me lu ju'i lobypli li'u 6 moi

From Lojban
Jump to navigation Jump to search

For a full list of issues, see zo'ei la'e "lu ju'i lobypli li'u".
Previous issue: me lu ju'i lobypli li'u 5 moi.
Next issue: me lu ju'i lobypli li'u 7 moi.

Ju'i Lobypli Number 6 - August 1988

Front matter

 Copyright, 1988, 1991, by the Logical Language Group, Inc.
 2904 Beau Lane, Fairfax VA 22031-1303 USA
 Phone (703) 385-0273

All rights reserved. Permission to copy granted subject to your verification that this is the latest version of this document, that your distribution be for the promotion of Lojban, that there is no charge for the product, and that this copyright notice is included intact in the copy.

Published by:

 The Logical Language Group
 Bob LeChevalier, President
 2904 Beau Lane, Fairfax VA 22031
 (703) 385-0273

Lojban gismu List Baselined

See Below for Details

This newsletter is a publication by and for people interested in the development of a 'logical language', based on concepts described in the June 1960 Scientific American article Loglan, by Dr. James Cooke Brown (JCB), and in later publications written by Dr. Brown and by others. The Loglan Institute, Inc. of Gainesville, Florida has obtained a registered trademark on the name 'Loglan' for its line of grammars and dictionaries. This acknowledgement is not to be considered acceptance of the validity of this trademark.

The use of the name 'Loglan' in these pages by the editor or by correspondents is generally not in reference to the trademark, or the products covered by it. Rather it is descriptive of either the movement of supporters of which this newsletter is the spearhead, or to the language development effort proposed and initiated by Dr. Brown, which has now culminated in Lojban. 'Lojban' generally refers to the current version of the language developed by The Logical Language Group. The editorial policy of this newsletter is to print references to 'Loglan' and 'Lojban' exactly as written by the authors to eliminate possible confusion of their intent.

The Logical Language Group is not connected with The Loglan Institute, Inc., and no publication of this organization is, or purports to be, approved of by Dr. Brown or The Loglan Institute, Inc. Dr. Brown has expressly disapproved of this publication effort.

Ju'i Lobypli is distributed for free. Donations are requested in order to cover our publication costs and to support our development work. We have implemented a voluntary balance system - you may maintain a balance from which we will deduct our production costs for materials sent to you. Details may be found below. The Logical Language Group is in the process of being incorporated as a non-profit corporation in the Commonwealth of Virginia by the time you receive this publication. The incorporated organization will be seeking tax-exempt status from the IRS as a 501(c)(3) organization. This may allow your donations, in excess of costs for materials you receive, to be tax-deductible.

Our intended charter requires us to inform prospective donors that we are limited to spending no more than 25% of our income on administrative expenses not directly related to the purposes in our charter. We will account at least annually in these pages to show you that we have met this limit. You are encouraged to make donations contingent upon our meeting this requirement.

This newsletter is being sent to nearly 400 people in about a dozen countries, 10-15% more than last issue, continuing our steady growth. This newsletter is the third consecutive issue to appear on a regular quarterly basis.

I'm happy to say that this issue of the newsletter is truly ours, instead of just mine. Much of the material this time was submitted by contributors from among our readership. See the special discussion. This is great; keep your comments and contributions coming.

WARNING - DO NOT BE INTIMIDATED BY HEAVY USE OF LOJBAN WORDS IN THE FOLLOWING. They are there to get you to look them up in the gismu list attached, and thus make either your money, or ours, well spent. A few technical terms are defined clearly in the glossary at the end. If you do these things as appropriate, what looks like GROSSLY TECHNICAL GARBAGE will actually make sense.

 'Capital Loglan Bulletin Board' (CLBB)
 Available to all in the Logical Language Community
 Local to Virginia/DC Area: (703) 391-8873
 (1200 Baud - login as 'loglan')
 'PC Pursuitable', Lojban Data and Shareware File Upload and Download Provided Courtesy of Joel Shprentz

At publication time, CLBB is down temporarily. See discussion in news below, and keep trying occasionally. Backup Bulletin Board services to the community are available via a Loglan conference on the AMRAD (Amateur Radio) BBS, also local to Virginia/DC (703) 734-1387; 300/1200/2400 Baud; It also has Lojban Data and Shareware, and is PC Pursuitable. Courtesy of Lawrence Kesteloot.

Send messages to the Editor via UseNet/ARPAnet/UUCP care of Joel Shprentz at his address:

'Capital Loglan Bulletin Board' Temporarily Off-line

As many of you have undoubtedly discovered, Joel has been having a series of problems with his computer which supports the 'Capital Loglan Bulletin Board' (CLBB). A lively discussion was in progress in April and May when suddenly Joel's monitor went out, and he could only use the system by dialing in himself from work. The system therefore went down a lot without him knowing it, and people had to wait quite a while for upload and download privileges to be granted. (Normally this was taking only a few days, except when Joel was travelling.)

The monitor problem was solved, and CLBB was up for less than a month when the main computer board failed. As Joel's computer is a UNIX-based system produced by a company that has folded, he has to send the board to the single repair outfit;they are taking a while to get around to his board. He does not expect the board to be repaired before September; October is a distinct possibility. (Joel is also looking for a cheap replacement, just in case.)

We urge you not to give up on CLBB. Keep trying every couple of weeks. Joel has made a strong commitment to support us with CLBB, and has gone out of his way several times to ensure this commitment was met.

Meanwhile, Lawrence Kesteloot, who operates another BBS locally for amateur radio and electronics types, has set up a Loglan conference on that board. He also supports file upload and download via XModem and YModem protocols, and you can do so without waiting for permissions - just register. Lawrence's board operates at 300, 1200, and 2400 baud; I've had trouble with phone lines at 2400 baud, but Lawrence says that others do fine.

I started having modem trouble myself, just about the time Joel started having problems, and I will likely have to get a new one. But we will find ways of getting new material, including the new gismu list, up on Lawrence's board, and then on CLBB when it returns on-line.

Phone numbers for both bulletin boards may be found on page 2.

Report on Logfest 88

Logfest 88 was even more successful than last year's. Not only did we attract more people, 24 altogether, but the participants were more knowledgeable and more able to get involved in our discussions. This presumably means that this newsletter is accomplishing its purpose. Attendance also remained high throughout the weekend and even into Monday. Given the delay in announcing the exact date, we were pleased by the turnout.

To prevent this problem from occurring next year, it has already been decided that LogFest 89 will take place one week later (to reduce conflicts with graduations and the like), the third weekend of June (the 17th and 18th). As usual, we will actually open for arrivals on Friday the 16th, and end on Monday the 19th. Make plans now and mark your calendars. Let us know early of your plans to attend, especially if you are bringing families (who are invited to participate or to use us as a base for Washington DC touring). Camping in the backyard will again be available.

Most important to the feeling of success was that the weekend was dominated by positive efforts towards the success of Lojban. Last year's effort, though well attended, was heavily overshadowed by the politics that birthed the Lojban effort. While organizational discussions were held at Logfest 88 to form The Logical Language Group, they were short and low key. The aura is now one where the politics of the Institute era is now behind us and we can move on to learning and using our language.

We did not make any attempts to have a sustained conversation in Lojban, as we did last year. While we passed around copies of the new cmavo list, the words were too unfamiliar to be located and used at conversational speed. Only a couple of the attendees who had been using LogFlash knew enough vocabulary to converse, so we emphasized more practical problems of teaching, learning, and writing in Lojban.

Bob Chassell led a group effort in a translation exercise. They attempted to translate and rewrite a composition that he wrote in an earlier version of the language several years ago (published originally in The Loglanist). Difficulties with the unfamiliar cmavo list slowed them down, but a portion of the translation was completed. (We had planned to include it in this issue; due to space considerations, we are delaying it until JL7.) Participants felt that they had learned a lot of the concepts behind the language, and had more confidence of their ability to write in the language. It has been decided to include short translation problems in the textbook to give the rest of you the benefit of this type of exercise.

There was also a discussion of teaching methods, especially on those to be used in the textbook. With the lack of a Lojban culture, or of Lojban idioms, teaching people the most elementary conversation skills is difficult. In most languages, simple conversation is dominated by idiom, and those idiomatic expressions are taught early to new students of the language. (Think of our greeting: "How do you do?", or our acknowledgement "You're welcome!", and their possible meanings to an English novice, and you will better understand the problem.) The approach to be used in the textbook, it was decided, will be to teach several ways of saying these basic expressions, in an order determined by their difficulty, and encouraging the students to devise their own additional means of expression. As a specific case in point, we had a group exploration of greetings and closings for correspondence. The results will be included in JL7.

A highlight was the demonstration of Richard Kennaway's MacIntosh version of LogFlash, which he is calling MAC Lojflash. It uses the MacIntosh icon and mouse user interface, and is thus a 'flash'-ier program than our 'glass typewriter' version for PC's and compatibles. Paul and Grelia Reiber donated the use of their new machine to serve as a demonstration device. Art Wieners also brought down his UNIX-based machine, and used it to demonstrate his flashcard program, which uses a Votrax box to say the word when it appears on the screen. Art also provided laser printing services, and he is responsible for the high quality of the gismu lists included in this issue.

Baselining of Phonology and Morphology, Lojban gismu List

The major accomplishments of LogFest 88 were to have been the baselining of the Lojban phonology and morphology, and of the gismu (primitive roots) list. 'Baseline', for those that have not been with us, is a systems engineering term. When an engineering product is baselined, it is set aside in what is called 'configuration control'. Under configuration control, changes can be made only at infrequent intervals, carefully chosen to allow the users of the product to be able to count on its stability. When changes are made, it is only after careful checking and justification. All users are then informed of the changes at the same time so that everyone is using the same version of the product.

Well - we got about 3/4 of the way there. The phonology and morphology seemed quite clear and stable to LogFest attendees. There was discussion of a couple of points needing clarification, and the Synopsis of Phonology and Morphology was approved as a baseline, with minor changes. These changes are given in an attachment to this newsletter, which is separate to allow you to incorporate it into that publication.

We started discussions of the gismu list. Unfortunately, only a few people had brought comments. One person who couldn't attend telephoned some comments, and we had a written submission from Faith Rich. One new Lojbanist, Athelstan, had been studying the existing list, attempting to learn the words. He was using a technique of categorizing the words by semantic type. For example, he grouped emotions together, plants, animals, etc. I had done something similar when we started rebuilding the words, using a classification devised by Paul Doudna.

Realizing only at Logfest that we seriously wanted comments from him and others, Athelstan heroically (the only possible word given how much was going on) stayed up all night on the computer, and loosely categorized every single word. On Sunday, he had a list of words for us that either seemed to overlap, or which had confusing English definitions that had made them difficult to categorize correctly. It took much of Sunday to complete both his and Faith Rich's list of comments, along with the few others we had received. About 20 of the gismu were deleted, and another 15 were significantly changed, with a few new ones being proposed.

As we completed these discussions, two 'camps' in the participants seemed to evolve: reductionists (who wanted to reduce the size of the list), and expansionists (who either wanted to keep the list as it was, or possibly to expand it slightly to fill in remaining gaps). It was necessary to reach a compromise, and we did! It was decided that Athelstan would do a complete review of the list in a little more reasonable timeframe. He would use Roget's Thesaurus as a standard basis for semantic classification. We would generate computerized sorts based on these classifications that would reveal either gaps in our semantic coverage or overlaps that could allow words to be eliminated. Gaps would be filled by choosing concepts that could be used in metaphors (tanru) that filled the void. A subset of the community: Athelstan, John Parks-Clifford(pc), Jeff Taylor, my wife Nora, Tommy Whitlock, and myself would then evaluate the changes - in some cases, actually building several words and choosing the ones which gave better rafsi (affixes) for turning the tanru into compound words (lujvo).

The effort was expected to take 2-3 weeks; as usual for our plans, it took about 6 weeks. The Roget's analysis was very time-consuming. I got involved, working from the Roget's Thesaurus back towards the gismu list while Athelstan continued classifying each gismu in one or more of Roget's multitude of semantic categories. We consulted regularly by phone with Jeff Taylor and pc, who are not local. When we analyzed the results, Athelstan and I went paragraph by paragraph through Roget's, comparing his results and mine. For each paragraph, I had selected a few key concepts that are commonly used, as well as listed the gismu that I believed fit in the category. Each key concept was built as a tanru, or possibly considered as a new gismu. (Athelstan found only one or two possible overlaps in his half of the analysis - and these led to changes in definition that broadened the use of some words in tanru.) By the time we had finished, we had made some 2000 new tanru, which will eventually become lujvo. This number is comparable to the number of lujvo found in the 1975 dictionary. (We estimate that we have between 7000 and 9000 potential lujvo altogether now; this includes some of the work of the old Word Maker's Council as well as a couple of thousand more tanru that I had assembled out of old issues of The Loglanist, built as part of my dictionary work, or devised while doing my earlier semantic categorization. The next dictionary will be much more complete than the last.)

The result was about 60 new words added after the 20 deleted at LogFest. But the added statistical data on tanru also affected our affix (rafsi) assignments. We had to change about 100 words either to assign them adequate rafsi or to eliminate sound conflicts with other gismu. We also uncovered some mistakes - two typos that had led to the wrong words being put on the January list, and one word - simzi - that violated the proscribed medial rules that were finalized after that word had been made. These were fixed, and rafsi were changed as necessary - the result was that 94.5% of all gismu proposed for use in building lujvo can be replaced by a short rafsi in accordance with the lujvo-making algorithm. This was difficult indeed, since over 1000 of the gismu on the list have been used in tanru, and there are simply not enough possible rafsi to cover them all. (For those who don't follow this discussion, you need the lujvo-making algorithm, and some of the explanation in the Synopsis of Phonology and Morphology. Some of the technical terms are defined in the glossary below.)

The analysis was completed on 28 July, and the proposed baseline was completed on 9 August. At this writing, pc and Jeff Taylor are reviewing the changes that the rest of us have agreed on, and the result will be enclosed with this newsletter.

What is the significance of all this?

A good question. These two are the first true baselines in the development history of Loglan/Lojban. We have committed to all of you that this is the best job that we are capable of at this time, that we have completed several analyses and convinced ourselves that our decisions are stable and academically defensible, given the known goals for the language held by the community. We believe that this phonology, this morphology, and this gismu list will stand up to the tests of time and usage, and will not need significant changes.

Natural languages change and evolve; no doubt Lojban will as well. But we have reached a plateau where we can claim that there need not be any further engineering changes. As such, the rate of change will no longer affect you of the community that choose to learn the language. No more will we have word lists and rules that change behind the scenes with such rapidity that no one truly knows what the current list is.

We have committed that there will be no changes in the phonology or morphology for a period of 5 years, and no formal changes are expected even then. We have set a shorter goal, with some restrictions, on the gismu list. There will be no changes to the gismu list until the textbook is completed, or 1 year, whichever is longer. At that point, we will have some usage to back our theory. We also will be attempting to baseline the place structures for the gismu at that time, which may point out some problems that have gone undetected. The rafsi should also have been baselined by then - after we have tested the lujvo.

At the time of this secondary baseline, we will permit specific changes to the list which will have been pre-announced for your comments. We will allow existing words to be deleted (unlikely given the short time involved), and we will allow words to be added. The latter will be added however with the provision that no existing words on the list are allowed to change. Thus, for such concepts, we will likely be choosing gismu that have relatively low recognition scores, and especially gismu that use the remaining unassigned rafsi to best advantage. I should note that we expect that such changes, if any, will probably be countable on your fingers. The far more thorough Roget's analysis only turned up about 60.

The second baseline will then hold indefinitely, again for a minimum of five years. If there are any later changes to the gismu list, they will fall under these new rules that allow only deletions and additions, and no changes.


Nora learned the pre-1981 list of gismu, and she and I both learned the last Institute-published lists using LogFlash. We have also had to learn an intermediate Lojban list prior to last January's published list. We have now started our last relearning. It is about time. We believe that you will agree.

LogFlash News: MAC Lojflash 1 being distributed, Updates for PC LogFlash 1, Status of other LogFlash efforts

Actually, Nora and I haven't been having much trouble on this relearning. I am doing maximum 99 word lessons, and Nora somewhat shorter ones. Nora has not done lessons since March, and I went through the list once since then, just before LogFest. Yet we have retention of over 90%, and the few new words and changes are being assimilated quickly. The following are notes that may help those of you who are using LogFlash repeat our success, or may encourage the rest of you with computer access to get started.

When you have completed all of the new words, the program will have nearly all of them in your 'Under Control' pile. They aren't. You actually will find that you know only about half of them - maybe even less. You will have been exposed to all of them, which means that you will have a much better idea of what each word really means, which is not identical to the keyword (See 'How to Use Your New gismu List' below).

By that time, also, we will have distributed a cmavo list, and depending on how quickly you've started, another newsletter and perhaps the first couple of textbook lessons. You will find that with a few hundred words vocabulary and a very little grammar that you will pick up easily, you will be able to write or say a variety of Lojban sentences. Learn some of the words and then do so! You'll start learning the words even more quickly when you use them.

The second time through the list goes a little quicker. There are a couple fewer lessons to cover all the words, and of course you know many of them. Furthermore, the first time through, you had a full lesson of words in each lesson pile. When Nora and I went through the first time with 40 word lessons, we were often being tested on 300 to 400 words per day by the time we finished the New Words - that's a lot to cover in an hour when you don't know them. In Maintenance mode, only the Under Control pile is a full sized lesson. The other piles consist only of words that you make mistakes on. The size of a daily session drops, and you will find that your time per day will slowly drop towards only a half hour per day.

Once you've made it through the list of gismu about twice, which takes about 2 months of daily 1 hour sessions, things suddenly get much easier. You spend less than 1/2 hour per session after that, and your rate of errors halves each time through the list. At that point, you will likely want to regroup to anywhere from 60 to 99 words per session - it is best to keep your number of errors per lesson high enough that the practices are meaningful (at least 10 is best). By the fourth time through the list, you will be able to do 200-300 words in 30-45 minutes by doing multiple sessions per day, and will thus be able to go through the entire gismu list in a week.

You will find that at the 75% level (probably by the end of your 3rd time through the list), you know enough words to support slow spontaneous conversation, especially with a gismu list handy. By 90% you will not need the list if there is someone else of equal skill to work with, because they will know any words that you don't know, and it takes little effort to communicate across the gaps in the other person's knowledge.

It took me roughly 3 months to reach an accuracy where I missed less than 10 words in the entire 1272 word January list (>99%), nearly all of which errors were typos. Indeed, by about the 5th time through, LogFlash is serving primarily as a typing tutor, since you will generally be responding with answers faster than you can type. (Tommy, who works in an office when not working on Lojban, has noticed that Logflash significantly helped his typing speed and accuracy.) When your error count on a 99 word lesson drops below 4 or 5, you will find the error practices themselves do not help your learning as much as your typing. Error lessons at that point simply serve to keep your focus on problem words.

The limitations seem to be twofold. If you skip more than about 1 or 2 sessions a week, your progress slows or even stops. You can get away with this once, maybe even twice. But LogFlash is designed for daily use. A couple of hours saved early on will cost you several hours in additional days. It is especially useful to get in the habit, even if you have to cut your lesson size down to 15 or 20 (which significantly lengthens the time to get through the list).

The second problem seems to be one of demoralization as much as anything else. It also can lead to taking too much time each day. As you proceed through the first pass through all the words, your rate will range from as low as 20% correct to as high as 60%. (Nora occasionally did better, and I have seen a couple of people do worse, usually because they were just starting or were heavily distracted.) As soon as you know where you stand in this spectrum, you should use the regrouping option to set lesson size so that you average around 15-20 errors per New Word lesson. This means that your New Word lessons will take about 20-30 minutes of the total hour, including review time. Feel free to regroup again later, if your percentage seems to change significantly for several sessions.

You will find that you miss a lot of words on the first Recall lesson (Recog 3 words) - typically about the same error rate as your new word lesson. You aren't reviewing these words first, so you are really doing a lot better. However, your Failure lesson pile will grow with these accumulated errors. YOU MUST KEEP GOING AT THIS POINT - THE PILE WILL SHRINK. You will be tempted to quit before starting a large Failure lesson, then will put it off for several days 'because it will take too long'. DON'T. Those Failure lesson practices that seem interminable are the major learning experience of LogFlash.

You have two better choices. The better one is to regroup your New Word lesson size to as small as 5 words as soon as you notice that your Failure or Recog 3 lesson errors exceed a reasonable number (say 30 or 40). With almost no new words being added to the list, the Failure pile will shrink in a few days, yet you won't completely stop your new word progress. Then you can again increase your New Word lesson size by regrouping again. We also recommend a stretch between Failure lessons and New Word lessons - but we do not recommend quitting there and starting a new day with the New Words. For morale reasons presumably, we have observed that people who do New Word lessons get that added feeling of continued progress that is necessary to keep learning. The second way is to start over again, possibly with larger lesson sizes. This is only worthwhile if you get hopelessly bogged down, or have to skip several sessions in a row. In the latter case, it is only useful to start over if you will be keep regular practices going thereafter.

When I had a word that was 'just off the tip of my tongue', I found it preferable to miss the word and practice it in the Failure pile along with other problem words till I got it right, as opposed to staring at the screen for a while till it came to me. 10-15 seconds is the maximum you should spend on a single word. When you have a bunch of words in your Failure pile, and none of them seem to ring a bell, just keep flashing through them at 5 or 10 seconds per word. After a couple of times through the set, you'll start being able to get a few. The rest will follow possibly after another couple of times through the lesson.

The last and most important suggestion is to use the words you have started learning, even if you have no other Lojbanist to talk with. Keep a copy of the gismu list (in English order) at hand. Throughout the day, attempt to label things and actions with Lojban gismu. You will find that this adds as much as 20% to your scores, and may cut weeks from your time to achieve mastery.

Remember, learning a new language, even an simple one like Lojban, is not easy for most adults. Children apparently are wired for it, but adults have lost this natural talent. Use every psychological edge available to keep going. Believe us - you will feel rewarded by the success of completion.

Now, assuming you've stuck with me thus far, I will describe what is going to happen with LogFlash now that we have a baselined gismu list. As soon as I complete this newsletter, I will start preparing diskettes. I will ship out a fresh diskette to everyone that I know has a copy of LogFlash PC. Since I will send them first class, some of you will receive your update before you receive this letter. It will have a fresh copy of the source, and English- and Lojban- order gismu lists. (The Lojban order list is the only one used by the program.) We will charge your voluntary balances $5 for this update to cover our costs. If you have LogFlash, and do not get an update by the end of September, let me know. But most of you should have yours well before then.

If you already have LogFlash PC, you will have to start over when we send the new gismu list. Keep working at the one you have until then - only 15% of the words are affected, and few of you have error rates that small yet. When you start over, try to notice any words that you already knew that have changed. You are likely to find that the number is small enough that you will learn these words easier than the others that you never learned before. Often the changed word and the old word are similar, which makes Recognition easier, although you will find Recall a little harder.

At the same time, we will be shipping MAC Lojflash to all community members that we know have MACs, along with the new lists. Richard Kennaway has prepared some MAC oriented documentation for you that will build upon the regular PC documentation. The MAC version has the additional option of allowing you to quit in the middle of a lesson without it requiring you to restart. I don't think that this is a good idea, but we await feedback from those of you who acquire the large error piles referred to above.

We also have ongoing efforts to port LogFlash to the Amiga, the Atari ST, and the Apple II. We hope to have one or more of these available by next issue. We are capable of producing CP/M versions with minimal effort, but there are problems with non-standard display interfaces, and the data load times are slow. Those of you with CP/M machines need to speak up if you want us to spend time on this now; otherwise it will be a back burner effort, since the number of CP/M users is decreasing rapidly.

The new gismu list, as many of you may notice, has about 40 of the cmavo included as well. These are the most stable cmavo, and the ones likely to be used in tanru. We have therefore assigned rafsi to these words, and included them in the baseline list, which will be used for LogFlash 2 (which teaches the rafsi) as well as for LogFlash 1. You will also find these words among the most useful to know. Although the cmavo list is not yet baselined, the comments at LogFest indicate that it is quite stable, and we are committed to avoiding changes in these 40 words in particular. See the discussion below on the status of the cmavo list.

Nora has had LogFlash 2 on the back burner for a while. While a lot of you - we believe over 50 - have LogFlash 1, we have no reports of anyone getting through the New Words except for pc, Jeff Taylor, and the three of us in this household. You will note that these are 5 of the 6 that are considered part of the current Word Maker's Council, the sixth being Athelstan (who doesn't have a computer, but has learned a lot of the words without one). The rewards for success are therefore more work, but a louder voice in what happens to Lojban. At any rate, there is no one ready to use LogFlash 2, so we have delayed completing it. LogFlash 3 will teach the cmavo. It has been designed in outline, but data needs to be generated. It also is on the back burner.

To spur you on, Nora and I will donate a $10 balance credit to anyone who reports being in Maintenance Mode before 1 November, when we start preparing JL7. We will also give a $25 balance credit to the first to so report. In addition, we will give $25 balance credits to each of the first three people who reach 90% recall on 10 successive Maintenance lessons. We don't expect any reports of this before November. Any reasonable evidence will be accepted, from a screen printing of your score screen to a hand listing giving your session number(s), and the lesson size(s) you used. We will trust you to be honest. (Word-Maker's Council members are not included in this offer, of course.) This is intended to encourage new LogFlash people to get started now, as well as to encourage those who have had the program for a while to finish. If we find that the 'old-timers' are taking all the prizes, we will add additional prizes reserved for those who are first starting with the baseline list.

