Copyright 1989, The Logical Language Group, Inc. 2904 Beau Lane Fairfax VA 22031-1303 USA 703-385-0273 Permission to copy granted for purposes of promotion of Lojban. Contact Bob LeChevalier at the above address or: lojbab@grebyn.com with any questions or comments. To all readers: Since there has been little interest in Loglan 1 and our review comments (2 requests after a total of 3 months), priorities demand that Athelstan and I not spend a lot of time on polishing this up. This, then, is the draft version of the longer review of L1, which we wrote before the short review that appeared in le lojbo karni #10. It contains more details, but unfortunately, not all the comments that Nora, Athelstan, and I took. To be succinct, there was hardly a page where we didn't find some error, contradiction, or simple confusion. In fact, we gave up taking detailed notes about halfway through Chapter 4. Much of the rest of Chapters 4 and 5 are relatively intact from the 1975 version, and thus is not really any better or worse than that older book. But the 1975 L1 had already been found wanting, so little more need be said. As an aid to understanding the problems, we recommend reading Chapters 2 and 6 together. The contradictions between them will be much more apparent. On questions of grammar of course, we are colored by our own experience in independently developing a grammar for Lojban. We also don't know exactly what Jim Brown's grammar is, since he is hiding it from us (certainly not to hurt us, since it is of little value to us). We have some clues from the Notebook 3 grammar. It took but a little time to find flaws in that version, either. For those with access to Jim's current LIP, we suggest attempting to parse the sentences in L1, particularly the ones in Chapters 4 and 5. A couple in passing seemed like that couldn't be grammatical by either his or our grammar, though we couldn't prove this. I didn't note the sentences with problems. The main problem with the grammar probably won't be solved with publication of L6, if it ever happens. The 'Preparser' contains many rules that are not LALR1, and are not encoded in a way that can be compared with the rest of the grammar. Jim's corpus approach to testing is inadequate. It verifies that selected grammatical sentences are parsed the way he expects. It does not test boundary conditions or supposedly ungrammatical sentences to see what happens. The last Parser I saw, broke so easily with test cases not in his corpus that it was useful only about half the time. Jim also begs several questions in L1 by labelling sentences as 'bad usage' that would parse. The label is meaningless from the standpoint of computer applications. If Jim cannot define the rules for determining 'good usage', he has not defined the grammar. The computer will not be able to tell 'good usage' from 'bad usage' without the rules. The Parser is worthless as a teaching tool without those rules included, since it will pass 'bad usage' sentences. If the 'bad usage' sentences are indeed problems, any attempt to do anything with them will cause the computer to choke in the Postparser processing, so it's worthless for computer application. To pc and the rest of us, the new Chapter 7 is the best thing Jim has written on Sapir-Whorf, but pc feels that this isn't saying much. Since pc is far better versed on the subject than the rest of us, I will rely on his judgement. If we can get pc to write a detailed critique, we will send it out to you or put it in JL. You can also write to him with specific questions. Otherwise, our notes in the review are probably enough to show its inadequacy. Considering how easily we found the flaws we did, I suspect that an expert would find many more problems. The final review includes a few things that aren't in here. We kept on finding more things every time we looked something up, till we finally gave up. A final note, based on pc comments to me by phone. Jim includes countless references to The Loglanist in his footnotes. However, in many cases, the referenced articles do not match the L1 text. pc, who of course was familiar with the TL material, was surprised at how many grammatical and usage points in TL that he thought had been accepted by Jim and the community were omitted from the new L1, or even contradicted. I suspect that the best way to check this is to read each of the TL articles that JCB references and compare it to L1. More importantly, compare the Supplement TL4/3, written by pc and approved by JCB, which contains things that definitely were added or changed, many of which are not in L1. I suspect that many of the sentence examples in the Supplement will also not pass Jim's LIP. Finally, check Jim's LIP against TL7. He probably has done this, but with what results is anyone's guess. A copy of this is going to Jim. If there ever is a reconciliation, we will no doubt offer to do a more detailed annotation of the problems so he can correct the book one last time. To others, I'm also including a copy of Paul Doudna's review comments. He touches on different issues than did we. Bob LeChevalier 19 November 1989 3 Book Review of Loglan 1 la lojbangirz. is perceived by Bob LeChevalier and Athelstan some, including Dr. Brown, to be operating in competition with The Bob: As those on the mailing Loglan Institute, Inc. We also, list of The Loglan Institute, Inc. as described below, are engaged in already know, a new edition of a trademark dispute with that Loglan 1, last published in 1975, organization. As such, our review has been released. This is Dr. might be suspected as biased or James Cooke Brown's description of self-interested. I'd like to his version of the language, of think that we have the which Lojban is another version. intellectual integrity to give a Since Dr. Brown was the original fair review, but will make the inventor and developer of the commitment to publish an opposing language, his ideas and opinions opinion next issue from Dr. Brown are especially of interest to or any other person who feels that those in the community. We feel this review is unfair. We also it is our vital obligation to seek to collect comments on L1, report such events as these, and both positive and negative, also to give a fair and honest ranging from editorial nits to review of these and other more extensive criticism, and de- Institute writings. You may then scribing both strengths and decide for yourself if they are weaknesses. We will assemble worthy of your time and interest. these and make them available to We're going to devote a bit of Dr. Brown gratis, and to other space to the review in this issue, interested parties for the price even though it might more properly of reproduction and mailing. We belong in JL than in a newsletter. are doing this because we believe This is primarily because a large ourselves to be more open to these core of loglanists/lojbanists are comments than is the Institute, sitting on the sidelines waiting and more willing to release both for the result of the perceived positive and negative comments to struggle between Dr. Brown and la the public. We also stand to lojbangirz. as to which language learn much from people's comments will 'win'. Most of these are at about L1, gaining insights which 'level 0' on our mailing list and will help improve our textbook and are similarly inactive in Dr. other teaching products. So Brown's organization. L1's please, let us know what you republication is a milestone in think. the project history. For those To help ensure that the review who view it as a competition, 'the reflects more than one person's battle has been joined' since we viewpoint, I have solicited com- have now both 'gone public again'. ments from Nora (who has long had For people like myself, however, a more positive attitude towards L1 is a chance to re-examine what the style of L1 than I have). I we are