Response to Steve Rice

[lojbab (Bob LeChevalier)]

In response to Steve Rice.

John Cowan answered many of your points, but we talked a bit about the
comments at LogFest and added some more ideas.

>Anyway, to give you some idea how petty and obnoxious "Institute Log6)n"
>looks to a logli, for the remainder of this message I will refer to
>"Group Lojban."  Again, I do this not to be testy, but to give you a
>mirror in which to see yourselves.  If you feel that it is justifiable
>to take over from TLI a term which it originated, used as a special term
>unchallenged for some thirty years, and has never ceased to use in that
>fashion in copyrighted works, your ethical system and mine are too far
>apart for us to even discuss matters.

Your comments betray your ignorance of the situation.  I have tried to
provide you with a) the information that is the basis of our ethical
position, and b) the justification of that position based on that
information.  I have also offered to put you in touch with unbiased
third parties that have attempted to mediate the 'ethical' dispute.  You
have labelled my attempts by asserting that you have "less than zero
tolerance for gossip".  This sounds quite testy to me.  Most people
recognize that ethical situations are generally cloaked in shades of
gray, and good judgement requires knowledge of the full circumstances.
If you would pass ethical judgement, you might want to iknow about what
you judge.

I offer two relevant analogies, with trivial changes in your words:

   If you feel that it is justifiable to take over from the British a term
   ("English") which they originated, used as a special term unchallenged for
   some several centuries, and has never ceased to use in that fashion in
   copyrighted works, your ethical system and mine are too far apart for us
   to even discuss matters.

   If you feel that it is justifiable to take over from Chomsky terms ("
   transformational grammar" or "Government and Binding" or "Universal
   Grammar" ) which he originated, used as a special term unchallenged for
   some several centuries, and has never ceased to use in that fashion in
   copyrighted works, your ethical system and mine are too far apart for us
   to even discuss matters.

Clearly your ethical claim leads to nonsense when applied to anything
linguistic other then Loglan.  Why is the term 'Loglan' so special to
demand unique ethics?

Note that as a language name, we assert that Loglan, like English,
refers to a variety of related languages extending both synchronically
and diachronically which are closely related in evolution and structure,
not all of which are fully mutually intelligible:  a family tree of
languages.

In any case, copyright does not protect words or phrases but complete
expressions of thought.  Only trademark law, and then in a very restricted
sense, provides protection for the association of a particular expression
with the expressor's intended meaning.

We use the term "Institute Loglan" to avoid confusion among the audience
of the several meanings of Loglan.  In addition to the language as named
by Heinlein in copyrighted works that do not mention Brown or TLI (or
give any clue that the referent was the same - Heinlein's language is
after all a computer language which does not have the expressive power
of English) and the several versions of Loglan that are part of the
Loglan/Lojban family tree, there is also an independently derived
object-oriented programming language named Loglan developed by Polish
researchers and used in their own copyrighted works.

We at LLG have occasion to talk about Institute Loglan, 1975 Loglan, The
Loglan Project (defined by JCB as the development and use of Loglan to
test the SWH, which we are still working on too), Polish Loglan,
Heinlein's Loglan, Carter's 1983-4 Loglan (also called 'Nalgol') because
according to pc he got everything backwards.  (But even you recognized
Nalgol as 'an archaic version of Loglan', according to Mark Shoulson of
the Planned Languages Server.)  And of course Loglan/Lojban.  Why is
choosing to be specific and non-confusing in our references so
'offensive' to "Loglanists" like yourself?  Is unambiguity un-Loglandic
all of a sudden?

The phrase "Group Lojban" is not particularly offensive, though
obviously a bit odd.  Since I know of know other thing or concept called
"Lojban", the qualifier "Group" has no restrictive value (and in
tanru/metaphor you will surely agree that some sort of restrictive
effect from the modifier is necessary to justify a tanru).  Note also
that it is common practice for TLI to use "Institute" as an adjective
for things pertaining to it like "Institute policy".  Why is it rude to
use the names people/groups choose for themselves?  I have never heard
anyone talk about anything specific to us as "Group ..." until you.  If
ever such a restrictive term is REALLY needed, I would hope you would be
respectful about it and call the version "la lojbangirz Lojban".  We
will happily use any name the Institute chooses for its version to
uniquely distinguish it from other versions of Loglan (Hmmm.  Would you
prefer "TLI Loglan"?)  "Loglan" does not.