Other status: grammar work, cmavo list, teaching materials, Lojban classes, current schedule

I'll be honest. With the massive effort on the Roget's analysis here, everything else came to a stop. We put off incorporation. The brief I am preparing for my lawyer on the trademark problem did not get finished. I stopped working on the grammar (though Jeff Taylor has continued, with some progress in some of the tough problems that were outstanding). I did finish the outline of a synopsis for the grammar, but it may not be written until we've turned out a few textbook lessons. I started on one textbook lesson, but no one here picked up the ball when I dived into Roget. (While I'm still not working - except full time on Lojban - Tommy and Nora are. None of us are getting any money out of Lojban to pay the bills. They therefore have much less time to give to the language on a steady basis.)

The discussions of the cmavo list at LogFest, however, have convinced me that it is nearly ready to distribute. While some concepts have not been tested in the grammar, we found that less than 5% change was needed since my previous draft in February. Since pc has looked much more closely at the cmavo than the gismu, they have received more scrutiny than the latter had prior to our baseline discussions. One change worth noting is that the number '4' has changed from *to to vo, which is a change from the list in JL5. vo, and the rest of the basic 10 digits are among those assigned rafsi, and found in the attached gismu list.

I expect to apply the latest changes and put out a working list by the end of September to all those receiving Level 2 information. (This is those with balances over $20, those who have LogFlash or MAC Lojflash, or those who have specifically asked for teaching materials as they are developed.) There will be examples using each word category, and explanations of each word - in effect, this will be a mini-dictionary and tutorial. It will also be long, possibly over 50 pages. (It is over 30 already without all the examples that I intend to add, and one criticism of the existing list has been a request to use a larger type-face that will lengthen the publication version.)

We also hope to have that first lesson ready to send out by then for review. We will then continue developing lessons as fast as we can. When we have 4 lessons prepared, we will start the local DC-area Lojban class. Unfortunately, this is still the couple of months off that it has been all year. I am guilty of having underestimated the demands on my time for just about every step in the process.

The bright side of things is that we continue to make progress on all fronts. We continue to improve our organization, and to grow in community membership. The report on finances below is full of good news. We also have demonstrated that the language can be developed by people working in consensus: Lojban will be the first successful committee intellectual effort since the King James Bible. And we have achieved baselines that allow all of you to finally learn a language that isn't shifting like sand beneath your feet.

Best of all, we can report that many of you have stopped sitting on the sidelines. The letters below show that many of you are thinking deeply about Lojban, and that your support is active. Several people have volunteered for tasks and completed them, some without my direct involvement (which leaves me time to tackle other problems). Our next baselines will come after many more of you have reviewed and commented on the material, ensuring an ever better language. The logical language community is alive and active again.

Incorporation and Organization of The Logical Language Group

LogFest 88 also served as the organizational meeting for The Logical Language Group. I had prepared a charter, Articles of Incorporation, and Bylaws for people to look at. These were based on the original versions used by The Loglan Institute, Inc., with some changes made to fit our modified circumstances and to correct the problems and abuses that occurred. We had a side meeting at which these were discussed, and initial members were selected. These are - you guessed it - the five who completed the LogFlash New Word lessons, Tommy, Nora, pc, Jeff Taylor, and me, although this wasn't the criteria used.

Our membership will be limited to those actually needed to coordinate Group activities; we do not want to turn it into a money-raising scheme, or as an excuse to shut out everyone else who doesn't meet our criteria. We also ensured that Nora and I were not ourselves capable of dictatorially dominating the group, while keeping a minimal quorum in Virginia. We selected initial officers, who will serve until our first annual meeting, which will be at LogFest 89, and thereafter on the 3rd Saturday in June.

I will be the initial President, pc will be the initial Vice President, and Nora will be the initial Secretary/Treasurer. As Nora has no accounting experience, and we have had an awful time trying to set up accounts and books that we are satisfied with, we will probably pay for an accountant to set us up correctly at our start.

I have modified the initial documents to reflect all of these things, and will be calling an attorney to review the documents and help us file the papers. I have established a bank account in the name of The Logical Language Group, and Nora and I have placed separate funds in this account equal to your outstanding balances. Upon incorporation, the account will be transferred to the corporation. (Right now, I am legally equivalent to The Logical Language Group, pending this incorporation.)

We have started operating as we will after incorporation, hence the notice about administrative expenses and donations on page 1. (The Loglan Institute, Inc. had a similar requirement, but with a 10% limit on administrative expenses. It has not met this requirement for notice; we are trying to do better.)

With any luck, we will finally be incorporated as a non-profit corporation by the end of September. Any of you are welcome to obtain a copy of our incorporation documents (at cost applied to your balance). We also will consider applications for membership, stating your reasons for wanting this, and what commitments you are willing to make in support of the responsibility of membership. (We don't ask much, do we?) We aren't really looking for 'membership'; we consider all of you 'members' of the community to be our constituency. Thus formal membership grants no especial privileges; just added responsibility.

Lojban Publicity Plans: video, USIA, the scientific press, science fiction/fantasy conventions, brochure distribution

People have asked about various efforts to obtain publicity for Lojban. We have thought about this a lot, and some plans have gone ahead. However, until we have the textbook well on its way to completion, we are expecting to keep our advertising low key and generally by word-of-mouth. We do not want to overcommit, and to find ourselves with thousands of inquiries, with nothing lasting to offer. Many of you experienced something like this in either 1960 or 1975. We need a stable organization and a cadre much larger than my household, one that is knowledgeable about the language, and which is capable of dealing with inquiries, before we can handle massive response.

As a result of this, we have separated our publicity goals into short term goals and long term goals. In the long term, of course, we want Lojban to reach world-wide audiences and have lots of speakers. In the short term, we are trying to build that stable organization, with small groups in each metropolitan area, which we are capable of supporting with our current scale of operations.

Guess where all of you come in. Even those of you who cannot take the time to learn the language now can learn about the language. When you know enough to answer your own questions, pick one other person and tell them about Lojban. Try to answer their questions, and obtain their interest. It may take a few people before you find someone interested. You will convince yourself that you know more about the language than you thought, and incidentally find someone local that you can talk to about the language. We of course will double in size if everyone does this, as well as prove our capability to meet the public needs for information. Then when people ask for information, we can pass them to the local people (you!) who can give better, more personal answers to questions, and to provide a support network.

Meanwhile we also have other means to spread the word about Lojban slowly through the likely community of interest. I have given presentations at the local Evecon and Unicon science fiction/fantasy conventions. At least one other presentation has been given, at a Blacksburg, Virginia convention. We have had brochures available at three other regional conventions, and I have reports of people who found out about Lojban at another convention (Lunacon) that I didn't know of. We get several new people in response to presentations, and more send in brochure registrations later.

We are planning for future presentations as well. Brad Lowry, one of the apprentices taught by Scott Layson in the 1970's, is now a professional in the audio and video industry, and has offered to assist us in producing a short video describing the language and its unique features. We're planning this for fall, and the resulting video should be available early next year. This will take the burden of telling the Lojban story off the presenter, leaving me, or you, free to answer questions. I have initiated contacts to try to enable our participation in the World Science Fiction Convention in Boston next year.

We also are spreading through the UUCP/Usenet/ARPAnet community. There apparently are several forums on artificial intelligence, linguistics, and language development, and every couple of weeks we get a new letter from someone who has heard of our work and our progress. Since I don't have access to these nets, it is you people who have given us this publicity. Apparently, you are saying good things about us. (These nets are worldwide, by the way. We have had a response from Switzerland, apparently from one of these nets.)

If you don't like trying to find people interested in Lojban directly, there is another way you can help, while keeping our growth rate manageable. Nearly everyone has a university or college nearby, and some of you are attending one. Let us send you a dozen brochures, or make copies of yours (including the registration form). Post them in the linguistics and/or computer science departments on a bulletin board, optionally putting your own phone as a local contact. Jeff Prothero brought us a half dozen people in a couple of weeks this way. Again, this boosts your local area Lojbanist community, making it more possible that a class can be conducted, or that you will find a local Lojbanist to talk with.

Now for the long-term plan. When we do have materials, we intend to be ready. We are trying to identify contacts in the scientific press who will review our activities. Articles in Discover, Science, Scientific American, or the New York Times Science section would give us an audience of millions. If but 1/10 of a percent respond, we must be capable of supplying tens of thousands of copies of our publications. Either we hire employees at that point, or our development work stops (it might anyway). We had better truly have the language done, by anyone's standards, at that point. I also figure that we need between 1000 and 2000 people before I can financially justify regular outside help.

A second possibility is enhanced by our location here in DC. We have interested one person at the US Information Agency in giving us an interview on one or more their overseas broadcasts when we are ready. This is especially attractive, since it would truly make Lojban an international effort. I've been told that over 10 million Spanish-speaking people listen to these broadcasts in South America alone. A world-wide audience, especially with a well-planned infrastructure to support the teaching of Lojban, would bring us to the forefront of linguistic attention (hopefully favorable).

I suspect these latter possibilities are months away. I'm learning, often the hard way, how to manage a mail-order shop with 400 customers, and still get some work done. When we hit 1000, we'll be ready to talk about 10000. So help us get there. Spread the word about Lojban.


Money being the eternal worry of a volunteer organization, I have started to research the most direct way to solve this worry. Being located in DC, I can easily obtain information on government grant programs. I'm also somewhat experienced in writing proposals, having worked on several as a defense contractor (though I'd like to find someone out there with better experience, especially in the non-profit grants field - any takers?). I also briefly approached my last employer, after being laid off, about a commercial grant or subcontract, but decided to put this off until I have more to show.

Let's face it. I will eventually have to go back to work. I do not have Careers to support me through the rest of my life, and Nora shouldn't have to do so. I will either have to find a job, or find a way to get income from Lojban. Since I have committed to Lojban being a public domain effort, I cannot expect royalties. As an officer of The Logical Language Group, I am not paid by that organization. Lojban is not going to be a profitable activity for me or anyone for a long time, if ever.

If I can obtain employment with one of the few firms that could have application for Lojban, that would be great. If I can get a contract or grant to support my (or our) work, we could proceed even faster in producing materials. Grants would also open doors for non-monetary support that are often closed otherwise.

I've identified several possibilities, and am investigating them. But I haven't had much time yet, as described above. Give us a few months and our non-profit status, and grants will be more likely.

News from the Institute

This section is inherently self-serving to our organization. I cannot claim to be unbiased, and the Institute's activities have often been directly opposed to our own. However, several things are going on that aren't totally of a political nature, so we will try to keep you informed. If nothing else, by contrast, our plans and policies will be more clear by comparison with those of The Loglan Institute, Inc.

Interestingly enough, JCB has recently rediscovered grants, himself. A recent Bulletin was sent to some members of the community, asking opinions as to whether the Institute should seek grants, attempt to obtain a professional grant-writer as the next CEO, etc. He also asked whether the Institute should grant licenses on the use of Loglan, presumably referring to his 'trademarked' products. There was also a piece in which JCB compares himself (favorably) and the Loglan movement to Zamenhof and the Esperanto movement; he describes the latter as much smaller than he had previously thought, and not growing, based on a recent book about the movement that he read.

JCB also discussed some of his own plans. He states that he will complete a revision to Loglan 1 (L1) this fall, and that this, along with a revised Notebook 3, called Loglan 6, along with some teaching programs that Bob McIvor (RAM) has been working on, will be sufficient teaching material for the community. Unfortunately, he sounds as far as we are from producing a dictionary, and neither NB3 nor L1 has been sufficient to teach the language in the past. (People who have received NB3 have described it as inadequate for self-teaching. L1 is an interesting language description, but also not a textbook. It is inadequate for teaching more than simple sentences in isolation.)

RAM is likely to develop excellent programs, but porting them to other machines is not trivial, as Nora and I have found (RAM has a MacIntosh). And only about half of you who have responded have computers available at home that could be used for such programs.

A written textbook, of professional development quality, is the only approach that will give lasting satisfaction, and JCB cannot do the work needed for such a quality product himself in the timeframe he has promised. (Chuck Barton has estimated that the job would take more than one person several months. He is a professional language instructor, and his estimates match what I am experiencing.)

JCB also has no infrastructure capable of supporting a post-publication influx of correspondence. He has alienated most of the best workers in the community, and he himself has proven incapable of timely response to correspondence. (I know - I should talk. But I don't think I've taken longer than two months to answer you, while letters to JCB or The Institute have taken 6 months or more for a response recently, and then they were inadequate for providing the requested information. I've been told that in 1960, pc never even got a reply from JCB, so I guess he has improved.)

The community will need a full-time public liaison office when we 'explode' upon the public. And people will not bother to learn the language unless they think that it has a future. JCB talks constantly of retirement, of the job being complete. But I knew when I started my newsletter two years ago, and am now even more convinced, that the job will have only begun. JCB has told us that we can't count on him being there, and he has provided no one else of comparable competency to fill his role when he isn't.

JCB also produced a LogNet about the same time that JL5 came out. It consisted of a little news - mostly irrelevancies about his ocean voyage - and a lengthy diatribe against my 'errors' in the previous issue (JL4). I'm not going to answer this for several reasons, most especially because it is a waste of time. I'm not likely to convince JCB, and I don't need to convince you. Since JCB has complained of my countless high crimes and misdemeanors, I have received not one letter or comment from anyone other than him that questioned my motives or activities. Not one of you has dropped out, to my knowledge, because of what JCB has said. A couple of you still want me to continue to try for peace and reconciliation to JCB, and I have kept my last proposal on the table.

A second reason, and the only other one I'll mention, is that JCB chose to respond, not in these pages, but in his own. I don't know who received LogNet. (JCB's mailing list for Bulletins and LogNets has not matched the membership lists of the Institute in several recent mailings. As such, I can't tell who has seen any given statement of his.). Even if I wanted to, I can't repeat his multi-page rebuttal for those of you who didn't see it; he claimed copyright on this, too. So, while his comments are filled with numerous misinterpretations of history and the motives of myself and others, they will go unanswered. I'm sure you all won't mind.

Enough of this. Back to Lojban!

How to Use Your New gismu List

Many of you are getting this list for the first time, and others who have received it have not gotten much explanation of what it contains. I'll attempt to remedy this.

There are 1316 gismu, the primitive 'building block' concept words of Lojban. There are also 41 cmavo, short structure words that interact with gismu, their borrowed relatives (le'avla), and compound complexes (lujvo) to form Lojban sentence-like constructs, called bridi. gismu, le'avla, and lujvo together form a category of words called brivla. (For those with the Synopsis of Phonology and Morphology, you will note that le'avla were called *na'evla, and brivla were called *ridvla. As the change document for the Synopsis attached to this newsletter explains, these words have changed due to the changes in rafsi as part of the gismu baseline.)

The list as you are getting it is in the form used by LogFlash 1 (and eventually by LogFlash 2), and by MAC Lojflash. There are several columns of data. The first column contains Lojban gismu (and the several cmavo, which are the ones that are shorter than 5 letters long). The next space, with three columns of data, are the rafsi (affixes) for each gismu, if they have been assigned. The first column gives a CVC(Consonant-Vowel-Consonant)-form rafsi where one is assigned to the gismu, the second column gives a CCV-form rafsi, and the third column gives a CVV-form rafsi. These are used in building lujvo; see the examples below. The rafsi also appear on your LogFlash or MAC Lojflash screen to help expose you to them - you will eventually need to learn them.

The next column gives an English keyword to help you associate with the meaning of the gismu. THESE KEYWORDS ARE NOT EXACT DEFINITIONS. In most cases they are only approximations of the Lojban meaning. Most English words have more than one meaning; sometimes these meanings are unrelated. The Lojban word has a single meaning, possibly a continuous range of semantic space. Thus klama includes the concepts covered by both the English words 'come' and 'go', when an origin and destination are specified, as well as 'travel' when it refers to a destination, as well as 'move', as in 'Move over!'. 'Come' is used as the keyword, because it is closest in both sound and meaning to the Lojban, and also has few other interpretations that are inconsistent with the Lojban (though most of you can think of at least one, I'm sure.) The concept associated with keyword 'travel': litru, on the other hand, refers to motion without necessarily having a destination, though a route is generally applicable. Thus you can litru around a circle endlessly, but not klama, which implies that you start and end somewhere. muvdu, with keyword 'move', is the transitive verb. You muvdu something from somewhere to somewhere. If the thing you muvdu is yourself, this will result in the event of your klama-ing, or in Lojban le nu klama.

So what are the keywords for? First of all, they are memory hooks. It is a lot easier to learn a single keyword (or a short phrase, in a couple of cases) than a complete place structure definition. This gives us something practical to work with in LogFlash - we couldn't reasonably expect you to type the complete place structure for each word in testing yourself.

In addition, the keywords help you pin down what a given Lojban gismu means relative to other words. In short, they limit the semantic space and give you an approximate meaning to focus on.

When you know all of the gismu, you (at least theoretically) have 1316 words that together broadly cover all of semantic space. tanru metaphors are generally used to qualify or restrict that space; they do not bring in new basic concepts. Thus, a knowledge of all the gismu gives you the capability, given Lojban's unambiguous grammar, to discuss any topic generally. le'avla are shortcuts, in effect a kind of name, that allow you to use specialized jargon without being long-winded. lujvo are encoded tanru; they will have a specific assigned meaning and place structure eventually, but again they are short forms for the longer tanru.

You will have to learn ALL the gismu to be an effective Lojban user. (I'll qualify that. There may be a couple areas of semantic space, such as the lesser used culture words or the words for metric units or quantifiers, that you will be less likely to encounter. See below for discussion of these words.) Unless you know the list well enough to know what each of klama and muvdu and litru and cliva and bevri and benji and their several relatives mean (look them up!), and are able to recall them quickly (using the keywords), you will often use the wrong word when communicating in the language. The person you are talking to will likely assign different meanings and values based on their different knowledge of the words from yours. Voila! Chaos results instead of communication.

Actually, this happens in everyday language. Words have different meanings and connotations to each individual based on your experiences. Tommy's favorite pair are 'royal' and 'kingly', which nominally mean the same thing. They are respectively from the Romance and Anglo-Saxon roots for the same concept, and at one time did mean the same thing. But you can probably think of examples where one word would be correct and the other incorrect in English usage, because we have over time assigned slightly different values to each.

Sometimes these meanings vary regionally. 'Soda' means a flavored carbonated beverage, where I come from (also for Nora); 'soda water' is the clear stuff. In England and other places, 'soda' is the latter, as I found out to my chagrin on my recent visit there. In other locations, 'pop' is the word for what I call 'soda', and 'soda' has ice cream in it. To me, 'pop' is the sound of a bursting balloon or tire. Where Tommy comes from, down South, everything is 'CocaCola', or 'Cocola', or 'Coke', regardless of flavor or brand-name. You can ask for a 'Coke', and they will ask you whether you want Royal Crown or Pepsi or A&W. The keyword for sodva is 'soda', and I will tend to think of my own definition for it. But sodva includes in its semantic space all of the above meanings - a tanru would be built to select a particular one: cream-ice-soda, chocolate-soda, without-taste-soda.

Lojban has no culture, and no regional variations (yet?!). Nominally, all words have a single meaning, and we want people to start from this one meaning. As Lojban builds a culture, the culture will affect meanings just as (if Sapir- Whorf is true) the system of meanings will affect the culture.

So the keywords help you keep track of this central meaning as you attempt to communicate. Once you learn Lojban well enough to think in the language, you will no longer need the keywords.

The last column is a proposed English definition of the word, giving the place structure. Only the actual gismu have been baselined, hence the word 'proposed'. The place structures need to stand the test of usage, though we believe we have caught most inconsistencies. The definitions given are not complete; we do not consider them adequate for a dictionary, for example. We were limited by LogFlash to 40 characters per definition (or some other short fixed amount), and this is adequate for many words. In some cases we have had to abbreviate heavily to fit the definition in. In some cases, a place is likely to make sense only with certain types of arguments (sumti), such as properties, events, or amounts - and we couldn't include this information in 40 characters.

The given place structures should be sufficient to clarify the keyword - to tell you which English meaning of the keyword is intended. If not, we expect (and need) some questions from you. They also should give a good idea of how we believe the place structure should be, which will allow you to test it in usage and let us know if it gives you problems.

We have followed a minimalist approach to place structure definitions. You can add clauses that act like extra sumti to Lojban predicates (bridi) using modal phrases. You cannot take them away. Thus klama always has a destination. Whether you are coming somewhere or going somewhere, the 'somewhere' means that klama applies. If you aren't going anywhere in particular, but started somewhere, then you can be said (in English) to be 'going away', 'leaving', or more rarely 'coming away'. The word cliva covers the place structure where there is no destination, and uses keyword 'leave'. litru has neither origin nor destination, just a route and a means. Because of the different place structures, you know that the words have different meanings, and what that difference is. You also know that if you say mi klama, that you are going somewhere rather than travelling aimlessly, or just getting away from the origin (as in 'running away'). Obviously you are going somewhere, but the fact that you are doing so is irrelevant to the claim of cliva.

We have similarly taken the comparative places out of color words. As some of you may recall, in older versions of the language, JCB said that blanu (which is unchanged in Lojban from the old language) implicitly meant that you were claiming that something was bluer than something else. Such an interpretation can lead to strange results. The negation of the old blanu is 'no blanu'. By the rules of the language stated by JCB, this effectively claimed that there was nothing less blue in the universe - which is not generally what you are claiming when you say something is 'not blue'. And when you want to claim something is blue by noting that it is similar to a standard of blueness, 'more blue' is incorrect. There is nothing 'less blue' that you are considering in your claim.

By removing the comparative, we eliminate these problems. You can add a comparative either as a tanru: blue-more (blanu zmadu), or with a modal phrase: 'more than...' (mau ...). You can also reference a standard of blueness in either manner.

What this means is that you have fewer mandatory places to learn in the place structure, and there are more modal operators in the cmavo to learn to add these places in when you do need them. The cmavo list will give them and the textbook and LogFlash 3 will teach them. This explanation is given here though, so that you can figure out why a place you may think is relevant to the concept is missing.

This should be enough to give you a running start on interpreting and using the baselined gismu list. If you learn these words, then by the time we have given you two lessons from the textbook, you will be able to speak or write thousands of bridi sentences in Lojban.

Explanation of Cultural gismu

Especially in the English-ordered gismu list, people immediately notice the metric prefixes that start off the list, and the large number of 'proper names' that follow. If they look further, they may find gismu that refer to things they've never heard of, like 'steradian', and others that refer to obsolete terms like 'cubit'. Why?

First of all, in one sense, Lojban is not culturally neutral (See the discussion in later sections for more on why this is relevant, and why it may or may not be acceptable). Designed in the late 20th century by members of a technological society for an audience that is also generally oriented towards technology, the vocabulary must be capable of supporting the needs of that audience if people will use the language. We are designing a language for real use, by people of many cultures, but all have contact to some degree with 'modern technology'.

Even if we weren't trying to reach a particular audience, we would need a system of measurement. We could invent our own, but that would be reinventing the wheel - and who would use it. We need to design the language for people that exist in today's world; we don't have the ability to design a world to fit this language in a vacuum of other influences.

The metric system is used to some extent by nearly all world cultures. The United States is actually the, or one of the, most backwards in this regard. The metric system is international and accepted, and its standards are set by an international organization. To be consistent with this standard, we have defined gismu for each of the basic units and prefixes of the International Metric System, even those that are little used by most people like steradian (stero) and candela (delno), atto- (xatsi) and exo- (xexso). We also have specifically chosen tanru for each of the dozens of other officially approved metric units that are not basic, and either gismu or tanru have been made for each quantity measured by each of these units.

The basic orientation of the language design has been to accommodate, rather than to exclude, ideas from the myriad of cultures on this world, as well as to enable them to discuss topics of current concern. For the latter goal, we have gismu like petroleum, which could be made as a tanru such as rock-grease, which is petr-oleum is etymologically in English and several other languages. In the 20th century, petroleum is used in plastics, in gasoline ('benzene' or 'petrol' in other languages and cultures), and many other products. It also, to some extent, has sparked wars (as well as fires). By making ctile a gismu, we enable it to be used in tanru of its own, something not too easy in English when you are starting with a four-syllable word.

To achieve the goal of accommodation, we have to recognize that not all cultures and peoples use the metric system for everything they do. We must provide a means for people to discuss the idiosyncratic units found in each culture, like inches, feet, furlongs, and fathoms. This is vital in translation; these units had significance to people who invented them. We can translate an inch as .0254 mitre, but this would seem like an odd number to someone unfamiliar with the conversion (he would miss any cultural associations relating to inches, as well). We could also use tanru based on the origin of the unit where it is known: inch = thumb-long-unit. The latter suffices for most purposes, but takes long tanru for some fairly simple concepts built off of such measurements.

Those who got issue number 2 of this newsletter will recognize some of the following, which evolved considerably to the baseline version.

The approach taken was based on JCB's original approach. The 1975 and later lists of primitives that he has used include words for several of these units, apparently based on whether they were common according to Helen Eaton's word frequency study that is the standard for much of our work. JCB also included about 30 words for culture specific words not directly found in Eaton: the monetary units of certain countries, and three words for each of several cultures. These cultures were chosen to represent the eight languages he used as sources, as well as a couple of others that were in Eaton such as 'Roman', 'Indian' (referring to the Amerind cultures), and 'American' which refers to the United States (to the offense of other countries in the Western Hemisphere). There was also a trio for the 'Loglandic' people, language, and culture (but not for the trademark).

JCB broke his own rules, by making 3 words for each culture that differ only in the final letter (not otherwise permitted in primitive words which could be confused in complexes). One word represented the culture, one the people, and one the language. For reasons he never explained, he also had a separate primitive for the primary monetary unit of the country that was the namesake for the culture, which didn't follow this pattern. Thus we had dalra for the 'dollar', and frani for the 'franc'.

