doubtless believed he
>had good reasons for managing change in Loglan in the manner that he
>did.

JCB never accepted the principle of baselining, for one thing.  His
language was and is the embodiment of what the French Academy would like
to happen with French.  Nothing new until approved by the Academy, but
the Academy is always open for business.  All changes and innovations
are null until approved from the top.

>This approach was reportedly somewhat autocratic, to the dismay of some.

To say the least.  It 1) doesn't work as the French Academy can testify,
2) excludes skilled voices from participation - your fuzzy logic stuff
would not even be discussed publically.  You would be obliged to write a
full proposal, which would be judged by an Academy without public
debate, and either accepted or rejected or adopted with modifications,
with no interaction with you or the rest of the community.  And 3) the
chages when approved are not effectively promulgated through the
community because of JCB's tight copyright controls, and so are not much
used.

>As I understand it, lojbab's approach is a participative leadership
>style until the language is baselined, and then an abdication of
>leadership, with change henceforth to be managed laissez-faire by the
>speakers of lojban, with no formal apparatus.

No.  Upon baselining, there will be a 5 year period where the ONLY
change permitted will be of the "language exploration" variety.  If
anything comes up that is so serious as to require formal change during
the 5 year period, the baseline experiment will be considered a failure,
and we will have to decide what to do then.  Undoubtedly the community
will be considerably different than it is now, and we cannot decide now
what we will do then.

After 5 years, I envision some kind of review of the state of the
language and a decision to continue the baseline or not.  The procedure
will be formal or informal as seems suitable at the time.  Almost
certainly the LLG voting membership will be consulted at the annual
meeting each LogFest on the status and plans for after 5 years.

>This approach will lead to serious problems, and we will eventually have
>to adopt some formal apparatus to deal with rebaselining the language
>periodically, as informed by usage and future insights as to the
>strength/weaknesses of the language.

I am unconvinced that this will be necessary, and if we get "thousands
of speakers", I am unconvinced it will be possible.  The French Academy
can control all the baselines it wants, but the people speak the
language they choose.

>For those of us now interested in lojban, this approach means we are
>each putting pressure on lojbab to include our favorite extensions to
>the language now before the baselining.

Largely to no effect since, I cannot consider the flood of proposals at
the rate they are appearing.

>If we could be reassurred that there will be an (eventual) opportunity
>to rebaseline the language, perhaps those with proposed grammer changes
>would feel less adamant about including those changes in the initial
>baselining.

I cannot make such assurance, since I am not sure that I will have to
power to fulfill such a promise.

>Change is the one invariant of natlangs.

True.

>We must think carefully about how change is to be managed for lojban.

The absence of effective managed change is a corollary invariant of
natlangs.

I am unconvinced that change needs to be (or can be) "managed" - i.e.
imposed from the top - once we have a large community of users.

>Lojbab has recognized the problem with leadership which occurred with
>Loglan.  It is to his credit that he wishes to provide better leadership
>for lojban.  But abdication of leadership after initial strong
>leadership is a mistake.

Who said anything about abdication of leadership?  It will take strong
and stubborn leadership to say "no" to all change proposals for 5 years.
And rest assured that if I get to actually use the language, that I will
be leading with the best of the rest through my actual usage.  Because
that is how natural languages are "led".

>The de facto result of this will be that those changes which lojbab
>believes ought to be included in the language at baseline will be
>immortalized, and the (possibly valuable) inclusion of others will be
>impeded.

We are essentially already past the point where inclusion of changes
that substantially change the language can be made.  Indeed LONG past
the point.  We have had a grammar baseline in effect for 3 years now.
The first one failed and we rebaselined a year later.  That baseline has
held, with a minimal number of changes making it past the gauntlet since
then.  I contend that the time for debating changes that have
substantial impact on the language are past, and that we need to finish
documenting what we have and start using it.  I have obviously NOT been
successful at getting the most vocal participants on this list to agree
with my policy, and language is essentially democratic.

