response on Lojban names 11/30/91

[Logical Language Group <cbmvax!uunet!pucc.princeton.edu!lojbab>]

And writes:
>But you have to know the language of a name before you can reconstruct
>it from the lojban.  I suppose that's better than the present system, in
>which it is impossible to tell what someone's name is when it is
>expressed in lojban.  (E.g. la djan. kau,n. could be John Cow.)

Not likely.  He would more likely then have Lojbanized it as "kaun" with
no comma.  You can often tell a lot about a name by what the speaker does
NOT do.

It is NOT impossible to tell someone's name when expressed in Lojban.
In your example, his name is "djan. kau,n."  Now as to whether you can
tell what his ENGLISH name is:  well, yes, that could be a problem.  So?
Nick points out that if you care, you can use foreign text quotes and
spell it the foreign way if that is important.

But no method will easily resolve "Marian" vs.  "Marion", either, or
"Joe" and "Jo", or "Mary" and "Merry" (of Tolkien fame).  If you are a
person to whom the distinction in spelling is important you will
Lojbanize so as to make the distinction clear, probably by matching the
spelling.

Likewise, some with the names "Ann" or "Anne" will Lojbanize their name
as ".an.", some as ".en.", but if they have acquired the diminuitive
nickname "Annie", they may alternate between ".an." and ".enis."

In another message, I said that I Lojbanize my lastname as
"leceval,ier."  Actually, I also sometimes do it "leceVALi,er." which is
slightly closer to my pronunciation.  But if I want to match my
pronunication in English, I would do more like "licyVALi,eis."  (wishing
that the person could treat the final as a triphthong as I pronounce it
rather than mandatorily splitting it by pairs per the Lojban rules).
But since I am pround of the heritage implied by my French name, I
prefer to use a more Francophone spelling.  Not accurate for the FRENCH
version of the name which would have ultimate stress, but still
preserving spelling and etymology.

If you are transliterating a name from another language and do not know
how it is pronounced, and cannot ask, it is wisest to preserve spelling,
or to make conventional changes.  I don't know how "Mr.  Khan"
pronounced his name, but I would transliterate it "genxis. xan." based
on my conventions of mapping /ng/ to /n/, and /h/ and /kh/ to /x/.

I would also do Edinburgh as "ednbrx.", or "Edinburx."  (note the cap),
and with no idea of pronunciation before it was discussed herein, iVAN's
address as "rel. byklyx. cmaklaj." which mixes two Lojban names with
one foreign one, or perhaps "pleis" if I were feeling timid.

Nora independently confirms my approach on 'Edinburgh'.

Ivan can drop the gender suffix on his last name - I could easily
imagine however that some Slavic speakers would prefer to retain it as a
sign of cultural identity.

Ivan:  by the way.  How do you feel the stress-shifts on vowels in
Slavic names should be handled?  I've been told several times that the
leader of the USSR should not be Lojbanized as "gorbatcof.", but rather
more like "garbaTCAF."  But this loses the Cyrillic visual recognition,
perhaps unnecessarily, in the pursuit of perfection in sound.  We did
this kind of thing in gismu-making, and have been criticized because the
result is actually lower in recognition because CONTRAST in vowels is
lost in pursuit of accurate phonetic mapping.  Chinese suffers even more.

Some have even gone to unnatural lengths to preserve their cultural
identity in their name.  Kim Pizer, from the southern US, preferred her
southern US drawl embedded in her name "ki,ym." to the 'obvious' "kim".
Paul Francis O'Sullivan, proud of his Irish heritage, insisted on
Lojbanizing his name as if it were Irish, giving something like
"polfranCIIS. osuliebein."  But that is NOT how he says his name when
speaking English, of course.

As to the final consonant in a vowel final name:  I have no objection if
people want to add a 'strange' consonant to make their name unique.  But
this seems unnatural to me, and also to those who have Lojbanized their
names so far.  Generally you WANT your name to sound like it came from
your native language.

One person whose name is Jay, has already expressed a preference to have
his name Lojbanized to suggest "Jason", rather than to add some unique
consonant that could be confused with some other name.

My bottom line, I think is that And's original question:  what is done
in the Lojban culture so far, is that no effort is made for unambiguity
with respect to the original name.  Generally, people have seemed to
prefer that their personal names conform more closely to their preferred
pronunciation, which may not be their true pronunciation.  I think that
people are far less dogmatic with country names, street names, and names
of other people that they are writing about and cannot expect to know
the correct pronunciation - generally in these cases preferring to err
towards visual recognition of the name.

There is a slight tendancy towards loan-translation in foreign names with
semantic content, as in the "Place" of Ivan's street address.

(By the way, as an alternate answer to Dave Cortesi regarding the street
number, I might use "la byklyx. cmaklaj. pe li re".  I would be
unlikely to use either cardinal or ordinal suffixes, reserving them for
"2nd Avenue").  Colin's answer, using ordinals, might be valid in some
limited situations where the numbers are actually consecutive and
indicate position on the street.  This is seldom the case in US
addresses.  Or maybe that should be "la'eli re po'e la byklyx. cmaklaj."
That respresented by the number 2 inalienably associated with Buccleuch
small-street.