doing and why. la lojban- have also asked Athelstan to give girz. supporters have as much to his impressions as a skilled gain from L1 as the supporters of linguist and as a relative Dr. Brown, in that we have the new newcomer to the project. perspectives of the man who has Athelstan has come to know the spent more time thinking about the language as well as anyone in the Loglan project and Sapir-Whorf project besides Dr. Brown, Nora than any other person. Lojban and me. will profit from his insights, as This review is a quick look, will all those who learn it, just primarily based on first as much as the few who might impressions. However, before this choose to learn Dr. Brown's review is done, I will have read version of Loglan. Similarly, the entire book. Nora is reading those who are sitting back and more for comparison with the watching in 'level 0' may gain previous edition, and is going some new perspectives that will slower. For this review, Athel- help them make a decision. stan is reading selective portions at my request and other sections 4 of his choice (we only have 2 is not totally a flaw; the details copies to share among us), and is are sufficient to cause ones concentrating on ensuring that the intellect to be prodded into viewpoint we present is balanced questions and analysis, chal- and reflective of the interests of lenging the various ideas that are the community, rather than of presented, thereby potentially individuals. causing one to truly understand With this background, here is the concepts rather than merely our review: the language implementation of those concepts. Alas, this strength becomes the Before opening the cover of L1, book's main weakness. Put simply, we already knew what kind of the moment one digs beneath Dr. review we wanted to write. We Brown's surface presentation and wanted it to be positive, to challenges his ideas, one is acknowledge Dr. Brown's enormous confronted by countless errors and contribution to the language inconsistencies, failures of development, and to stress his analysis, and even the abandonment ability to present clear some of the project goals that in- exposition of the underpinning spired many of us when we first ideas of the language, while heard of Loglan. maintaining an intellectual level One doesn't need to be expert in that challenges and defends the the language in order to find Loglan project so as to increase these problems. Indeed, we who our academic respectability. know the language better can often By the time we had completed a figure out what Dr. Brown really brief skim of the book, means without serious difficulty. concentrating on new sections, as But a newcomer, especially one well as on areas of the language with any knowledge of languages, known to have changed signif- will quickly become confused and icantly since 1975, our hopes were dismayed by the shoals in Dr. dashed. Brown's thinking and presentation. Dr. Brown has indeed continued Even worse, in his forward, Dr. in that excellent style that Brown identifies dozens of people attracted many people to the who have contributed to the lan- language after reading the 1960 guage over almost 35 years, Scientific American article and/or including professional linguists the earlier editions of this book. and experts in other fields who he Possessing only a modicum of says advised him and "helped me intellectual curiosity, a reader find the errors in this is easily drawn into a world where multidisciplinary work". He fails new perspectives abound, where to give the traditional notice much that is unfamiliar to English that "all remaining errors are speakers (especially those who solely his responsibility", have never learned another thereby throwing some of the onus language) is made familiar. on these reviewers for missing He continues his approach of blatant linguistic errors and a explaining each idea separately, general sloppiness in verifying giving several examples of each the quality of his details. (Note point before moving on to the next - all quotations in italics are one. He has added pronunciation from the text; other quotations aids and many new examples, are our own stylistic expressions, rendering the book far more paraphrases, or hypothetical effective at teaching these translations.) concepts than it was before. Alas, Dr. Brown did not include The book still falls short in on his 'panel of reviewers' any teaching the language itself, competent loglanists, especially because it fails to cover the those known to challenge some of pragmatic concerns of how the Dr. Brown's viewpoints. As such, various rules and concepts many of those viewpoints are pre- presented interact with each other sented indefensibly - Dr. Brown in non-simple expressions. This simply doesn't acknowledge that 5 there might be other points of that "a pattern of relationships view on the issues he has decided, exists between the cultures of and hence leaves those ideas open certain peoples and the structures to question by each new reader or of the individual languages with loglanist who independently raises which those cultures were the same questions. associated", that "structural Of course, since most of these features of languages are seen by dissenting loglanists are the core Whorf as limiting the domain of of la lojbangirz., Dr. Brown used the possible for minds shaped by his 'trade secret' claims as a them" and (actually from American basis for avoiding the productive philosopher F. S. C. Northrup) "a dissent he needed. In a private facilitory one ... notational note to Nora a couple of weeks advances ... that replace other ago, he expressed regret that notations are facilitory", or in "Pity you [Nora] and Bob weren't other words, that a language with around to be among its critical a more efficient or effective way readers". We were right here and of expressing certain kinds of willing, Jim; it was you who chose ideas will cause its associated to exclude us. culture to be richer in the realms To truly see what is good and of those ideas. Dr. Brown cites bad about he book, let us examine the "enabling" effects of new it from the perspective of the mathematical notations, and the several goals Dr. Brown gives for relative poetic richness and the book and for the language. We "deeply esthetic orientation of will use his Introduction table of the Chinese civilization" due to contents as an outline: Chinese grammar and the "virtually limitless interplay of verbal "The Scientific Strategy" categories" found in Chinese. Dr. Brown then goes on to claim The main discussion of the that Loglan can be used to test scientific rationale for the lan- either of these formulations, and guage takes place in the outlines a fairly detailed introduction and in the new approach to such a test. Along Chapter 7 on "Testing the Sapir- the way, he points out that the Whorf Hypothesis". The introduc- nature of modern science is such tory section is lucid and presents that experiments must be designed the purpose of the language quite to disprove a theory or model as- well. In Chapter 7, he then goes suming that it is true, rather on to explore a variety of than to try to prove the model, formulations of the Sapir-Whorf and that scientific advance comes hypothesis as he perceives it, and when such experiments lead to discusses the implications of refinements of the model that these formulations for testing. allow it to more accurately In so doing he nicely avoids the describe reality. controversial aspects of Sapir- Dr. Brown also indicates that Whorf, including the politically the 'Whorfian effects' perceived and philosophically controversial by current loglanists are not sci- issue of 'linguistic relativity' entifically useful, although he that Ralph Dumain discussed in then goes on to use those effects JL7, as well as the various as a means of identifying true attempts to divide