>Leadership styles run the gamut from rigid and autocratic to constantly
>shifting and anarchic.  There are successful examples of leadership all
>along this spectrum; some endeavors require more heirarchy, others
>require near chaos to work effectively.  The key point here is that the
>nature of the endeavor has a lot to do with whether a given leadership
>style will be successful.  Thus a person designated as a leader must
>carefully consider how he will establish a style which will make success
>possible.  If a leader fails to tailor his leadership style to the
>endeavor, then the success of the endeavor will be imperiled.

I would contend that I did this 8 years ago, and we have a language now
because I *DID* establish a leadership style appropriate to the effort.

>It is useful to consider three common leadership styles:  participative
>leadership, consultative leadership, or collaborative leadership.

All 3 are used within the Lojban community.

>In collaborative leadership, the leader is first among equals,
>decision-making is by consensus among the members of the team.  The
>collaborative leadership style seems most appropriate to manage
>post-baselining change.

And who decides who is a member of the team?  Specifically, what about
the majority of the community that is not on Lojban List?  That
percenatge will certainly grow if the language "works" - imagine what
Lojban List traffic would be like if we had 5 times the current
membership.  One could say that Lojban List participation is
self-selecting, but it favors those with computers, and money and time
for lots of net access.  Any argument that says that the List members
should be the primary deciders excludes, for example, Nora, who enables
me to keep up with the list to the extent I do mainly by refusing to get
sucked into reading it herself.  It excludes Nick, our most fluent
Lojban speaker, and Colin Fine, the longest-experienced skilled speaker
outside of my household (Colin and Nora and pc have all been active
since the mid-70s).  It excludes all but me of our top 10 donors that
have funded the project.

>>From Jorge's and lojbabs discussion on change:
>>>This is a matter of public relations.
>>
>>And one that JCB managed horribly, and which we will manage horribly as well
>>if we cannot present the image of standing very firm against change.
>
>I strongly disagree.  Standing firm against change is not the right
>lesson to learn from the JCB-Loglan experience.

I dare say that you were not involved in the JCB-Loglan experience and
do not know what effect constant change had on the community.

>Rather, matching the leadership style to the project at hand is the
>right lesson.  JCB failed to choose the most effective leadership style
>for Loglan;

It is true that JCB chose an inappropriate leadership model (though then
again, it might have been the appropriate one for the first 20 years -
it just became inappropriate once he opened the language to the
community.

>Lojbab now proposes a different, but equally ill-advised leadership
>style, that is, initial participative leadership followed by the absence
>of leadership.  This is not a good idea.

And I say that you are failing to understand my intent.

>I recently proposed a set of priniciples by which change in lojban could
>be managed. lojbab pointed out that it has already been decided that
>once the baseline is established, no formal changes will be made for 5
>years.  This is fine.  It is probably a good idea to allow some time to
>pass before re-baselining.  But a constitutional language convention
>ought to be scheduled for the end of the five years.

No promises, one way or the other, as to what we will do or how we will
do it.  We do not need to decide now, and any decision now will be
patently unfair to those who will join later.  If we had made a decision
2 years ago, you would not be being listened to now.

>If there is a broad consensus for formal recognition and inclusion of
>the various "slang" extensions and changes which have entered the
>language, then the implications of these extensions and changes should
>be carefully worked out, and then a "lojban language academy" consisting
>of interested experts should deliberate.  If there is consensus on a
>point, it should be adopted; if not, then the change should be deferred
>until the *next* constitutional language convention.  The genius of
>documents like the US Constitution is that they have a built in
>apparatus for amending the document.

The other genius is that they made the process difficult enough that
such change seldom occurs.  They also had the good luck to have the
document become sacred enough that changing it lightly is considered
sacrilege.

Your description of the way things "should be" is precisely the way they
are now - periodic consideration of change proposals by a group of
deliberating experts - the periodicity in this case dependent on our
time and outside activities - which will continue to be the case until
Lojban can support full time workers (not in my lifetime!).  So you are
arguing for continuing the status quo.

