[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lojban] Response to Robin's "Essay on the future of Lojban"

Long, but I suspect sufficiently comprehensive that we won't have too much more to discuss to achieve whatever resolution you want to achieve NOW. 

Robin Lee Powell wrote:

On Thu, Apr 08, 2010 at 01:55:23PM -0400, Robert LeChevalier wrote:


I call upon everyone involved in this discussion to reread the
policy http://www.lojban.org/tiki/Official+Baseline+Statement


I think the success of that policy speaks for itself.


I think it has been quite successful in achieving a stable language.

Obviously not in getting the baseline done.


The policy explicitly says that "let usage decide" is NOT
applicable until AFTER the byfy completes the 4 tasks assigned
under "THE LANGUAGE DESIGN COMMISSION" is the policy statement
(read the second sentence under INCORPORATION OF PRIOR STATEMENT).


One place where we differ is that I no longer think usage should
decide *ever*.  Unless the BPFK becomes totally controlled by a pack
of drooling morons, of course usage will be acknowledged and
respected, and the BPFK may *choose* to promote usage to the status
of officialness, but we should not let the language drift via usage.
It should be well specified.

I also don't think that describing the lanugage is what we
should be doing.  We should be *declaring* the language.
At this point, I agree with you. The long term is where we still are apart. But we don't need to solve the long term yet.
With regard to the stuff that is presently covered by the baseline documents, I can live with your position, which is easier to sell to computer people and harder to sell to linguists. I am also sure that if it gets to be a problem, you are a reasonable person and would reconsider. 


The problem is that:

1.  defining the cmavo requires making sense of them, which
routinely leads to discovering contradictions and having to make a
decision

2.  no-one wants to do work that will shortly be made irrelevant; if
we *know* there's a problem, defining the current state and then
discussing the change at some future date feels like a waste of time
As you defined the procedures (and as I intended it), defining the current state was supposed to be part of the job of proposing a change. If there are contradictions, explicating them might be sufficient for task 1, and it serves as a major justification for change in task 4.
The point is that defining the current state *should* be relatively easy to get agreement on, ideally a 100% consensus. The language is what it is, warts and all. Can we agree on what the warts are?
Once we've agreed that there is a problem (and an explicitly stated existing contradiction is such), even "conservatives" have no basis to oppose a fix, though we might have differences on how large a fix is necessary.
If task 1 had been done relatively quickly, as I had intended, the delay of discussion would have been minimal, and the clearer agreement and understanding of what the problem is would make discussing the necessary change more efficient. 

(But I accept that you aren't willing to try this now).


I shouldn't have to say this after 7 years, but: it's not going to
happen.  I explicitely refuse to try to document things we are later
going to change without working on the changes as part of the
documentation, and other BPFK members have told me they feel
similarily.

In other words: I refuse to do part 1 of
http://www.lojban.org/tiki/Official+Baseline+Statement by itself.
If you can find people who do not so refuse, let me know.
Can you get people include what is necessary for part 1 as part of their writing the change proposal? (And I am thinking something far more minimal than correcting CLL to describe something that we know we are going to change; just an agreed upon description of what is. If out of all this, you can get change proposals that match what Cowan wrote in justifying changes to the YACC grammar, that well-define what is being fixed, I'll be quite satisfied)
If so, then my position becomes no more than having byfy vote that the part 1's accurately reflect the status quo before trying to finalize agreement on the solutions. Ideally, we'd have a complete set of task 1 description before we start formally trying to approve the changes, but I can bend there.
The status quo has been that checkpoints aren't final until the whole thing is done and voted on as a package. That would not be necessary under the above. If we have a set of task 1 checkpoints complete, then I (and possibly others) have far less basis to oppose a change considered in isolation of other changes, provided the change is fully documented. 


If what you want is a simple description, written by people who are
going to simply ignore the CLL's contradictions and actual usage and
so on, umm, you need to find a completely different group of people.
I'd be satisfied to simply have us agree that the contradictions and any actual usage exists and what they are, before trying to fix them. 


After byfy finishes the four assigned tasks, it (actually the
membership, because the byfy has only delegated authority, but I
can't see the membership overruling byfy) is supposed to declare a
"final" baseline, whereupon the "let usage decide" policy would
take effect.  At that point, no change proposals would be
considered for at least 5 years.  The question is left open as to
what would happen after the 5 years, so the apocalyptic "forever"
of Robin's complaint really isn't "forever".


It's been, what, 15 years since the first baseline was declared?  I
think "forever" is quite an apt description.
But the current policy requires only 5 years after the next baseline. Unlike the present situation, there would be no procedural reason why one or many changes couldn't be voted on, on the 5th anniversary of the baseline. 