the hypothesis 'Whorfian effects' in a test into a strong and untenable form population. This is but the first (that due to different language sign of several that Dr. Brown's structures, people can never truly remarkably cogent presentation of understand the culture of another the hypothesis and the problems in language community) and a weak and testing it did not open his mind trite form (different languages to countless problems that he did give different perspectives, so not see. The result is that his learning a new one broadens your detailed experimental scheme is mind). fatally flawed. The formulations he chooses as A few brief examples are needed; best exemplifying Sapir-Whorf are this section is the major new 6 section of the book and covers an This is actually a good choice; it important part of the scientific is too likely that any attempts to justification for the language. measure development of logical First, Dr. Brown suggests that a thought over a short period will useful test of Sapir-Whorf can be be colored by the teaching conducted with second language methodology used and the fact of learners, specifically American actually teaching of logic, since university students (not the most most people are ignorant of logic, homogeneous of intellects or especially as it applies to their cultures) who study Loglan, linguistic expression. He French, and Chinese in a "Summer suggests that linguistic Workshop" 'total immersion' creativity be the basis for environment. He then contradicts measurement, citing the 'Whorfian himself by deciding that this effects' mentioned earlier. workshop should last 8 months for Unfortunately, he never addresses Chinese, 4 months for French, and the issue on a basis of cultural two months for Loglan (he does not development. The tests he propose any standard for setting proposes measure individual these times, like a pre-experiment creativity, and not cultural cre- to determine how long it takes to ativity. A far better test might reach a measurable level of be the comparative group solving proficiency using the proposed of problems requiring significant teaching methods; nor does he creativity and linguistic recognize that time taken to learn interaction in the target a language may be unrelated to language, and measuring either time needed to evoke 'Whorfian speed or quality of results. effects' - the difficulty of This example shows how Dr. Chinese is due more to its writing Brown's elucidation, while in system and tones, while the itself severely flawed, is useful grammar is relatively easy to in pointing to a better approach; learn - so 'Whorfian effects' the above alternative test was based on grammatical structure may composed 'at the computer' as this show up relatively sooner). is written, merely on rereading Also, the summers may be the relevant section. awesomely long in Gainesville (8 Another problem is Dr. Brown's months?), but even so, few univer- failure to take into account the sity students are going to be able bias imposed by the teachers of or willing to take 8 solid months these workshops. Recalling the of their educational program to difficulty in motivating study a single subject, even for disinterested students in learning pay, unless that subject was their a language, one cannot imagine primary interest - and he has trying to teach them Loglan specifically excluded choosing without engaging in the type of students on the basis of their word-play and linguistic interest in the language. (Money experimentation that the doesn't motivate everybody, Jim!) experiment is supposed to measure. Dr. Brown never really addresses This would be especially heigh- the question of holding a disin- tened if the teachers are today's terested subject's interest in loglanists, as Dr. Brown proposes; learning a language (probably the we all recall what inspired our toughest problem in foreign passionate interest in Loglan and language education), except to would be unable to teach it to hold out to the non-Loglan others without passing along that students the offer to teach them which we found inspiring. It is Loglan for free in a workshop the unlikely that Chinese and French following summer. Wow. teachers could be found that could Somewhat surprisingly, he emphasize the same mix of skills chooses to measure 'creativity' as and interests as the Loglan the area for detecting short-term teachers of this generation, and Sapir-Whorf effects, and not the Dr. Brown, while addressing logical thinking that is the several other biases, does not original basis for the language. appear to recognize this one in 7 his attempt to convince people of that would generate 'Whorfian the relative immediacy of a pos- effects' to be tested, such as sible Sapir-Whorf test. logical competence, it is unlikely Perhaps the strongest evidence that any such effects will be of Dr. Brown's inadequate analysis noticed. of the problem is his method of As an comparative example, how eliminating what he calls "host many English speakers can identify culture effects". He chooses the the implications of a restrictive three languages to be taught on vs. a non-restrictive relative the basis that French is a lot clause even a week after the like English, while Chinese is grammar test, even though that is different, in ways he presumes are a structure basic to English like Lojban's (He never says why grammar (as well as Loglan's). Lojban is more extreme than (One of our respondents indicates Chinese in ways that would that most college graduates don't increase linguistic creativity). even clearly remember the dif- He predicts that this comparison ference between nouns, verbs, and would cancel effects not due to adjectives, five years after the languages themselves, with college. This may indicate why succeedingly higher 'Whorfian many Americans can't write effects' in comparing French to cogently, and why these parts of Chinese to Loglan students. (He speech are blending together; doesn't explain why this test e.g., using 'impact' as a verb. couldn't be done with three Dr. Brown suggests, while natural languages first, if only discussing metaphor relations, to prove the methodology.) that English and European To eliminate host country languages may be evolving towards effects, he then suggests that this loglandic ideal.) French students be given the same Interestingly, Dr. Brown doesn't program, with the expected results mention any testing implications to be increasing from English to related to the 'metaphysical Chinese to Lojban. He also pro- parsimony' that is a basic design poses a similar test for Chinese principle of the language. What students, but never indicates the 'Whorfian effects' might be ex- results that would be expected pected by the optional tense (which are indeterminate with approach? Dr. Brown doesn't regard to his methodology, since discuss this, nor does he suggest English and French are supposedly a comparison with the Hopi equidistant from Chinese, while language and culture that led Loglan is an unknown difference Whorf to his hypothesis. from Chinese - presumably less, These reviewers believe that Dr. but skewed in a direction that Brown has made no case that would increase 'Whorfian effects' suggests that Sapir-Whorf testing as much the 'short distance' would is plausible without at least minimize them). second-generation bi-lingual For long-term effects, Dr. Brown speakers who can interact in a proposes a five-year period of significantly loglandic relatively low levels of moni- environment. In fact, the prob- toring and control. He indicates lems he foresees suggest that it some of the difficulties of is harder than originally monitoring such long-term studies, perceived to isolate variables but forgets the obvious one. Why that would invalidate test would the students, by design not results. However, by at least especially interested