It isn't working.  Too many changes get proposed; there is too much
indecisive deliberation, and too many changes are slipping by as "well,
I can't see anything wrong with it, and it is only an extension to the
language, and person X who has contributed mightily to the effort is
fighting hard for it".  It is safe to say that of the primary recognized
experts on the language of a year ago:  Nick, myself, Nora, Colin,
Cowan, pc - only Cowan and maybe pc are taking enough time to evaluate
the proposals Jorge has made, and Cowan does so largely by filtering out
all of the List discussion and deciding on his own.  Colin and Nick and
Nora have not even seen the proposals.

Now take a proposal that is more controversial that Jorge's "simple
extensions" like your fuzzy proposals.

>A. No formal lojban revisions.  If this is true, I will lose interest in
>lojban, and probably not spend much more effort on learning or speaking
>it.  (Unless I can convince myself that there is some other slick method
>of doing the same thing that doesn't take 15 syllables)

We have Peter and you who have called each other names, and an enormous
volume of technical discussion.  And I feel that after it all, I have
ABSOLUTELY NO useful information on which to base a decision.  We have
And's proposal for "xoi", never written up formally so I don't even know
which one it it is, and your agreement that it would satisfy you.  WHERE
IS THE USAGE?!!!  A couple of examples of how it might be used is NOT
"usage" - there is no context, and no sign that anyone looked at
alternate methods to communicate the same ideas.  You have made a
nebulous implication that without "xoi" what you want to say might take
15 syllables, and that therefore you would lose interest in the project.
Can you provide me ANY evidence that the Zipfean demand to use the
construct JUSTIFIES being able to say it in less than 15 syllables,
should it really take that many?  And what am I to make of such implied
threats:  make the change that I want or otherwise satify me, or else I
will likely leave.  Now multiply that across the 100 people on Lojban
List, the 1000 people on our snail mailing lists, the 50 *new* people a
day that hit on the Lojban Web pages, and the 1000+ speakers of the
language we would all LIKE to see exist in 5 years.  How can we POSSIBLY
manage to control and evaluate change when it is proposed so loosely,
yet so forceably, and presented with the dire style of argumentation you
have used in this post?

>B. Formal revision procedure (every five years?) with a lojban language
>academy established.  I will of course be disappointed that <xoi> is not
>in the language, but I will redouble my efforts to learn the language
>with the hope that through better understanding and rational discussion
>I will persuade the members of the language academy to adopt mechanisms
>for handling fuzziness, guttman scales, etc.

Would you accept a formal revision procedure if it REJECTED making any
changes relating to fuzzy logic because you were UNABLE to convince
people?  How about if the academy refused to even consider your
proposals and allow you to "try to persuade" because there are 20 or 200
other proposals on the table, and discussion of the changes alone would
take another 5 years if done at the cvolume and intensity we have seen
on this list.


>>>None of the changes that occured
>>>or were seriously proposed in the last three years (which is the time I've
>>>known Lojban for) has or would have had much effect on Lojban as used.
>>
>>If they have no effect on Lojban as used, then they are unnecessary.
>
>False dichotomy. Jorge said "none...had much effect" not "they had no effect"

The question is whether they were IMPORTANT - important enough to effect
or prevent the success of the language.  That is the standard we need to
be setting to have language stability.

>My amended recommendations for managing change after the baseline:
>
>1. Add a formal version number to the name of lojban
><la papinomoi lojban> or <la pasosoxajoi lojban>
>
>"lojban 1.0 or lojban 1996"
>
>After a five year initial period, if there is a broad and deep consensus
>about changes to be made,

I guarantee that there WON'T be.  There are people even now who vote
"no" on all changes as a matter of principle.

>where the academy of right-thinking lojbi will eventually give the
>stamp of approval to the new version.

whereupon the language promptly collapses and dies in schism because
most of those who did not participate in the debate see no reason for
changing, and feel put upon by calls for them to relearn what they have
put months and years of work into.

Remember that Esperanto survived largely because after 7 years or so,
they had just such a revision proposed and deliberated and voted upon.
The vote of the community rejected the new version (I personally believe
Zamenhof set up the proposal to be self-defeating by including enough
change that most people found something to object to).

