Because of questions that arose on Lojban List and in other discussions regarding the nature and status of the Lojban baseline, the LLG Board of Directors has examined the topic in light of what has been said in the past. We discovered to our chagrin that prior statements, and indeed the official baseline announcement, were to some extent contradictory. Rather than trying to sort out what we meant in prior statements, it has been decided to issue a new statement of LLG policy on the baseline and freeze to supplant all prior statements, and to specifically answer questions that have been raised.
Furthermore, because progress on the dictionary and other critical endeavors has been stalled for too long, the Board of Directors proposes to implement a new plan to complete documentation of the baseline language design as defined by the above policy.
This policy and plan are hereby submitted to the community (not merely the voting membership of LLG), seeking its mandate of support. Those members of the community who are subscribers to the yahoogroups edition of Lojban List will have the option to vote using the Yahoo poll system; those members of the community who are subscribers to the lojban.org mirror of Lojban List, and those on the Lojban Beginners List who are not also on Lojban List, are asked to vote by email submission to email@example.com
This statement specifically supplants the official statement on the baseline which may be found at
http://balance.wiw.org/~jkominek/lojban/9703/msg00012.html and approved motions regarding the baseline in the 1996 and 1997 annual meetings, and well as impromptu comments by Bob LeChevalier over the years.
by Robert LeChevalier
President of The Logical Language Group, Inc.
on behalf of the Board of Directors
It has now been 5 years since the publication of The Complete Lojban Language, by John Cowan and the accompanying baseline of the Lojban grammar.
When that baseline was declared, it was expected that a collection of word lists would be published shortly thereafter in lieu of a dictionary, that upon that publication, the documented lexicon would be baselined, and the combination of lexicon and reference grammar would be frozen for a minimum of 5 years while language usage grew. That collection of word lists has not yet been published; our statement, that the language design was "complete", was therefore premature.
The Board of Directors intends to remedy this situation and reach a definite ending point for the "design" phase of the language as soon as possible, with the goal of having the LLG membership declare the language baseline to be complete at the annual meeting of the LLG in the summer of 2003.
Upon that completion, the language design (baseline) will be frozen, excepting only for the addition of new lujvo, fu'ivla, and cmene, the experimental use of cmavo within the experimental space, and rafsi fu'ivla and compounds in accordance with the description in the reference grammar.
The actual baseline and freeze initially will be on the English language text. Translations of baseline materials to other languages as well as enhanced English definitions may be added to the baseline at a later date having been reviewed for consistency; translations need not be literal translations of the English. Translation of the baseline documents is encouraged, whether or not the translation is intended to be reviewed for formal incorporation into the baseline. See the policy below on LLG support for baseline compliance.
During the baseline freeze, LLG will not consider any change proposals to the language, and will encourage anyone seeking to have a proposal considered after the freeze ends, to formally write their proposal up in Lojban. Any such proposals will be collected in a controlled "wiki" format and accessible on the lojban.org website, but will not be formally considered. If errors in the baseline documents are detected during the baseline period, they shall be similarly documented in Lojban, and may be formally considered by the appropriate language standard group of the time (which by default will be the rump form of the commission established below).
The language design freeze will last for a minimum 5 years after its initial declaration, and may be extended on a year by year basis until such time as discussion of language problems, and potential changes to fix those problems, is substantially discussed in Lojban by skilled users of the language. When that situation is deemed to occur, the question of maintaining the baseline, or terminating the freeze in the language prescription will be turned over to the Lojban-using community.
The specific requirement to complete the design phase will be the preparation for publication of a baseline version of the following: gismu list with rafsi, cmavo list, a limited number of lujvo (100-500) with place structures to serve as examples, and if time permits, a sampling of reviewed and defined fu'ivla borrowings and cmene names, also to serve as
With a rising tide of usage, accompanied by substantive debates about the logic and semantics of the language, we have learned a lot in the last 5 years. The baseline being assumed, most of these debates have been resolved as semantic enrichment of the documented grammar and word lists. But some inconsistencies were noted; some errors were recognized in the published texts; and some questions arose that were not answered in the word lists, where those answers are fundamental to the meanings of words. The new policy endeavors to provide official resolution to these questions, and to correct noted errors.