>Now, if you wish to
>entertain the notion that "Loglan" is a generic term, fine; though it is
>ambiguous and quite unnecessary.  The generic term "predlang" already
>exists, and is more precise.  After all, there have been several
>"logical languages" in the past few centuries.  The distinctive feature
>of Loglan and Lojban is use of predicates.

I can say lots of things about Loglan, as the general term encompassing
all of these meanings except the Polish language and possibly
Heinlein's.

By contrast, what could I say useful about your generic concept of
"predlan".  Only that all "predlans" have predicates.  So what?  And
since 99% of our audience has no interest in any predicate language
other than Loglan/Lojban and its linguistic heritage in other languages
named Loglan, that is what we talk about.  Since most of us don't know
much (or care much) about either guaspi or your unidentified language,
to use a term inclusive of them might lead to false or misleading
statements.

The fact that 'Loglan' is acronymic of "logical language" does not mean
that "Loglan" refers to all logical languages.  After all, "CIA" doesn't
refer to all central intelligence agencies - only one.

If you feel that the only distinctive feature which separates Lojban and
other versions of Loglan from other languages is the use of predicates,
you have tunnel vision.  The tanru "logical language" is much more
complete in describing those features than "predicate language".  The
fact that there are other interpretations should not bother you.  You
support the language version that considers "dead-make_from" a valid
tanru for "kill".

>I should also point out that, contrary to what you may have
>heard, the legal status of "Loglan" is as yet undecided.

The legal status was decided.  That decision is being appealed.
                     -------
Please look the term up in your legal dictionary if you are going to
use it in a legal context.

>Phonemics and transcribing the digraph: ... I still think mapping
>three English phonemes onto one Group Lojban phoneme is asking for
>trouble.
 ...
>On a similar note, it's unreasonable to expect an ordinary
>English-speaker to hear a /i/ in "later", as is apparently required for
>"balvi" to work.

Indeed.  We also map several Chinese phonemes into 'c', and we drop all
aspirations in Hindi, thus mapping at least two phonemes to one in every
case.  There are numerous systematic decisions that we made in making
the gismu/prims, some of which might be considered systematic errors.
Our mapping is undoubtedly less erroneous than JCB's original mapping
for Chinese.  Look in L4/L5 and you will find that his Chinese mappings
were usually based on Wade-Giles transliteration, which is not
phonemically correct, and he occasionally used Pinyin and Yale
transcriptions in word-making, leading to an unpredictable mishmash.

In the case of "the digraph" (which has a proper English term, as you
should know if you are studying Anglo-Saxon), it is phonetically between
Loglan 'a' and Loglan 'e'.  How to map it is certainly debatable.  It
happens that your assertion is not supported in Institute literature.  I
refer you to TL 2, page 42, which maps the sound to Loglan 'a'.  There
were two articles in TL2 by Chuck Barton (whose specialties are
phonology and language education) on mapping of sounds in all of the
Loglan source languages, and JCB has several times acknowledged Chuck's
skill and accuracy in making prims.

JCB mapped the sound to a dead phoneme with no score in his original
prim-making (reference Loglan 2).  We chose to use Chuck's mappings as an
accepted and superseding set of rules that was more widely known in the
Loglan community.  Our aim was to keep the language in the tradtion of
'Loglan' which we did.

In names, JCB has been inconsistent in mapping the digraph.  He has
written "Alice" as "la alis", not "la elis" (both offer minimal pair
confusions with other English names), and "Ann" is "la an", not "la en".
But this is besides the point.  Names are mapped as the namer, or the
named, wishes.  I note the comparable discussion for the name "Jim" on
page 90 of 4th edition Loglan 1, which states that the vowel, not found
in Loglan, can map to either 'e' or 'i', giving "djem" or "djim", as Jim
himself chooses.