This was idiosyncratic, but would have been only a minor problem if only these countries existed in the world. But it really was preserving an American (United States) world-view at the expense of other cultures. 'Americans' rated a trio of words and a monetary unit, but 'English' referred to all English speakers, not to Britain. There was no way to refer to the British, nor the Canadians, etc. who also spoke English. There was also the implication that Americans didn't speak the same language as the rest of 'English-speakers'. Similarly, there was the implication that everyone in India speaks Hindi, even though that country has 15 official languages, and there are hundreds of others spoken there. Similarly, there are other countries besides the U.S. that use the dollar as a monetary unit, or besides France that use the franc, etc. This type of bias is known to lead to cultural stratification, and is unacceptable in a language that purports to be culturally neutral.

It was easy to come up with a solution that solved most of these problems. The word trios could be eliminated by making them two-part lujvo based on tanru such as German-language, French-people, etc. We could then devise gismu for each of the world cultures.

Oops. Cultures do not equate to countries. Languages equate to neither. And then there are countries like the Soviet Union (and the U.S.), which consist of several states which are to some extent nominally independent. We ended up with gismu for the following concepts: people (prenu), culture or society (kulnu), and language (bangu). We also added the following, each to fill a particular semantic space not identical with its English usage: country (gugde) to refer to the incorporated geographic region associated as a unit, territory (tutra) to discuss the actual land in that region, state (jecta) to discuss a unit that was a single political entity (a polity or government), and nation (natmi) to discuss a people united by a culture independent of any political boundaries. We also added monetary unit (rupnu), then discovered that most countries have two (or more) monetary units, and often neither predominates in all aspects of the society, so we added a second for a smaller unit (fepni). The latter two are true hybrid words. We built them from the names of the several units of the countries speaking our six source languages, and not from a single word as translated into each language.

Some examples of the above: The U.S., Canada, England, and the U.S.S.R are all gugde, and you can talk also of their prenu, their kulnu, etc. To the extent that they have a common culture, they are also natmi. They are also jecta, since they are incorporated with a government. Quebec is not a gugde, but it is a political unit, a jecta. It also can be described as a separate natmi - there is a distinct culture due to language differences. The states, counties, and incorporated cities of the U.S. are also jecta, as well as, to some extent, water districts, religious parishes, etc. (These would be probably be defined as used as lujvo built off of jecta, but are part of the semantic space of the basic concept behind the gismu as well.) Ireland is a natmi divided among two gugde. Northern Ireland is a jecta in a limited sense, especially for those who are partisan to one side or the other in the strife there. Actually, the Irish natmi can be extended to include Irish-Americans and other expatriates who still consider themselves tied to the 'Old Country'. The various Amerind tribes in the U.S. are each a separate natmi to the extent that their kulnu remains intact; the reservations are definitely jecta, and could even be described as gugde in the same limited sense that the Soviet Union considers its 'states' to be semi-independent countries. A similar comparison could be made with the South African tribal homelands, which are nominally independent. (I am not particularly partisan on any of these issues, so please don't attempt to read my views into these examples. I am pointing out how the division of semantic space built into the Lojban gismu allows distinctions and similarities to be noted that are not obvious in English.)

The latter pointed the way to solving the problem of non-metric units. We could eliminate the couple of token English units and build composites from the local units of each of the source language cultures. In some cases this made no difference; the long distance unit is still minli, bespeaking its heavy origin as 'mile'. The reason? The Chinese translate mile as their equivalent local unit name, even though it is a different size. They then use their own concept of metaphor, similar to Lojban's tanru concept, to tell which unit they mean. (This made us feel very good. One of the world's oldest cultures had solved our problem the way that we had independently reinvented, confirming that our solution was 'natural'.)

We thus have local unit gismu now for long distance (minli), short distance (gutci), weight (bunda), volume (dekpu) and area (kramu). The place structures of these units were defined so as to allow for subunits. This is an exception to most place structures, in that we have reserved places for subunits without implying the place is there if not mentioned. In effect, we have variable numbers of sumti. (There is one other gismu with this exception: jutsi = species - which carries as its sumti the entire taxonomic classification hierarchy, by whatever classification system is being used, of that which is being categorized.)

As part of the Roget review, we added 'legend/myth' (ranmi), and the beings of those legends (crida), to the existing word for 'religion' (lijda). Each of these concepts has strong ties to individual cultures and belief systems; the words for these can thus be used in tanru to refer to specific ideas that are unique to the culture.

With apologies to those not of the United States, we used the most applicable or frequently used 'American' keyword for most of these concepts. rupnu does not mean 'dollar'. (LogFlash users from other countries can use a plain text editor to change the keyword in file LOGDATA.FLA to something that is more appropriate to them, as long as the file layout remains intact, before you start using LogFlash. British users especially may wish to change American spelling for words like 'gray/grey'. For our British users, remember that 'pound' is the keyword for bunda; you may wish to use 'sterling' for rupnu.)

Having built all these gismu, we realized that many of them are now useful in less cultural circumstances. Thus jecta can be combined with tcadu to talk about an incorporated city. tcadu jecta (city-state) refers to such an incorporated city in most contexts. You might use the reverse tanru, jecta tcadu (state-city) when referring to such as city as distinct from or compared with its suburbs or metropolitan area. Both are valid; Lojban has (at least) these two words (the lujvo built from these two tanru) to make a distinction, where English has but one. (The tanru 'tcadu gugde' = city-country is a better representation for describing the ancient Greek city-states or Vatican City.)

There remained only the problem of the words for the cultures, countries, etc. These, like chemical elements, and the various specific types of animals, plants, and foods, are generally 'names' for things. In one sense, they aren't really bridi predicates. But in another, they (and all names) are perfectly valid as bridi. Thus we have incorporated the cmavo 'me' to turn any proper name into a bridi, as in me la kraislr, which translates as 'pertaining to Chrysler in quality/property...', a two-place bridi. (me actually works on any sumti to turn it into a bridi, not just names.)

(We have specifically made gismu with this place structure for the sun - solri, the moon - lunra, and the earth - terdi. The 'elements' that have been made into gismu have somewhat different place structures; they represent the materials (often impure or intentionally mixed with other substances) that many common items are made of rather than the chemical stuff: an American nickel has never been made entirely of that metal, but metaphorically the label is valid.)

As JCB did, we generally chose to make the words for the culture/nationality names from the local populace's name for themselves. Because of a statistical paucity of the letter o among the gismu, all such words that were coined instead of algorithmically derived were made to end in o. Cultural names are thus among the easiest gismu to learn. If it ends in o, it probably is a culture/nationality name, and this name is pronounced close to the Lojban by the people of that culture/nationality.

In this sense, 'American' is legitimate for people of the United States. That is what we call ourselves. We distinguish ourselves from the rest of the hemisphere, when it is relevant, by calling the others Hispano-Americans, North Americans, South Americans, or Central Americans, and use the term Pan-American in some contexts to refer to the whole. The former has been adopted by some portion of the culture that it describes; they call themselves Hispanic, and we have used that word as the basis for xispo. This can be modified via tanru to cover all of the Spanish-speaking countries and cultures south of the United States; we also have kadno for the Canadians. Of the major western hemisphere peoples, this leaves only Brazil (which speaks Portuguese) uncovered. We added Brazil, Argentina, and Mexico, as the major population and resource centers of Latin America.

Returning to the main problem, there are hundreds of countries, and thousands of languages presently in the world, not to mention formerly extant states like the Roman Empire. We decided that a fair limit would be to include the 12 major languages that were originally considered for the Lojban gismu-making effort. As mentioned in an earlier issue, we limited our efforts to 6 languages primarily for logistic reasons, and because of idiosyncrasies in the gismu-making algorithm.

We then took a list of countries where one of these languages is the primary language spoken by the country. We arbitrarily set a minimum of 10 million speakers as a cutoff for significance. In some cases, such as France, the language is clearly associated with one such country, even though other peoples use French. Arabic is similarly associated with Saudi Arabia. But unlike French, several other countries which use Arabic have more speakers than do the Saudis. So the etymology does not reflect the true nature of the Arabic-speaking culture. In such cases, we have a separate word for Saudi (the country and its people), and Arabic (the language and culture). We thus have several gismu for the several Arabic countries.

For Malay-Indonesian, we have two countries that speak a closely related language which is being standardized between the two governments. The language is not even spoken by most Indonesians, even though it is the most common of the languages spoken there and the official language, so we separated 'Indonesian' (bindo) from 'Malay-Indonesian'. The only appropriate source word for the latter language is the language's term for itself, which is 'bahasa'; this word, of course, means language. Hence baxso, though 'bahasa' also can refer to other languages as well. (We consulted with Lojbanist Steve Ladd, who has studied this language extensively.) For Hindi, we have the opposite problem. There are two names for the language: Hindi and Urdu, depending on whether you are in India or Pakistan (not all speakers of these languages accept this equation due to the political situation in those two countries, and they use two different alphabets). We had the problem with India that, although Hindi is only one of several languages used, they had no common name for the country that could serve as a basis for a 'nationality' gismu, separate from the language, primarily because of the 15 official languages. (The Hindi word 'Barat' would have been used, but brato and barto, the only obvious adaptations, conflicted with other gismu. No rafsi were available for assignment in that region of the alphabet either, and we had a policy of treating the cultural gismu as a lower priority than words built from the six languages. Pakistan (kisto), of course, could be easily separated from Urdu (xurdo). In this case, the gismu 'xindo' must refer to the language as a united whole, although when referring to monetary units or government, it will generally be metaphorically used to mean 'India', the country. 'France' will have the same correlation with 'French', using fraso.

Something similar will work with Spanish and Hispanic (which was added for the reasons given above). There is some difference between the two varieties of Spanish to support the distinction. We also added Portuguese (porto), since we had briefly considered including it in the Spanish weighting in gismu-making as a major dialect of Spanish, as some linguists classify it. To talk about the greater Spanish language you use spano bangu; if you want to distinctively refer to the variety spoken in Spain, you must use a tanru like spano ke spano bangu (Spain type of Spanish language), or possibly ropno spano bangu (European Spanish language).

We added words for Asia, Europe, Africa, and Polynesia to aid in locating places, cultures, and languages that aren't gismu, as used in the example above for Spanish. They also may be used with le'avla borrowings, which are another way of turning a name into a brivla for use in Lojban sentences. Thus ronlnede /ro-nl-NE-de/ might be an acceptable le'avla for the Netherlands culture/nationality and the Dutch language. It is built from the rafsi ' ron-' attached to '-nede' representing the name, with the l as a vocalic hyphen to glue the two pieces together into a legal borrowing. The le'avla can be used in lujvo with rafsi, such as ronlnedbau (Netherlands-language), or if you want to separate the language from its European association, you can use the rafsi for 'language' to make a separate le'avla: banlnede for Dutch. This could have been done with all cultural words, but it is believed that there is too much metaphorical value to culture/nationality names to deny the 'important ones' rafsi. As such, all cultural terms have rafsi, regardless of their importance. Some culture/nationality words were 'bent' more than others to ensure that a rafsi could be assigned.

We also added a few extra cultural words for three major cultures that did not qualify on the basis of their language populations: Greek, Hebrew/Jewish/Israeli, and Latin/Roman/Italian. tanru may be necessary (but are not mandatory if context is sufficient) to identify whether the present or past culture/nationality is intended. These three were justified also on the basis of their mythologies/religions, which are more widespread than the language, and the use of their languages or alphabets for certain international scientific purposes. (You probably know examples of Latin and Greek. Hebrew aleph is used in the mathematics of infinity, among other uses.)

Trying to be sensitive to the cultures involved, we then rounded out the list by adding culture/nationality words usable for the major religions of the countries covered. We used muslo for Islam because there is a cultural aversion on their part to the bending of the names 'Allah' and 'Mohammed' that would be necessary to make them fit Lojban morphology and phonology. jegvo serves to cover the entire Judeo-Christian mythos, as well as to serve as a reference to the deity of both religions. We added xriso as a counterpart to xebro to allow separation of the two. The Arabs and Jews are similarly united as semto, the Semitic peoples/cultures/languages. semto tutra is a reasonable tanru for the 'Middle East', which is of course not 'east' to them, and is 'west' to the Chinese and Hindi. We then added dadjo for the Tao, and budjo for Buddha to allow discussion of the major eastern religions.

This is a lot, but a price is to be paid for cultural neutrality, especially for those who want to see Lojban considered as a world language.

This is possibly more than you wanted to know about these cultures and peoples, but should enhance your Lojbanic education and make the studying of the 80-odd culture/nationality words seem a little more relevant. If not, please note that they are only 6% of the total gismu; a small price to pay for the power implicit in these words.

Examples of lujvo-making and some basics on tanru

We've talked of tanru and lujvo throughout the above discussion. I'm not going to write the textbook lessons on these two subjects just yet, but to do much with Lojban you will need to know how to use rafsi to make lujvo from tanru. To do this you need to know what tanru are, and to start practicing in making them. So here goes.

First of all, to make a lujvo, you need a tanru. But you need to know what a tanru is first. Eventually, when we have a dictionary, you will make lujvo more rarely - you will more often look them up (when you haven't memorized them). But if you want to learn Lojban, you MUST be able to make and analyze tanru. And if you are truly good at it, you may never need a dictionary. (Well. Maybe when you are communicating with people who aren't as good at tanru as you, you might want to 'follow the book'.) I've said it before - Lojban is not a complicated language.

What is a tanru? If you haven't seen a definition before now, or figured it out yourself, a tanru is a Lojban metaphor. It is made up of gismu representing concepts that are related to the concept you are attempting to communicate. The relationship is not necessarily unambiguous. A blue-nest (formerly a blue-house, for those who read my exchange with Paul Doudna last year), in some way nests someone or something. It could be a nest for blue eggs, or blue people, or it could be a house painted blue either partially or completely. It takes a more elaborate tanru to distinguish these less ambiguously (if such is important), or you can use non-tanru methods to expand your communication unambiguously. Most of the time you won't: it is long-winded, unnecessary to most understanding, and a pain.

tanru are something like English adjective-noun and adverb-verb combinations. They go beyond these concepts by combining and expanding upon them. In this, they are similar to Chinese metaphor more than to English. In general, the gismu on the left modify those on the right, and all groupings are in pairs from the left. Thus broda brode brodi brodo brodu is a 5-part tanru. You would interpret its meaning by grouping gismu in pairs as: (((broda brode) brodi) brodo) brodu.

If you need to change this unambiguous grouping, you would use specific cmavo that allow unique expression of the possible groupings. The cmavo 'bo' causes two adjacent gismu to group together before any other groupings. Thus we get ((broda brode) (brodi bo brodo)) brodu. If there are two bo's in a tanru, the leftmost takes precedence, but this is unlikely to occur in normal usage. You can also use the cmavo 'ke' to change the grouping. It causes everything to the left of it to modify everything to the right, with one exception that I'll get to in a minute. Thus results broda ke (((brode brodi) brodo) brodu). If you don't want to go to the end with the right-grouping, you close it off with kei, the right grouping terminator cmavo: (broda ke ((brode brodi) brodo) kei brodu. These complexities may take some close examination. Don't worry too much for now. No quizzes until the class starts.

You can also use the logical connectives and negation to modify part or all of a tanru. The cmavo are ja, je, jo, ju, and nai. You can also use the mixed connective joi, and the four main abstraction cmavo: nu, ka, ni, and jei to do funny things to the components of tanru. Again, you'll learn all this later.

I'll give one real example of how grouping is important and move on. In the unmarked (by cmavo) tanru: nakni cinse ctuca (male-sexual-teach), you have the potential for a highly embarrassing label; the unmarked grouping is (male- sexual)-teacher, suggesting that what is being taught pertains only to male sexuality, or possibly even that the curriculum is something inappropriate for an educational institution. With the ke grouping cmavo: nakni ke cinse ctuca = male-(sexual-teacher); this matches the intended English of 'male sex-education teacher' more closely, without implying anything more than gender about the teacher's sexuality, or suggesting anything odd about what is being taught. This is of considerable concern if your daughter will be in his class. (Nora notes that there are other possible interpretations of both tanru, but I think this will serve for the example.)

Now, I would like to say that you now totally understand tanru. You don't, because no Lojbanist will have learned all that will be needed for fluent tanru-making for quite a while yet. tanru-making is one of the most important skills in speaking Lojban, because tanru are the primary source of ambiguity in the language. Accidental puns like the example just given could get serious if your life depended on someone understanding your communication.

To learn the essence of tanru interpretation, you must learn to imaginatively think of possible meanings (using some simple conventions to limit the possibilities), and then determine why the speaker used this particular tanru as opposed to some other to help weed out the interpretations that are not intended by the speaker. To make a new tanru, the reverse process is used. Think of a few possibilities, then try to analyze how a listener might misinterpret each possibility. Thus making tanru gets to the true essence of human communication: putting yourself in the mind of the other person, and figuring out what that mind is thinking. It's a neat concept. It is obvious, though, why no one is good at tanru yet. Until we have many people thinking Lojbanically - seeing the world through tanru - we won't have the variety of thought needed for fluency in this form of mind-reading.

(Tommy notes that when we have multiple cultures involved, the making and interpreting of tanru will be even more challenging. Both German and Chinese use a form of metaphor-making similar to tanru to express complex notions. In many cases, the metaphor they choose will be different. For Lojban to bridge multiple cultures, it must in some way build its own culture, which cannot be too alien to those of other languages, but rather a true blend or hybrid culture. A difficult goal, but I believe Lojban is capable of reaching it.)

When you've learned even a few gismu, start trying to describe things, actions, qualities, and behaviors in terms of tanru. When you know all the gismu, you will have the full power of Lojban at your disposal, but you can start much earlier. If you have someone to share with, tell them your tanru, and see if they figure out what you intended. You will learn as much by their wrong answers as by their correct ones. (If you don't have a Lojbanist to try this with, use an English speaker, and the English keyword equivalents, asking 'What does this phrase suggest to you?'

I will give two of the most basic of the tanru conventions, so you don't make a false start. In a tanru, where a gismu such as 'bajra' (which means 'x1 runs from x2 to x3, etc.') occurs, the keyword (which is an English verb) should not be used in interpreting the tanru. The sumti 'le bajra' refers to 'the runner from...', and not 'the run from...'. You know this because x1 in the gismu definition would be replaced by the one who is running. Similarly, in tanru, you should use the x1 place in building your meanings. A tordu bajra is a short-runner, not a short-run. To get the 'verb' as in the latter, use nu bajra (the event of x1 running) in the tanru, giving tordu nu bajra. You can also get an 'adjective' which is otherwise hard to convey by using the quality abstractor cmavo 'ka' (This is a little harder to come up with good examples for, so I won't. Let it be an exercise...).

The second convention, and relevant to the following discussion, is that tanru have the place structure of their final gismu, and not a combination thereof. Thus a pikta lebna (ticket-taker) is one who in some way related to tickets takes something from someone, since this is the place structure of lebna. There are ways to get the places of pikta involved, but these are much more complicated. Using the principle of 'keep it simple', you need to keep your mind on the place structure you want to end up with.

lujvo-making Examples

The publication on the lujvo-making algorithm tells you everything you need to know about this subject, but there are a couple of problems with it, which is why we haven't sent it to everyone yet. (We will be sending it to all level 2 recipients that don't already have it within a couple of months, but I needed to write this article before I felt that it would be useful to everyone. The current publication has only minor changes from the attachment to our 2nd issue which was on the same subject, if you have that issue.)

The first problem is that the algorithm is just that, a semi-mathematically described procedure for performing and evaluating lujvo. JCB and I wrote it two years ago (together - the good old days!) so I could write programs in support of the dictionary work. For those of you who aren't technically inclined, it is a bit dense. I hate to turn people off with stuff that intimidates them. Such stuff kept me from getting active in the community for over 6 years. And it needn't. Lojban is easy to learn and understand.

The second problem might solve the first (but I doubt it). There are no true examples. This is because we didn't have a baselined gismu and rafsi list, and since I'm only part of this show, I expected changes before the baseline. There were, and I'm glad I don't have to redo all the examples I never wrote.

(While the rafsi list is not yet baselined, I am sufficiently convinced that changes will be minor, now that the gismu are baselined, that I'm taking my chances.)

OK. Let us say that you have a tanru which expresses an idea that you use frequently. How do you turn it into a lujvo?

First of all, when the dictionary is out, you should check it if possible, to make sure that we haven't used your tanru for some other purpose. I'll explain this in a moment.

What you will do in building a lujvo is to replace each gismu with a rafsi that uniquely represents that lujvo. These rafsi are then attached together by fixed rules that allow the resulting compound to be recognized as a single word and to be broken down in only one way. It is that simple, and the rules are pretty simple, too. The complication is that you need to know the phonology rules pretty well, or you'll make mistakes. But they're pretty simple, too.

There are three other complications; only one is serious. The first is that there are usually more than one rafsi that you can use for each gismu. Which do you use? Simple. Whichever one sounds best to you. I'm serious. There are many valid combinations of the possible rafsi They all are equally valid, AND ALL MEAN THE SAME THING. The lujvo-making algorithm gives two fancy scoring algorithms, one by JCB, and one by me. One of these algorithms, or something like them, will be used when we take your tanru and put it in the dictionary; we need some fixed standard of 'aesthetics' for deciding which form is 'best'.

Both existing scoring algorithms try for short forms and tend to push more vowels into the words to make them easier to say. The Japanese, Chinese, and Polynesian speakers will prefer this; the Russians will have a different aesthetic, since they are used to saying consonant clusters. I lean towards vowel-rich forms because it is easier for a Russian to deal with more vowels than for a Japanese to deal with more consonants. Also, from what I've read, children learn to say vowels correctly before consonants, and especially before consonant clusters, and a vowel-rich language may therefore be easier to teach to the young. But these are not necessarily the criteria you wish to use.

The lujvo-making algorithm also gives a lot of fancy flags to allow you to override any scoring algorithm's choice. This is because the algorithm is designed for a computer, and we need to tell the computer when not to follow the rules. But as I've said, you never need to follow the scoring algorithm rules. (Even after the word is in the dictionary in one form, the other combinations will still be valid.) When using a word with a new Lojbanist who doesn't know the rafsi very well, a longer form may be preferable to a short form. We can't include this factor in the dictionary, except as an override flag if it turns out that most people prefer a form different from the one given by one of the scoring algorithms.

The second complication is the bad one. Remember that a tanru is ambiguous - it has several possible meanings. A lujvo, or at least one that would be put into the dictionary, has ONE MEANING. Just like gismu, a lujvo is a predicate which encompasses one area of the semantic universe, with one set of places. Hopefully this is the most 'useful' or 'logical' of the possible semantic spaces. A known source of linguistic drift in Lojban will be as Lojbanic society evolves, and the concept that is most 'useful' or 'logical' changes. At that time, it might be decided that we want to redefine the lujvo to assume the new meaning. The lujvo must not be allowed to retain two meanings. So the 'Lojban Academy' that maintains the dictionary will be ever watchful of tanru and lujvo usage to ensure this standard is kept.

You also must be aware of possible meanings of your lujvo, especially if you are writing for 'posterity'. If you invent a lujvo which involves the same tanru as one that is in the dictionary, and assign it a different meaning (including a different place structure), you are directly causing such linguistic drift. Which isn't all bad. Every natural language does it. But in Lojban, when you use a different meaning, someone else may use the dictionary and may therefore misunderstand you. This is why you need to look a potential lujvo up in the dictionary, when possible.

This is the other major reason for having a standard scoring algorithm. With several possible arrangements of rafsi to consider, you need to know which one to look up in the dictionary. Actually, it is logical that at some time, possibly even before the main dictionary is written, we will have a second type of dictionary giving tanru that have been made into lujvo and their 'official' meanings. This may turn out to be more useful than the main dictionary both for writers and for readers, given the Lojban philosophy, but it is only useful to readers who know the rafsi well enough to disassemble any lujvo into its component tanru. This is why LogFlash 2 will teach you the rafsi. They may prove vital, especially in the early years of the language, to being able to use Lojban effectively.

As with tanru, there are ideas for guidelines on how to choose the right meaning to assign to a lujvo. I won't even start on them here. For now, do the best you can. Most often, you will be as right as any of the rest of us. The essential nature of human communication is that if the listener understands you, then you've done well. As you practice, hopefully with another, let this be the ultimate guideline for choosing meanings.

In conversation, you aren't going to sit there and look up every word in the dictionary, or you'll never get anywhere. Nor will you instantly, or even quickly, have the vocabulary to converse intelligently at the level you do in English. Most college-educated adults have vocabularies in excess of 25000 words, or 20 times the size of the gismu list, and English is estimated by some to have 3 million words. Never will they all be in a dictionary, especially in one that you could use. Nor will you want to use the long-winded forms of tanru when you use a concept a lot. So you will need to make lujvo, as will all fluent Lojbanists, as well as those on the road to fluency. This is the advantage Lojban has over ANY auxiliary language in existence. Between its unambiguous grammar and it simple rules for making tanru, and from them lujvo, a MASTERY of 1316 words and these simple rules gives you a communication power that NO OTHER LANGUAGE CAN EQUAL with a comparable level of learning.

The third complication is also simple, but is the scariest. Based on the concept called Zipf's Law (actually a hypothesis), which says that the length of words is inversely proportional to their usage. The shortest words are those which are used more; the longest ones are used less. The Lojban corollary that comes up all the time is that commonly used concepts will be abbreviated by natural language users AT ALL COSTS. In English, we have abbreviations and acronyms and jargon, all of which are complex ideas that are used often by small groups of people. So they shortened them to convey more information more rapidly.