in Loglan to discussing the topic, after years begin with, be expected to of silence, Dr. Brown has made it continue to grow in fluency and possible for others to build on usage? Where is the 'cultural his work, perhaps solving the environment' wherein they would problems that seem to make a interact to grow in this manner? Sapir-Whorf test impractical in If their lifestyle and the near future. interactions with others do not Our conclusion is that Dr. Brown demand use of loglandic features has made a start towards a Sapir- 8 Whorf test strategy, but that his Resolver was embedded in approach is academically weak. He programming language code that had embarrasses his 'advisors', not been verified for unambiguity. including scholars noteworthy in (Lojban's 'resolver', more stan- the fields of linguistics, anthro- dardly called a 'lexer', is pology, and language education, by verified for unambiguity as part suggesting their presumed approval of grammar development, although of this approach. It is little we then have to use programming wonder that the NSF proposals in language to implement parts of it, the mid-70s did not impress the then test thoroughly to ensure linguistics experts that reviewed that the programming is correctly them, who were possibly more com- implemented.) mitted to a thorough peer review We can also make guesses based of the program than was Dr. on what Dr. Brown says is not Brown's team. there, and by what he does not discuss. He specifically says "Loglan as a Logical Language" that he does not handle other logics besides elementary There are two major senses predicate logic, a distinct wherein Loglan is described as failing in this era of 'fuzzy being logical. The first is the logic'. He mentions almost no incorporation of the trappings of mechanism for metalinguistic predicate logic, including truth- references beyond a couple of functional connectives and logical 'pronouns' and indirect reference scope. The second is the with "lae", a limited set of unambiguity of the grammar. Dr. discursive operators with no Brown has modified his claim in capability for expressing scope the latter sense, now claiming other than sentence and word level only that the language is scope, and in the introduction "conflict-free", though he never indicates that there is no me- defines this term. In computer chanism for expressing and resolv- parser development, the term ing paradoxes. (Lojban has suggests the absence of detected greatly expanded all of these ar- 'conflicts' by the YACC compiler eas from Dr. Brown's last public tool. design.) Dr. Brown has changed very Dr. Brown has removed the little of his discussion of Appendix that showed the logical structures in the language derivation of the logical from the 1975 edition, which is connection words used in Loglan, surprising since many of the major making the book more difficult for changes in the language related to one not versed in logic. the question of clearly defining Unfortunately, he didn't remove logical scope. More likely, he the reference to the Appendix in has deferred the question into the the Chapter 3-35 footnote. promised Loglan 6, which he has promised elsewhere is likely to "Loglan as a Laboratory take at least a year to develop, Instrument" and then only if enough income is generated from selling L1. Dr. In the discussion in this Brown seems to be holding his section, Dr. Brown refers to language design hostage for the Loglan's "small size", its ransom needed to publish it. "metaphysical parsimony", and its In the meantime, his grammar "cultural neutrality". We can simply cannot be analyzed for examine each of these. whether it handles logic usefully. His size claim seems to be based Those of us who were familiar with only on the 'parser grammar' and the grammar before he started not including the 'resolver treating it as a secret can make grammar', unless he has guesses. When it was last public, drastically shrunk the former Dr. Brown's grammar was not since Notebook 3. Lojban's 'lexer verified as unambiguous in its en- grammar' is roughly half of the tirety - what he calls the entire grammar, suggesting that 9 Loglan and Lojban are roughly metaphor around: "da matma go comparable in complexity. In any gudbi de" which 'must' mean "X is case, it is safe to say that Log- better-as-a-mother than Y". But, lan continues to be much smaller he forgets the place structure of than natural languages. "gudbi", from the appendices as His claims of metaphysical "is better than .. for ..". Using parsimony are realistic, although the logical interpretation of this he occasionally only pays lip place structure (a purpose), he is service to the concept, and saying that you have no frequently violates it in his flexibility in what you can fill examples. into this place - it must be For example, a long-standing "being a mother". This might be criticism is his treatment of acceptable in setting the place colors as comparatives. He goes structure of a complex predicate from "X is bluer than Y" to "X is word, bu t not for the deriving bluer than" to "X is blue", noting metaphor, which always implicitly that Loglan makes the elliptical has these extra places left open. omission plain and "invites Of course, he contradicts his completion". He makes the standards in a footnote (3-23) in assumption that in the incomplete which he states that there are sentence the speaker has omitted a many "modification pathways" for standard of blueness, which is metaphors. But even given this, itself a metaphysical assumption, Dr. Brown goes a bit far in some and one contrary to linguistic of his defining metaphors for com- evidence, which suggests that plexes. Basing "kill" on "dead- colors are decided not by being make" or "morsi madzo" is an exam- 'more than' a standard, but by ple; he clearly is using the being closer to a blue standard English association of "make" with than to a 'not-blue' standard. "cause", an unwarranted A better example is Loglan's assumption. To people of other supposed tense-optional system. cultures who see this metaphor and After making it apparent that a go back to the place structure speaker does not need to express definitions of these two to figure tense if it is not important, Dr. the meaning, "morsi madzo" might Brown proceeds to make it mean "dead-for-a-maker (of ... important by deciding that the from ...)", or even thinking time-free predicate only conveys liberally, "maker of dead things the weak sense of potential. Thus (from ...)", but not likely "kill" he says that "da cabro" means that which might be "dead become "X is flammable"; to make the cause". claim that X is burning requires Which brings us to cultural the present tense, and his tense neutrality, perhaps Dr. Brown's optional system goes away. most assertively-stated goal for Carried to logical conclusions, the language in the last 30 years. the flaws in this thinking are Sadly, almost all of the changes multitudinous; this means that his that Dr. Brown has added in the imperatives, being time-free, are 1980's are not culturally neutral, only commands to have the and usually severely violate this potential. "ditca", his example goal. This is in addition to the imperative in Chapter 5, should culturally biased metaphors that mean 'Be potentially a teacher are embedded in many of his older (teacher-able)', and not "Teach!". complexes. Similarly, his metaphor-system The examples are numerous. has hidden metaphysical Adding 'h' as a consonant is not assumptions. His example in 3.11 in itself entirely bad, but it se- and 3.13 of "da gudbi matma", is verely complicates cultural translated as "x is a good neutrality compounded when Dr. mother", with the requirement that Brown chose to use 'h' as the it always be interpreted as "x is primary means of creating good-at-being-a-mother", and "not consonant clusters in scientific necessarily good cooks or good borrowings. Few English speakers, wives". He then turns the and probably no French speakers, 10 will clearly pronounce the 'h' or resolve some problems in word always hear it in "athomi", Dr. forms, thus often causing one of Brown's word for "atom". The the three diphthongs with final Hindi will of course hear the 'th' 'i'. He then uses this biased as a single sound, an aspirated sample of borrowings to reverse a 't'; Americans don't distinguish decision to group vowels in pairs between aspirated and unaspirated from the left, now requiring the 't's, and are likely to aspirate (he admits) counter-intuitive them in places that will cause a right grouping that causes the Hindi speaker to presume a 'th' hypothetical name consonant cluster. "Loioioioioioioios" to be read as In choosing his permissible [loh-yoh-yoh...yohs] and not [loy- initials and medials listed in oy...oy-os]. Chapter 2, Dr. Brown used taste Incidentally, he also heightens tests performed entirely by the Latinate bias by using none English speakers. As such L1 has other than an English-Spanish voiced/unvoiced consonant pairs dictionary as the basis for permitted in medials (e.g. 