Our primarily American Lojbanists have generally chosen to map the
digraph as 'a'.  For example, we use "la .atlstan", for Athelstan, whose
name is pure Anglo-Saxon, and who first letter is the epitome of the
digraph.  Athelstan, who is skilled in Anglo-Saxon, Lojbanized that
sound in his own name as "a" by intention so that it would NOT be
pronounced as "e".  I can generalize though that Loglanists usually
place as high or higher value on visual recognition of names as on aural
recognition.  Thus we have a "tomis" for "Tommy" and a "garis" for
"Gary", both used by polyglot professional translators who are more than
familiar with the IPA and phonemic analysis.  On the other hand, Paul
Francis O'Sullivan, proud of his Irish heritage, uses "polfranCIIS, and
a student used to longer, drawled vowels Lojbanized her name "Kim" as
"Ki,ym".

>In referring to problems in deriving primitive predicates, I relied on
>your own gismu list.  If "censa" neither contains nor resembles the
>English word "sacred", then the English word is worthless as a mnemonic.

John covered this, but I want to add that in all versions of Loglan, the
English keyword is not necessarily the mnemonic.  Institute Loglan (sic)
uses for example "goltu" for "throat", from "gullet", and "gomni" for
"adhere/stick to" from "gummy".  Our keywords were chosen to maximize
distinctness of meaning, thus we use "fly" for sfani, the insect and
"flight" for vofli" the English verb "fly".

But you chose a bad example - while "censa" has no score for either
sacred or holy in English, we chose sacred because it is used more
generally in discussing religions including non-Christian ones, but also
because of the two it offered some visual recognition in lieu of aural.

In any event, JCB recently said in Lognet that the engineering tests
demonstrating the relevance of his mnemonic algorithm were conducted
informally and the results were never published.  We have instrumented
the new version of LogFlash and intend to determine whether aural
mnemonics even have any relevance to word recognition in a way
consistent with JCB's word-making algorithm.  I suggest scientific
research before making unscientific claims.

>When I saw "baxso", I thought you might have taken it from "bahasa"--and
>almost died laughing.  Then I said to myself, "Nah, they wouldn't do
>THAT!  They must be using some other word."  I'm sorry I doubted you.

Almost all of the words various cultures use for themselves are derived
from the local language word for "people", "country/nation", or
"language".  Since the commonalty of Malay-Indonesian is primarily
linguisitic, not territorial or tribal, and this is reflected in the
words of their language, the Malay-Indonesian word "Bahasa" is far more
respectful of the local culture than any other choice.  I am more proud
of it than of "bindo", since "Indonesia" is from the European root
stock.

>Loglan is a real language,

It is?  You are the first 'linguist' to say so, then.

>so it and
>its grammar are found not on some computer but in the minds of
>Loglanists.
 ...
>grammatical in terms of The Grammar as known by speakers.  Otherwise,

Where are the native speakers qualified to make grammatical judgements?
In fact, where are the speakers period; JCB claims that there were a few
back in 1977 but not since.  And if JCB is qualified, then howcum 1977
Loglan and 1991 Institute Loglan have such different grammars?  For that
matter, JCB changed his human grammar so that the machine could parse it
unambiguously.  He just failed to finish the job.  We have.

>It's true that the computer sees "bi" as a type of predicate LEXEME,

The correct term is "grameme", one of many JCB errors in terminology,
and in this case, we only recently have discovered and corrected it.
Others include "primitive" (or are 'billiards' and 'football' truly
primitive concepts to you) and "metaphor".  He apparently assumed
"lexeme" because the YACC 'Preparser' is called by professional computer
scientists a "lexer" - it identifies individual words from the text
string.

> but
>that's not to say that "bi" is a predicate in speakers' minds.  It isn't
>in mine, for example.