Mark Shoulson writes in response to John Cowan
>> I forgot to mention that "gicbau" has to be "gicybau", and even "gicban." is
>> suspect; there is an evolving policy (which may become firm) to avoid
>> impermissible consonant clusters even in names.
>
>Um, I don't think that's such a hot idea.  I know that there's a lot of
>discussion about cmene going on now (I'm a litle behind in reading my
>mail), so here's what I think:
>
>1) Lojban should place as few restrictions as possible on cmene.  YHou
>should have the right ti name yourself as close as possible to whatever
>you want.  Apart from conforming to orthography and those silly
>"la/lai/doi" restrictions, we should really have nothing else.  There is
>some justification to doing "gicyban.", since that's a lojbanic (as
>opposed to lojbanized) name, but if my name had an impermissible
>cluster, I think I should have the right to use it.  After all, that's
>why we have buffer vowels.  Now, this method would probably allow people
>to pick names which are thoroughly unpronouncable by human equipment.
>Which brings me to the other point:

If your name has an impermissible cluster, then people will try to
pronounce it will often succeed - incorrectly, assimilating per the
rules of their native language.  Marks's position seems especially
strange, since he objects to doubled 'e' and to syllabic 't' in:

>Several problems. first off, Lojban has no definition for the
>pronunciation of two consecutive "e"'s (or two consecutive anythings,
>for that matter, with the exception of "ii" and "uu").  Also, you cannot
>have a pause inside a name.  If you have a pause, it's a two-part name,
>and each part must end in a consonant.  Using a comma doesn't help in
>this case, since a comma is a syllable break, and thus implies a sylable
>consisting only of the "t".  That sound is not considered syllabic in
>Lojban, and I suspect you'd be hard-pressed to find a language in which
>is *is* considered syllabic.

By his argument:

>2) Contrary to popular belief, the *real* authority in name-giving is
>the *namer*, not the namee.

So you SHOULD be able to do any of these if there are no restrictions on
names.

I feel that Mark's answer on "ee" is applicable to all such instances.
Lojban has no defined pronunciation for an impermissible medial cluster.
Therefore you shouldn't use one.  If you do, you should be prepared to
have you Lojbanic listener attempt to map what he/she hears to a valid
Lojban phoneme string, thus by audiovisual isomorphism to a spelling
different than you use, which will then be read differently by a third
speaker.

(Actually 'ee' has the defined pronunciation 'e,e', in Lojban.)

I would say the same thing to a speaker of a language with aspirated
consonants that argued for putting 'x' after each aspirated consonant.
The speaker might argue that this comes closer to the way the speaker
says the word, but how would a non-speaker deal with Hindi "bxarat."
Buffering it loses whtever recognition benefit is gained by accuracy in
contrast.

I >DO< favor allowing some relaxations, including some additional
syllabic consonants, strings of permissible medials that are not
permissible initials at the beginnings of words "cprents." for Lojbanist
Joel Shprentz, or "lau,dz." for Lao-Tse.

Note that And's buffer vowel analysis that you can say ANYTHING as long
as you buffer it is NOT official Lojban, but merely a proposal.  THus
you are NOT allowed to leave out /y/ hyphens between impermissible
medials in lujvo even you could always add a buffer vowel.  The idea is
to MINIMIZE the use of buffering, for the very reasons And has
criticized the buffer in previous postings (mostly on conlang):  its
phonetic value is dependent on the phoneme mappings of the speaker and
listener.  We are NOT trying to make the buffer the equivalent of the
other vowels.

>It should be noted that in the interests of courtesy the namer ought to
>defer to the namee, but when push comes to shove, it's the one
>speaking/using the name at the moment that is... well, speaking/using
>it.  If I chose to call George Bush "Murgatroyd", that would be somewhat
>rude and very confusing to everyone, but that particular kind of
>ambiguity is permitted in Lojban, and who has the right to tell me what
>to say?  I'm taking my chances of having you misunderstand, but that's
>my risk to take.

I agree, but with the additional caveat that in Lojban instruction we
try to emphasize that it is the speaker's obligation to make statements
understandable to the listener, who may be from a different cultural or
metaphysical background, and only secondarily the listener's obligation
to puzzle out the speaker's intent.  I recognize that in reality,
communication requires both parties to cooperate, but the speaker
obligation paradigm is vital to learning to speak another language, as
different as Lojban is, with hope of communicating.

There will be some lists of names and countries and cultures in whatever
constitutes our first 'dictionary'/'reference work', to help serve as
examples.  I am watching this discussion and will try to include some of
the examples and conventions you people propose, especially those of
you, like Ivan, who can give us authoritative views on what to do with
names from some other cultures.

Ivan writes:
>From: "61510::GILSON" <gilson%61510.decnet@MIL.NAVY.NRL.CCF2>
>>  In British English, perhaps [{j} is absent as a word-final consonant].
>>  Words like "garage," "camouflage," and
>>  "sabotage," in America at least have that sound word final.
>
>Same in British English, but these are not proper names.