So what happens if the right-thinking lojbo say "no"?

>2. As part of the new version release, assure that there is a
>well-defined, nonambiguous translation algorithm from la papinomoi
>lojban -> la papipamoi lojban (or whatever).  Thus all well-formed
>extant texts can be translated.

What if there isn't and cannot be, because of the nature of the changes?
(For example, a simple realignment of "lo" based on the discussions of
the last year or so would alone make this impossible - there IS no
algorithmic way to decide which gadri is to be used if the semantics
change.

3. Agree to a standard notation for lojban version specification; for
example, at the beginning of an utterance, the version of lojban to be
used could be specified.

In short, FORCE schism, since such a notation presumes that ther weill
be people who will NOT use the current standard version.

>4. Emphasize that there will be considerable tolerance to experimental
>additions/changes to lojban among the lojban community.  But these will
>be uncertified "slang" usages until & unless a broad and deep consensus
>builds as to incorporation of the "slang" into the latest release of
>lojban.  Such changes will then be considered at the next meeting of the
>lojban academy

Status quo, except that we have no formal "Academy".  We do have a
voting membership that could establish one if it wanted to (which I am
sure it doesn't).

>5. Maintain a list of recognized problems in lojban.  As solutions
>appear, propose them for inclusion in the next version of lojban.
>Maintain a specification of "proposed, but still under consideration"
>changes/additions/extensions as part of the formal definition of the
>language.

You would be surprised at how large that list is.  Last January, I made
a list based primarily on the le/lo/any discussions of a couple of dozen
"issues" and "problems" nestled in that discussion.  AFAIK, all are
unresolved.  Indeed.  EVERY discussion of the last year that has posed a
"problem" is unresolved, since my "due consideration of proposals" is
stagnated back at 2 September 1994 which is where my old mail backlog
rests.  And I have a couple dozen issues unresolved even from before
that time, though Cowan considers all to be resolved and/or unimportant.

I think you MASSIVELY underestimate the number of issues that will have
to be considered at the time the 5 year baseline concludes.  If we have
ANY requirement to formally consider proposals, then unless we make the
strictures on making a proposal quite onerous, there will be no end to
the deliberations.  Indeed, I suspect that the campaign for the next 5
year baseline change set will start the day the previous one concludes.

>>And as John just mentioned
>>to be today, the problem in writing these thinsg is not the writing itself,
>>but in deciding what to write about.  That last year's discussion of lo/le
>>and family will have ENORMOUS impact on what he evenutally writes on the
>>logic paper, i am sure.
>
>Perhaps John and lojbab are trying to do too much for the baseline. If we
>build in a mechanism for eventual revision, then it will be easier for them
>to get the first version done, as they will be less concerned about making
>everything optimal.

If it is not documented, it is not formally part of the language.  As
simple as that.  People will be learning from what we write.  Any usage
that is not covered will not be widely promulgated throughout the
community, and hence will never get used.

Indeed, this is the bottom line problem with continuous change - all the
formal procedures in the world do not help unless there is widespread
and voluminous propoagation of all changes when they are approved, AND
such changes are accepted and adopted by the community.

The moment Lojban xi 2 comes out, all existing copies of the textbook,
dictionary, and reference grammar are immediately obsolete.  In order to
establish a new baseline we need to have new versions of all those
books, incorporating all the changes, not to mention new parsers and
other tools.  We then have to achieve suffcient promulgation of the
changes that the preponderance of the community almost immediately goes
over to the new version.  And we have to do this in the face of the fact
that any such changes require relearning in order to be so used.

The problem JCB faced with his continuous change philosophy was just
that - there was NO ONE, JCB included, who had a current document that
incorporated all of the changes.  Indeed, my first fight with JCB came
as a result of a summer of looking at his new gismu list and coming back
to him with HUNDREDS of questions and commented regarding
inconsistencies of his new list with things that he had said along the
way in the previosu 12 years.  JCB's current edition of L1 has so many
self-contradictions that it is embarassing.