And are you a fluent Loglan speaker, or even a minimally competent one?
Who do you speak it with, and do they understand you and use the same
grammar.

Our machine grammar is intended to unambiguously define the bounds of
grammatical Loglan/Lojban sentences.  It is vital that it do this if it
is to be used as a standard in teaching the language (as TLI does with
its LIP program, too).  We have just rebaselined the machine grammar as
a prescription for the language until we have a body of fluent speakers
that can serve as the basis for the model you speak.

>I would suggest that, while you may be in a position to pontificate
>about Group Lojban semantics, before doing so about Loglan, you should
>bother to learn the language.

Looking at the list of basic errors you have made by not knowing the
sacred writings of JCB, I suggest that you take your own advice.  After
all, YOU are the one on the 'Loglan Academy'.  I can't help it if Nora,
John Cowan, and myself all know Institute Loglan as well or better than
some of its official pontificators.

>In fact, based on this and other instances of pseudo-Loglan, I can see
>why Brown's bugged.  I thought it was a mere matter of having the
>Institute's work ripped off.

'The Institute' has done no work.  JCB did much of the design work.  So
did Jeff Prothero, John Parks-Clifford, and several dozen-to-hundreds of
other individuals, including myself (responsible for the 'y' hyphen -
see the second entry on page 597 of 4th edition Loglan 1, and the
reference itself.  The Institute could be considered to have attempted
to 'rip off' our work, since neither I nor most of the others intended
the Institute to assume 'ownership' of whatever intellectual property
rights might exist.  Most of those Loglanists who did the germinal work
on Loglan are now Lojbanists, or have dropped out completely in disgust
or protest; JCB and Jenny and McIvor are among the few exceptions who
stuck with TLI.  I could consider accusations like yours to be
slanderous.  Please desist.


Since JCB testified in the trademark action that he had never looked at
Lojban to determine whether it was 'logical', are you accusing him of
lying?

>But now I think rather that he's
>(justifiably) afraid that you're contaminating the experimental
>area--spitting in the test-tube, so to speak--by misrepresenting Loglan
>semantics and metaphysics.
 ...
>2. The tags create a duality in the language.  So what?  Optionality is
>the name of Loglan's game, and I rather like having the choice.  I think
>more languages have redundancy of this sort than anyone's aware of; if
>Group Lojban doesn't, my condolences.

You do not understand Loglan to make the latter claim, and if JCB agrees
with you, it is he who spits in the test tube.  Far more basic than
'optionality' in the design principles of Loglan is metaphysical
parsimony.  Reread the Scientific American article, Chapter 1 of Loglan
1, and almost every other thing JCB has written before you assert who
misrepresents 'Loglan metaphysics'.

Only on page 50 of 4th edition L1, do I know of a reference that suggests
your point of view.  In the first paragraph thereon, we find:

   "In these last ten years of engineering optionalities into Loglan, I
   have formed a view of language which I suspect is very different from
   the one with which I began."

It is we who are true to the traditional Loglan metaphysics.  JCB and
Institute Loglan are apparently not.

John Cowan mentioned the shoe-horning of inconsistent semantics into the
Institute Loglan case system.  Every assignment of an argument place
into a fixed case imposes an unnecessary metaphysical assumption that
the two places are semantically identical.  Thus English speakers make
metaphorical and metaphysical assumptions about objects that are
labelled with the preposition "to", although some are destinations, and
some are beneficiaries.  The >use< of the case system in Institute
Loglan is optional, but its very existence causes unnecessary patterns
among predicates that are supposed to be logically independent and
unclassified in a predicate language.  If I, a Loglanist, read and
understand a case tag to identify a similarity between two supposedly
unrelated predicates, I am inherently led to see a closer tie between those
two predicates than between others.

>1. Linguists can't agree on how many cases there are.  True from the
>standpoint of universal grammar, false from the standpoint of a given
>language (e.g., Loglan).  It's not too hard to figure out how many cases
>a language needs to account for its syntax.  The trick is coming up with
>a system which works for everybody.  Now, Loglan is A language, not ALL
>languages, so there's no problem.