'gt'), determining that a root is in spite of significant linguistic international. Not to denigrate evidence that one sound will cause the Spaniards, but German, French, the other to assimilate the others and Russian are generally con- voicing or non-voicing. A few sidered to be the other Americans thought they could hear international languages of science the difference in carefully besides English. One could controlled environments and in exclude the French because of careful speech, so Dr. Brown chose their policy against borrowing, to ignore linguistic theory. but German would have been a On the other hand, Dr. Brown better choice given a propensity omits a large number of for metaphorizing complex permissible initials that are no concepts, just as Loglan purports problem for Slavic speakers. Why to encourage. Better still would doesn't he permit the voiced be Russian, which has minimal ties equivalent of 'st', which is 'zd', to either Greek or Latin compared just as he allows 'zb' as the to the others. If a root has been voiced equivalent of 'sp'. They borrowed by Russian, it is more are equally distinct in a likely to be international. culturally neutral sense. Another possible choice would be Similarly, he omits the voiced Arabic, which is the true origin equivalents of the cX initials, of many words in mathematics, and and the unvoiced equivalent of is not Indo-European, although the 'zv', which is 'sf'. Interest- borrowing process would be more ingly, Lojban speakers are having difficult and less trustworthy due little trouble either hearing or to its relatively unfamiliar saying these unfamiliar-in-English alphabet and a heavily British and initial clusters, and the Slavic French influence in the 19th and speakers might have had higher 20th century, when most scientific word recognition scores if these borrowing took place. In any initials has been permitted. case, to decide that a word is Readers of JL6 will recall that international because its root is Ralph Dumain accused Dr. Brown of shared only by English and Spanish 'scientism'. While the definition is ridiculous. of that word as Ralph used it may A related bias is shown in his not be applicable, there is no limited set of sample doubt that Dr. Brown has strongly translations. Although the Table biased the current language to- of Contents implies 18 pages of wards a Greco-Latin scientific translations, there are but four base. He has developed a bor- partial articles from Scientific rowing policy based primarily on American, averaging 200-250 Loglan scientific words with Greco-Latin words apiece and a lengthy inter- roots, especially Linnean binomi- linear translation. Even this als. This borrowing policy adds overstates the volume, for each 'i' to the ends of words to text is only about 10 sentences 11 long and is heavily populated with is done with computer languages borrowings which are little more possessing a predicate grammar than an exercise in word remaking. like Loglan's. He mentions the One can get little sense of the subject of machine processing and flavor of the language from such a artificial intelligence a couple small sample of homogeneous and of times in his discussion within non-challenging text. By com- the book, but his examples seem parison, it took Athelstan, a like afterthoughts that try to relatively new Lojbanist, only a make the subject relevant, rather day to translate 400 words of more than carefully though out challenging literary text into applications. Lojban (part of the translation of An interesting point is that an a Saki short story that will be in artificial intelligence project of JL10), and Jim Carter translated the type Dr. Brown envisions is several longer texts of a variety going to have to deal with Loglan of styles into a version of Loglan as it is designed, as well as in the early 1980's, thereby Loglan as it is actually used in giving a much broader sense of speech. There is little evidence what worked and did not work in that Dr. Brown has thought much the language. about pragmatics of actual lan- As a comparison, we recently guage use. Except in a couple of examined an artificial language footnotes, he seems to presume proposal called Frater, devised by that all Loglan speakers will a Vietnamese in the 1950's. speak both grammatically and with Whether the language proposal was accurate pronunciation, a forlorn any good or not, the author demon- hope. strated its completeness and its wide potential by using it in "Loglan at the Man-Machine dozens of short translations from Interface" several cultures and in several styles. Dr. Brown either doesn't This concept is another long- know his language well enough to time dream of loglanists, one use it with versatility, hasn't which Dr. Brown pays lip service tested that versatility, and/or to, while totally demolishing it has ignored the potential with his changes in morphology and weaknesses that may exist in a sloppy pronunciation rules. language that hasn't been used in We've already mentioned his odd as wide a variety of ways as choices for permissible medial and possible. So in the end, Dr. initial consonant clusters. Not Brown has NOT proven that his mentioned is the fact that he 'small language' is a complete one either ignores of forgets them in nor a useful one. some names and borrowings (e.g. "svero" for "Swedish", an extra "Loglan in the Artificial sin on top of the fact that, if Intelligence Laboratory" 'sv' were permitted as an initial cluster, the actual Swedish This point refers to Loglan's "Sverige" is otherwise a perfectly potential as a model of a natural valid borrowing under his rules language that could be used to that should not have been study the human linguistic shortened.) process. Dr. Brown gives little We've also mentioned places evidence and states that there is where his pronunciation rules are no recorded body of spontaneous practically designed to cause Loglan speech that could be errors in speech, as in his 'h' processed (although he elsewhere inserts into borrowings. A long- refers to tapes of an older Loglan standing grievance is that his version). We agree with the goal, numerical digits, especially in a but Dr. Brown doesn't develop the noisy environment, have too little idea enough that we can do more redundancy. For example "fo" and than acknowledge his mention of "fe" differ from "vo" and "ve" it. He doesn't even mention that only in being voiced. Coupled much artificial language research with his rule that you cannot 12 pause inside of a long number, we be legal under his current rules. would challenge Dr. Brown to read His example in the Appendices a long financial report in such a suggests that "Xai-kre" and "Xai- way that a computer could under- kreni", and in 2.18, he says that stand all of the numbers within "mrenyclika" and "mrecli" are two unambiguously, much less forms of the same word, but his communicate with an air traffic discussion of "mekykiu" in 6.4 controller on a noisy radio link. suggests that "mekykiu" might be Of course, computer processing used for "eye-doctor", while a of speech requires that his less reduced form might be used algorithms be stated for easy pro- for "ophthalmologist". Note that gramming. Most of his speech, the two are NOT identical in grammar, and word formation usage; most Americans consider an algorithms are to be kept secret optometrist an 'eye-doctor' when until L6, and are not covered in they get their eyes examined, and the text, including his formal certainly an optometrist fits the statements of the definition of "kicmu": "Slinkui"/"Spektri" test (he uses "treats..for..with..". both names at different places in As an example of his sloppiness the book - only an old-timer will in algorithm description, we point recognize that they are the same out that his almost incompre- thing) and the "tosmabru" test. hensible footnote description It is known that his rules stated defining the structure of in earlier texts, up until and predicate words does not appear to including Notebook 3, are allow for 'y' hyphens. This erroneous; we detected the error hardly matters, since one of his for Lojban only a few months ago. main examples of a complex He also has had a history of not violates his defined rules for testing his grammars properly. He what constitutes a predicate word. has used a corpus consisting of It is unlikely that he can explain simple, often ideal, sentences, how "mekykiu" follows the rule that seldom involve rules that that "Predicates must contain at might conflict. He has never least one pair of adjacent unlike tested his 'Resolver' algorithm, consonants, a CC.", as stated in nor published it. His PreParser Section 2.16. We have no doubt approximation of a Resolver does that there is a valid rule that not de-compound little words can be stated describing the before determining their grammar, formation of predicate words, but and that portion of the grammar in Dr. Brown hasn't stated it. As the programming code, as opposed such, new loglanists and computer to the YACC grammar, has never programmers will be at a loss to been rigorously tested. More figure out his language. seriously, he has never given Another error is in his rules evidence that he has used his for names. Dr. Brown credits grammars in reverse, to generate Robert A. McIvor with successfully possible sentences that he might arguing that "lapla's" is a valid not want to be permissible. name form for "LaPlace". This Without doing, so, even Dr. Brown does a disservice to McIvor; he does not truly know his language gets the blame, when Dr. Brown grammar; without doing so, a should have taken only seconds to computer will never be able to devise the counterexample that process Loglan. renders this unworkable, if he had It is also unclear whether truly understood his own different forms of a complex are algorithms. Just how should permitted at a speaker's option Pierre LaPlace be described? How 'in noisy environments' or when about "la pi,e'r, lapla's, mrenu", talking to a computer. At one which every computer will point he mentions "matmymatma" as recognize as two arguments before the complex for "mother-mother", the predicate. Or if you are or maternal grandmother. Earlier talking to someone else named publications used "matmaa" for the Pierre, you get "Pi,e'r, lapla's, same concept, which should still mrenu", which might or might not 13 be interpreted as the command "Be diphthong 'ei' as "eigh-ee" or (potentially?) a man, Pierre 'ay'. Bob has heard Dr. Brown Laplace." as well as telling himself pronounce these two - they Pierre that "LaPlace is are indistinguishable. So what is (potentially?) a man". the pronunciation of the three Dr. Brown has abandoned the 1960 vowel 'ei,i'; how is "ay-ee" Scientific American claim to different from "eigh-ee", such audio-visual isomorphism. In this that the computer will know to book, he first uses doubled produce the extra 'i'? Similarly, letters to stand for vocalic 'eo' can easily be mistaken for pronunciations of 'l', 'm', 'n', 'ei,o'; he even says in the text and 'r'. Not only does this that this is pronounced like the violate his consonant cluster sound in "mayo". rule, he makes this use mandatory in borrowings, optional in names, "Loglan as a Translation Medium" and never mentions it for the vocalic hyphen 'r' and 'n' found The main requirement for Loglan in complexes. to be used as a translation medium A final problem is his is an unambiguous grammar, which definition of vowel pronunciation. cannot be evaluated until L6 is At one point, he allows the written. Otherwise the ideas are British sound of 'uh', identical reasonable, although a bit less to his 'y' hyphen AND his buffer clear than Patrick Juola's dis- as an alternative to the other two cussion in JL8. suggested forms of 'a'. The Interestingly, while Dr. Brown possibilities for confusing a gives copious credit to other computer with such speech are end- people in his forward, footnotes less, not to mention humans. To and bibliographies, including both even suggest that the pronuncia- Nora and Bob, he fails to mention tion is permitted encourages that Nora had a working partial people to devise all sorts of language translator on a TRS-80 mutually unintelligible dialects back in 1981 that could handle of Loglan (but then we already most of the basic grammar and all have two). of the 1975 dictionary, and More serious, since he at least produce quite understandable, if points out the problem with 'uh' stilted, English. This feat is pronunciation, is his treatment of more noteworthy than all of his 'e' and 'o'. He flat out words in demonstrating the contradicts himself every time he feasibility of the concept. describes 'o', including his pronunciation guide at the "Loglan in Information Storage and beginning. Before 'i', he says it Retrieval" is pronounced [aw], but the diphthong 'oi' is pronounced [oy]. This topic rates only a paragraph, He ignores the fact that half of and is a subcategory of artificial all Americans recognize no language research, and a combi- phonetic difference between the nation of that with man-machine sound of 'aw' in 'law' and the interface, and with machine sound of 'a' in 'father'. This is translation. The same criteria as much a teaching problem as relate to all of the topics, and anything else; these people will the same potential problems with make the sound he wants when Loglan. The main difference is pronouncing 'or'; they just don't that Dr. Brown suggests that this make that sound in 'law'. application would be based on Most Americans pronounce the 'o' written text, whereas his earlier in 'note' as his 'ou', but he discussion focuses more on speech. fails to note this problem. The He seems to miss the point that computer won't. this application is merely a Finally, his 'e' pronunciation variation on machine translation. is ambiguous. He says that before If a machine uses Loglan as an vowels that it is to be pronounced interlingua for translation, it is "eigh" as in "eight", thus in the in effect storing the text in 14 Loglan for information retrieval. outrageous comparisons. In the Existing AI techniques already can second footnote 1-27 from Section search a body of data that is 1.10, he has the audacity to already stored in predicate form; compare the halting Loglan speech the main problem is one of trans- of four people in his living room lation. in 1977-8, speech that he elsewhere acknowledges was so non- "Loglan as a Planetary Second fluent as to be dominated by Language" pauses while searching for words as opposed to difficulties