When it came to new usage proposals, new grammar constructs, new cmavo,
the community was in total asynchronicity.  Changes would be published
along with texts written using the pre-changed language, and newly
proposed changes would appear in the same discussions as approved
changes.  The bottom line was that no two people followed the same
version of the language.

Your suggested method cuts this down only slightly.  There would be a
spasm of indefintiely long deliberation of change and codification of
that change, and then we would have 2 language versions.  But I
guarantee that the new language version will not have fully propagated
by the time the next baseline review date rolls around.  Meanwhile the
academy of experts will be expert ONLY in deliberating, and not in USING
the language, because the volume of proposals will be too large to
consider, decide, and write books for.  Remember that it has taken *8
years* to get to our current state of still not having books published.
How long will it take to rewrite the books each time we go through this
spasm?

>>>> John and I have found, and presume that others find it as well, that the
>>>> image of instability affects those of us working on learning or using the
>>>> language FAR MORE than the changes themselves.
>>>
>>>Then let's change that image. You should not react so violently to changes
>>>that have no effect whatsoever on learning
>
>Agreed. Change in and of itself shouldn't be absolutely forbidden. However,
>once baselined, changes in the formal definition of the language should be
>*hard* to make. They should not be *impossible* to make!

How hard is hard?  Or how impossible is impossible?  I think it is safe
to say that Lojban changes will be imposable by a group of like-minded
leaders on the community with at least as much ease as the Constitution
is amended today - at least at first.

>Lojbab's proposal of letting the users of the language develop their own
>slangs is fine, but we still need a way of specifying what "standard"
>lojban is.

Every person that writes a parser or other tool for the language will be
specifying their own "standard".  Because LLG is NOT retaining
intellectual property control on the language, we cannot and will not
stop people from promulgating such "standards".  If any of them finds a
market either in terms of an economically remunerative application, or
in terms of acceptance by the community, their version will BECOME the
new standard, and all the academies in the world will not stop this.

I thus envision that standardization of Lojban will have the fate of,
for example, the C computer language, where all of the community pays
obeisance back to Kernigan and Ritchey as the original standard of the
language, but the "real" standard is the current compiler
implementations promulgated by Borland and Microsoft in the PC world and
by whatever groups are promulgating Unix versions.

I also note that the K&R language standard has survived something like
25 years and an enormous explosion in usage.

>Lojban is not a natural language, and I believe we loose something by
>allowing it to turn into one if we don't have some kind of standards
>mechanism.  The French and Brazilian Portuguese academies provide use
>with an opportunity to learn from the experience of others.  There is a
>reason why such language academies exist.

They exist to provide a force for tradition.  They also have some legal
clout.  LLG has no legal clout, and if we need a force for tradition to
preserve language stability, then the effort is hopeless.

English has NO academy, and yet has the widest international spread of
any natlang (even if you argue total numbers of speakers).  There is a
reason why no such academy exists.

>I think lojbab should be the president of the convention, and that the
>academy should consist of (maximum) 15-20 people.

This will be decided by the LLG membership, probably at a Logfest 5
years down the pike.  The LLG membership is almost at that 15-20 person
count right now.

There is no guarantee that I will be alive in 5 years or that I will
WANT to be leading such a convention.  (It is safe to say that if I am
leading such a convention that it will be VERY conservative.)

>>This must be done, of course. But most of the problems and solutions
>>will probably not require as much modification of the grammar rules,
>>as modification in how we interpret those rules.
>
>There are solutions and then there are solutions.  As you noted in your
>discussion of ju'e+stag+bo vs stag+bo as logical connective, some
>utterances can get unwieldy, and a shorter utterance may be preferable.
>Same thing with <xoi> and fuzziness.  This is the sort of thing I mean.

If they are unwieldy but still are not frequent, then Zipf's Law will
argue against change.  I doubt that either of these will see enough
usage to warrant significant shortening.  And I daresay there will be
other MUCH more actively used parts of the language that WILL exhibit
Zipfean effects that will have to be dealt with in 5 years, if anything
is.

lojbab