Because 5 years have passed, and there has been some usage of the language, it is appropriate to consider usage and the issues that have arisen, in a strictly controlled and very conservative review of the design, as part of the big push to get the last of the documented pieces in place. If this review results in consensus agreement that certain changes are appropriate, to be consistent both with actual usage and the engineering goals of the language, they will be incorporated as part of the discontinuity between the existing provisional baselines and the actual design baseline covering the full language.
The Board of Directors intends to commission a review of submitted issues to be considered as part of the process of preparing the final baseline. This commission will complete the following tasks:
- The primary task of the language design commission (banpla fuzykamni) will be to complete brief definitions of the cmavo. The target date for this effort shall be 15 May 2003, in order for consideration by the members at the annual meeting of 2003; if that target date cannot be met, then the member ratification of the final baseline declaration will be delayed until the following year. It is intended that this effort shall take priority over work on other tasks charged to the byfy.
- The secondary task of the byfy will be to define, with place structures, between 100 and 500 lujvo including those most frequently used in actual text. Some portion of these lujvo should be words fitting the patterns of the jvojva as described in The Complete Lojban Language, and some should be explicitly chosen as examples of naturalistic or metaphoric lujvo that cannot be predicted based on jvojva. The format for place structure definitions will use conventions in the manner of the jvojva examples in The Complete Lojban Language.
- The tertiary task of the byfy will be to define with place structures a small number of example fu'ivla from multiple fields, and to prepare a small list of validated Lojbanized cmene from a variety of source languages.
- The final task of the byfy will be to consider other proposals for justified changes to the various baseline documents based on the following principles:
- Because all portions of the language (except for the lujvo, fu'ivla and cmene lists) have been under a preliminary baseline, the status quo is presumed by default.
- In the event of inconsistencies, the published printed text of The Complete Lojban Language, will take precedence by default.
- The published gismu and cmavo lists will be presumed as valid, by default. The byfy can choose at its discretion whether to abide by the intent of earlier language designers or by the strict wording used, and can add clarification or modify the wording based on its decisions.
- The byfy will have the power to correct typographical errors, to resolve contradictions and inconsistencies (including any that are internal to the reference grammar), and to expand upon the text of word definitions in order to clarify meaning.
- If usage has established a pattern clearly inconsistent with the existing documents, but consistent with the design goals for the language, the byfy reviewers will have the power to approve changes to the baseline to reflect that usage. A critical goal is to preserve the fundamental design goal that Lojban words have a unitary and self-consistent meaning. Thus, if multiple meanings for a word have emerged in actual Lojban text, the byfy shall select one meaning, justifying any change from the default. Usage based on alternate meanings shall not be acknowledged in the baseline documentation, and are formally discouraged by LLG. Usage patterns cannot be established on the basis of only one or two individuals, however prolific; they must represent the practice of a plurality of competent Lojbanists.
- If formal logical analysis is inconsistent with either usage or the documented status quo, the byfy may consider adding a brief note to this effect.
- Changes will be documented, including the text of the change, the nature of the change as typo or other correction, expansion of definition, conflict resolution, elimination of inconsistency, or reflection of usage, and the rationale justifying the change (including examples of usage). A collection of multiple related changes (such as typographical corrections), provided that all are accepted by consensus, may be combined in one change record.
- Reflecting the desire of the community for language stability, the byfy will prefer a position as close as possible to the status quo in case of uncertainty.
- Changes must be approved by consensus, with specific procedures to determine consensus decided by the byfy subject to Board of Directors review. In general, a single objector shall not be presumed to deny consensus. If the byfy cannot reach a consensus, and the status quo is unacceptable, the Board will make a decision among alternatives presented by the byfy after consulting with the community, or may choose to delegate the decision to another body.