I envisioned, in effect, that byfy would be supplanted or
supplemented by the entire Lojban-using community.  Whether some
sort of official body like byfy would make a formal decision is
something the community can decide then.  We don't need to decide
this now, do we?


We need to do *SOMETHING* now.


Agreed.  If only because you, as byfy jatna, want something done now.


I conflated two issues: (1) the forever-in-the-future when the BPFK
is done and usage rules the land, which I think is a bad idea and
You've convinced me more is needed than is promised in the statement. But I still would rather not decide what it is until we reach the point where we have a baseline.
Give me a baseline first, and I might be willing to grant a blank check for whatever some sufficient number of people are able to debate and document solely in Lojban. And I see no problem with byfy as answerer of questions and documenter thereof. (But the documentation should be in Lojban, shouldn't it? What level of documentation of what kind would be necessary for the answer to be sufficiently authoritative? I don't know the answer, and I am not prepared to think about it). 


(2) the built-in respect for the current state, implicit in the
ordering of the line items in section 4 of
http://www.lojban.org/tiki/Official+Baseline+Statement ,which I
*also* think is a bad idea.
Other than the statement of defaults in 1-3, I don't think that order was supposed to be significant. It was merely a list of possible things that the byfy is authorized to do, and the factors we wanted to be considered. The defaults basically say that you have explicitly approve a change - what is unchanged stays the same. Respect for the status quo is explicit (point 8) rather than implicit, but also says it applies "in case of uncertainty".
Minimizing relearning is certainly an argument that can be made against a major change, but nothing in the section 4 points actually requires that as a factor - it was more or less assumed that the need to achieve consensus would ensure that the sorts of things covered by your goals are considered to a degree people find appropriate. Item 8 simply provides a tie-breaker if all else fails. As does item 9's passing the decision to the Board in case of deadlock. I think NO ONE wants the Board to *ever* have to decide such things, so that comes into play only if the byfy-internal process is desperately broken. 


We're basically demanding that every newbie have a gigantic level
of dedication just to use the language effectively.  We might as
well put up a sign that says "Warning:  hard work within".
which seems to argue for a minimal specification 

Absolutely the opposite; it argues for a complete as possible
specification, so newbies do not have to work to understand how to
use the language.


They have to understand the specification to use the language competently.
Two year olds learn natural language without any specification. Two year olds don't learn programming languages that are specified in detail. The fewer the formal rules, the easier it is to know them all.
(And this ignores the problem that the specification is written in English, and the larger it is, the less likely it will ever be translated into Lojban, which I think people agree is how it should be). 


I expect there to be varying levels of documentation; a beginner's
book, the CLL, the online super-spec, that sort of thing.  How much
people choose to learn is up to them; you can speak perfectly good
Lojban without ever opening the CLL, even now.  But when people *do*
have a question, there should be an answer, somewhere.
I'd like to accept all this. But to say that "you can speak perfectly good Lojban without ever opening the CLL, even now" implies that someone who doesn't ever open CLL and is still speaking perfect Lojban even if he violates CLL.
If you can beat someone about the head and shoulders and say "you are doing it wrong" if they don't follow some part of the prescription, then they have to know that part.
In language, older people are less willing to take risks that might be seen as mistakes, which is a major reason why adults have trouble achieving the fluency of young people. Rather than say something wrong, they say nothing. And NORMALLY in language, if people don't know that there is a rule governing something in the language, they wing it (though they may ask for an opinion later).
You seem to be turning this around, suggesting that if there are no rules people tend to say nothing, but ask for the rule, but if there are an infinite number of rules people will be willing to use the language more because they know there is a definitive RIGHT WAY (even though they probably don't know it and would have to look it up in the super-spec to find out). 


Robin and others booted me out, apparently because I wasn't
delegating that power enough (or fast enough).  Robin now seems
unhappy with the result.


I am unhappy only with the stagnation of the BPFK, which you have
not participated in at all to my knowledge.
Certainly not under your jurisdiction. My form of participation was part of what drove Nick from the job, and I don't want to do the same to you.
I tried a few times to take some chunk no one was working on, go off on my own and try to come up with something that others could "shepherd", so I wouldn't be obliged to work under the dynamic of the web-based byfy procedures, but I found I couldn't do it by myself, Nora didn't have the time, and while Cowan said he was willing, it never seemed to happen.
(Perhaps what I should have done was work one-on-one with Jorge via email, who has the patience to deal with me even though we disagree on many issues. I can deal with email, just not with web-interaction.) 


Everything else is
going just fine, IMO.


Good.


If I want natural, growing, living languages, I know where to
find them.