in Dr. Brown does a complete understanding the grammar, to the reversal on Loglan as a 'world between 30,000 and 100,000 language' since earlier editions. speakers of Esperanto. The latter He considers the current design figures probably refer only to adequate for such a language, even true bilinguals who use Esperanto as he notes in various places that regularly; Esperantists claim 1 his version of Loglan has limita- million speakers in China alone, tions in expression. (He mentions and 5 million worldwide. Even if in this section that English has those numbers are inaccurate as to no "epistemic tense" like Hopi, fluent speakers, it is likely that but neither does his version of there are at least that many who Loglan. Lojban has a usable can speak Esperanto as well as the equivalent, although not as a four apprentices spoke Loglan. tense.) His main problem in this "Loglan as a Linguistic Toy" discussion is one of pomposity: "Everybody who has thought about This is probably the only aspect the matter seems to be convinced of the language which we find no that we're eventually going to fault with within this text. Log- need one on this planet". He lan, in all its versions an cites no references for this bald variations, is a marvelous statement (who is 'everybody'?), linguistic toy. Other aspects of and it is a most controversial the language cause people to one. The concept of a single investigate Loglan; only this as- world language offends at least pect has proven inspirational one prominent scholar (Professor enough to get people to learn the James Yorke, U of MD, Applied language. Physics - expert in chaos mathe- It is too bad that Dr. Brown matics) who has communicated with does not truly believe what he us over our much milder statements says here. We have long memories in the Lojban brochure; the sci- of Lognet, wherein Dr. Brown entific community does not neces- fervently wished that Jeff sarily share the idealism of the Prothero "would play in some other international language community, sandbox", and suggested that Bob, and this scholar thinks that instead of choosing to listen to Loglan is hurt by association with Dr. Brown's fervent request in the that movement. Among modern old Loglan 3 to "make Loglan your foreign policy specialists, own", instead "could have gone to Esperantists and others of the the movies". international language movement Ignoring this hypocrisy, the are considered relics of 'One- book is useful at encouraging Worldism', a defunct idealism. those who wish to play with Dr. presumably these people have spent Brown's version of the language, as much time as Dr. Brown has in and it even suggests ideas that thinking about international could be inspiring to Lojban relations issues. students as well. We will make Dr. Brown sidesteps the issue of liberal use of Dr. Brown's ideas, competing with Esperanto, in keeping with his hope on page suggesting that Loglan has a 15, the 4th paragraph of the 1969 different path to success. But he Forward, that people borrow his does see it as a competition, work "for unexpected which leads to one of his most 15 applications". I doubt if he had the end; the two never inter- us in mind, though. change. There is NO other relationship between 'h' and 'x'. "Learning Loglan" - 'x' is not the 'rough breath' Greek, 'h' is. The name of Greece In the final analysis, it is in itself uses this sound: 'Hellas', teaching the language that most and Dr. Brown recognizes this by will judge L1. Dr. Brown has building his borrowing for Greek definitely improved the book in as "helno". Russian has no 'rough this area since 1975, at least in breath' sound. The only true part terms of explanation and examples. of the statement is that x equates He has removed many of the to the Greek chi, which offers discussions in the 1975 edition little teaching aid to an already that prompted questions; removed confused and misinformed reader. them, not answered the questions. He also badly mis-spells the These areas range from the lengthy name of the ex-premier of Russia, appendix on "Pretty Little Girl's causing him to mis-Loglanize it as School" to the discussion of "no "xrustcyf". In Chapter 6, he nu" as a superlative (which he almost gets it right, giving the still uses in one of the SA correct form as "xructcyf", but he translations; he just never there gives "kructcyf" as a valid discusses it in the text). alternative, which would not have We've mentioned the happened if he had spelled the multitudinous problems with pro- name correctly as 'Khrushchev' nunciation above. In teaching he (the shch is a single consonant in has a few problems that are not Russian). The latter related to computer understanding, Loglanization has no validity but could cause tremendous except to English speakers who re- confusion to a reader. In one fuse to learn how to pronounce sentence, he makes so many 'x'. linguistic mis-statements that we If these problems weren't truly find it hard to believe that enough, he then uses such three linguists reviewed his work. pronunciation examples as to "x [kh] is the voiced version of encourage mislearning in 2.17: h; it is the "rough breath" of "SHLAWR-roh" for "cloro" doubling Greek and Russian, and the sound the 'r' and suggesting 'claro'. spelled with the letter chi in Then there is "FLOOR-roh", which Greek." he goes on to call it a most people will pronounce as "distinct but dry sort of gargle 'floro' instead of "fluro". which will sound very Russian". Finally there is "FOOT-boh" for His fallacies, based on our "futbo". If he ever had any assumption that he means by 'x' chance of convincing Americans what we think he does: that this refers to 'international - 'x' is NOT a voiced 'h'; the football' or 'soccer', and not latter is called a pharyngeal American football, he blows it consonant, is found in Arabic, and here. (He doesn't indicate this is similar to the sound you make meaning in his Appendix when you gag; voiced 'h' is not definition, although it is found in Russian, Greek, German or mentioned in passing during the Scottish. A voiced 'x' is found text.) in Arabic; it is the sound spelled Which brings us to the as Gh or Q in the name "Moammar contradictions. Remembering that Ghadafi", and may approximate to a in "Learning Loglan", he says that gargle (which is a voiced sound to this book will have to stand alone most people and is not particu- until a dictionary is written, is larly Russian). Voiced 'x' also "nucleus" "nukle" per the Appendix is not found in the other or "nukli" per the Chapter 1-21 languages. Dr. Brown may be footnote? Better yet, is "tulpi" referring to 'x' as an allophone the primitive for "tulip", or is of 'h' in German. 