It is this interpretation of Zipf's Law that some believe leads to the plurality of meanings in the natural languages, especially of short words. (Ever looked 'run' up in the dictionary? Even worse, try the Oxford Dictionary which shows how it has evolved through history. Pages upon pages of meanings!) This law obviously fights directly against the rule that all lujvo, indeed all brivla must have one meaning. The assumption that this corollary is probably true leads to the Lojbanist's resigned acceptance of the third complication: A perfectly good and clear tanru may have to be abbreviated, if the concept it represents will be used so often as to cause Zipf's Law to take effect.

I just said a few paragraphs ago that you will be forced to make lujvo up on the fly, especially when you are first learning the language, in order to make conversation practical. As a defensive measure against people misusing lujvo when they aren't following a dictionary, as in normal conversation, the Lojban design takes some preventive measures. To some they are poetic ones as well.

The third complication is that, given a tanru with grouping markers, abstraction markers, and other cmavo in it to make it syntactically unambiguous, in most cases you drop some of the cmavo to make a shorter (incorrect) tanru, and then use that one to make the lujvo.

'Gross!!', you say. 'How illogical?' 'How can you have a logical language if you do this?' 'And what about unambiguity?'

We have a natural language first, then a logical language. Or it isn't a language at all.

It actually isn't all that bad anyway, and not really illogical. It doesn't lead to ambiguity. A given lujvo still has exactly one meaning and place structure. It is just that more than one tanru is competing for the same lujvo. But more than one meaning for the tanru was already competing for the 'right' to be used for the lujvo. It means that someone has to use judgement in deciding that one meaning is to be chosen over the other for the dictionary definition. And this judgement will be made on the basis of usage, presumably by some fairly logical criteria. What would be illogical would be to try to fight human nature.

If the lujvo made by a shorter form of tanru is in use, or is likely to be useful for another meaning, the decider then retains one or more of the cmavo, preferably ones that set this meaning apart from the shorter form meaning that is used or anticipated. Generally, therefore, the shorter lujvo will be used for a less complicated concept, possibly even over a more frequent word. If both words are needed, the

simpler one should be shorter. It is easier to add a cmavo to clarify the meaning of the more complex term than it is to find a good alternate tanru for the simpler term.

And of course, we have to consider the listener. If he hears a word he doesn't know, decomposes it, and gets a tanru that makes no sense, what will he do? If he knows that you may have dropped out the grouping operators, he may try alternate groupings. Or he may try using the verb form of the concept instead of the first sumti, inserting an abstraction operator if it seems plausible. Plausibility is key to learning new ideas, and evaluating unfamiliar lujvo.

Common sense should dictate the rest. The listener has limited possibilities to consider. If one meaning is obvious given the context, he can assume this one is correct. Take the (valid) Lojban word 'lojycfikabysicyciskinmla'. If he can't figure out the meaning, or if no one meaning clearly fits, or if he believes it important to be sure he understands correctly, he asks 'lojycfikabysicyciskinmla ki'a'. ki'a is the most useful cmavo for the novice Lojbanist. It means 'I don't understand. Please clarify.' After any brivla, it is specifically asking the speaker to clarify the meaning of the brivla, which in this case should be attempted first by the speaker giving the full defining tanru, and then elaborating as necessary if that still isn't clear to indicate the desired meaning. As for 'lojycfikabysicyciskinmla': I hope it doesn't mean anything. I tried for something intimidating by stringing hard to say rafsi together. We'll find out what it means (both of us) in a few minutes.

As you can see, it has minimal effect on you when you make lujvo, especially at first. You do exactly what I said above in conversation. You make your tanru, and if you don't know of a conflicting meaning, you build your lujvo from that tanru. If you aren't sure, or if you don't think you will use the word a lot, you can retain some or all of the important cmavo. This is why we have assigned rafsi to just those cmavo that are most significant in this regard. To be safe early on, you may choose to be overdefining in your tanru. It doesn't hurt, just makes the mouthful bigger by a couple of syllables.

All this is presuming that you need to make a lujvo, which is what this section is about. As you learn and use the language, and when we have a dictionary for emergencies and for careful writing (that's when we use them in English, too!), what seems like a chaotic and illogical mishmash will quickly come together. After all, you learned English, didn't you?

You've survived the traumatic part, if you are reading this. I hope you haven't been scared off by discover that natural language is more complicated than arithmetic or BASIC. The rest is truly simple and I'll now give you those examples I promised you. I will put myself in your shoes, and I will not use either of the scoring algorithms. (It saves time and paper, and there are several examples of the scoring algorithms in the existing text.)

For my first trick, I will take the tanru that I used above. (As I write this, I do not know what the rafsi are, or what my answer will be.) The tanru is nakni ke cinse ctuca.

First, look up the rafsi (or memorize them - I haven't yet either, though I sometimes make guesses since I know no one else knows them any better. Obviously this will have to change, but I like to get in the practice in making lujvo on the fly.) We have, from the gismu list:

nakni : nak na'i ke : kem cinse : cis ctuca : ctu

In addition, each of the gismu (but not ke) can be used in an 'unreduced' form. In the last (ctuca) position, this unreduced form is the gismu itself. For the other positions, you drop the final letter of the gismu to get the unreduced form (but you will be replacing this letter by a hyphen y, as described below). You also cannot use a CVC rafsi in the final position. These rules ensure that the word does not end in a consonant. If the word ends with a consonant, it would be a name rather than a lujvo.

Now, assuming you have no instinct for what will work best (which you acquire fast), you make all possible combinations:

nakn-kem-cins-ctuca nakn-kem-cins-ctu nakn-kem-cis-ctuca nakn-kem-cis-ctu nak-kem-cins-ctuca nak-kem-cins-ctu nak-kem-cis-ctuca nak-kem-cis-ctu na'i-kem-cins-ctuca na'i-kem-cins-ctu na'i-kem-cis-ctuca na'i-kem-cis-ctu

Here's where the phonology comes in. You will easily see that some of those conglomerations are unpronounceable as is; others are illegal by Lojban rules. There are five important rules, and no exceptions. That's it.

1. Between any two impermissible medial consonants, stick a y. 2. Where there is a consonant triple formed where two rafsi join, if it is impermissible, stick in a y where they join. 3. If you have a CVV rafsi at the beginning, add in an r hyphen after it. An n is used instead if the letter after the hyphen is also an r. In a tanru with only 2 parts, do not add a hyphen when the second rafsi is a CCV-form. 4. Always stick in a y hyphen after a 4-letter form, which is the gismu without its final letter. 5. Starting from the left in any set starting with a CVC rafsi, if the first consonant pair (which occurs where the first two rafsi will join) is a permissible initial, you must stick in a y if there isn't one there already.

These are easy, if you didn't get turned off by the nasty words 'permissible' and 'impermissible'. If you have the phonology synopsis, you should know what these mean. If not, these are simply the rules that determine what letter combinations are allowed in Lojban words. They are rigidly defined. Absolutely no exceptions. 99% of the rules for determining what is valid are based on intuition: if it is too hard to say, or if it is too hard for someone else to figure out exactly what letters you've said, it isn't allowed.

These problems happen under two conditions. If the consonants do not flow together well and must be said in one syllable, it is hard to say. If the consonants are too much alike, it is hard to separate them and figure out what was said.

There are 48 pairs of 'permissible initials' that say what pairs can always go together in Lojban. A couple are unfamiliar for English speakers, but you will find them easy with practice. You need to memorize them.

There are 5 rules for 'permissible medials' which dictate which sounds are too much alike. You need to memorize them. They aren't hard, and most are intuitive if you try to say the forbidden pairs. I won't repeat them here for space reasons.

The impermissible triples rules are combinations of the last two sets of rules. A couple of trios not covered by these rules are added because some people thought them too hard to distinguish from some other combination. You'll rarely hit these exceptions, but it helps to memorize them so you don't make 'on the fly' mistakes. You can look them up at first.

Now, I apply the rules, and insert letters as necessary in place of the dashes, and drop the dashes where no letter is added:

naknykemcinsyctuca naknykemcinsyctu naknykemcisyctuca naknykemcisyctu nakykemcinsyctuca nakykemcinsyctu nakykemcisyctuca nakykemcisyctu na'irkemcinsyctuca na'irkemcinsyctu na'irkemcisyctuca na'irkemcisyctu

These all look bad, and as intimidating as the mouthful I coined up above. But they are all pronounceable. Try each, remembering penultimate stress, but ignoring syllables with a y vowel in counting syllables. A good idea at guessing syllables is to group letters in threes and pronounce them, since that is the typical size of a rafsi, but don't take the consonant after a y into the y syllable. Always attach an r hyphen to the previous vowel, and an apostrophe (pronounced like a quick /h/), into the next vowel.

Each one is a valid lujvo for the tanru we started with, which was a 4-plex, so you wouldn't expect a short word.

The number of syllables in each word is 6 or more. If you had to choose one, you'd probably choose one with 'only' 6 syllables. If you want to guess the scoring algorithms' results, you will take the word that you inserted the fewest letter 'hyphens' into, and then in case of ties, take the shortest one in letters not counting the apostrophes, and then the one with the most vowels. You get the last one as being best (no I didn't set it up that way).

You then spend no more than 10 seconds thinking about usage. Is this word going to be used a lot, so that people will insist on a shorter form? Probably not, but people hate 20-letter words. You can either use the mouthful, or drop out pieces such that it makes the word more manageable.

You can drop the ke term if you think it is unlikely that the alternative interpretation is likely to ever be needed as a single word, which is the case. Repeating the above, I find the best resulting word then as either: na'ircisyctu, nakcisyctu, or na'ircinsyctu. The scoring algorithms would probably take the middle one, since it only has 1 hyphen. You choose.

The word is now short enough to say, and unlikely to be Zipf'd into something shorter. But an alternative solution to the length problem exists that I think is better. See if you agree. We observe that 'sex-education teacher' is a useful word, irrespective of the sex of the teacher, which is seldom important. You then use only the last two terms to form the lujvo and leave the 'male' a separate 'adjective' in a tanru. Since there are only two terms in the tanru, you no longer need the ke, so you've made the word only two terms. Repeating the above, which is easier since you have only 2 terms, gives nakni cisyctu or nakni cinsyctu. The scoring algorithms prefer the former. I prefer the latter because I have trouble saying lots of c's and s's in the same word with no padding.

We now have a choice of short words that are totally safe from Zipf, much easier than the original tanru to use, and yet they completely and unambiguously retain the full structure of the source tanru for a new listener to recognize. Magic?

For my next trick, I'll use the lujvo in the title of this newsletter, and go both directions. Boy will you be surprised. This one is much shorter than the last problem, since it has only two terms. The word is lobypli. Since y is always a hyphen, except in names, it is best to start your breakdown by replacing y's with hyphens. This gives lob- pli; the word easily breaks into two three-letter rafsi with a y hyphen between. Most Lojban words will break down just as easily, because as we have seen, more than 3 or 4 terms is unusable as a lujvo.

If you've memorized the rafsi, or have a gismu list sorted in order by rafsi, you can interpret the pair easily. (I do, and I will make them available shortly; I may even include them into the lujvo-making publication.) If not, you have to hunt, which is a lot easier with the Lojban-order gismu-list. Many rafsi start with the same letter as their associated gismu. In this case, we get lojbo pilno, or using keywords: 'Lojbanic-use'. But remember, we use the 'x1' place as a noun, rather than the keyword in interpreting the tanru. Thus we get: 'Lojbanic-user'. This has at least two possible interpretations. Either the user is in some way Lojbanic, or the thing used is in some way Lojbanic (or both). Although the lujvo can only have one meaning, the source tanru is ambiguous; both meanings are valid. With no other clues and no dictionary available, you would be justified in asking ki'a to get me to explain which I meant.

In context, I am vocatively calling (ju'i is the attention-getting vocative call); you are the audience. If you are actually learning and using Lojban, one interpretation makes sense, and this was my intention. I want you all to be in the category of Lojban users. However, if you are sitting on the fence watching the rest of us, you are still a member of the Lojban community, and you are using the newsletter to keep an eye on us. In the latter case, both you and the newsletter are in some way Lojbanic, so the other interpretation is valid as well. This is thus a type of Lojban pun, in this case unintentional. I didn't analyze it this carefully until now. But my source tanru, it turns out, is singularly and unavoidably appropriate.

What is the place structure? We can make a good guess, though I've never used the word with its places. We will also assume the meaning I originally intended, a user of Lojban. Start with the places of 'use'. x1 uses x2 for x3. Given our meaning, we know what x2 must be so we fill in the place and renumber them. Our place structure is obvious: x1 uses Lojban for x2. Most Lojban lujvo will have place structures derived in this straight-forward manner from the source tanru. They will thus be easy to figure out without carrying a dictionary or memorizing each.

Going the other way from lojbo pilno, I will show the other possible lujvo we could have used for our title. Your reward for reading this is that you get to vote on keeping the current title, or changing it to one of the alternatives. Since the set of rafsi changed with the gismu baseline, the scoring algorithms would no longer give our current title the best score.

For lojbo, we can use rafsi: lob, lo'o and jbo. For pilno, we have only 'pli'. We also have the unreduced forms lojb- for the first term and pilno for the second. The choices are:

lojb-pilno lojb-pli lob-pilno lob-pli lo'o-pilno lo'o-pli jbo-pilno jbo-pli

After applying hyphenation rules (try them yourself and compare), you get:

lojbypilno lojbypli lobypilno lobypli lo'orpilno lo'opli jbopilno jbopli

Did you note that you need no hyphen in lo'opli, since the CVV rafsi is followed by a CCV rafsi, and has only 2 parts (rule #3)?

All of the above are valid forms for the lujvo. (You can see that this may make poetry in Lojban much easier. Not only can you choose a wide variety of tanru, but each word can have several forms which can be selected for rhyme and meter.) You may vote now. The scoring algorithms would prefer any of the last three to our current word, since they use no hyphen. There would be a slight preference for lo'opli over jbopli. What do you think?

Now, I will apply lujvo-making to two recent problems. First I will try the name of the language, which is derived from 'logical language'. The tanru is thus logji bangu, and the possible lujvo are:

logjybangu logjybau lojbangu lojbau

Obviously, for the lujvo form of Lojban, we will use the latter. The place structure might be: x1 is a logical language of people x2, with rules x3, and for purpose x4. Since there has been more than one version of logical language, this place structure may be adequate to distinguish them. We might choose to add a name place probably instead of the listed x2, since the name is the current primary distinction between versions, not the people. We could either drop the 'people' place or move it to the end as: x1 is a logical language named x2, with rules x3, and for purpose x4, used by people x5.

For the name, we could have used any of the above lujvo after appending any consonant on the end (n and s are used a lot, but there are no mandatory rules; names are at the discretion of the namer). We also might have dropped the final vowel if this left a nice consonant as final, but this isn't preferred in this case; most English speakers have trouble pronouncing ng without adding a vowel sound afterwards. For this tanru, we have an extra option, we can use the CVC rafsi for bangu in the final position. This adds logjyban and lojban to our list of possibilities, and you know which we chose for our name.

My next application is the Lojban name for our organization. If you read JL5, you may think you know the answer. Unfortunately, we had to change the word for 'group', and the tanru is now logji bangu girzu. The options for the first two places are the same, and we can include the CVC for bangu in the lujvo now. The rafsi for girzu are gir and gri. The possible lujvo are therefore:

logjybangygirzu logjybaugirzu lojbangygirzu lojbaugirzu logjybangirzu lojbangirzu logjybangygri logjybaugri lojbangygri lojbaugri logjybangri lojbangri

For the name, we can use any of the above with any consonant appended, or a vowel dropped. We add the possibilities built with the CVC final:

logjybangygir logjybaugir lojbangygir lojbaugir logjybangir lojbangir

Based on our name for the language, the two best choices seem to be lojbangir and lojbangri(z/s/n), although the scoring algorithms would probably prefer lojbaugir. Given English sound associations, we don't want to seem either green or greasy, suggested by the possible consonants I've listed for the second choice. So I think we'll take the first choice.

Now I will challenge you with a worst case scenario. I will coin a three-place and a four-place tanru with 3 rafsi for each component plus the unreduced 4- or 5- letter form. I won't give all the possibilities, since a three-term tanru of this type has 48 combinations, and a four-term tanru has 192. You can list all of them out as an exercise, or you can cheat like me, and leave out the unreduced rafsi, which are almost never preferred. This cuts your possibilities to 18 and 54, respectively, a significant reduction. Then eliminate those that use the most hyphens and check your answer(s) against mine. I will use a mediocre tanru for submarine, or perhaps a diving bell: deep-more-boat, and a place where they are kept when not in use: deep-more-boat-store. Have fun. My answer will be after the next paragraph. (hint - there is at least one lujvo that needs no hyphens)

I will close with an analysis of a new unknown lujvo, to show you how the 'expert' thinks (or if he does?). As I write this, I don't know the answer. I will use the monstrosity I gave above: lojycfikabysicyciskinmla. I won't pronounce it. Lojban phonology is unambiguous. Follow the rules. Remember to stress penultimately. Breaking it into rafsi, we have loj-cfi-kab-sic-cis-kin-mla. Now, first of all, a 7-term tanru should never become a lujvo. You can see why. The rafsi expand to logji-cfika-karbi-since-cinse-skina-mlana. We seem to have a logically-fictional-comparingly-snakishly- sexually-cinematic kind of side. Whoever invented this concept had some truly strange ideas. Lojban allows you to express strange ideas in exactly the same way you express 'normal' ones. You have to analyze the lujvo and figure out what it means. Then you can evaluate the idea. After all, this last one could have been a ringer. I could have set one up this long that really made sense, like: fe'urkemlojbaukezyfitcme. (Do you believe me? Can you come up with a better way to express this concept?)

My answers for the 'submarine' problem are:

conzadblo conmaublo conzbablo conzadlo'i conmaulo'i conzbalo'i cnozadblo cnomaublo cnozbablo cnozadlo'i cnomaulo'i cnozbalo'i

and for the 'submarine reserve port':

conzadblosro conmaublosro conzbablosro conzadlo'isro conmaulo'isro conzbalo'isro cnozadblosro cnomaublosro cnozbablosro cnozadlo'isro cnomaulo'isro cnozbalo'isro conzadbloso'u conmaubloso'u conzbabloso'u conzadlo'iso'u conmaulo'iso'u conzbalo'iso'u cnozadbloso'u cnomaubloso'u cnozbabloso'u cnozadlo'iso'u cnomaulo'iso'u cnozbalo'iso'u conzadlotsro conmaulotsro conzbalotsro cnozadlotsro cnomaulotsro cnozbalotsro conzadlotso'u conmaulotso'u conzbalotso'u cnozadlotso'u cnomaulotso'u cnozbalotso'u

As you can see, there are a lot of possible lujvo, all valid and all meaning the same thing. The scoring algorithms would probably pick cnomaulo'i and cnomaulo'iso'u, although my version might choose a different answer than JCB's (I'm not going to check), especially for the 4-term lujvo. I myself prefer cnomaulo'isro for the latter at this writing; my tastes tend to change with time.

A Challenge

I've now gone into tanru and lujvo in significant detail. I could tell you more, but there really isn't that much - some lessons of experience. I need to know how understandable my presentation has been. So I need your help.

About 400 people are getting this newsletter, and every single one of you is getting a gismu list. So you have all that is necessary to make tanru. If each of you made just 20 tanru for common words and concepts and sent them to me, we could double our tanru stock for the dictionary. I would also learn how 400 people tackle the problem of making tanru, based on the material I've just given you. This would in turn help me develop it into the textbook lessons on the subject.

So I'd like each of you to do just that - make up 20 words. Take words, ideas, and concepts that you use in daily life. No need to be fancy. You can make tanru for common articles of furniture (based on nilce), for tools (based on tutci and cabra and minji), for kitchen utensils (from the same words listed for tools and from jukpa). You can also work from your job: teachers can make words for common classroom items, etc.

The above examples are all noun-ish, but you could make other kinds of words as well. One technique is to write, in English, a short paragraph describing what you do at work - getting into a little detail: a secretary might do typing, filing, take-dictation, and operate-a-copier to make-copies. As I've shown, you will find several concepts that would have to have Lojban words to translate that paragraph. So make the tanru. And you will quickly see related tanru that can be derived from these: typewriter, file cabinet, file drawer, file system, file folder, copy machine, copy paper, toner. You might try some adjectives too, like letter-quality (a kind of printer). Some of these have obvious metaphors, the tanru-making consists merely of looking up the Lojban words for the components. But for others, the component words won't be gismu and you will have to hunt through the list to see what you can come up with. And of course, always be wary of how someone unfamiliar with what you are trying to get across might misinterpret the tanru. (No tanru have been made for any of the above, and someone can feel free to tackle them for a start, but I hope you secretaries will go beyond this list.)

The technique I learned with while editing JCB's dictionary material was to take a gismu, any gismu. I would look up in the old dictionary what lujvo had been made from this word to avoid duplication and to gain inspiration. Then I would see what I could do with that word. Although I did not yet have LogFlash, and did not know very many of the old gismu, I was often able to come up with 20 to 50 ideas per gismu. I made it through old gismu starting with the letter b. By then, I had the trick well down.

You can do the same if you have an old L4/5, though you would have an extra step of determining the old dictionary word for the new gismu. In many cases there won't be one. In 1975 there were less than 900 gismu. If you don't have a dictionary, just make up words from scratch.

Speaking about dictionaries, you can open your Webster's at random. I would suggest sticking to common concepts (no - no one has made a tanru for antidisestablishmentarianism yet, but do you want one?). And remember, if multiple definitions are given, each might have to be a separate tanru.

You may also want to check your dictionary briefly when trying to make tanru for any concept you've chosen to make. I had to do so when rebuilding gismu, and in going thru the Roget's analysis, and I can't guess how many times I learned that my conception about what a word meant was limited in some way.

If you have the lujvo-making algorithm publication, you can make the lujvo from the tanru. You can use either scoring algorithm if you wish, or just pick your favorite form. If you have only the Synopsis of Phonology and Morphology, you can use the rules listed above with that publication, and don't worry about using a scoring algorithm. If you don't have the Synopsis, just put the rafsi together according to the rules given above, and try to guess where additional hyphen y's might be needed due to consonants that don't go together well. You'll probably be right. (Or you can send to us for a copy, but don't wait for it to come to make your tanru.)

If you can't do 20, do 10. If not 10 then 5. Of course we'll be happy to look at more than 20. Send whatever you come up with. Try to get to it early, certainly within a month after reading this. (You can take longer, but if you wait that long, you are likely to never get around to it). Send what you can off to us, as well as any comments you have - especially let us know what publications you used and any problems you had. Use your writing to us as an excuse to comment on anything else that you've seen or heard about in these pages. We still haven't heard from half of you.

You needn't spend a lot of time on this, and your results don't have to be gemstones of Lojban - though I suspect we'll find some among them. I will assemble your creations, and print a number of them next time. I'll comment on the problems I see, and perhaps suggest alternatives. In short, I will use these as a basis for the 'next lesson'. The good tanru and lujvo will be saved up for inclusion in the dictionary.

Let's see what you come up with. Don't hesitate. Make tanru NOW.

Jerome Frazee's New Logic Symbology

Long-time community member Jerome Frazee has sent me a copy of his recently published paper, which has relevance to the teaching of Lojban, and possibly to some of our cmavo. His paper describes his proposal for a new set of symbols for the basic truth-table operations of symbolic logic. These operations are the ones which are embedded in our logical operators, and are a key to why Lojban is called 'logical'.

Since many people have trouble learning the operations, and especially the relationship between the operations and their truth table equivalents, his approach might be valuable in getting Lojbanists to think 'logically' enough to use these connectives fluently. He claims that his symbology renders logical manipulation as easy as arithmetic, and it appears so.

When you have two statements, each may be either true or false. There are four possible combinations of true and false terms that are possible with two statements. Logical operations assign a true or false value to each of these four pairs of true-false possibilities. For example, 'logical AND', or 'conjunction' is true only for the possibility where both component statements are true. 'logical OR', or alternation is true if either term is true, or if both are true. In English, we represent this by 'or' or by 'and/or'.

There are 16 possible operations which yield a different truth value set (the set is called a truth table) for two component statements. Lojban has cmavo or compound cmavo for each of these. Standard symbolic logic has symbols for 14 of them, excluding the always true tautology and the always false contradiction. The standard symbols that are used have no system; a logic operation symbol was chosen because it was used in English typography to represent a similar concept in language (& for AND), or in mathematics for a concept that is metaphorically similar, or in a couple of cases, they are arbitrary. You cannot tell anything about the truth values of the operation from the symbol, which makes logical manipulation difficult to learn.

Jerome would change these symbols to a logical set. Each symbol represents the truth table of the operation. I can't work graphics with my word processor so I'll have to draw them, but they are easy. There are four dots in a square or rectangle. The two on the left represent the truth value of the 1st statement; the two on the right represent the truth value of the 2nd statement. The top dots are the true condition; the bottom dots are the false condition.

For each of the four truth relations in a truth table, there is a line that you can draw from one left dot to one right dot. This line is drawn if the relation is true, and left undrawn if the relation is false. For example, if the line between the two top (true) dots is drawn in a symbol, it means that the operation is true when both component statements are true. There are 16 symbols possible, each uniquely representing one truth table. The symbols and their correlation to Lojban and the current system is:


Those with some logic background will note that the converse of an operation is determined by flipping its Frazee symbol upside-down. Other manipulations, and the resolving of more complex propositions, can also be described by symbol manipulations. The lines of the resulting symbol can then be read off to determine the truth values resulting from the manipulation.