If you might Lojbanize "Place" rather than translate it, you might need
to Lojbanize any other English word.  I took my car last week to
"la djoz. graj."


And again:
>Nick:
>> From: And Rosta <ucleaar%UCL.AC.UK@pucc.PRINCETON.EDU>
>> Subject:      onomastic etiquette in lojban
>> >When using a name in lojban one has to convert to lojban orthography.
>> This is desirable, but not obligatory; one can use {la'o} and foreign quotes.
>
>Could you explain la'o, goi, gy.  (& zoi, which I thought was used for
>non Lojban quotes), or refer me to an explanation elsewhere?  Do foreign
>quotes allow
>
>(a) nonlojban romic characters
>(b) non-romic characters? (E.g., Greek, Chinese)

1. Lojban quotes

a. zo quotes a single morphological Lojban word (not orthographic:  zo
lenu only picks up the "le", so be careful)

b. lu ... li'u quotes one or more Lojban word in text which is presumed
by a parser to be (a) grammatical utterance(s)

c. lo'u ... le'u quotes text which properly lexes into Lojban word(s), but
is not presumed to be grammatical utterance.  Treat like the linguist marks
? or * for questionable or incorrect grammaticality.

d. zoi .[word]. text .[word]. delimited quotes, where [word] is any
valid single Lojban word, often a name, or the first letter of a word.
The text inside the quotes need not be Lojban lexically, grammatical or
otherwise, but cannot contain [word] with pauses surrounding for obvious
reasons (similar conflict resolution rules exist for other quotes).  The
"gy." in this case is the letteral "g", standing for "glico".  Also oft
used for English is "gic.", or "kuot."  text can include things like
whistles, music, animal sounds - whatever.

e. la'o .[word]. name .[word]. effectively the same as "zoi" except that
the result is interpreted as a name, rather than a quoted text.  This
was primarily introduce to resolve in written Lojban, such problems as
giving Ivan's address, or "Goethe"s name in a form that it can be
recognized visually without worrying about translation.  Since I'm sure
the UK postal system would have no idea what to do with "byxlyx.
pleis.", this kind of thing is pragmatically vital.

All of the above have the grammar of sumti

2. goi - is used for anaphora assignment of sumti, e.g. to free
variables.  It is like an assignment statement of a computer language,
whereas its grammatical equivalent "po'u", the restrictive appositive is
an equals sign or a definitional statement. goi works bi-directionally,
in other words one side is an anaphora and the other is its assigned
definition, but which is which is pragmatically determined.

>> >(4) Is there some cmavo to mark nonce lojbanizations of cmene? I think
>> >this could be used for politeness.
>> The cmavo {za'e}, used to denote arbitrary word coinages, can readily be
>> extended to this function.
>
>I thought so. I hope this will be noted by the grammarians at lojbocrat HQ.

This has been part of the intent for za'e from the start.  It is usable
ANY time you make up or use a word and wish to flag for your listener
that it may be nonstandard in definition or possibly meaning/place
structure.  This is poart of our solution for living free of
dictionaries, even when one is written:  if you suspect that a word
might be in the dictionary and/or might have a place structure you don;t
expect, or if you are Lojbanizing a name or borrowing in a potentially
non-standard way, you can warn the listener.

>The following ideas might be considered:
>
>_la_ is matched with a terminator, say _la'a_ (for argument's sake,
>ignoring the real _la'a_, whatever it means).  The *only* constraint on
>names is that they can't contain /laha/.  It follows from this that an
>entire sentence, even one including names, can function as a name (e.g.
>a title of a book), providing that la-la'a can nest.
>
>This idea is not only incontrovertibly brilliant [is there an UI meaning
>'smiley'?  - insert it here] but also characteristically lojbanic.

"zo'o"

>
>No - wait - what would be even more lojbanic, though less elegant, would
>be to keep the current arrangement but also add two new cmavo
>"open-name" and "close-name".

Congrats:  you just reinvented "la'o" which has open and close
delimiters of your choice since any delimiter we might choose could
theoretically be found in a totally unconstrained name.  Instead, you
merely have to have a name that does not consist of ALL POSSIBLE Lojban
words.  I HOPE this is an acceptable limit zo'o.

I will accept your compliment as to its brilliance zo'o, and compliment
you for again thinking of it zo'o.  This is one Lojban word you might
remember xu?

Ivan says to Colin:
>>  >  if you need a standard, isn't {j} a much better one?  Its chances
>>  > to be mistakenly stripped off the end of a lojbanised name are much
>>  > lower than those of {s}.
>>  >
>>  <...> There is something in me which is repelled  by
>>  all those English names turning into
>
>But that is exactly how it must be.  If a lojbanised English name sounds
>drastically unEnglish, you'll know for sure that you must do something
>(like stripping off the "repellent" last consonant) to it to get its
>real form.

But how do you know in LOJBAN text that it is a Lojbanized English name.
If you want its Englishness identity preserved, use la'o quotes.  If you
want a LOJBAN name, then Lojbanize it, and thereafter try to forget that
this name translates to English, while that one to Bulgarian while the
other is a purely Lojban coining.

"la" names are for LOjbanized, i.e. >Lojban< names.

lojbab