Ah!  But in having 11 case tags such that each place of a given place
structure associates one-to-one with a case tag, Institute Loglan can
have no predicate with more than 11 places (probably not too bad a
constraint), but also no predicate such that two different places need
the same case tag.  Hence "sau" for example means that no predicate can
have both a source place and a reason place.  "beu" means that none can
have two places that encompass the 'semantic field' of "patients",
"parts", and "properties" (whatever those three have in common other
than initial letter in English).  But you also prevent me from, for
example, seeing myself as a Loglanist/Lojbanist who "cirna (Lojban
cilre)" from seeing myself as a "doer", a "beneficiary", and a "part" of
the Loglan community.  And why must it be in the Loglan world that one
of the participants in "kitsa" be active and the other passive, and
which one is which?  Is Institute Loglan sexist?  Judge for yourself.
It includes the metaphor "mrenu durzo" (man-do) as the origin for "to
man a ship".

"It's not too hard" - but since you have no complete dictionary of all
valid compounds (and cannot by definition in Loglan), how can you
possibly conduct the semantic analysis.  Indeed, though additional work
may have been done since, the cases were done from analysis only of the
prims.  Indeed, the analysis was probably done by a native English
speaker, thereby ensuring that the semantics of the Institute Loglan words
will encode English semantics.  Certainly you can see that in some
cultures, "kitsa" has reflexive cases.

"The trick is coming up with a system for everyone" - indeed.  And since
JCB and Jenny did so with no speaker base, they were guilty of imposing
a specific world view on their Loglan, making it less likely to be
useful in a test of SWH.  And tell me, is it only the use of a computer
that causes you to see our use of a formal grammar to define the human
grammar as different from the formal preassigning cases to define the
semantics of places?

>I'm also informed that Group Lojban doesn't have the full spectrum of
>ethnic forms found in Loglan.  Using *lojb- as an example, we'd have
>
>         lojba     is a part of the Group Lojban language
>         lojbe     is an area associated with Group Lojbanists
>         lojbi     is a Group Lojbanist
>         lojbo     is a feature of Group Lojban culture
>         (No -u form exists as of now.)
>
>In Group Lojban, such concepts are handled with complexes ("lujvo" is
>the local shibboleth, I think).  Now, I can write (and even say, but not
>in a mail message)
>
>         lojbyleu       is a part of the Group Lojban language

or given the above 4 forms, a language used in an area associated with
Group Lojbanists, any language spoken by Group Lojbanists (obviously
this includes English more likely than Lojban), or a language associated with
the Group Lojban culture (definitely including English)

>         lojbysia       is an area associated with Group Lojbanists

or an area associated with the culture, or with the language

>         lojbypeu       is a Group Lojbanist

or a person associated with the area associated with Group Lojbanists, etc.

In all of the above, you cannot tell the source metaphor for the
compounds; the affixes are ambiguously resolve to use any of the four
'declensions', AND EACH RESULTS IN A DIFFERENT INTERPRETATION.
Therefore the whole purpose of GMR was violated by this kludge.

> and even lojbykau ("is a Group Lojban dog") if I wish.

(And the new animal declension is worse.  Can I use the animal
declension on 'logla' if I am referring to such a Loglandic dog?)

>In effect, you're telling
>me that as a Loglanist, I have more choices than you Group Lojbanists
>do.

No, IMHO you just have more sloppiness in your language.

------------------------------------------------------------------------

Now can we cease being mutually rude, and work to end dispute rather
than cause more?  I suggest that you study our language a bit more
before condemning it, and that you study your own version and the words
of JCB a bit more before challenging people who have been working with
Loglan a bit longer and also more extensively than you.  And finally, I
suggest you get both sides of the story before rendering summary ethical
judgement.

----
lojbab = Bob LeChevalier, President, The Logical Language Group, Inc.
         2904 Beau Lane, Fairfax VA 22031-1303 USA
         703-385-0273
         lojbab@snark.thyrsus.com

Note: I am also posting this to lojban-list