It seems like a simple, but elegant idea, but I am not a logician. I've asked Jerome what the reaction to the paper has been, but I haven't gotten a response yet. Jerome also suggested that Lojban cmavo could perhaps be arranged to better match or suggest these symbols. I've asked him for a proposal; the arrangements of Lojban cmavo are, however, among the hardest things to change, because there are really 3 cmavo for each logical connective.

Logicians and others interested in the paper will find it in History and Philosophy of Logic, Volume 9 (1988), pp. 87-91. The title is A New Symbolic Representation of the Basic Truth-functions of the Propositional Calculus.

If you cannot locate the paper, you might try writing to:

Jerome Frazee, 402 Richmond Road, Susanville, California 96130, USA.

I'd be interested in comments on his proposal, and on whether you would like to see use of his symbols in the textbook, especially if we try explaining some of the mechanisms of symbolic logic.


A long time ago, I started getting comments from several sources asking this question in several forms. Some versions were related to misconceptions about the language, its 'logicalness'; some were about its potential as an international language or as a computer language. Since our emphasis was on remaking gismu, there have been a lot of comments asking why we were bothering - why we didn't just use English roots. On the other hand, I heard from people who were primarily interested in the Sapir-Whorf question, who took the opposite standpoint, sometimes seeking to make the language more un-English-like. There are grave misunderstandings about the purpose and nature of Lojban's 'cultural neutrality'. Lay people have one set of misconceptions, and scholars and linguists tend to have another. The questions from inside the community tend to be different from those outside.

Why are we bothering? Why should I go and spend hundreds of hours to learn a language that isn't spoken yet by anyone, and which has had so many false claims of being 'almost done' as to make the claim itself an object of ridicule? How will you use the language when so few people speak it? How is this language better, or worse, or whatever, than Esperanto and many other artificial languages?

The other side of the coin is a different set of questions. Why has this language survived two major public announcements that fell flat? What has kept JCB going all these years? What has kept people interested in spite of the grueling politics, and the years of silence between publications? And why do we come back for more?

Everyone seems to have their own answer to these questions. Most people seem to be interested in the language for only one of the several reasons that will be presented. They tend to look at the project and the movement from this one perspective, and see things going on that they don't understand.

There are people who actually want to learn Lojban. There are others that believe it to be a pipe dream, and fair game to talk endlessly about. There are those interested in the grammar who couldn't care less how we build the gismu, and others who don't understand why an unambiguous language is important.

The discussion has reached a crescendo, and certain themes keep reappearing. Some are based on misconceptions about the language and its purpose; others seem to be based on assumptions that one particular goal is more important than the others. All seem related to the basic question 'WHY?'.

I have attempted to give off the cuff answers at times. What came out wasn't always the best. I recall with particular chagrin when I got very defensive about questions regarding the use of 'English-like' words as gismu instead of our 6- language composite words. To all of you who haven't gotten satisfactory answers, I apologize. This special section is an attempt to do a little better job, by drawing on the opinions of many.

It is vital that the questions of why we are doing this be understood, and that disagreements be settled now. As we expand to an ever larger public, the misconceptions of a few turn into rumors that are difficult to turn aside.

Lojban exists for a large and diverse set of reasons. It has survived and grown, and its community has stayed together, under stresses that have killed many other artificial language movements because of this diversity. Lojban has something to offer people in several different lines of interest, and a feature or problem that has turned off one person usually has attracted another. Thus, these problems have been solved rather than being put aside, and the language is much more robust as a result.

As a result, our phonology has been gone over so many times, and so thoroughly that it is understood probably better than that of any natural language, even though the language isn't really spoken yet. Our gismu have been beaten into cultural neutrality by critics who are on the watch for little flaws. Our grammar has 'climbed the pinnacle' of unambiguity, because a few people knowledgeable about computer techniques applied their special talents in this area. My contribution, of course, is to bring a systems engineering discipline to all the diverse efforts that have gone on too long, and to see the strengths and weaknesses of all the conflicting points of view, somehow arriving at a whole that satisfies everyone.

Of course, JCB provided the creative spark that got us started - and he stuck with it long enough that we can see it through to completion even without his help. I will always honor him for this, in spite of our well-known disagreements, and we should all recall that he has sacrificed a lot for the sake of the language that has been his dream for over 30 years.

To build on his dream, we must understand his reasons and goals for inventing the language. I will attempt to summarize these based on his writings, primarily in Loglan 1. Then I will turn to you, the community, for your viewpoints. I have received a lot of correspondence, some of which is excerpted below. I also sought pro and con statements on two issues that seemed to come up most often, the use of English roots and the Sapir-Whorf hypothesis. I received answers from each person that I asked for a submission, a 100% response. This makes this issue rather larger than I expected, but it is definitely no loss to anyone, given the quality of the responses.

I can make no statements about the intent of correspondents whose work appears or is quoted herein. Their usage of the various names for the language and the movement is always given verbatim, to let you make your own interpretation of their intent. The opinions of the correspondents is their own, and are not necessarily the opinions of myself as individual or as publisher, nor of The Logical Language Group as an entity, and certainly not of The Loglan Institute, Inc. Since several of the comments reflect upon JCB's or The Institute's present or past policies, and sometimes are quite critical, I would have liked to have offered him the chance for rebuttal. Given our current state of disagreement and non-communication, this is not possible. I will therefore reiterate the offer to JCB to rebut any of the material in this publication for printing next issue. I also will speak in defense of JCB in my responses, or will give quotes or paraphrases of JCB's writings, where I think the writer has misinterpreted too strongly. I will clearly indicate my writings as separate from those of contributors.

The result of all this will result in some revision to the 'Announcing Lojban' brochure when I run out of copies and reprint it, which won't be for several months. There are a lot of misconceptions that should be cleared up when people first get involved. I would like feedback on whether (and which of) these contributions answer your questions.

What Hath JCB Wrought?

What has JCB said his purposes for the language are (or were)? By my understanding, JCB's original purpose for the language was solely to test the Sapir-Whorf hypothesis. Several of the following contributors will give their interpretations of this goal; I am trying to convey JCB's sense of it.

The first problem is to identify the hypothesis. It has been phrased several different ways, which is one of the problems. There are extreme versions, such as 'if you don't have a word for war, people won't ever think of fighting any'. Nice idea, but it seems unlikely. This is NOT a version of Sapir-Whorf that has any relation to Lojban.

JCB's formulation is key to understanding what his intent for the language was. I quote his version: "the structure of language determines the boundaries of human thought" (L1, pg. 1, all quotes hereafter are also from the first chapter of L1). A concept that is more difficult to formulate in a language, will be used less, or not at all. The relationship with Zipf's Law, discussed above under lujvo, should be apparent. The boundaries of human thought are not necessarily fixed. Sapir-Whorf is dealing with cultural phenomena, and not with individuals. An individual may transverse the normal cultural bounds, and see new ideas, and may even be able to rephrase them in terms the rest of the culture can understand. But Sapir-Whorf says that the culture, as a whole, will tend to pattern its philosophy and modes of thought around its language.

According to JCB, the "Whorf hypothesis is essentially a negative one: language limits thought". JCB's plan was to attempt to build a language tool that would have the major features of natural languages, but would have some strong warping in its structure that was deviant from all other natural languages. This warping would attempt to take normal structures that presumably set limits on thought, and "push them outward in some predictable dimension". His language tool would be an extreme case, not a "typical language", but "a severely atypical one", in order to enable any 'Whorfian effects' to be more easily seen. He attempted to put "decisive but non-essential differences" into the language; he still needed the language to be speakable.

I believe JCB's original goal was not to build the full language we are completing today. He speaks of trying to "import ... a human language into the laboratory" by "reduce(ing) its scale". Certainly, for the first 10 years of its existence, the language was only the nucleus of what it is today. In 1960, there were only a few hundred words, but JCB was already trying to teach the language, and was promising a dictionary and grammar 'shortly'.

The structural extreme he chose was to model the grammar on the well-understood structures of symbolic logic. There are no natural languages based on a predicate grammar, yet logicians are skilled at analyzing the structural relationships between natural language and formal logic. If you've taken a logic or philosophy class, you probably have experienced this.

Loglan/Lojban is 'logical' only in this limited sense, according to JCB. It "purport(s) to facilitate certain limited kinds of thought" - those based on logical transformations. The language was not meant to be a "deductive system" since it isn't wholly self-consistent, nor "reasonable", nor "self-evident". The essential logical concepts that he imported into the language were (abbreviating somewhat) the functional calculus including connective scope, quantification theory including a clear distinction between bound and unbound variables, clear distinction between modes of designation and description, and the treating of all predicates indiscriminately (no nouns, verbs, adjectives, etc.)

Everything else in the language is the baggage believed necessary to make the above differences "non-essential", or needed to make the language usable as "a laboratory instrument". Most of the work has been done in the latter area, bringing in the concepts of "cultural neutrality" and "metaphysical parsimony".

The essence of these concepts is that "it forces on its speakers a reasonably small set of assumptions about the world ... perhaps the smallest possible set". "Any speaker, from any culture, should find it possible to express in Loglan what he takes for granted about the world ... without imposing ... or being able to impose - these assumptions on his auditor".

As such, the language was designed with a simple grammar, with as few obligatory arrangements as possible, but with a lot of optional ones. Thus, there are no obligatory tenses nor genders nor cases nor inflections that are mandatory to understanding a sentence. The only obligatory concept was the fixed place structure of predicates, the alternative to which is probably some sort of case system with lots of prepositions (case theory was not developed when he started on the language) - linguists still haven't agreed on the essentials of case structures, so it was wise of JCB to choose the path that he did.

The language was always intended to be spoken by people of many cultures, and of many different language backgrounds. Each source language speaker was intended to be able to hear "many obvious clues to meaning", the source of our method of building composite gismu.

In order to be usable by many cultures, the language design has to be centrally focussed on the concept of "accommodation". We must be able to accommodate all that a speaker might wish to say, either in the language or from previously spoken languages. Thus, we are adding, even at this late date, some form of observationals, as Donald Simpson discusses below.

Other key concepts included keeping the size and complexity of the language small, striving for unambiguity in the grammar, which has been achieved, minimizing the plausibility test that I've discussed a couple of times, and ensuring that one can think and speak nonsensically, which is vital for 'logical thought' as practiced by the logicians.

As Ralph Dumain notes below, JCB did not go into detail about how the actual test would be conducted. It is, however, definite that he thought about it, as shown in these writings, and he communicated these ideas to others. In any case, we have a couple of visions of the Sapir-Whorf test presented below. It is not clear to me whether JCB presumes the Sapir-Whorf hypothesis to be true; his attitude seems to include the scientific approach of trying not to let his opinions on the matter govern the experiment design.

The language was built to attempt to remove some limits on human thought; these limits are not understood, so that the tendency is to try to remove restrictions whenever we find the language structure gets in our way. You definitely can talk nonsense in Lojban. It turns out that other concepts such as tanru, may be more significant than the logical structures in demonstrating Sapir-Whorf.

One thing is to be made clear. As a scientific instrument, Lojban must be built as if the Sapir-Whorf hypothesis were true, or it will never be seen. This does not mean that a Lojbanist must believe in or agree with Sapir and/or Whorf. However, the question must still be thought relevant.

That Sapir-Whorf is still a relevant concept is debated by linguists today. Because it is vaguely worded and generally thought not-testable, it is no longer in the mainstream of linguistic thought. Followers of linguist Noah Chomsky are especially critical of concepts like the Sapir-Whorf hypothesis, as they attempt to probe 'deep structures' in human grammar. They also tend to feel that Loglan/Lojban is presumptive, since it assumes that enough is known about language that one can be developed. In general, therefore, Lojbanists tend to be less sympathetic to Chomsky's ideas than mainstream linguists. But Chomsky's ideas are far from universally accepted, especially outside the United States. Chomsky's theories presume that one can intensively study one language, English, and thereby determine and understand deep structures of all languages. To students of other languages, including Lojban, this seems not only naive but arrogant - since the language they are studying is their own.

JCB also discussed the possibility of the language becoming an international language, after the manner of Esperanto. JCB specifically states that he did not intend the language for this purpose. He indicated that "the uncompromising logical character" of the language might make it unsuitable for such a purpose. He notes, however, that others in the community were interested in this purpose. JCB clearly favors an international auxiliary language, and thinks that even if his tool cannot be that language, it could be useful in developing one.

There is no direct mention of computer applications for his laboratory tool. JCB is not particularly knowledgeable about computers, and expressed surprise when the computer community was the source for most interest in the language. Perhaps the fact that he foreshadowed predicate-grammar computer languages by decades, though, suggests that in some ways, he was ahead of current computer thought. In any case, pc ended up adding a section on 'Loglan as a Computer Language', when he wrote the 1979 Supplement to L1.

This is what JCB said, at least as I understand it. Now lets hear from the community.

Viewpoints of Contributors: Donald Simpson, Doug Loss, Ralph Dumain, Jeffrey Kegler, John Parks-Clifford

Donald Simpson has written me several good letters. He made some useful suggestions on colors, and in a letter last February, made several comments, some of which are given in the correspondence section after this one. More to the point, Don accused us of retaining some metaphysical biases that needn't be there. I quote from his February letter:

"The Little-Word Problem: There are not enough available "little words" in Loglan. Loglanists are heard muttering that perhaps another vowel or two would be useful. Yes, Little-words get used up fast. One reason is that people hate to see unused members of a finite set. Given a command language that uses only single letters of the alphabet, and which requires only twenty functions, most people will immediately invent six more functions, so that they can use up the whole alphabet. Another is that English (and probably any natural language) has enormously many more functions which could be made into little-words, than there are little-words available. Any function which can be served by a predicate instead of a little-word should be. We should look closely at every little-word with the intent of doing what it does with a predicate instead. If there are not several little-words left over, the language is mis-designed, and more work is required.

Metaphysical Bias, 1: As a result of using little-words for math terms, over-use of little-words as solutions in general, and unthinking habit, ten as a number base is currently built into the metaphysics of Loglan. The language should be able to handle at least base-sixteen without weird expedients. But this is a minor problem.

Metaphysical Bias, 2: I believe that Loglan would benefit from a set of predicate modifiers that are as easy to use as those specifying space and time relationships, but which, instead, indicate the relationship of the speaker to the utterance. For example, is it memory, observation, inference, gossip, expectation, etc. I expect that you will correctly guess that my inspiration for these "observer-based" relationships is articles about Hopi. I opine that the power and general usefulness of such a set of modifiers (and their Whorfian significance) is obvious."

Briefly, with his points. I don't know of anyone who wanted to add vowels to make more room for cmavo (little-words). We were happy to move the letterals out of cmavo space, and to gain the added cmavo that are available due to the ' sound between vowel pairs. But this is because a large number of cmavo were intended to be made easy to learn by making them resemble gismu with related concepts, and the job is easier with more space. Not easy enough, it turns out. There also is a shortage of two kinds of cmavo, the CV-form and VV-form ones. The former is reserved for the most used words, and we tried to make them fit patterns that allow them to be easily learned. The latter are reserved for attitudinal indicators, of which there never would be enough using JCB's original system. We've done better, we hope. But in CVV- forms we have over 100 unused words, and don't expect them to change much. I am currently working to reserve a block of cmavo that all start with one letter (x), so that people can invent personal use cmavo on an experimental basis.

Hans Havermann and at least one other person mentioned the base-16 concept, and we added those words, though in CVV-form without rafsi. Yes, we do have this bias, but then so do pretty near all other languages. I agree that concepts should be expressed as predicates wherever possible. However, since the basis of the language predicates is logic, it must be remembered that in logic all predicates claim a truth. Most of the cmavo discursives and the like are meant to enable a speaker to make the claim of a predicate without cluttering it up with predicates that are irrelevant to the claim, or which make unrelated claims. Thus "observationals" belong as cmavo - they are ancillary to a claim - and they must be optional, or they would be a bias themselves. From a Sapir-Whorf standpoint, we have no need to include any more new features. That which is to be tested for Sapir-Whorf is there. Additional things like "observationals" are valuable in that they increase our accommodation of other ways of thinking.

On the basis of his letter, I asked Don to elaborate on his proposals for "observationals". He went beyond what I asked, and produced the following, which also comments on other aspects of Sapir-Whorf.


Well, I've gotten myself into deep water again, arguing that Loglan, a test for Whorf's hypotheses, is Metaphysically Biased by lack of "observationals", and agreeing to do a short paper on what a good set of such would be. Observationals and evidentials are parts of speech that comment on the source and reliability of the speaker's statements. Instead of saying that lemons are sour, you might say that you have tasted a lemon and it tasted sour to you, or that someone you trust has tasted a lemon and she says it tasted sour to her, or that from experience you deduce that lemons are generally sour. Instead of saying that something happened, you might say that you remember seeing it, or remember reading about it. And instead of saying that something will happen, you might say that you expect it to happen.

It is possible to say these things in English, as I just did. The trick is to say them as easily and compactly as we use the past, present future, and other tenses, so that we may say them consistently. The advantage of saying these things consistently is that our statements become more informative and are, in a sense, unarguable, for example: I propose that though we believe that the Earth turns, we still say conventionally that the Sun rises, because that is what we see. I accept from logical argument and authority that the Earth turns, but I consistently see the Sun rise. I presume that a hypothetical Flat-Earther accepts from logical argument and authority that the Sun revolves over the Earth, and that he sees the Sun rise, just as I see it. I expect that if he were to tell me that he sees the Sun doing loop-the-loops in the sky, I would probably not believe him, but that I would doubtless have no way of knowing what he saw.

The example is long-winded, awkward, and has a number of imperfections that result from the constraints of English grammar. But I thought that Loglan, with its simpler structure and optional tenses, could easily assimilate a set of little words that would do the same work, and do it right.

And then I was asked to define such a set.

Now, over the years I have seen a few such sets proposed for artificial languages. They had at least three categories: First, direct perception. Second, hearsay. Third, deducing, surmising, imagining, etc. Sometimes there were additional categories for dreams or traditions, modifications for the speaker's estimate of the probable reliability of the information, etc.

It seemed to me that those three form a good basic division, and in the right order. Either you see it happen, or someone tells you it happened, or you think about what you've seen and been told and decide something about it happening.

So I got out the only book of such proposals that I had on hand, Suzette Haden Elgin's A First Dictionary and Grammar of Láadan (pub. 1985 by SF3, Box 1624, Madison, WI 53701-1624), which I recommend highly to any serious student of language construction. And there were the evidentials:

1. known because perceived externally or internally 2. known because self-evident 3. perceived in a dream 4. assumed true because source is trusted 5. assumed false because source is distrusted 6. imagined, invented, hypothetical 7. no knowledge of the validity of the statement 8. no comment on the validity of the statement

Additionally, Láadan has seven tenses: present, near and far past, near and far future, optative, and hypothetical. And six degrees, five duration types, five repetition types, six speech modes, nine emotionalities, nine states of consciousness, a pejorative, and many other features. Most of these are optional, but if used carefully can give a statement the informational density of some of the Native American languages. However, the use of absolute time tenses means that these are not what I consider true observatives. So I went to the source: Language, Thought, and Reality - SELECTED WRITINGS OF BENJAMIN LEE WHORF - MIT Press, Cambridge 1956.

Though I claim that this book is required reading for any language designer, it had been years since I had read it, and I had forgotten just how rich in concepts and examples it is. It is a stunner. Whorf discusses linguistic orientations that make Loglan look barely more exotic than French, and a number of his concepts (particularly the idea of covert grammatical categories, or cryptotypes) should probably be applied to Loglan.

I began to feel that I was perhaps being foolish in pushing the observational concept for Loglan. Loglan is based on formal logic, which is the very heart of IS-ness. A long, long way (Whorf comments on the Latinate use of space as a metaphor) from an observer-based grammar. Loglan's strength is in its formality, its emphasis on structure and elegance, the application and testing of modern concepts of consistency, completeness, and resolvability. Do I really want it to be more like Hopi?

There are three Hopi assertives: Reportive (it happened or is still happening). Expective (it is expected to happen). Nomic (it happens). But, as in Láadan, there are also Modes and Modalities, types of Status, and other interacting and modifying categories to complete the picture. One cannot just haul pieces of a language out of its context and have them work right.

So I decided that I would have to accept Metaphysical Bias, and see what I thought would be a good set of words that would complement and combine with the other little words in Loglan.

Loglan already has forms for questions, hypothetical cases, requests, and orders, which are some of the Modes. I do feel that there is philosophical use for the Expective, though, as an alternative to the Future. I would also like alternates for past/present which distinguish: 1. things directly sensed. 2. things heard about. 3. things inferred/deduced, guessed/intuited, imagined (all crammed into one to save little words). And some form of indicator for how much credence the speaker has assigned the statement. Perhaps this would be best done with a predicate, to save little words, something of the form "X is the source or type-of-source with credence Y of statement Z" or "X is the credence of source or type-of-source Y for statement Z. This would also allow the speaker to be more specific when needed, or just use the four or so little words.

By combining these forms with the absolute time tenses, one might get a good set of tenses for time travelers ('I recall that in the future", "I expect that in the past").

Anyone really interested in trying to shake some of the Metaphysical Bias out of Loglan should read Whorf's writings (and the introduction and preface) thoroughly, carefully, and thoughtfully. The book on Láadan should help, too. And for some fiction along the same lines, I recommend:

  • The Languages of Pao by Jack Vance
  • Always Coming Home by Ursula K. LeGuin

Both of these authors are also very aware of language and culture in their other works.

Bob again:

Don claims that "observatives let you make statements that are unarguable. This is the antithesis of a logical language, in that each statement that is a predicate makes a claim, and is thus, in a sense arguable. But several of the ideas are good, and I am investigating them.

I have finally located and obtained Carroll's collection of Whorf, and Nora is reading it. She is not done yet, but is not nearly as impressed as I expected. Nora thinks that a lot of Whorf's observations and conclusions seem to be him imposing structures and concepts that he wants to find on the languages he looks at. He also seems to generalize examples into universals.

Expressing tense in Lojban is optional, and I personally don't think it will be the easiest thing to use or to explain, especially after reading some of pc's essays, which show just how complex the expression of tense is in English. Not having read some of the material Don mentions, I don't correlate tense with "observationals", but this could be my lack of knowledge of the concept. It seems to me that observationals are more like discursives or attitudinals than tenses, at least in Don's examples from English. Of course attitudinals are easy to use in Lojban, so this may be what he wants. Also attitudinals do not affect the truth value of the predicate claim. We'll look the possibilities over fairly carefully, and you'll see the result in the cmavo list.

I've ordered the Láadan book, and we'll see whether it is useful. It should be interesting, at least. I also have LeGuin's book, unread, and Nora and Tommy have both read Vance's book, though we don't have it here. My reading is terribly backed up. I'm taking the equivalent of several course in linguistics in my spare time, and am reading books on grammar, semantics, phonology, language universals, and of course, the writing of language textbooks and other teaching techniques.

I'll throw in my own reading recommendation: The Cambridge Encyclopedia of Linguistics. It is encyclopediac in scope, but it isn't a dry alphabetic enumeration of concepts. The author touches on every aspect of linguistics, and gives countless examples. He has a broad viewpoint of what constitutes linguistics, and goes into sign language, problems of the handicapped, and many other sidelights. Heavily loaded with sidebars discussing key points, and a lot of pictures and graphics, this book is excellent for browsing, as well as for cover-to-cover reading, as I've been slowly doing. It is expensive, however. I paid $50 equivalent in England, and I'm not sure of its availability in this country, but I strongly recommend it.

Our next letter was the first response that I actually solicited. Doug had been referred to me by someone else who told me that Doug was a long-time community member who had been turned off. He reportedly thought that we had drifted away from testing the Sapir-Whorf concept, which was the only reason he was interested in the language. I wrote to him saying that this was not true, and soliciting his thoughts on the Sapir-Whorf test for this issue. I'll include some preliminaries from his letter to give more context, and because it they are also relevant to the greater topic that we are discussing. Here's Doug.

Doug Loss

Dear Bob,

... thoughts on why I became disillusioned with the original language were pretty much on the mark. I made an early attempt to learn it on my own (alone), but the constant fiddling with the structure made that virtually impossible.

I made a go of translating the Lord's Prayer into Loglan (that being traditionally one of the first things translated into a language newly come upon by Westerners; missionaries, you know). The criticisms I got were well meant and well taken, but told me that there was an in-group of hobbyists who were making changes to the language casually, if not whimsically, and that those changes weren't getting out to the Great Unwashed (me and anyone else in the logical hinterlands).

After that, I watched the debates in TL with a slightly different viewpoint. A lot of neat stuff got proposed, but nothing ever got accepted as the official way of expressing a thing. Evidently the idea was that usage would sort it all out.

About that time I also began to notice a certain vagueness in the point of it all. Pretty obviously, from the premises on which the original language was constructed, the original point was to test Sapir-Whorf. As interest grew and the language attained cult status (I mean that in the same sense that The Lord of the Rings attained cult status) the direction sort of "vagued away." I started to see mention of how great Loglan would be for communicating with computers, or how it could become a world language. It was at about that time that I pretty much lost interest, although I continued to read TL on its very occasional publication.

Let me quickly give you my view on those last two points. I'm totally unconvinced that either the old or the new language is radically better for communicating with a computer than any random natural language. The whole point of getting a computer to understand a humanly speakable language is to make using the computer easier and more accessible to more people. Having them learn a new language to talk to the computer with does neither of those things. Furthermore, you won't get people to do it. Even furthermore, the point is moot; devices to recognize continuous speech have been demonstrated, and I expect they will be commercialized within the next five years or so. The accuracy of these devices is pretty good (about 95%, if memory serves); Loglan (generic term here) might increase the accuracy, but even if it doubled it (to 97.5%), I doubt that the benefit would justify the cost of learning a totally new language.