If we never reach that stage, Lojban will never be more than a
toy.



I disagree.  You can have a large, popular language without it being
subject to random, unfettered linguistic; see, for example, French.


French isn't a "natural, growing, living language"??
And yes, it is subject to random, unfettered linguistic forces, especially in regards to the lexicon. Borrowed English words are used frequently despite official opposition.
When people are speaking a language fluently, they don't look at specs or check items in the dictionary. While I like the concept that Lojban might have a better defined spec than any natural language, I know that if the language is successful, then people won't generally be following it religiously. (The ways that they end up NOT following it will likely be linguistically significant, especially when their "incorrect usage" doesn't reflect their native language) 


The fact that xorlo was approved, and then the membership
reinforced that approval, and yet people are apparently still
asking "Are you using xorlo?", 

I don't know where you got that impression; I haven't heard anyone
say that since the approval.


Okay. I misunderstood what you were saying.


Actually, you have.  From me (and to some extent from Nora).
Which is why I have never made an attempt to learn xorlo.  I'm
relying on xorxes claim that I don't need to learn it %^)


You weren't speaking the language *before* xorlo, so that's not
really interesting or relevant to me.
Actually, I was. I never achieved Nick's fluency, but for many years Nora and I were among the only people who could sustain a conversation without a word list handy. Nora and I spoke ONLY proto-Lojban for several hours during our honeymoon in October 1987 when neither of us knew more than 300 words (and they weren't the same ones). One Logfest we called up Jorge and had a short conversation, and we several times did so with Nick.
No, I didn't try to do this at LogFest, except for brief periods. I was wearing too many hats to do so. 


Task 4 explicitely orders the basis for such decisions.  That
ordering *is the problem*.
It states criteria and gives byfy permission to override them when appropriate. Again, I don't think I intended any ordering of those criteria. 


The baseline policy regarding "let usage decide" agrees entirely
with this, with the explicit provision that after the baseline is
declared, no such vote will be even proposed for 5 years (and the
implicit provision that when such a vote takes place, actual usage
will be considered).


That was very much not my understanding; my understanding was that
after that 5 year period, usage *always* wins, even if it's
illogical or stupid.
After the 5 years period it is up to whatever Lojbanists speaking the language decide, which could include any amount of formal change. My vision incorporated a great deal of recognition of what will really happen when the community has a large skilled base of users. It never hurts to have my vision for the language match what will really happen, since it makes me look wise and prescient %^) 

> That at that time the BPFK is not empowered to

over-ride usage in the interests of the overall design goals of the
language, whatever those might be.
We did not explicitly provide for a byfy at that time. If the Lojban-speaking community wants to have an academy working only in-language, and is willing to abide by their decisions to whatever extent, that is fine by me. My vision doesn't require it. With only relatively fluent users making the decisions, it will be "usage deciding" with the usage of the "deciders" being dominant.
"Let usage decide" never meant that all usages, and all users, are equal in influence. That isn't true in natural language (ref. "the Queen's English", "Ebonics", etc) and it won't be true in Lojban. Whether I agree with xorxes or not, his usage is probably at least as strong a model for others as any words in CLL could be (which means that even now, "usage is deciding"). 


If I am misunderstanding, then I assure you I am not the only one to
so misunderstand, and clarification is needed regardless.


Agreed.


A page of approved typo corrections and clarifications doesn't bug
me. The lack of a multi-year period without prescriptively imposed
change, (and without any discussion of such change, except
in-language) would bother me greatly.


I really don't understand why you have this hard-on for the freeze
period?  What difference does it make?  People are coming into the
language all the time; some will be during the freeze, and some
won't.  How does it help anybody?
1) There have been and still are people who will not invest in learning the language because things aren't nailed down. You've more or less said so yourself in this discussion, but think that making a decision solves the problem. But especially in this community, there will always be more proposal being considered and as yet undecided, and your proposed change means that the situation you described for xorlo will be constantly recurring, from the date a change is proposed to the date the formal change pages to CLL and other documents are approved.
(It also means that any copy of a second edition CLL is obsolete the day the first change is approved. For the sorts of people active now and buying the 1st edition CLL even with known problems in it, this isn't a problem. But there are others for whom the obsolescence of CLL means that they won't use it at all, and won't bother to learn the language.)
2) Regardless of point 1), a gap when there is no formal change means that there will be a significant period of time for the numbers and level of fluency to grow, so as to enable what I think we both want - that change ideas be discussed in Lojban rather than English, and the user base becomes large enough that xorxes alone isn't enough to "decide" usage.
As long as it is common for people to announce or propose changes in English, we lose credibility as an effective language, the sorts of changes being considered will tend to have a strong English bias (either to be very much like English or very much unlike English), and not necessarily well informed in the dynamics of Lojban communication.
2a) Lots of talk of change and reform in English makes things look unstable for newbies. If discussion occurs in Lojban, they have to learn the language in order to even understand that change is being discussed.
The 5-year period is somewhat arbitrary, but more or less reflects how long I think it would take for the Lojban speaking community to be large enough and skilled enough (and for the lexicon to be well-enough developed) to discuss change without recourse to English (and maybe to enable CLL and the baselined word lists to be translated into Lojban)
(Actually, the history of byfy suggests that the latter would take MUCH longer, because people don't exist willing to do that much work - CLL took 4 years to write in English with Cowan at the peak of his productivity. But I'm willing to say that if we can't get that far in 5 years, I will support a Plan B that more or less moves things in the right direction.) 


