Message-Id: <199310300813.AA22645@eli.CS.YALE.EDU>
Date:         Sat, 30 Oct 1993 04:11:08 -0400
Reply-To: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
To: Erik Rauch <rauch>
Status: RO
Subject: Lojban

UC> This holds if the deep structure is based on some kind of predicate
UC> logic (maybe the Chomsky kind) -- which, at least in my case, is true.
UC>
UC> >What this means for linguistic theories if Lojban develops 'native
UC> >speakers' i.e. children who can learn the language through natural
UC> >processes, is to me an interesting question.

In Lojban, we have the privelege of DEFINING our deep structure.
Whether this deep structure has anything to do with human internal deep
structure (assuming such really exists) is always open to question and
if the two are too different, the language will either not be learnable,
or will be learned by a native learner is some way different from our
predicted deep structure; i.e. there will be noticeable differences in
surface structure unless by some chance there is an isomorphism or
direct transform between our designed deep structure and the human deep
structure.

UC>  Lojban, in comparison, takes more
UC> mandatory arguments than most (natural) languages -- which in turn
UC> makes usage of case tags less often.  I think case tags in Lojban are
UC> like prepositions in English?

Lojban has predicates so far with between 1 and 5 arguments.  All of
them are 'mandatory' in that they are presumed to have some value in all
sentences, but all of them are optional, i.e. they may be ellipsized if
obvious from context, or irrelevant.

The Lojban case tags are very much like English prepositions except that
each tag is very narrowly defined in terms of semantics.  The set of
tags is theoretically open-ended, in that any predicate can be used as a
case tag - thus resulting in a subordinate clause of an odd sort.

UC> So when we say "the sheriff of Nottingham", we mean "that which is
UC> the sheriff of Nottingham".  ("is the sheriff of" should be a single
UC> word in Lojban, I guess?)

Close "the thing that sheriffs Nottingham feels better to me, since it
is always safer to look at Lojban predicates as verbs without even an
implied copula "is".  The verb "sheriffs" could be a single word in
Lojban (no one has suggested on, yet - we haven't talked much about law
enforcement %^)

UC> I wonder if there is still
UC> something corresponding to the English "of" -- or do we have to say
UC> "that which is-the-A-of B"?  If it's the latter case, then "the
UC> sheriff" would become "that which is-the-sheriff-of zo'e".

There is no exact equivalence of "of".  The words be/bei are just
grammatical separators to attach any arguments onto the'sheriff'
argument.  If you actually did do the sentence as a relative clause,
then there is an optional relative pronoun that wouldn't likely be used.
"zo'e" is seldom expressed except when a place is omitted, and other
methods tend to be used more for that situation than zo'e, too.  There
is an indefinite case tag which can only be translated as "of", and "of"
is also the proper case tag/preposition to use when you stick an
argument on with no case tag that is beyond the defined number for a
predicate (the 3rd argument of a 2 place predicate, for example).

Other than this, the implied link that is represented by "of" is not
lexically represented in Lojban:

la djan. pulji la NATinxem.
John     polices within-the-organization-labelled-Nottingham.
(straining the definition of 'pulji' just a little)
[enforcing laws ...(unspecified).]

UC> I'm sorry about my ignorance in linguistics, but what does
UC> "appositive" and "relative clause/phrase" mean?

Appositive, is when two alternative ways of describing the same argument
are opposed:  John, the sheriff, came to town.  (where John IS the
sheriff) Relative clause is similar, but a sentence/clause is used to
define the argument.  Usually explicitly or implicitly marked with a
relative marker, (in English most commonly "which", "that", or "who").
A relative phrase is halfway in between, having a second argument which
is related to the first argument linked to the first one:

relative clause: John, who rode a white horse, came to town.
relative phrase: John, more-than Jim, likes me. (where "more-than" in Lojban
is a preposition: la djan ne semau la djim. cu nelci mi

UC> As long as exchanging the places of arguments (both mandatory
UC> and optional) is allowed, this special role of the first argument
UC> makes no difference.

It is allowed, but fairly highly 'marked'; i.e. unusual and wordy unless
you have a really good reason.

UC> Yes... Actually I suppose that Lojban can even be SOVO, in the case
UC> of >=3 arguments.

Yes.

UC> What is the difference between the word-making processes
UC> of Loglan and Lojban, which made the two vocabularies different?

We used newer population data, and the much higher population of China,
and the greater literacy among the entire population in Mandarin since
1950 enormously increased the relative weight of Chinese.  Similarly for
Hindi in India, whereas the number of native speakers of English may
actually be declining since the British Empire collapsed (2nd language
speakers are not weighted fully, and in any case, a 5% increase in
native speakers of Chinese can outweight a 15% increase in English
speakers because the raw number base is so much larger.  Using newer
data meant that we dropped German, French, and Japanese out of the
word-making entirely, and added in Arabic.  If we had gone to 7 or even
8 or 9 languages (Loglan originally used 8 and we used 6), the next 3
languages nowadays would be Malay-Indonesian, Portuguese, and Bengali.

We also used a good Pinyin dictionary with official Chinese
interpretations for the IPA values of each letter.  Jim Brown used a
hodgepodge that was usually Wade-Giles, and presumed that the
romanization was accurate as to pronunciation, from what I've been able
to tell.  He also either didn;t have a good dictionary, or failed to
check himself, because there are a lot of strange word choices fro his
Chinese (I say this since I did the Chinese dictionary work for Lojban
myself.  I don't know Chinese, and therefore checked doubkly hard to try
to get words that meant the right thing - oftentimes the Loglan choice
backtranslates into a less-common denotation of the English word that
has nothing to do with the intended meaning.  Of course, we probably
have some silly choices, too, and since I didn;t KNOW Chinese, I may
have chosen words that are the rarely used variant over a common word.
I personally am looking forward to someone evetually looking at what we
did and systematic- ally telling us how bad we really did, but I'm sure
it was better than the original Loglan work by a large margin.

UC> I do have several friends who are interested in Loglan/Lojban, and of
UC> course I will send them copies of the materials.  Translation into
UC> Chinese should be easy.  I don't know much about linguistics research,
UC> but I am surely looking forward to distributing information in both
UC> China and Taiwan.

You will have us at your beck and call if you can do this.  WE REALLY
WANT native Chinese speakers to learn Lojban, given the weight we put on
the language.  I'm cc.ing this message to John Cowan, our chief
grammarian, who has done some serious reading on Chinese, though he
doesn't really know the language (he did a translation of a fable from
hakka Chinese a couple of years ago, though, if people are interested
and know that dialect).  John may have occasional questions about how
Chinese works for you and others.

lojbab