The international/world language is a particular bugaboo of mine. Loglan/Lojban will not become a lingua franca. It's that simple. I've done a fair amount of study on the artificial language movement, and I feel quite confident in making that statement. Very few Americans (the only group I've had enough contact with to make a statement about) realize that Esperanto wasn't the only artificial language to gain large groups of adherents. There have been many, but the first that I know much about was Volapük. It meant "world speak" in Volapük, and was extant in the last half of the 19th Century. It fell into disrepute when groups of enthusiasts disagreed on various aspects of the language and fragmented it into many different "dialects." The same thing happened to Esperanto in the first few decades of this century. I don't off hand remember all the variations, but here are some of the names I do: Ido, Mondial, Latino Sine Flexiones, Interlingua, Panglossa (or maybe just Glossa).

You've probably never been to a meeting of an Esperanto society. For all they claim millions of speakers worldwide, their meetings seem lifeless. Esperanto seems very much a hobby and not a viable, vital means of expression. I'd really hate to see Lojban become like that.

I hope that all that doesn't seem too down. I really think that Lojban could be a very valuable tool in testing Sapir- Whorf. Who knows, out of all the energy invested in doing a scientifically valid test might come another area of application where Lojban would be the obvious choice. I just don't think machine communication or international language are such areas.

Bob asked me to write something about the use of lojban in testing the Sapir-Whorf hypothesis. I don't know that I'm the best one to do that, but here goes.

Edward Sapir is considered by many to be "The father of American linguistics." He did most of his work with American Indian languages during the early part of this century. If you want to see just how different a language can be, try studying Hopi. This study brought Sapir and one of his best pupils, Benjamin Lee Whorf, to their hypothesis.

The hypothesis goes like this (I quote form An Introduction to General Linguistics, by Francis P. Dinneen):

If we define "culture" as "what a society does and thinks" (Language, Edward Sapir, p. 219), then the thought aspects of different cultures are strongly conditioned by their particular languages--not, of course, by the formal side of language, nor even directly by the conceptual type, but rather language determines culture through the particular contents of the concepts that make up the world of things in which the culture is interested... In the view of both Whorf and Sapir, it is illusory to think that "experience" can occur without the formative guidance of the linguistic habits of the person experiencing, and that the world we live in is first and foremost one "to a large extent unconsciously built up on the language habits of the group" (Selected Writings, Edward Sapir, p. 162)...

According to the so-called Sapir-Whorf hypothesis, language differences introduce a new principle of relativity, according to which men are not led by their experience to the same picture of the universe unless their language backgrounds are the same or similar.

This hypothesis isn't widely accepted among linguists today. This is partly because the fashion in linguistic theories has moved to discussion of Chomskian deep structures and the like, but also because the hypothesis hasn't been testable. Try as you might, you can't find a natural language against which to measure the others.

This is where lojban might come in. We might hope that the grammar of the language, based on symbolic logic, makes the fewest assumptions about the nature of reality possible. By the same token, developing the vocabulary from weighted phonemes from the six most widely-spoken languages is probably the closest we can come to culture neutral words. Such a vocabulary is needed because if a word looks and sounds too much like its cognate in one of the six languages, it's likely to acquire some of the cultural meaning that the cognate word carries, to the detriment of its neutrality.

There is, of course, no guarantee that the Sapir-Whorf hypothesis can ever be shown to be accurate. That's why you do experiments. The point is that no one has ever been able to test the hypothesis before.

Setting up such a test would be no mean task. I'm not the one to determine what would constitute a valid test, but I'm pretty sure that one would include some of the following things.

You'd need a fair-sized speaking community of fluent lojbanists, made up of people from as diverse of cultural backgrounds as possible. You'd further need homogeneous speaking communities from each of the cultures represented among the lojbanists, as control groups. The lojban group would have to use lojban at least as much as any other languages (ideally, they'd use only lojban) in their daily activities for an acclimatization period. I'd guess that would need to be at least 3-6 months. At that time, tests would be administered to the lojbanists and the control groups. These tests would be specially designed to test the hypothesis somehow; I don't quite know how that would be.

Perhaps we should administer a bunch of tests, only one designed to test the hypothesis, so as to keep the participants from knowing what was being sought and skewing the results. Actually, the best way would be to test them in such a way that they didn't know that they were being tested.

As you can see, setting up and running a Sapir-Whorf hypothesis test using lojban would be no small undertaking. It would, however, give us a fluent speaking community that could demonstrate other desirable features of the language to the world at large; by its very nature it would extend lojban from its mostly American, natively English-speaking community to a more diverse group; and it would allow us to try to establish a viable community of lojban-users rather than the (let's face it) hobbyists we all are at present. I think using a test of the Sapir-Whorf hypothesis as our goal could be just the push we need to make lojban a usable alternative language rather than just another interesting oddity.

Bob's notes: On Doug's inspiration, a few people made an ad hoc effort to translate The Lord's Prayer. If I have some extra space below, I'll print it.

On the computer interface language concept - I don't think that speech recognition is the justification for Lojban in this application, though it helps. Nor is the purpose to enhance general access to computers. Lojban's advantage, if any, is its simple, unambiguous grammar, which might make it easier for computers to 'understand' what people say to them. Its predicate grammar could presumably perform all the functions of the LISP or PROLOG languages, but the accommodation principle, and Lojban's usability as a natural language, give it a far greater power than those languages to communicate or store human thought in a computer.

I'm not sure that any language would be accepted as an international auxiliary language. If one can be, Lojban should be in the running. What Doug says about Esperanto, by the way, seems to be dependent on the group. Other local groups are more active. But in the U.S., we tend to view other languages in general as a 'hobby', not as a vital means of expression. That is because everyone here speaks English or is forced to learn. This isn't true in other countries. I believe that the U.S. in particular is responsible for the stagnation in the international language movement, simply because of our failure to recognize that not everyone wants to learn English. But we have held economic power for several decades, and "he who pays the piper...'. We will see what happens when the Japanese are calling the shots, if things continue as they have recently. In any case, Ralph Dumain has a different perspective on Lojban, both as a means of testing the Sapir-Whorf hypothesis, and as an international language.

Ralph's first letter to me was provocative, and I'll quote a couple of portions before his main article.

Ralph: "... My personal feeling is that the acceptance of any constructed language by the world at large has nothing to do with its inherent linguistic qualities. People have been trying to improve upon Esperanto for the past 80 years, rather pointlessly in my view because the acceptance of such a language is dependent purely on political and economic factors, not on linguistic efficiency or cultural bias.

"I certainly respect Loglan as a hobby, as a brilliant intellectual exercise, or as an experiment. Has anyone seriously tried to act on Loglan's purported aim of testing the Sapir-Whorf hypothesis? Personally, I think Whorf was a charlatan and that his view represents a superficial and retrograde racial mysticism.

"... I understand that some Loglanists want to teach their children Loglan (as one of their native languages, I presume)... It will be interesting to see how this works with Loglan, as Loglan seems to be radically unlike human language as we know it. If such an experiment succeeds it could have drastic implications for language universals and for Chomsky-like restraints on the properties of possible human grammars."

Bob: As you can see, Ralph's point of view is not too sympathetic to Loglan, although he has admitted that he knows very little of the actual language design. This has hurt him, because some of his statements are wrong, simply because of what he doesn't know. pc has reported that he has been able to manipulate the language grammar using Chomskian transformational techniques, and has found nothing extraordinary. I've been reading on language universals in the last month or so, and found that Lojban is probably more normal in these terms than English. In being accommodating to the concepts of other language, Lojban has no place in trying to prove how different it is.

In any case, Ralph's and Doug's letters started me to realizing just how many misconceptions there were about Lojban and its goals, and hence led directly to this section.

I asked Ralph to write up his viewpoint more completely, since it was contrary to Doug's. I had also received from an Esperantist a draft chapter on the history of the international language movement. He had included material on Loglan/Lojban, and wanted me to review it. He also referred to the Sapir-Whorf hypothesis and its relation to Esperanto. I consulted with Ralph, who had attended LogFest and who seemed to know something about some of the points raised. He agreed to write his thoughts on Sapir-Whorf and Loglan/Lojban, which follow.


Benjamin Lee Whorf, in attempting to justify his thesis of linguistic relativity (or, we should rather say, linguistic absolutism), tried to prove that language manifests (and controls) a world view in both its syntax and lexicon. I want to focus on the syntactic aspect of the question, both because it is where Whorf's obscurantism really shows itself and because the proponents of Loglan/Lojban have also concentrated on this aspect.

Loglan's claim of metaphysical neutrality based on its syntax is a point best taken up by philosophers of logic. That first order predicate logic is ontologically and culturally neutral (as if there were no controversies in logic) and that somehow in its cognitive effect eventual speakers of Loglan/Lojban become freed from metaphysical presuppositions, are suppositions that deserve closer scrutiny. The scientism of such thinking is very revealing of the subculture of the Loglan/Lojban community.

But here I am interested in the reciprocal assumption that the syntax of any natural language embodies any metaphysical viewpoint whatever.

Whorf went to great lengths to argue how the aspectual system of Hopi verbs reflects the Hopi concept of time. He never backed up such a claim by actually trying to find out how Hopis think of time. Furthermore, he got away with such nonsense because after all how many linguists knew anything about Hopi? Whorf claimed that Hopi is more akin to the world view of modern physics than the Indo-European languages, a preposterous statement in view of the fact that no Hopi ever received a Nobel Prize for physics but a whole lot of Europeans did. Some other obscurantist anthropologists followed Whorf's lead, but their papers are full of contradictions as well.

Whorf could never have gotten away with this had he chosen to analyze Indo-European languages in this way. For example, considering the aspect system of English as contrasted with that of other Indo-European languages, does the use of a simple past tense where in other languages there is a distinction between imperfective and perfective verbs, or the use of the simple present (e.g. "I run") to show customary action and use of the present participle (e.g. "I am running") to indicate real present action, reveal metaphysical differences between English speakers, French, Russians, etc. in their concept of time? Whorf would never have gotten away with that, and in fact he lumped them all together as "Standard Average European." It is quite clear that our ability to analyze and conceptualize time, and to come up with identical interpretations of statements regardless of specific linguistic forms, gives the lie to Whorf's ethnic mysticism.

No one was able to put Whorf's claims about syntax to the test, and many questioned his infantile assumptions. In fact linguistic theory was not remotely sophisticated enough to deal with such issues. However, there were experiments to test for possible effects of the lexicon (such as the domains of color and shape) on perceptual discrimination, memory, attention, "codability," yielding contradictory, inconclusive, or methodologically questionable results.

The semantics of syntax is a subtle issue, having been addressed by many linguistic theories in the past twenty years: e.g. generative semantics (now thankfully deceased), case grammar (dead), relational grammar (great results at first, started to wane after a while), Montague grammar. Linguists have learned to distinguish the roles of syntactic and semantic interpretation. Government-binding theory (originated by Chomsky) has attempted to parametricize grammatical features in a way that would relate the constructible alternative syntaxes of human language to one another. Because of the modular thinking of linguistics in the 1980s, arguments are made as to whether a particular linguistic phenomenon should be considered proper to syntax, imputed to the lexicon, explainable by external psychological or cognitive factors, etc.

Any monolithic view of a metaphysics directly encoded into syntax is hopelessly out of date, a possible ironic achievement of a theoretical tradition seeking to discover universal grammar. The concept of autonomous syntax not synchronically motivated by external cultural factors, e.g. the gender systems of various languages (say French and German), renders ridiculous any facile imputations of world view, even if historically a syntactic phenomenon such as a gender system was originally semantically motivated.

The silliness of Whorfianism is masterfully dissected by Frederick Newmeyer in his book The Politics of Linguistics. And even pre-generative linguists and anthropologists were smart enough to recognize that much in language could not be taken literally - was it John Carroll who wrote of dead metaphors? The human mind has a capacity to operate on a more abstract and sophisticated level than the jerry-built machinery of semantics and perhaps even syntax that language has diachronically bequeathed to us. Let me speculatively suggest that in its origin, development, and material basis human language in its entirety is metaphorical, which thought stands upon but which thought easily transcends by its abstractive power.

Furthermore, the generative view of language acquisition recognizes that individuals actively construct grammars in their minds based upon their genetically inherited cognitive and linguistic machinery in combination with linguistic input from others. The child language-learner can only deal with a language synchronically and can not know without formal study the history of the language or all the information that was encoded into it through time (e.g. few people know the original meaning of the word "rude" though many understand the double meanings of "vulgar" and "common"; no one alive knows why any particular language developed grammatical gender). Thus the classic Whorfian view is obsolete.

Newmeyer believes that the only direct influence of cultural particularity is in the lexicon, although even that is largely a trivial issue. The old shibboleth that the Eskimos have many separate words for different types of snow is an uncontroversial reflection of their environment and in no way implies that Europeans are any the less capable of perceiving distinctions in snow or could not as easily describe them using adjectives.

Of course, any language is syntactically biased in that it embodies a particular syntactic system and thus excludes others (left vs. right branching, prepositions vs, postpositions, analytic vs. inflectional, etc.) In this respect Loglan is biased as well, and metaphysics does not necessarily even enter into the question.

The naivete of the Loglan/Lojbanists is reminiscent of the foolishness of Esperantists who argue that Esperanto is not an Indo-European language because it is agglutinative and hence could be considered an Asiatic language. Besides the fact that there is no such category as "Asian" languages, besides the fact that agglutination is not the be-all end-all of syntax, exists in varying degrees in many languages, and has no particular correlation with culture or geography, besides the fact that historically the creation of Esperanto was influenced by no non-Indo-european language with the possible exception of Hebrew: from the point of view of universal grammar the (non-) Westernness of agglutination is completely senseless.

The Loglan/Lojbanists' philosophical worries about syntax are just as pointless, but they do not have much of an excuse, for they unlike Esperantists consist entirely of highly educated people. This proves that a little linguistics is a dangerous thing. Mathematicians, logicians, philosophers, and computer scientists may know what syntax and semantics are, and may even know what morphology and phonology are, but all too frequently have no idea of what human language is all about.

The real issue is semantics, and the interesting issue of bias is not merely among distinct languages, but within subcultures and competing ideological systems within the same language. The most interesting aspect of bias concerns the systems of terms that reveal, conceal, and conflate aspects of reality, the most obvious cases being political, philosophical, and other ideological terminology.

So far Loglan/Lojban has concentrated on etymology, ie., its original and clever a posteriori method of using an algorithm to generate essentially a priori words that purportedly suggest their natural language origins. I don't know what has been accomplished in the realm of meaning. I know that the a priori language inventors of the 17th and 18th centuries, such as Bishop Wilkins, cast their lexicon in the form of a taxonomy. Taxonomy was good enough for Linnaeus, and is useful for botanists and zoologists, but I think it is a relatively shallow phenomenon as far as cognition is concerned.

I shall be interested in seeing how the Loglanists and Lojbanists manage to free themselves from their cultural biases in constructing the semantics of the language. Since the subculture of Loglan/Lojban is extremely narrow and homogeneous, unlike that of the Esperantists or of any national or ethnic culture, it may turn out, as Paul Doudna has suggested, that Loglan/Lojban will be more culturally biased than English or any other language.

Loglan/Lojban, like its Enlightenment predecessors, is an attempt at realizing the old dream of making the outward form of language isomorphic to its inner content. If this is at all possible, look to linguistics for suggestions as to how it really might be accomplished.

Human cognition can function very well without such isomorphism, yet if there is any cognitive issue related to Loglan/Lojban, it is efficacy of expression. Perhaps as a tool for thought training, the precision of expression that Loglan/Lojban purports to make possible might be a selling point.

Some Lojbanists envision teaching the language to their children. Hopefully, they won't drive their children crazy as some logicians have done. While this might be an interesting experiment, the question remains, also perceptively raised by Paul Doudna, whether Loglan/Lojban is an autonomous language or whether it is parasitic on English or whatever other natively-spoken natural language is used. This is an especially crucial question because Loglan's grammar seems to be radically unlike human language as we know it and therefore there may be doubt as to whether Loglan is a possible human language. Some question the autonomy of Esperanto, although Esperanto is unquestionably a human language and certainly exists in its own right even taking linguistic interference into account. But can Loglan stand on its own, or like formal logic or mathematics, could it be only an ancillary tool employed for specific purposes and unlearnable as a native language? And as logicians are as biased as any illiterate, what do the metaphysical neutrality and cultural impartiality imputed to Loglan really amount to?

For those who envision Lojban as a future international language, let me remind them that bias is a political issue, not linguistic.

I have seen no evidence that James Cooke Brown ever seriously considered how to use Loglan to test the Sapir-Whorf hypothesis. I think he is guilty of bluffing, of waving the banner of Sapir-Whorf as a magic wand. This is scientism, the appearance of science without the substance. His unabated ignorance of the development of linguistics and his failure to learn from fair and warranted criticisms of his work voiced in book reviews or in response to grant proposals represent a fundamentally anti-scientific attitude and a crackpot mentality. His followers and ex-followers seem to have learned nothing as well.

'Language and cognition' is still an open-ended issue. Loglan/Lojban could prove a worthwhile tool of thought training. The issue must be efficacy and precision of expression, not the metaphysical jailhouse of linguistic relativity. If Sapir-Whorf is very much alive for the Loglan/Lojbanists, so much the worse for them. The vampire will feed upon their living flesh and spirit them off to the realm of the dead.

NOTE: This article appears by special permission of the author, who retains all rights and may republish it as he wishes.

--Ralph Dumain, July 20, 1988

Bob again: I have to say that I disagree with both Ralph's opinion and his style of expressing it. He did, however, raise points that need consideration, and I have no aversion to printing opinions that disagree with my own. I wish Ralph had expressed his views more constructively, though. He seems to attack everyone, from Whorf, to JCB, to those of us working on the language today, and he attacks the Esperantists as well. One wonders if Ralph finds anyone's viewpoint reasonable other than his own. Does he have any better idea of what human language is all about than those he accuses? If so, he should express it.

I could go to some length rebutting Ralph's paper, but he does a fine job of it himself. After going to great lengths to demonstrate that Sapir-Whorf hasn't been tested, and wasn't all that interesting or relevant, he states that "bias is a political issue, not linguistic". This statement assumes that the Sapir-Whorf hypothesis is wrong, making his reasoning circular. Later, he says that he doesn't know what we are doing in the area of semantics (not that much, really), then describes and lambasts the taxonomic systems of earlier languages. This is irrelevant to Lojban, which is in no way taxonomic. And of course, his claim that the reciprocal of the Sapir-Whorf hypothesis is an 'assumption that the syntax of any natural language embodies any metaphysical viewpoint whatever' seems false. This type of assumption would be relevant to a hypothesis that cultures (which tend to have a metaphysical viewpoint) determine the bounds of language, but not to Sapir-Whorf.

In Ralph's invective, however, there are some words of wisdom. Indeed, there is a bias in Lojban's designers - what he calls scientism, though he uses the term derogatively. There is embedded in Lojban the assumption that the expressions of predicate logic are relevant to metaphysical outlook, that making the tools of predicate logic easier to use would in some way bring about the Sapir-Whorf effect by removing some constraints on human thought. This assumption may not be true; no one has ever built and spoken a language based on predicate logic before. But it is an open and admitted assumption. We also presume that the scientific method is valid, that accommodation is a satisfactory way of reducing cultural bias, etc. There are a lot of assumptions behind the project we have undertaken. Some may be false, and may invalidate the experiment. But Ralph has no more knowledge than we do as to which assumptions might be wrong, and which can be modified without changing the language.

The implicit fact that a language is regular, that it has structural rules, itself imposes a bias. Now, according to Ralph, bias is political and not linguistic, but he spends a good deal of his time picking out some potential biases. Since the main danger in using this language for experiment is unrecognized bias, this is valuable, although I don't think Ralph actually brought up any that were unknown. Loglan/Lojban was designed to be biased, but in a recognizable way, one that can presumably be accounted for in tests to see if those biases transfer to, and/or limit thought.

I think that Ralph is not correct that the subculture of this community is narrow and homogeneous. Yes, we are mostly Americans, and effectively all of us are English speakers, and all of us are highly educated by world standards. I doubt if even these admissions make us homogeneous in thought, though, and the varieties of opinion on why we are building this language reveal that diversity. Even if Ralph were correct, however, there are two ways to be narrow: you can have people all alike who steep themselves in their own thoughts, or you can have people of similar background looking outwards to achieve greater understanding, and multiplying their capabilities to grow by building on their common background. Which will our community choose? That is your decision.

A last general criticism is on Ralph's accusation that Loglan/Lojban will be 'parasitic on English or whatever other natively-spoken natural language is used'. If one uses his logic, however, no one should ever bother learning a 'foreign language', as opposed to a native tongue, because you are only using it parasitically on your original language. The corollary is that one can never truly learn to 'think' in a foreign language, which seems to be false. If Lojban never achieves sufficient autonomy as a language that it can be taught as a native tongue to children, and that it can be used across the full spectrum of human experience, then the Sapir-Whorf test can never be conducted. This is why we need you to start learning and using Lojban now.

At least one other fact of Ralph's article is wrong by my knowledge. Case theory is far from dead. I have recently received from England a new book entitled The Case for Lexicase, which is a summary of the current theory as it has been used (evolving over about two decades) at Cambridge University, which has an extensive linguistics research unit. We got this book through our follow-up of a chapter sent to us by Paul Doudna from an out-of-print book called Inferential

Semantics, which was written by a member of that research team several years ago. Both books seem interesting; the recent one will certainly prove valuable in determining what (if any) rules are needed for defining place structures.

I personally feel that Ralph's statements about JCB are out of line, and based on a lack of knowledge. I have my disagreements with JCB, but I do not question his scientific integrity. There have been other criticisms of his methodology, on the language as a whole, on his knowledge of the field of linguistics, and on specific techniques used to evaluate linguistic data. To the extent that these criticisms are intellectual and are used to strengthen the eventually quality of the experiment, they are valuable. When they are used purely to denigrate the man or the experiment, as Ralph seems to be doing, we are going beyond science to personalities. Ralph does his beliefs a great disservice by these attacks, as his article does as well. He expresses interest in having this republished later. I suggest that he rework it so that his comments are constructive rather than destructive.

The next submission came unsolicited via the 'Capital Loglan Bulletin Board'. Mr. Kegler has also written to JCB previously on other subjects, and was published in LogNet last year. While we have had communication from several proponents of views similar to his, this message summarizes the position most effectively.

From: Jeffrey Kegler to: Bob LeChevalier Subject: Anglan

I hope some suggestions from a very infrequent participant in your group might be of interest.

I believe that attempting to create an entirely new vocabulary is counter-productive. I will attempt the argument for this as follows. First, I will try to show that English is extremely popular world-wide as a language of commerce--a de facto make-shift universal language. Second, I will point out difficulties on coming up with a new vocabulary. Third, I will briefly point out English's problems as a universal language. Then I will sum up with an argument for an Anglicized Lojban (AngLog).

Note that implied here is that my goal is creation of a universal language, not testing the Sapir-Whorf hypothesis.

First, English has a lot of momentum as a universal language. "The Story of English", a recent PBS series and accompanying book, makes this case very well and I will just try to list a few provocative facts. A firm named IVECO is headquartered in Turin. The money is French, German and Italian. Corporate business is conducted in everyone's second language, English. English is the language of air flight. A Malayan pilot landing in Saudi Arabia speaks English to the tower. Attempts to spread the native languages in South Africa and India are resisted as impositions of the leading language group to wipe out the others. The language the minorities want to be taught as a second language--English. Probably, if Mandarin Chinese were to be mandated as a universal language, the resistance would be strongest in China itself. Any of you familiar with a specialized scientific or technical literature will know that much of this first appears in English, regardless of the nation where first published or the native language of the author.

Note that the English embraced by the world's people is often adapted, with local vocabulary and pronunciation variants. We Americans would be considered by the English to speak such a local dialect.

Second, it is sometimes argued that Lojban words should not have a one-to-one correspondence with the English equivalents. I believe this is unavoidable. All the literature about Lojban is in English, and Lojban dictionaries define their words in English. If you have ever tried to translate between two languages whose speakers have a lot of contact (between English and French, for example), you will notice how often the usual translation of a word in one language acquires all the implications of the word in the other, so that they tend to become precise equivalents in almost all contexts. And where a language with a vast literature and long history, like Italian, starts sprouting words like "debuggare", to debug, how will Loglan fight off these influences?

For any group of people to develop a vocabulary sufficient to cover all present needs, much less future ones, seems to me very unlikely. Adaptation of a successful technical language's vocabulary is really the only hope. Even with the English language, evolving technology puts strains on the huge vocabulary available.

Third, I can be brief in pointing out English's problems as a universal language to this audience. Its irregular spelling, worse than most other languages, and ambiguous syntax, in common with most natural languages are two difficulties for new human learners and machine understanding. I can outline English's problems further, but I do not think anyone in this forum needs me to do so.

Fourthly, then I propose creation of an Anglan. Combine the syntax of Loglan and that of English, adding Lojban particles as necessary. Create a set of rules for the transliteration of words from English (and perhaps other languages). The result should have the momentum of English with none of its difficulties. The dictionary for Anglan would be short--just the new words necessary to implement a logical syntax. For the rest, the interested party should consult an English dictionary.