- Any decision of the byfy which would affect the text of the reference grammar (including the formal grammar and EBNF in the appendices), shall be documented as an erratum to that document. When the final baseline is declared, a complete set of approved errata will be published by LLG, and insofar as it is possible, LLG will consider producing a set of PDF-form pages which could be inserted in a copy of the existing book. If the byfy does not complete its work by the 2003 LLG members meeting, then the Chair will, as part of that meeting, report on the progress of the effort and present a revised schedule for completion which will be subject to ratification by the members.
The Board of Directors appoints Nick Nicholas as Chair of the commission, to report regularly on the progress of the commission to the Board and LLG membership, and to have other executive powers as agreed to by the commission members by consensus.
The Commission will be established by special mailing list at lojban.org. The Board invites all Lojbanists who believe that they have demonstrated some skill in using the language to voluntarily participate in the commission activities. Participation will require being responsive to requests by the Chair, and being willing to actively seek consensus; level of participation will be determined by the individuals. If this open membership becomes unwieldy, or factions develop that are preventing progress, the Chair will seek a modification of this open membership policy from the Board, but our stated preference is to be as inclusive as possible of all Lojbanists who are willing to help in the critical effort of cmavo list documentation.
Nick Nicholas has some specific intentions to automate the efforts of the byfy as much as possible, which he will communicate to the community and/or the byfy members at his discretion. The LLG Board calls on members of the community with appropriate software and web experience to support his plans with the highest priority, so as to get the effort off to a rapid start. He has agreed that if the automation he desires is not in place by a date specific (within 3 months) that he will communicate, that the documentation effort will proceed manually, or with whatever limited tools are available at that point.
The Board of Directors calls upon the webmaster of lojban.org to provide software, mailing list, and web support for the byfy and its activities, as requested by the Chair, and to coordinate volunteers who are providing software and web design. The byfy shall be an official committee of LLG, as documented on the appropriate web page.
Because the Lojban community that has emerged comes from a variety of language backgrounds, and because different members of the community have their own goals and priorities for use of the language, LLG must establish a baseline compliance policy that respects the rights of individuals using the language, one which is not too heavily tied to the necessary reliance on English in the original language documentation.
LLG formally reiterates that the Lojban language design is in the public domain, and any person has the right to use the language in any way he wishes, towards any goals. On the other hand, LLG must strongly encourage people to adhere to the baseline as defined, in the interest of unity of the language community.
Following the formal declaration of the baseline, the LLG Board will assign a committee, probably a rump form of the byfy, with the task of supporting baseline compliance. Any person or group who creates a text in Lojban, or who writes or translates language description materials or language teaching materials in any language, may submit that effort to the committee for review; the committee may also choose to review an effort on its own without formal submission.
In the long term, this committee should become an independent entity not under the jurisdiction of the LLG Board or membership, although the Board or members might retain the right of consent to proposed members of this committee. The implementation of this independence should await the capability of fluent language communication on all matters regarding the language, since discussion in Lojban should be a fundamental principle for the operation of such an independent standard entity.
The purpose of this review will be to assign one of the following levels of declared compliance with the baseline:
- I. Baseline-Standard
- This standard will be reserved for texts and materials that have been certified to be compliant with the existing baseline by multiple people recognized to be skilled in the language, and then approved by the LLG Board or its delegated committee to be officially incorporated in the baseline standard.
An example might be an official dictionary prepared from the baseline word lists, or translations of baseline materials into languages for which we have several fluent speakers among the skilled Lojban community.
All materials that are part of the baseline standard will be set aside on the official web site in a manner clearly distinguishable from other materials.
- II. Baseline-Compliant
- This standard will be for texts and materials that have been certified to be compliant with the existing baseline, but which cannot be declared standard either because
- the material is in a language such that there are insufficient reviewers who competent in both Lojban and the target language to certify the highest level of compliance;
- the text, while thoroughly reviewed, is of a nature not appropriate to serve as a language standard;
- the text, while thoroughly reviewed, includes material extraneous to the baseline (such as experimental cmavo or speculative usages) that the committee does not wish to become a standard for the language by assigning the highest level of approval. An example of the first sort would be translations of the gismu list or other documents into a language for which we cannot get multiple fluent speakers skilled in Lojban to review it, but at least one skilled Lojbanist reviewed it beside the original author/translator.