People who want to learn to speak the language seem to deal just
fine with changes; people like you and Nora who don't don't seem to
be able to catch up even over the 5+ years the language *was*
frozen.


???
We haven't tried to catch up. There is nothing to catch up to until the baseline is done. The language remains that of CLL except for a supposedly transparent xorlo. 


I explicitely refuse to spend time writing definitions based on the
current documentation without stopping to consider whether that
documentation makes any sense internally.


By all means, so consider.


This means I refuse to do tasks 1-3 without also doing task 4.
As stated above, as long as you finish task 1 before asking me to *vote* on task 4 changes, I can probably live with this. (And if I don't personally try to participate in byfy, I can probably live with less even while not approving. I really do think as jatna you can do what you want unless overruled, and I can't foresee the Board even trying to overrule you if things are getting done. Your being 1/3 of the Board makes this even less likely %^). 


Task 4, and its ordering of requirements, was what caused the deadlock.
Can you produce a rewording of task 4 (or a statement of byfy policy/procedures that would override the text) that eliminates this apparent ordering of requirements that I don't think we intended, and which would end the deadlock? 

If you can do so, will you try it?
Since I think byfy has the power to prioritize requirements on its own, and in any event the Board has the interim power to interpret the baseline statement and you can easily get a Board majority, this aspect of your issues might be resolved quickly and without acrimony. 

Then, if you are correct, no more deadlock?


You seem to think that everything is fine with the BPFK's charter as
is, yet clearly it isn't, because we've been stuck for 5 years on
{.ai nai} (yes, really, on *just* that issue, for that long).
Alas, Nora and I don't really even know what the issue is (and didn't even know that there was an issue till you mentioned it), so I'll have to get back to you on that. You can be as nasty as you want about my failure to keep up, but I simply haven't been able to follow the discussion, or even to distinguish that there WAS a discussion amidst the overwhelming volume of Lojban List traffic that I have also been trying and failing to follow.
Something you said to another about restarting the byfy mailing list may be part of it. If it stopped, I immediately ceased to recognized that any discussion was going on. in many years, I haven't learned how to read and follow a wiki discussion (or indeed any sort of web-based forum) effectively.
For task 1, I would be happy with a clear statement of the issue and the relevant contradictions involved.
For resolving the issue, this seems inherently *much* smaller than something like xorlo, and I'd likely go along with anything that was clear and not egregiously weird. I doubt that there is enough usage of .ainai for a usage or relearning argument to count for much.
As for byfy being stuck on one issue for a long time, the way it ideally would have worked would have been to drop it and get all the noncontroversial stuff out of the way and come back to the tough issues. You had a new person come in today and write a cmavo definition. I don't know if it is good enough, but I imagine a couple dozen people could do something similar for all the remaining cmavo in a few weeks, even if some of them only do a couple of words. Then debate and vote on individual definitions if necessary rather than entire sections. Especially at a preliminary stage, I suspect that the number of issues unresolved would rapidly drop.
95% of the recent health care reform legislation was done before the meaningful negotiations on the sticking points really got moving. And what the Democrats had to do to pass the legislation over 100% Republican opposition such that they couldn't lose more than a trivial number of votes on their side make it a excellent example of how consensus politics works (at least within the Democratic caucus). 


I'm very, very close to being past caring, but: what do *you* want
to do to fix it?  (and would this involve you yourself actually
doing anything?  because it if would, you might as well keep it to
yourself; I simply don't trust you to actually do any BPFK work.
Too many broken promises)


I won't make any promises because of that.
When this discussion ends, I WILL look at the cmavo definition someone new proposed and comment and I will try to dig up the Eaton stuff some want to work on. Again, not a promise. Just my intent. I am satisfied to be judged on my results and not my promises. 

lojbab

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to lojban@googlegroups.com.
To unsubscribe from this group, send email to lojban+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.