As a side effect, this should eliminate the Lojban legal entanglements. Neither English vocabulary or the laws of logic, whether existing or invented, are copyrightable, patentable or protectable as trade secrets, even where someone can prove they invented a field of logic, or an English word.

I realize the fast outline of a program for Anglan given above glosses over some difficulties. I am simply trying to argue it is far easier than the alternatives.

I will let pc answer Jeff. pc's experience and credentials as a linguist far exceed my own, and he has additional credibility as a spokesperson for the goals of the language, having been the only other person JCB has ever allowed in that role (i.e., in the Supplement to L1). pc then gives his own viewpoint on the nature of the Sapir-Whorf test.

John Parks Clifford (pc)


One of the most frequently made suggestions for developing a new language (or changing one already in development) in the present pattern, is that we should draw our vocabulary directly and exclusively from English, rather than making words up from some apparently random scraps of words from some set of different languages. The usual justification for this suggestion is that anyone, from whatever native language, who is likely to get involved with Lojban, will already know English, the de facto world intellectual language. Further, the supposed recognition aid that Lojban word derivation gives to native speakers of the source language does not, in fact, amount to much, since the pieces are so distorted and hard to find as to be as often misleading as helpful (and we can all think of cases in point). It is, after all, the grammatical structure that is the important thing for Lojban, not the lexical items that fill its slots.

It would, of course, be possible to make a language with Lojban grammar and English vocabulary, call it Anglan, say. But it would not be Lojban nor even of the Lojban pattern, for the derivation of the vocabulary is as much a part of the pattern as is the grammar (indeed, more so, since vocabulary construction antedates unambiguous grammar as a part of the program.) As long, then, as we are faithful to the whole project, we cannot shift to a single language (or even a set of closely related languages) as source.

But beyond this appeal to tradition, as it were, for not constructing Anglan, there are a number of other, external, reasons for thinking it would not be worth doing within the less specific parameters of the Lojban project. Here is a brief list of some of the major ones.

1. While the source of the words may not be important to its purposes (a claim we will examine below), the shape of them is. For almost any purpose, the fact that an utterance is uniquely decomposable into words at the phonological level plays an essential role. While there may be other patterns than the present one for doing this, whatever pattern is used will involve some fairly severe restraints on word shape. And these constraints, whatever they are, will not be met by words of English, taken unaltered. Thus, even in Anglan, the source words would at least sometimes (and probably usually) be distorted and, in a dense language like English, therefore misleading. Much of the supposed advantage would thus be lost even for native English speakers, without any compensating advantage for speakers of other languages.

2. One of the purposes of Lojban is to test the Sapir-Whorf Hypothesis (vide alibi [ed. note - this means see below]). For this, the test subjects would be monolingual speakers of various languages. Thus, the claim that they would all know English would not hold in this case, a central one for the project, at least in its inception.

3. Another possible (though certainly not central) use of Lojban would be as an international language at some level. Again the assumption has to be that the users would not know English, since, if they did, they would already have their international language. And, of course, if we were going to bother to teach them a language with an English vocabulary, we might as well teach them English, which would offer them a wider access to the world than the limited one - whatever it is - that Anglan would make available to them. Even the claim that Anglan has simpler grammar or is machine parsible could be met with reasonable subset-Englishes, which are slightly odd only from the point of view of full English, and allow access to most of it. Much the same sort of comments apply to the use of Lojban (or Anglan) as a language of machine control.

4. Note that in either of the latter cases, Anglan or subset English as either international language or machine control language, the big problem is that of importing things from the full or source language into the derived or subset one, where they do not fit. By the nature of Lojban (hence Anglan) grammar, the words of the language, even if they were identified with English sources, would rarely mean exactly the same as those sources. What the place structure of a given predicate is, is largely a matter of usage or is someone's best guess and is not something that can be figured out from just the English (or even from all the languages taken together.) Thus, knowing the English source would be misleading again, even if it were the only source. We already, being entirely an English- speaking group, have problems with bits of English usage creeping into our Lojban, and this would multiply if there were no constant reminder that the language is not just coded English.

5. Even though English may be a de facto international language, it is not always welcomed in that role. To attach yet another item, especially one which became important for some purpose - machine control or conference abstracting, for example - to English would make it harder for the proposed language to come to use, even after its essential virtues (parsibility in both these cases) were apparent.

Testing the Sapir-Whorf Hypothesis

Another common question is, How do you test the Sapir-Whorf Hypotheses? This, after the question whether it is testable at all, has always been a major point about the Hypothesis. We can only answer about tests in which Lojban is to play a role.

In broad terms, the Hypothesis claims that the limits of the language one speaks are the limits of the world one inhabits (also in Wittgenstein), that the grammatical categories of that language define the ontological categories of the word, and that combinatory potentials of that language delimit the complexity of that world (this may be Jim Brown's addition to the complex Hypothesis.) The test then is to see what changes happen in these areas when a person learns a language with a new structure, are they broadened in ways that correspond to the ways the structure of the new language differs from that of the old?

The ideal test, to which we may hope eventually to approximate (after several reasonable successful tests with, say English speaker) would require at least 1000 young monolingual native speakers of each of the base languages (and maybe a few others, to be on the safe side.) These would all be given a battery of tests to define the nature of their world views, as sketched above (limits, categories, combinations.) Each language group would then be divided into a number of subgroups, each as representative of the group as a whole as possible. One of these subgroups from each language group would be taught Lojban. Other subgroups would be taught variously another language closely related to the native language, a language totally different from the native language, logic and advanced mathematics, world literature, philosophy, a social science, a life science, and a physical science. All of these courses would be intense and thorough - the language course by the total immersion method. Finally, one subgroup from each language group would be kept out of all these courses and merely entertained or left to their normal life for the duration of the experiment (about a half a year, say.)

At the end of the experiment, all the participants would be put through the battery of world view tests again. The Hypothesis would be supported if the second test for those who studied Lojban differed from the first test in ways that correlated with the ways in which the structure of Lojban differed from that of the native language. Further, these differences would be different from those which could be accounted for simply by learning a new language (factored out on the basis of the groups who learned languages similar to their native one) or learning a radically new language (the other language learning subgroups) or learning new facts or a new discipline or, indeed, simply getting through another half year. Further confirmation would be found if the results of the various Lojban-learning groups came to be more similar on the second test than they were on the first and these similarities were also correlated with Lojban structure. Negative evidence would come from failure anywhere along the line (this is a pretty risky test): no change or change that is not correlated with Lojban structure or not distinct from that for learning another language or discipline or from maturation or change that does not bring different groups closer together.

The part of this outline that is least clear (and probably what the question is really about) is the world view test. So far as we know, there are no such test, though many existing tests may do a bit of what is needed. It is for this reason that every scenario for the experiment has included a clever social psychologist, who might either devise such tests or select from the available or, perhaps, find the right analysis of available (and created) material to do the trick. We may assume that the earlier experiments mentioned in passing would have served to refine and validate the instrument used. A further improvement would result simply from the pretesting phase of the grand experiment, when the various language groups each revealed its common core in distinction to all the others (or, at least, all the significantly different ones.) To be sure, this process may seem a bit circular, testing an hypothesis by using an instrument chosen because it gives confirmatory results in previous tests of the hypothesis, but this kind of circularity is typical of the rebuilding-the-boat-while-sailing-on-the-water approach needed in social sciences and does not invalidate the results, since there is a significant chance of getting no correlations at all at each stage, and thus disconfirming the hypothesis.

pc's view on the Sapir Whorf test is new to me also, so I'm glad he wrote it up. Most of us have presumed that native Lojban speakers would be needed, and he does not mention this as a requirement. It could even be said that for pc's experiment, we needn't even have a language demonstrably capable of serving as a natural language, since it is a second, possibly auxiliary language to all participants who speak it, and he mentions no requirement that these participants speak only Lojban during the test.

Bob's Response to Jeff

pc didn't really address one of Jeff's points, and it is one that I've gotten useful information on. As Americans, we see the English language from one perspective only, and we don't often hear the opposing views. I don't want to fault the PBS series, which I haven't seen, but it was written by an English-speaker. The arguments whether English is becoming, or should become, a (the) world auxiliary language, go both ways. Much of my information comes from Esperantists, who obviously tend to share a different point of view.

Is English truly a world language? First of all, which English? One source, author Anthony Burgess (Language Maid Plane), who incidentally supports English as a world language, notes that the language has diverged so much that it can barely be understood. He gives an example from a Tamil (in India) phrase that had to be written down before he could understand it. In another place in the book, he tells of when he moved to Lancashire England and his wife could not understand several local expressions, even though she was had been born only a short distance away. From an Esperantist, I have a story of how the British tried to import skilled labor from India into Kenya, and had to send them home because the two English dialects were mutually unintelligible.

Then we turn to foreigners who learn English. Our spelling and pronunciation is among the more obscure of the world languages. Any attempt to internationalize English would have to change this, and there is perhaps heavier resistance to such change than to learning other languages. Jeff mentions English as the international language of air flight; actually it is the one for air traffic control. But the UN organization governing air traffic control has the highest translation budget of any of the UN organizations in the attempt to make sure that the controllers know what they are saying. They don't really speak English; it is a code to them. The results can be catastrophic, since there have been air mishaps caused by misunderstandings of the 'English'. Will you rest comfortably flying over foreign territory knowing that your pilot might hear 'Go round!' as 'Ground!'?

The reason English is used, by the way, is twofold. Firstly, there is generally no better alternative known to most controllers and pilots; in Latin America, though, it is Spanish and not English that is used in the control tower. Secondly, when the international air traffic control system was set up, right after World War II, American planes and pilots were the vast majority of those in the skies; this is no longer true for pilots, and American dominance of the aircraft industry is being reduced.

This is just a sidelight of the wider reason why English is so widespread. It has nothing to do with scientific literature; most such material was written in French and German until World War II, but Latin was more widely taught in American schools in that era. England had a colonial empire wherein most locals were illiterate, and there was no written form of their language. Imperial England thus spread its language to the educated of all of its colonies, and these are the bulk of the countries where English is widely spoken in some form. We did likewise with our few colonies, such as the Philippines.

These countries have been independent only a generation or so, but English is already losing ground. The British generally left behind a written form of the local languages, which has allowed those languages to become the primary ones taught. Jeff is correct about groups of South Africans and Indians resisting Afrikaans and Hindi. They aren't winning, though. At best, they are gaining English being taught in addition to the favored language. In 1940, in India, you had to speak English to get ahead. Now, although perhaps a third of the populace knows some English, you must know Hindi to succeed. In China, Mandarin has been made the official language, and is mandatory in all schools. The massive increase in Mandarin literacy in China is what caused the major increase in Chinese weighting in Lojban gismu-making, as compared with JCB's 1950's data.

Hong Kong is shortly to revert to Red China. The natives are being taught Mandarin now, in addition to English. In the Philippines, English is still widely understood, and is one of the official languages. However, a side effect of the Aquino revolution is that English is no longer permitted on their television in the evenings; only Tagalog is allowed. Students have launched major demonstrations at universities when their professors tried to lecture in English, shouting them down with chants of 'Tagalog'.

English is the most widely taught foreign language in other countries, by the way. This does not mean that it is taught well, or that it is widely known or used. How well do you know a foreign language you learned in school? Probably better than most Japanese know English. Esperantist Donald Harlow told me of a Japanese Esperantist friend who has credentials to teach English in Japan - she was certified for 30 years; yet she cannot speak English, and can minimally get by reading it. Apparently, she needed only to pass a written test to get the credential. She says that 90% of English teachers in Japan have her skill level. How well do you believe their students will speak English after the class? In Russia, things are a little better. 70% of all students take some English, and apparently the instructors are better qualified. However, while an American author visiting Russia found that the places he visited on the state tour to always have English speakers, whenever he went out to talk to people on the street, they used German to mutually communicate.

A final reason for English's widespread use in the world is economic. To work in an American or British-run factory, one had to know some English. Now the Japanese are buying those factories and building new ones. Another documentary tells of the Japanese in Singapore (formerly a British colony), who had to learn English in order to communicate with their workers. A few minutes later, though, they show factory workers in a classroom with headphones on, learning Japanese. Americans are losing are total world dominance, and I believe that English has therefore reached its peak as a 'world language'. If you live in the Southwest U.S., the odds are good that you speak Spanish as well as English. In some parts of the U.S. where the Japanese have bought factories, that the managers must learn Japanese or lose their jobs. The next generation may see people arguing for Japanese (or more likely Chinese) as a world language. Or perhaps Lojban?

Jeff also mentioned the problem of languages transferring semantics when a word is translated. This is a problem, and one that concerns us. This is why I keep hitting people over the head with cultural neutrality. A translator is aware that there may be a difference in meaning between two words can, by being careful in the translation, defeat the tendency to read more into the words than is really there. This is especially viable in Lojban, which can make arbitrarily fine distinctions in meaning, and which makes it difficult to transfer certain assumptions about time. In short, if Sapir-Whorf is true, Lojban's structure will cause a difference in the way Lojbanists look at ideas, even those transferred from English.

We also can defeat this tendency by gaining greater representation among non-English native speakers. While at this stage, these people also must speak English in order to learn Lojban, they presumably would tend to write things in Lojban in translation from their own languages, or to draw on their native experiences, rather than to borrow from English speakers. We are continually exploring ways to bring greater participation from this group, and we will also be translating Lojban materials into at least 1 or 2 non-English languages as soon as we can.

The last of Jeff's points that I will deal with in this essay is the borrowing of technical vocabulary. Jeff makes the example of 'debuggare' from Italian. If the concept he mentions has a limited technical meaning which is identical in Italian usage as it is in English, there is no problem with borrowing it. Lojban will do the same. We have no fear of the semantics transfer associated with these words because the transfer is pretty much independent of the culture, and is only used by the segment of the populace that programs computers. There is no Italian word that this is displacing, and the English word being borrowed has partially Latinate etymology in the first place, so it doesn't even seem strange. I use his example, the English jargon word 'debug', in a discussion of le'avla vs. lujvo following this special section.

Now, it is true that there may be the capability to coin a word from totally Italian roots that would mean the same as 'debug', and would be different from 'debuggare', but who is to gain. Jeff implies that the Italians resist this type of borrowing, but not in any systematic manner. Now the French, on the other hand, would be less likely to accept such a borrowing, and would coin a 'pure' French word. They have an Academy to enforce this policy, and enough state control to make it hold. The Germans, are prone to make their own metaphors rather than to borrow, as are the Chinese. The Japanese borrow freely, as do the Arabs, but they heavily modify the words to fit their language structure, using syllables to override consonant clusters in the former case, and using only the consonants as meaningful semantic indicators in the latter.

Lojban will follow the latter approach, borrowing freely but heavily modifying words to fit our unambiguous morphology. There is no reason to doubt that this will hold true while the language is first being spread, since that morphology offers one of the selling points of the language.

pc's was the last of the contributions solicited specifically for this topic. But I have heard from many more of you on the subject of why you are interested in Lojban, and on what you want and need from us. The next section consists of excerpts from just a few of those letters, which I will generally include without comment. The letters are presented in chronological order.

Excerpted Viewpoints of Correspondents

[Bob: The first letter is from one of the few 'drop-outs' I've had - actually he had given up before we started - who tells us why the language failed him. Even in dropping out, he helped us learn about what people wanted. He also donated his collection of materials, a spare set that has proven handy since then.]


"This is in response to your Loglan prospectus. I will not be taking part. "I've forgotten how I learned about Loglan; perhaps it was through the Scientific American article. My initial enthusiasm peaked on receiving a copy of L1, the waned fairly rapidly, and I never did make any concerted effort at studying it. There were always excuses, among them dabbling in Swahili, Quechua and a couple Polynesian languages. "Eventually, this was several years ago, I came to the realization that I dislike the language -- the Chicken MacNugget-like primitives are ugly, its expressions too dense... So I haven't responded to any Institute mailings since then.

"Intellectually I'm as much convinced as ever that the concept is worthwhile and that the effort is much superior than any previous constructed language. Then too, my sense is that most of us are lamentably indifferent to matters of language. But that doesn't make me want to spend my time on Loglan. "...Good luck in getting your SIG going."

William J. K. Harrington


"I am a long time passive member of the loglan interested set and I just want to thank you for activating the language and for the recent mailing...


"I grant all due respect to JCB for creating loglan so long ago, but now it is his private obsession. This would be all right too if he hadn't involved the public (me) by selling books, using us for word "taste" tests, etc. We have been used, and have a right to expect reasonable progress, not including arbitrary halts for JCB to retain control. "Once loglan ceases to be JCB's private game, it faces new dangers. The loss of standardization is perhaps the main one. After all this is one of the appealing things about a language built by design for consistency, ease of use, and above all to facilitate communications. We have the precedents of Forth, Pascal, Prolog, etc. before us. We have to realize that once the cut is made between the concepts and ideas of the language (unpatentable) and the words themselves (copyrightable) there is nothing to stop others from doing likewise. And Babel will arrive. This is where we came in. I do not have the answer, but hope for one main loglan-type language ultimately. Meanwhile, the sooner we can speak Loglan-X the better."

G. L. Koenig


"... I do, however, have some interest in the premise around which Loglan was constructed, and I have read the 1960 Scientific American article and have started to memorize the 112 words. My interest in the premise revolves around the fact that my youngest daughter is profoundly deaf and communicates in sign language. She is 13 years old, is mainstreamed in a total communications program and seems to have an extremely well developed ability to handle problems of logic.

"... Sign language is based on English, but does not have the structure for math or logic. At least it does not seem that it has the needed concepts and structure."

Claude Van Horn

[Bob: It occurs to me in typing this that Jerome Frazee's symbology might be highly adaptable to sign language, given its pictorial content. Claude, you might want to try this idea with your daughter, and to contact Jerome for more ideas. Let me know.]


"In 1977, I bought a copy of Loglan Grammar and the Loglan dictionaries 1 & 2. I studied hard and practiced sounding out the words. I had no one to speak to, but I was willing to study and wait. Later, I heard of revisions to grammar and word lists and I thought I was fighting a losing battle. I then quit, and the process of forgetting has been painless. "Now you come along and stir up old dreams and make it look promising. Can you deliver the dream?

"What I am looking for in lojban is simple. I want a language I can speak, write, and teach to my son. Something fairly easy to learn. Something dynamic but not changing every day because it isn't finished. Your promise of a baseline may give me what I want. I hope so, because I am willing to give it another try; particularly now that I have someone to talk to."

Dan Cheek


"Glad to see someone getting on with the job. In Cambridge Loglan classes run by Chuck Barton and Scott Layson, I had advocated revolt against the nitpickery in The Loglanist. I soon came to realize that no workable universal language can be as simple as I'd imagined on first reading the Scientific American article. As the saying goes in R&D work, everything takes longer. Nonetheless, the world will ultimately need Loglan, so keep up the good work. "In 1957 I began a book on war-prevention ethics which offered a cultural immunization program for children. Recent findings now allow completion of the tough part, the adult section. A culture-free language, initially a universal second language, is one item to be called for."

Milton W. Raymond


"Thank you for including me in your mailing. I am certainly willing to support your efforts. "My main interest in Loglan (or Lojban, or any similar language) is in learning to use it. I would like to see some personal effects related to S-W, and I would also like to eventually be able to have my computer take loglan input and produce loglan output. ...

"As far as loglan-type languages and computers are concerned, my personal feeling (based on not much more than gut feel) is that a more workable (and perhaps useful) first application is as a database query/maintenance language than as an artificial intelligence language.

"Although I have been associated with the Institute off and on since 1975, I have never managed to learn Loglan. I am therefore interested in Lojban, since you appear to have a better handle on teaching than does the Institute. You have also published a schedule, which is the first I have seen from anyone on any GPA. I agree that the Institute appears moribund. ...

"What I have been hoping for from the Institute and have not yet received, is a way to "bootstrap" my ability to read and understand Loglan. ... I have not been able to learn in isolation.

"On the subject of humor in Loglan, I haven't seen any. I feel that anecdotal humor should have few problems. I think double entendres and puns would be almost impossible, due to the intent of the language design. That is, in fact, the only thing my wife likes about Loglan. Vickie approves of anything which limits my ability to make puns. I'm not at all certain that metaphors can fill the etymological niche occupied by puns.

[On the letter from Paul Doudna in HL3:] "I understand plausibility tests by the listener to be against the spirit of the language. Loglan 1 stated that being able to speak nonsense unambiguously and intelligibly was one of the features of the language... Perhaps the limiting of assumptions is not total, but I doubt that it could be. If a language made no assumptions, period, then it would not be a language, since there could then be no common referents among speakers. It seems to me that 'house-containing blue objects' and 'house inhabited by blue-skinned people' are more complex concepts than 'house which appears blue', and deserve a more complete exposition to get the idea across. Perhaps Ockham's Razor should be made a part of the language definition under 'resolvability of metaphors and predicate pairs'. Of course, that assumes that speakers and listeners share enough of a word-view that Ockham's Razor will bring them to the same point."

Steve Wheeler

[Bob: Steve is an 'aficianado', and wrote this letter after having received NB3. His comments on the quality of available teaching materials thus are especially relevant.

On Steve's comments re. Doudna's letter: Plausibility tests are valid for the listener; metaphysical parsimony means that the speaker must make the fewest possible assumptions about the listener's metaphysical outlook. This would make the use of Ockham's razor invalid, since, as Steve pointed out, the listener may not have the same world view. Given an incomplete exposition, the listener is therefore allowed to infer whatever he/she chooses about the ellipses. Pragmatic considerations in normal communication will lead to plausibility tests on these ellipses. If the speaker leaves no ellipses (and a tanru is inherently elliptical), the need for plausibility tests goes away.

The three interpretations of 'blue-house' are equally complex according to both English and Lojban grammar. All three are expressible as relative clauses with one level of subordination. (Relative clauses are a possible topic for JL7.) Using comparable wording in English to match the Lojban we have 'nest such-that it contains blues (or blue-objects)', 'nest such-that blues (or blue-people) inhabit it', and 'nest such that it appears blue (or blue-out-surfaced)' It is our world view that causes external appearances to be more important than internal details, which of course causes worse problems than plausibility tests in other fields of human endeavor.]


"I think more than ever that we need a maximum language for world use. However I believe Loglan in both forms will be most useful in the eventual construction of a maximum language. Unfortunately I am 79 and doubt if I live to hear such a language."

Faith Rich


"I'd like to thank you for having 'kept the faith' and left me on your mailing list ... even without hearing from me in any way. After reading JL4 and JL5, I feel a bit of the old spirit and once again get the impression that Loglan (whatever name we call it) has some vitality, an impression I haven't had for some years even while things were ostensibly happening - I think that what was missing was a feeling of growth, not just continuation along a logical path."

Rohan Jayasekara

The Editor's Viewpoint

Hopefully, these letters gave you some sense of the variety of goals and desires of the members of the community. You also may glimpse how I might feel in receiving them; feelings that have compelled me to continue in spite of JCB's opposition. It is hard to say no to letters like these, and this is but a sampler. There are several other opinions out there as to why Lojban is interesting and/or important.

So why Lojban?

I believe that ALL reasons for creating, learning, and using Lojban must be accepted as valid, since each plays a part in making Lojban interesting to some segment of our audience. For Lojban to be a language and not a linguistic toy, it needs the broad base of speakers implied by these various interests. In any case, Lojban is already a composite of all these goals, whether or not they are fully realized. The goals are valid to the extent that each motivates people to contribute to the language. Without all these people, and therefore without all these various goals, the language would never have made it this far. We who value Lojban owe each viewpoint respect and understanding. I hope these contributions and letters cause you to value some aspects of Lojban that you had not considered before.

As for me, I will strive to enable the language to fulfill as many of your goals as possible. Some tradeoffs are going to be necessary; not all people will be satisfied by any of our decisions. It is my job and my skill as 'systems engineer' to recognize and try to balance these tradeoffs to maximize the quality of the resulting product. I'd like for Lojban to live up to Mr. Harrington's evaluation of its quality; he might eventually decide that our result 'makes him want to spend time'.

I believe enough in my abilities to allow me, in good faith, to promote the language as strongly as I have been. I believe enough in the value of the work I am doing, that I have continued unemployed to allow me to put my full effort into achieving these ends. I am determined to set a standard of quality and thoroughness for this effort, given the constraints under which we have worked, that no one will ever consider 'splitting' off because they believe they can do better.

I had no training in linguistics when I started to work on Lojban last year, when it became clear that JCB would not allow me to work with his version of the language on terms that I could accept. In fact, probably like most of you who received them, I couldn't understand 90% of the stuff in TL when it came out. I haven't read Whorf's writings, nor Chomsky's. I learned a smattering of Spanish in elementary school, the same for German in high school. I remember almost nothing of either. While I have some competence in computers through professional experience, I am learning about artificial intelligence, the field where Lojban has most likely application, through my work with the language. While I have trained 'junior' professionals in my own field, I can claim no formal training in education to justify my ability to write a Lojban textbook.

Who am I therefore to judge that one reason for working on or learning Lojban is better than another?

For that matter, do any of us have the credentials in all these fields to give us that 'right'? I doubt it. Of the people I've talked to in the community, pc seems to come closest to that combination of training and experience. But I haven't heard him claim to know the answer. Still he believes enough to keep working.

None of us are qualified to 'invent' a natural language from scratch, something that hasn't truly been completed in recorded history. JCB got us started with his creative spark and determination. As a group we can do what no individual can do, what no individual nor scientific group can define how to do: create a language sufficiently robust that it serves all of the requirements that we assign to language as human beings in society. For this truly is the goal of Lojban.