An example of the second sort would be a significant translation or writing piece in Lojban, which has been certified as compliant by multiple reviewers, but because it is a translation and/or a work of literature, does not actually define a standard for others to follow. Because we have decided not to have a baseline-standard textbook at this point, the textbook by Nick Nicholas and Robin Turner fits this category.
An example of the third sort might be the Proposed Interpretive Conventions page on the wiki.
- III. LLG-Approved Author/Editor
- This standard indicates that the work has been written or edited by someone recognized by the compliance committee as skilled enough in Lojban to perform the task, with the author/editor indicating his own level of compliance with the baseline as part of the work. This allows for works where we have no capability for independent certification, and also recognizes that a certain level of competence in the language provides its own authority. Experimental usages and non-compliant dialects can be tolerated, but the author/editor must explicitly note that the work is not baseline compliant (and preferably will indicate which usages are not compliant).
Ju'i Lobypli and other LLG publications shall be produced to this standard of compliance, with the editor being approved to serve as the compliance authority or delegating that authority to others as she/he sees fit.
- IV. LLG-Sanctioned
- This standard is reserved for efforts such as the Lojban wiki, and jboske discussions, where LLG wishes to encourage public activity in and about Lojban, but where the editor/author, project director, or contributors do not wish to guarantee or enforce baseline compliance.
- V. LLG-Accepted Non-Compliant
- This standard is for works known by the author/editor, or determined by the compliance committee, to not comply with the baseline, but which provide substantial and helpful assistance to Lojbanists. This provides recognition and respect to useful works, while warning users to beware of errant usages. The two versions of the draft textbook, having obsolete and therefore incorrect information in them, would fit in this category in their current form. Writings of language inventor James Cooke Brown, written in and about other versions of Loglan, would also fit this category.
There has been confusion over these terms in the past, which are mostly resolved by setting forth this new policy. However, in the event that confusion again arises, these terms are more explicitly defined and distinguished.
A language "baseline" is a snapshot of a portion or all of the language design as documented at a particular point in time. Baselines are declared so that everyone has a single image of what the standard for the language is, for purposes of discussion and usage. While we might like all usage to be compliant with the baseline, usage will inherently involve some trial experimentation, and in any event people learning the language will not always write in a manner consistent with the formal language description.
Baselines are not necessarily "frozen", but are changed only under formal procedures with all changes being documented. A "freeze" is when the organization managing the baseline commits itself for a period of time wherein it will not consider such changes because of the need for an extended period of stability in the baseline. Most of the language design has been under such a freeze for approximately the last 5 years, but people have been unable or unwilling to work on completing the documentation of a baseline lexicon under the freeze conditions. The efforts of the byfy described above will necessarily amount to a lifting of the freeze to ensure consistency among all the baseline materials, but any changes to existing baseline materials will be documented and justified, so baseline practices will continue. After the baseline is declared complete, a new freeze will then be imposed.
The following material from the 1997 baseline declaration is reiterated in the current statement, with a couple of notes as indicated. These statements apply after the language baseline is declared complete and the freeze is again imposed.
Presuming that an active community forms, further language evolution will NOT take place by prescriptive changes produced by the Lojban designers.
LLG's long-term role will be to promote the language and to serve a research and description function in analyzing how Lojban speakers actually use the language.
Design points not covered in the reference grammar and/or the dictionary are intended to be resolved by actual usage, preferably by skilled speakers of the language. LLG may report on such usage when questions arise, but will not rule prescriptively on whether the usage is "correct" or not. If a usage occurs in natural communication and it is comprehended and accepted by other Lojbanists, then the usage is de facto "valid".
Some Lojban supporters have asked that LLG make specific plans for a language review following the 5 year period, that it establish a committee and/or plan a formal procedure for proposing and discussing proposals during that period for consideration and implementation following the period. The LLG Board of Directors believes that this is contrary to the intended purpose of the baseline and freeze, which is to turn over the language to the users of the language, and to move LLG from ANY prescriptive role which constrains the natural evolution of the language.