'h' is found at "tulpe" the borrowing, both in the beginning of words, and the separate Appendices. For that remnant of 'x', the hard 'ch', at matter, is there a difference 16 between a primitive-form borrowing to small in volume, possibly to and a primitive borrowed from one distinguish it from "short". he language? There seems to be no doesn't make the same restriction clear pattern as to how words are on "groda". But what do we do categorized between the two with two dimensional areas, and sections. Is there supposed to how do we say a "small number"? be? We could go on, but the book The there are the place simply doesn't teach the language. structures. He gives three dif- It is merely another progress ferent versions for "ketpi", two report on the ever-changing versions differing over whether Institute version of the language. the 5th place is price or At the end of Chapter 6, he makes accommodation, and one (in a this clear: "In the end some footnote) solving the problem by change in the fundamental putting price in an optional 6th structure of the language will place. (Not to mention that these almost certainly be necessary. place structures totally excludes This may seem unfortunate; but the the more natural and older usage alternative policy of freezing its of ticket as a license for early forms could well mean its admission to anything in general, early death". With his policy, which would make a travel-ticket a however, who has motivation to good complex justifying the learn the language? complicated place structure.) (He makes a historical error For someone who is trying to here - or perhaps is trying to promote Loglan as a predicate rewrite history - suggesting that language, Dr. Brown pays very the objections to Notebook 2 were little attention to the from people who didn't want to determination of place structures. give up the old forms. One objec- In the old Loglan 2, he begged the tion was to the vowel-rich short question, stating that the places forms like "matmaa", as opposed to were the obvious or natural ones the "matmymatma" longer forms - that clearly identified the expressed by Robert McIvor himself concept. These were so obvious - notice which form is in the that several place structures Appendix. Other objections were change every time he publishes. to the fact that the list In this new, improved text, he continued to change even after gives examples ad nauseam of bor- Notebook 2 was published. Of rowing Linnean binomials, probably course the ultimate factors were not the most common thing someone the questioning of Dr. Brown's will ever do, but he never gives scientific integrity in cooking any explanation of how to the data from the taste tests, and determine their place structures, the sheer political chutzpah that vital if he wants people to use the Board of Directors of The the language as a predicate Loglan Institute, Inc. had in say- language. ing that it, and not Dr. Brown, He also doesn't make clear when was the final arbiter of the one should borrow and when one language. We also know who won should make a complex. His bor- that one. rowing list includes But there was little rebellion hydrotherapist and heliotherapist, over the remaking of the words, both with place structures just as there was nothing but identical to that of "kicmu". So encouragement when we again remade why not "?cutkiu" and "?solrykiu"? the words, this time all of them, He doesn't say, but does talk in recreating Lojban. Nor have we about "Institute policy" that non- met with complaints about the two science words are generally to be grammar changes that have affected made as complexes while scientific the DC Lojban class while it has words are to be made as continued.) borrowings. To conclude this section, we ask Not only do place structures "Is the 4th edition of Institute change from time to time, so do Loglan learnable?" With the meanings. "cmalo" is now limited copious errors, inconsistencies, 17 and flaws in the new edition, not - Major Editorial Flaws - such to mention the changes that will as his inconsistencies in Loglan come about when Dr. Brown writes words and place structures, the L6, will we wait 14 more years for complete lack of an index (vital the 5th edition, or just a couple. either in a teaching book or a scientific description), and the Summary omission of several bibliographic entries (e.g., many listed in the Dr. Brown is guilty of a whole various Chapter 1 footnotes) that range of sins and errors that will make it difficult to check his make the new edition of L1 an references. embarrassment to the Loglan - Insufficient clarity which may community. After all, he lists suggest errors of scientific fact, half of us as credits for the such as his pronunciation guides. development, and suggests the - Errors of Omission - His incompetence of the several quali- omission of controversial fied scholars who read the text questions and their answers. and helped Dr. Brown "find the - Errors of Scientific Fact and errors". Yet the responsibility Principle - such as his "voiced for the errors must lie with Dr. h", mixed voiced/unvoiced Brown, and not with us, nor with consonant clusters, and his the reviewers. He is the only one comparison of Loglan speech to who really knows his version of Esperanto, errors which suggest the language, and he has made it that Dr. Brown needs to learn some that way on purpose. more linguistics before he writes There is a good deal of high about his language design. quality work in the book, and Dr. - Failure to Keep Current - He Brown is of course the originator cites, for example, Fillmore's of many of the most brilliant original paper on case grammar, ideas behind the language. We the giving no indication that he developers of Lojban will learn realizes that case theory has much from the book which will make significantly evolved since then. the eventual language even better. - Inconsistencies in his Rules But how about everyone else? and Practices - especially errors To academic scholars, the book in his algorithms. is a disappointment. It does not - Failures of Analysis - such as defend his ideas, merely states the implications of his theories them. The height of this behavior of potentiality of tense-less occurs when Dr. Brown states his usage, his use of "eigh" as a disagreement with Chomsky's trans- permissible pronunciation for 'e' formational grammar, thereby before vowels, his right grouping alienating most of American of vowels based on his skewed set linguists, giving as reason only of scientific borrowings, and his that he has not seen any support permitting of "la" in names. for Chomsky's position. Dr. Brown - Bombast - in his attitudes has not published this copious towards defending his academic research that suggests that scholarship, in his statements Chomsky is wrong; he just states about "Institute Policy", his 1000 it. This will convince no words of scientific text serving linguists who support Chomsky's as a sample corpus, and the secret views. grammar that he is keeping as a Dr. Brown has committed errors trade secret. at every level of his language - Major Conceptual Errors - design and his writing. Let us including his use of 'h' to solve list them, exemplified primarily borrowing problems, his use of a by problems described above: Latinate language as a source for - Editorial nits - such as his identifying International duplicate footnote numbers in Scientific Vocabulary roots, and Chapter 1 and his type-face that his inadequate analysis of the makes it almost impossible to find Sapir-Whorf testing problem - es- the footnotes in the text. pecially in seeing it as a short term phenomena and one detectable 18 primarily in individuals instead of cultures, leading finally to: - Failure to Live up to the Loglan Project Goals - by abandoning audiovisual isomorphism, incorporating several European and scientific cultural biases, and teaching that word- making is only a morphological problem, and not a question of place structures, in a predicate language. Dr. Brown has put a lot of hard work into this book and into the language. It bears the signs of a rush to publish, probably resulting from his perceived competition with la lojbangirz. This is sad. Dr. Brown's good ideas are lost amidst the confusion. Athelstan summarizes: "Imagine the Mona Lisa: beautiful hair, eyes and enigmatic smile - with a crooked nose and blackened teeth, and structures in the background that are out of proportion. It wouldn't be a masterpiece; a scholar would recognize the promise, but the errors would be to glaring to accept the painting as a work of art. In the same way, the L1 errors are too great for the scope he intends. If he had written a small book that didn't intend such comprehensive coverage, some of these could be ignored. But in a work of this scope, they can't be. If an artist can't, or won't, learn perspective, no version of the painting will ever be a great work of art." Bob adds: "Dr. Brown clearly intended this work to be his Magnum Opus. Alas he achieved only a big work."