As a research effort, the Loglan/Lojban project has been an oddity, especially for 20th century science. It is being conducted by a team of people, most of whom are neither professionally trained in, nor working in, the fields most relevant to the project (which are linguistics and sociology). Kieran Carroll told me that one thing that inspired him to work on the language was the spirit of volunteers working together to conduct scientific research. We are most of us amateurs, in both senses of the word. We are not 'employees' nor 'entrepreneurs' in that we seek no financial reward for our effort; but furthermore, we work on Lojban for our love of it, and of the concepts and goals that we and others hold for it. Can we ask for a better reason?

Other Correspondence - Jeff Prothero: Public Clarification and Update on PLoP

(14 July 1988) Public-domain Lo--an Parser (PLoP) Update:

Editor, Ju'i Lobypli:

I write this letter at the behest of Jim Brown, to clear up a misunderstanding which I may have inadvertently propagated. You will recall that shortly after my release of PLoP, Jim retained a lawyer to write me a letter alleging that my parser infringed unspecified copyrights, detailing the various fines and prison sentences attendant on conviction of such an offense, and demanding that I cease distribution of PLoP and recall all outstanding copies. On advice of legal counsel, I then wrote a reply requesting details of the alleged infringements, and offering to modify the parser as necessary to satisfy Jim.

After nine months and a change of lawyers, Jim has finally replied, stating that he and his lawyer "consider the case closed," that he intends no legal action against me to prevent distribution of PLoP, that he never intended any such legal action, and finally suggesting that I knew he never intended such action and should make a public correction on this matter. I will admit to having been under the impression that Jim was considering legal action, but I am happy to take this opportunity to correct the public record.

To forestall possible future accusations of misrepresenting my relationship with Jim, I also note for the record that Jim has informed me that, due to my communication with his "enemies," I am no longer numbered among his twenty remaining friends in the world, and that in fact we were never friends in the first place. In the course of a voluminous and cordial correspondence over more than a decade, I have found Jim to be, on the whole, a genial and sensible fellow, and I am sure he can be trusted on this matter.

On a more positive note, I can report that version 1.13 of PLoP is now available. Versions earlier than 1.12 had bugs so serious as to make the program basically unusable. As before, anyone who wants a free copy of the program may contact me at the below address. Since PloP has, unfortunately, not yet been converted to the updated vocabulary and grammar, its immediate utility as a teaching tool is fairly limited. Also, I am only able to provide 5" DOS-format floppies, the program can take a minute or more on really complex sentences (due to backtracking), and really needs some attention by a hacker with more time than I can currently spare. (That said, it's still a neat toy and a working prototype!)

Regards, Jeff Prothero 221 SW 153rd #194 Seattle WA 98166

[Please contact Jeff for your copy of PLoP, or you can download it from CLBB when it is again on-line. The grammar used is similar to the Trial.55 version of the Institute's grammar, and it uses Institute cmavo lists. C programmers can substitute the new cmavo for the old; Jeff includes source in his distribution. The new grammar will not be all that different from Trial.55 provided that you have substituted the new cmavo. - Bob]

Filler 'Page'

At pre-publication, I had almost a full page left over. So I've added a couple of short items inspired by the special section. Please pardon the compressed spacing.

I will first take Jeff Kegler's example, the word 'debug', and analyze how it can be transferred to Lojban. I believe this will reinforce pc's arguments against English roots. It will show some of the problems we face in borrowing, and how tanru and lujvo can sometimes resolve these problems.

'Debug' is not valid Lojban; the vowels aren't pronounced 'correctly, there is no consonant cluster, and the word doesn't end with a vowel. Once you Lojbanize the spelling to 'dibag', you've already lost a lot of visual recognition. Were we to permit the 'easy' change of turning into a gismu-form word, as was permitted in language versions prior to 1984-ish, we could make this dibgu, and it would be a valid word, though hardly a winner for recognition. In fact, I think dibgu still is valid, but this is happenstance. gismu word space is small, and the language doesn't permit words that are too close together, so someone would have to argue long and hard that 'debug' was an important concept to have it buy off some of this valuable space today.

We could instead move it into borrowing space. The algorithms for borrowing are still a little up in the air in Lojban. By pre-Lojban versions of the language, we would have to add at least one useless consonant to make it a legal borrowing, and possibly a full syllable, assuming that we don't allow it to occupy gismu space. We could model the Italians and make dibgure, or add a consonant as dribgu. I believe these would both pass the old tests for valid borrowings. Neither is particularly recognizable to English speakers, nor likely to Italians.

The current Lojban plan would be to recognize 'debug' as a computer jargon word, and build the word off the rafsi assigned to skami, for which the keyword is 'computer'. This rafsi is sam, which would be attached to a suitable form of the borrowed word, to result in something like samdibu or samdebu. These are valid borrowings, or le'avla (which derives from the tanru 'take-word'; we're too honest about the process to call it borrowing in Lojban), since they cannot be analyzed or broken down into either gismu or lujvo. These are more recognizable than the old type of borrowing to an English or Italian speaker who knows how Lojban builds these words. But they aren't useful at all to an English or Italian speaker who doesn't know Lojban. Borrowing buys far less in Lojban than it does in Italian. A tanru can be made for 'debug' in Lojban, one which is easily analyzed by a Lojbanist who isn't familiar with the word, using the principles I described earlier in this issue. Using the tanru built from computer-rule-repair, which I believe captures the gist of 'debug', we can get a lujvo such as samjvacikre /sam,zhva,SHI,kre/. or you can use computer-problem-repair: samnabycikre. Of course, if you are talking to another computer person, the 'sam' is unneeded (it is needed in the borrowing to make the word a valid form). This gives you the choice of jvacikre or nabycikre, or possibly even just cikre; you will be understood by your listener. And you know how to do this with lujvo now, whereas I have a tougher job teaching you how to make a valid le'avla; the algorithm is less obvious and requires a better feel for the morphology than even I have, in order to be able to make them spontaneously.

In Lojban, I'm sure you'll agree, the tradeoff between the two approaches is strongly weighted against borrowing when a tanru can do the trick. This is probably why German and Chinese, which have good techniques for metaphor making, seldom borrow. But, when you have to borrow, Lojban's method gives some clue to meaning, unlike other languages which borrow. The Lojbanist has the best of both systems.

Nora, visiting Lojbanist Gary Burgess, and I did the following quick translation of The Lord's Prayer in about 45 minutes at a weekend picnic just before publication. It is certainly is open to theological debate about its subtleties. Athelstan and Tommy Whitlock assisted by checking some passages against the original Greek, which is even more ambiguous than English due to an absence of punctuation. Alternate interpretations exist for the scope of 'on Earth as it is in heaven', and we have an alternate version for lines 4 and 5 in parenthesis that covers one of these. The second parenthetical text is the final line that is not found in all versions. As usual for our translations, we give each line three times, in Lojban, literal English, and smooth English. We have no space for detailed commentary.

Note that the translation uses current cmavo and grammar and may have minor errors or changes; I'll try to note any in the next issue.

le xisyctu jdaselsku be la jegvon The Christ-teach religious-expression to the one named YHVH (Jehovah). THE LORD'S PRAYER

foi cevrirni .iu ne zvati le do cevzda ku ku O divine-parent (affection) who (non-restrictive clause) is at your divine-nest, O Father who is in Heaven,

.aicai .e'ecai lo do cmene ru'i censa (Intention!!) (Obligation!!) Your (true) name is continuously holy. (Amen!!!) Your name must be hallowed.

.i .aicai .e'ecai le do nobli turni be la ter. se cfari (Intention!!) (Obligation!!) Your noble rule of Earth be initiated. Your Kingdom come.

.i .aicai .e'ecai loi do djica ba snada mulno vi'u le cevzda .e .a'o la ter (Intention!!) (Obligation!!) The mass of Your Desires will be successfully-completed throughout the divine-nest, and (I hope) Earth. Your will be done throughout Heaven and Earth.

(.i .aicai .e'ecai do nobli turni le cevzda .ebaza a'o la ter.) ((Intention!!) (Obligation!!) You nobly rule throughout heaven and soon (hopefully) (throughout) Earth.) (Your Kingdom become on Earth as it is (now) in Heaven.)

(.i .aicai .e'ecai loi do djica ba snada mulno vi'u le cevzda .e .a'o la ter) ((Intention!!) (Obligation!!) The mass of Your Desires will be successfully completed throughout Heaven and (I hope) Earth.) (Your will become done on Earth as it is in Heaven.)

.i .e'o ko dunda ca le cabdei ku le ri nanba ku mizu (petition!) Give, on the now-day, its bread to us. Please give us today, today's bread.

.i .e'o ko fraxu mizu lo ri zu'o palci (petition!) Forgive us for the mass of our Activities-Of-Being-Evil. Please forgive us our sins.

.ijo .e'e mizu fraxu da poi pacyzu'e xrani mizu if and only if (we should!) forgive those who (restrictive clause) evil-actively injure us. As we (should) forgive those that sinfully injure us.

.i .e'o ko lidne mizu vavinai loi pacyxlu (petition!) Lead us from-near-not-into the mass of Evil-Influences. Please lead us not-into temptation.

.i .e'o ko sepri'a mizu loi palci (petition!) Separate-cause us from the mass of Evil. Please separate us from evil.

(.i uicai ni'i loi se turni .e loi vlipa .e loi mi'orselsi'a me le do romei) (Happy!!) The previous logically follows from this: the mass of the Governed, and the Powerful, and the Famous-Esteemed pertains to your all-some. Because that which is governed and powerful and gloried is all yours.

Financial Status, Price List and Ordering Instructions

I am pleased to report that your contributions since JL5 were larger than in any previous 3-month period by a large factor. This was due to primarily to some large contributions, including one donation for $500, and several for over $100. Before this issue went to press, we topped the $5000 mark in contributions this year, and are closer to $5500. This means that we have received enough money in advance to pay the $1000 (or more) that it will cost for this issue, and still not exceed our income for the year. By cutting back on expenses, and with our bulk mail permit for newsletter postage, I have kept our expenditures this year to some $4400.

Nora and I have contributed matching funds as promised, and with the help of those large donations (which were generally non-balance gifts), we have managed to set aside about $4000 in a separate account to be turned over to the incorporated organization. This amount will be maintained equal to outstanding balances and obligations of the organization. We consider your subscription balances to be loans to the organization for operating expenses; we assume the obligation to pay back in goods (of value) or in funds, the money which you have loaned us. Matching funds, volunteer credits, and similar awards we consider to be obligations to you in services, though we cannot pay you back these amounts. The incorporated organization will maintain these policies.

We feel disappointment in that, while we received sizeable contributions from many people, a large number of you still have never responded. We don't mind continuing to distribute materials to all of you if you truly value them. But we'd like to know that we are fulfilling your needs, and that our money and your time is not being wasted. So please respond, even if you cannot afford to maintain a balance or even to cover the cost of your publications. Our balance system continues to be voluntary, and no plans exist to change this policy.

Ordering Policy and New Products

We are continuing our policy to keep overhead down by classifying you into one of two groups. Level 1 people get newsletters and materials that are baselined. Level 2 people get all of this material plus new draft materials as they come out. Faced with a lack of clear response from many of you, we are classifying the following into 'Level 2': people who ask for this level, people who ask for newsletters and teaching materials, people with balances in excess of $20 who have not asked for 'Level 1' status, and people who are known to be actively working on Lojban development. We have also included most DC-area people who we've kept waiting for a class in this category. The rest of you we classify in level 1.

For those who haven't been following, the time I was spending on scattered orders was starting to affect my ability to keep the work progressing. Furthermore, I was having trouble keeping to my goal of responding to all requests within 6-8 weeks, which is considered minimally acceptable for mail order services. I hope I am doing better now; please let me know if you have fallen through the cracks and have waited longer than that for something.

To a limited extent, I am accepting individual orders, most especially for more costly items like LogFlash or MAC Lojflash. We also have limited quantities of back issues of this newsletter, but I have to charge a higher rate when we run out of original print copies and need to reprint.

The top line of your mailing label gives the status of your voluntary balance. A typical balance line looks like:

C 50.00 -D 10.00 -S 21.25 =B 18.75

This says that we have $50.00 in Contributions from you, including matching funds and volunteer/award credits. You have Donated $10.00 of this as a gift not to be repaid. We have Sent you $21.25 in materials, resulting in a net Balance for you of $18.75. For foreign recipients, this balance does not include the postage for this newsletter, which we won't know until we take the envelope to the post office for shipping.

In addition to the material shown last issue, we now are offering MAC Lojflash at the same $20 Shareware voluntary balance price as we are offering LogFlash 1 PC. It will be sent automatically to those that we know have MACs, since nearly all of you have asked for the program. We are sending baseline updates for LogFlash 1 PC to all known recipients of the previous version at a balance price of $5 to cover our costs for diskettes, postage, etc.

We also have available calligraphic pin-on buttons (in several colors) which say e'osai ko sarji la lojban. This slogan, created by Claude Van Horn (with grammar corrected to its current state), translates as 'Please support Lojban!!' (See JL5). These buttons were distributed to all Logfest attendees; we have a quantity left and can get more if there is interest. Calligraphy is by Lojbanist Nancy Lebovitz, who offers an enormous selection of serious and humorous expressions in various calligraphic scripts, as well as special order buttons and other calligraphic work. We are charging $1.75 each, which includes U.S. postage, while they last. There is a discount for quantity orders. Nancy has had to raise her prices, so if we run out, the price will be 25 cents more. You may write her at the following address for her current catalogue:

Nancy Lebovitz 410 Wollaston Ave. C6 Newark DE 19711

We are including copies of the English-order and Lojban-order baselined gismu list in this mailing at a lower price than we can offer them separately, due to reduced bulk mailing postage. We are charging $1.00 total for both in this mailing, but will be charging $1.00 each (net $2.00) in separate mailings. The price of this newsletter at 58 pages is $3, plus the $1 for the two gismu lists.

Revised Price List

  • ORDERING INFO FOR RECRUITERS - A new respondent will be sent a brochure and JL6 with a balance charge of $4.00. If he/she sends a registration form, we assume they've seen a brochure and have knowledge of what Lojban is. We will send either level 1 material ($8.50) or level 2 material ($9.75), plus teaching programs ($20.00) if requested. The level 2 price will increase when the draft cmavo list is published.

Level 1: Pages Price UL1 44 $5.20 UL2 38 4.60 HL3 34 4.20 JL4 40 4.80 JL5 54 3.75 for the next 25 or so copies 6.00 reprint price when we run out

  • JL6 58 3.00 U.S bulk mail price

(Foreign recipients add postage) 4.00 new respondents and non-bulk US Mail 6.60 reprint price (after JL7 or if we run out)

  • Synopsis of Lojban Phonology

and Morphology 32 2.50 subsidized by us a bit

  • Baselined gismu lists

Lojban order 14 1.00 (50 cents with JL6 bulk mailing) English order 14 1.00 (50 cents with JL6 bulk mailing)

Level 2:

  • lujvo-making algorithm 9 $1.25
  • cmavo draft list 40+ ? unknown price, expected prior to JL7

Special Orders Only: LogFlash 1 for IBM PC, diskette and printed manual $20.00 Baseline update diskette for previous recipients $5.00 MAC Lojflash for MacIntosh, diskette and printed manual $20.00 LogFlash manual 15 $1.35 Answers to HL3 Questions 34 $4.20 Calligraphic button $1.75 e'osai ko sarji la lojban

All of the above materials may be copied and distributed freely. Our publications and software are generally copyrighted, with unrestricted permission to freely copy and distribute any publication or program as long as you include any copyright notices, and do not charge for the product. Any derivative works must subscribe to this license.

Contributors Since Last Issue

The following are contributors who have aided us since the last issue. Due to space considerations, I cannot list each contribution, whether money, volunteer work, or advice and suggestions on how to solve various problems. It is all of you that makes Lojban possible, and worth all the work we are putting into it. Let this list be our brief word of thanks to each of you.

Brooke Albert, Athelstan, John Atkins, Steven Angart, Rebecca Bach, Chuck Barton, Alan Beatty, Jamie Bechtel, John Bridgeman, Vincent Burch, Gary and Diane Burgess, Carl Burke, Sam Cardman, James Carter, Heidi Cassidy, Bob Chassell, Dan Cheek, Paul Doudna, Marc Drexler, Ralph Dumain, Pam Falk, Gyorgy Fekete, Colin Fine, Jerome Frazee, Mike Gunderloy, Donald J. Harlow, William Harrington, Jay Hart, Hans Havermann, Milagros Hermoso, Andy Hickmott, Arthur Hlavaty, John Hodges, James Hughes, Rohan Jayasekara, Burt Johns, Patrick Juola, Toivo Kalervo, Richard Kennaway, Jeffrey Kegler, Lawrence Kesteloot, Andrew Koenig, G. L. Koenig, Charles Korbas, Bob & Nora LeChevalier, Nancy Lebovitz, Bill Lee, Greg Lee, Doug Loss, Brad Lowry, Keith Marshall, H. Michael McGaw, Robert McIvor, Dick Metzler, Todd C. Moody, Leroy Nelson, Paul Francis O'Sullivan, John and Sara Parks-Clifford, Michael Pique, Jeff Prothero, Jon T. Radel, Keith Ramsay, Milton Raymond, Paul and Grelia Reiber, Faith Rich, Maureen Rosch, Rick Sakamoto, Jeff Saxe, Dave Schrader, Jerry Shifrin, Joel Shprentz, Donald Simpson, Terri Sisley, Steve D. Smith, Steve G. Smith, Guy Steele, Mark Stowe, Arthur Tansky, Jeff Taylor, Jonathan Tite, Guy Townsend, Claude Van Horn, Bruce Walker, Steve Wheeler, Richard A. White, Tommy Whitlock, Art Wieners, Erick Wujcik, Michael F. Yoder.

There are over 90 names this time, many here for the first time. If I missed you, please accept my apologies for the omission, and my appreciation for your participation.

Glossary of Technical and Lojban Terms Used in this Issue

These are brief definitions. The Synopsis of Phonology and Morphology gives more detailed explanations of many of them. For Lojban words not defined here, see your gismu list.

the principle that Lojban should make it possible for speakers of other languages to express the ideas of their culture in Lojban
see rafsi
a type of language much different than our own; these language attach pieces of words together in great complexity (much as English users attach such particles as 'un-', 'dis-', '-able', '-ed', etc.). The results are words that may be 20 or more letters long, but which carry the comparable meaning of an entire English sentence. Turkish and Finnish are examples of this type of language. Lojban and Esperanto have some features of agglutination. attitudinal - also often called 'indicator'; these expressions convey the speaker's emotional attitude about the claim being made, independent of whether the claim is true or false.
see pages 4-5
see le'avla
Lojban predicates (with arguments included). These convey the basic claim of a Lojban sentence. A simple sentence consists of a single central bridi. The bridi is a kunbri with sumti attached which give the specifics of the claim. Each sumti in turn may be composed of subordinate clauses with their own bridi; usually these are simple kunbri. The basic sentence syntax in normal form is:
central kunbri - sumti - sumti - sumti ... In this structure, the central kunbri corresponds to the English verb, and the sumti to the attached noun and prepositional phrases. There are other orders and structures possible.
Predicate words. The basic components of Lojban semantics, found in kunbri. brivla in a sense can be said to subsume several English parts of speech: nouns, verbs, adjectives, adverbs, and prepositions, depending on how they are used in the sentence. brivla are both singular and plural unless specifically quantified. case - a system of classifying the types of semantic relations that each sumti of a bridi has to the other sumti in the place structure.
CCV, CVC, CVV, etc.
abbreviations which indicate a pattern of consonants and vowels that is required in a Lojban word or rafsi
Noam Chomsky, a leading American linguist and leader of a school of linguistic thought that suggests that all natural languages can be described by a universal set of transformation types that allow us to generate all possible utterances in the language. The results of analysis by his methods is called a transformational (-generative) grammar.
'Capital Loglan Bulletin Board'; see page 3
Little Words (LWs). The structure words, usually only one to three letters long, which convey the organization (syntax) of Lojban utterances.
Names that have been Lojbanized into a proper word form.
see lujvo
A study of word/concept frequencies in 4 European languages produces in the 1920's/1930's by Helen Eaton. No more recent work of this type is known.
see Chomsky.
Individual primitive root words that are the most common non-structural concepts. These are the basic brivla. They are used individually, or combined in tanru. They are also used to build compound brivla: lujvo. gismu are of three types: xregi'u (C-Prims), lobgi'u (L-Prims) and klogi'u (N-Prims). Most are xregi'u, which were built algorithmically from six common languages. The other two types are, in effect, types of le'avla that are in gismu form. All gismu are of either form CVCCV or CCVCV.
the rules of syntax, or sentence formation, in a language.
In Lojban, a letter y, r, n, or l used to attach word pieces, such as rafsi, together. The hyphens ensure that the word doesn't fall apart into its components, which might have different meanings.
Dr. James Cooke Brown, inventor of the concepts behind Loglan/Lojban, and of the board game Careers.
An English word or short phrase which conveys the sense of a Lojban word for easy word-association. A given keyword is assigned to only one Lojban word, even though the English may have multiple meanings, or even pronunciations.
Bare bridi without arguments attached; these may be individual brivla, or strings of multiple brivla combined in various ways.
Loglan 1, the language description by JCB published in 1968 by Xerox University Microfilm, and then republished in 1975 by The Loglan Institute, Inc.
Loglan 4/5, the language dictionaries edited by JCB and published in 1975 by The Loglan Institute, Inc.
Borrowings (L-Prims). Specialized terms, often technical, that are taken from other languages and Lojbanized into an acceptable word form.
Letterals, or symbols for the letters of the alphabet, now conceptually expanded in Lojban to include all non-numerical symbols that have no defined Lojban cmavo.
see cmavo
The house organ, originally a member-run newsletter, of The Loglan Institute, Inc.
Complexes (Cpxs). brivla that are formed as compounds of gismu, each representing a particular tanru. Whereas tanru can be ambiguous, there will be a specific meaning assigned to each lujvo and specific place structures for associated sumti.
the interaction between two or more words to suggest a new concept that is not an exact combination of the source words. See tanru.
Mathematical expressions and their Lojban grammar; a longstanding problem. In Lojban: mekso.
The rules of proper word formation, which are rigidly defined for Lojban.
a grammatical means of expressing the speaker's observational relationship to the claim of a
by what process he/she has come to express the claim. See page 33; c.f. attitudinal
The determination of the syntax of a string of text or symbols.
Dr. John Parks-Clifford, professor of Logic and Philosophy at University of Missouri, St. Louis. pc also has a degree in linguistics, and has consulted in artificial intelligence processing of natural languages. He specializes in the study of tenses and tense logic. As the editor of The Loglanist during its existence, pc became the second-most authoritative person on the language after JCB. He is Vice President of The Logical Language Group, and my main technical expert on the language.
permissible initial/medial
Lojban's morphology rules only allow certain consonants to be found together in 'clusters' at the beginning, and in the middle, of words. These two sets govern several areas of word building, and indirectly, pronunciation. As such, the rules that describe these sets should be learned.
the detailed study of the sounds of a language
place structure
the fixed ordered set of sumti that are associated with a bridi; the bridi gives the semantic relationships between each of these sumti.
plausibility test
the process by which we figure out what something means based on context. Given several possible meanings, we use the contextual clues to eliminate implausible meanings, hopefully leaving a single answer.
see bridi
see gismu
These are affixes, combining forms of gismu and common cmavo that each serve to represent a single word. They are assembled together by specific rules to form lujvo, which can in turn be broken down uniquely into the represented components. Those components then indicate a tanru that may suggest the meaning of the lujvo, making new lujvo comparatively easy to learn.
Robert A. McIvor, retired Canadian Lojbanist who is currently aiding JCB by developing computer teaching programs for the Institute version of the language.
the recall of a word or phrase in a foreign language, when prompted with a keyword or clue in a more well-known language. This is much more difficult than recognition.
recog, recognition
the recognition of a word or phrase in a foreign language, and the ability to associate that word/phrase with a more familiar word/phrase in a well-known language. the opposite of recall.
Sapir-Whorf hypothesis
see the discussion in the special section for definition of this concept semantics - the study of the meanings of words, either separately or in combination
The arguments of a bridi. Each sumti is related in a specific way to the brivla in the kunbri to which they are attached, and to the other sumti that are attached. The most important of these sumti are defined to exist in a particular order for each brivla, and there are rules to determine which set is used.
syntactically unambiguous
The existence of a single correct word grouping (parse) for a given word order. In proving Lojban unambiguous with YACC, we must further be able to do so by a set of technical rules called 'LALR(1)', which basically means that the language groups from the left, and that it is possible to parse the language taking one word at a time from the left.
The structure and word order of a language, independent of the meanings of words. Similar to 'grammar' in meaning.
Metaphors that are made by combining brivla in various ways to create new, more complex, or more specific meanings. Lojban tanru are ambiguous in meaning, with the capability to expand upon and define that meaning more clearly to an arbitrary degree. However the syntax - the order and grouping of the brivla and cmavo that make up the tanru is rigidly and unambiguously defined.
the study of classification systems, especially used to describe systems which classify things by successively breaking groups into smaller groups. The most well-known example is the scientific names for plants and animals.
The Loglanist
the technical journal edited by pc and published by The Loglan Institute, Inc. from 1975 to 1984.
transformational grammar
see Chomsky
a word or phrase which is used in calling to gain attention, or in direct address
A computer software program written by Bell Labs (AT&T) that is used to generate parsing programs for computerized text. It also serves to verify that the syntax rules specified for these programs are unambiguous according to 'LALR(1)' rules.
Zipf's law
a hypothesis covering many fields, that observes that complexity is inversely proportional to frequency of occurrence. See page 21.