LLG makes the commitment, however, that should it ever decide to establish any such procedures after the 5 year period, that all discussions of possible changes will occur solely in the Lojban language, ensuring that only actual users of Lojban will participate in any decision process.Editors of publications sponsored by LLG shall have their own discretion to accept or reject or edit Lojban text according to their own criteria, which may include conformance with the language prescription. If they do so, however, their decisions will be personal, as individual Lojbanists using the language, and based on their own competence in the language. Editors shall not prescribe language usage on behalf of LLG. Similarly, members and officers of LLG ... who are actively speaking and using the language, may express opinions about Lojban text and issues, but shall do so as Lojban users independent of their LLG roles.
Any editor may of course consult with the committee described above for certifying baseline compliance, in deciding whether a Lojban text conforms with the language prescription.
The Lojban design includes language space specifically for trying new ideas in language usage. All cmavo (structure-words) of the form "xVV" are formally undefined, as are all cmavo-form words formed by adding one or more apostrophe and additional vowels onto the currently-used forms VV and CVV. There is also a ... concept for a specific subset of 6 letter fu'ivla (borrowing) space to be used as experimental "fu'ivla gismu"; these would have the ability to be compounded into lujvo (compounds). The lujvo of course would also be experimental.
Note: the byfy will, as part of the process of documenting the cmavo list, have the authority to assign meanings to unused cmavo. These assignments will be for two purposes:
- If an experimental cmavo has seen significant actual usage by multiple skilled users of the language during the last 5 years, the byfy may assign a cmavo in non-experimental space to that usage, recognizing that the usage is no longer experimental.
- If multiple conflicting interpretations of the meaning or usage of a particular cmavo exist, the byfy may select one of these meanings as the standard one, and assign the other meaning(s) or usage(s) to other unused cmavo in non-experimental space.
Recognizing that most experimental cmavo in current use have been of the CVVV form, the byfy will have the authority, if it runs out of unused cmavo space in implementing a and b above, to move the xVV cmavo space out of the experimental realm and assign that as well. This however is a more significant change than merely adding cmavo, and should be thoroughly justified.
Newly assigned cmavo shall utilize only existing selma'o, since a baseline change to the formal grammar such as would be needed to handle a new selma'o would require the highest level of justification, and shouldn't be considered unless a serious error or conflict is identified as part of the work of the byfy.
We note that there is one case that has been identified where the formal grammar and the text of the reference grammar conflict in a minor way. The byfy will have the responsibility to resolve such conflicts. The Board leaves to the discretion of a consensus of the byfy whether to consider additional grammar rules that are consistent with usage and with design goals, do not conflict with any existing standard, and can be successfully YACCed, such as the ongoing discussion of ka'enai; a new official parser must be created if any changes are approved (additional cmavo assignments will require a new parser as well). The requirement for consensus and full documentation will make such changes difficult enough that the Board does not fear that any changes approved will be excessive or unwarranted.
Experimental usages will not appear in the baseline lexicon. The editor of the official dictionary publication will have the discretion to include an appendix documenting experiments that have seen significant usage, noting the fact that these experiments are not officially approved.
Lojban allows for unrestricted creation of lujvo compounds and fu'ivla borrowings in accordance with rules and conventions discussed in the reference grammar. Users are encouraged to coin new words in these forms as needed in speech and writing.
Users are strongly encouraged to provide documentation of such coinages, with definitions and place structures, when they first use them, so that their intent is known. Most lujvo and fu'ivla that have been coined during the last few years have not had such documentation; as such any effort to add them or document them in dictionary form may assign a meaning or place structure different than the person who coined the word might have intended. LLG will provide a means to document such words, accessible via the lojban.org web site.
People using Lojban are encouraged to send archives of texts to
The work of the byfy commission is dependent on having as complete a corpus of Lojban usages as possible, so please submit any existing texts as early as possible in order for them to be used.
Please Contact Us with any comments, suggestions or concerns.