3.613 multilingual bibliographies (130)

Willard McCarty (MCCARTY@vm.epas.utoronto.ca)
Fri, 20 Oct 89 21:27:46 EDT

Humanist Discussion Group, Vol. 3, No. 613. Friday, 20 Oct 1989.

(1) Date: Fri, 20 Oct 89 00:18:00 EDT (15 lines)
From: Martin Ryle <RYLE@urvax.urich.edu>
Subject: RE: 3.607 queries (117)

(2) Date: Fri, 20 Oct 89 14:40 EDT (61 lines)
Subject: Multilingual bibliographies

(3) Date: Fri, 20 Oct 89 15:12 EDT (30 lines)
From: Ivy Anderson <ANDERSON@brandeis.bitnet>
Subject: Re: Multilingual Bibliographies

(1) --------------------------------------------------------------------
Date: Fri, 20 Oct 89 00:18:00 EDT
From: Martin Ryle <RYLE@urvax.urich.edu>
Subject: RE: 3.607 queries (117)

Re David Birnbaum's question on proper bibliographic citations. Twenty years
ago, when most manuscripts were typed, Cyrillic typewriters were rare, indeed,
and scholars had little choice other than transliterating quotations and
citations that were written in a Cyrillic alphabet. There was even a time when
virtually no scholarly journals used Cyrillic at all. Now that word processors
can handle Cyrillic, it seems only sensible that transliterations should be on
the way out. I vote for Birnbaum's solution.

Martin Ryle
University of Richmond, VA
(2) --------------------------------------------------------------67----
Date: Fri, 20 Oct 89 14:40 EDT
Subject: Multilingual bibliographies

David Birnbaum's predicament in co-indexing Latin and Cyrillic scripts is
an intriguing one. I myself find that there is no solution. However, as I have
never produced a book using several alphabet scripts, I speak from vast inex-
perience. Probably my comments could be treated accordingly.

Of his five putative reasons for not interfiling Cyrillic and Latin, rea-
sons numbers 1, 3, and 5 I agree are nonstarters. Reason 2 is also difficult
to take, but it has something hidden in it, I think. The problem is exactly
where to alphabetize Cyrillic words. Prof. Birnbaum wishes to place them ac-
cording to their "(unwritten) transliterations". The result is bound to con-
tain a significant anomaly: Cyrillic words spelled correctly but alphabetized
incorrectly. In Russian, for example, the equivalent of the English "V" pre-
cedes the equivalent of the English "D". If I am looking for a Russian word in
an index, I want to find it in its proper Russian place, not in its English
place while still in Cyrillic. I am guessing that this is the reason for the
two solutions Prof. Birnbaum mentions. In one, each language is printed and
alphabetized properly according to its own sequence; in the other both langua-
ges are alphabetized according to one model, but one of the languages must,
of course, appear dressed in the other's script.

This may be the reason also why interfiling Cyrillic words as if they were
English is not considered uniform (his reason no. 4).

There are actually more problems than he mentions for some people, be-
cause there are difficulties even in transliterating Russian. Not only are
there several systems (which admittedly differ usually only in small details),
but often there is use of transliteration characters which don't belong to
English either: then how do you alphabetize? The answer is sometimes not tri-
vial. The other problem I have run into, which I'm sure everyone is familiar
with, involves the issue of transliterating or not transliterating names
(sometimes words which are not names) from Cyrillic -- usually Russian -- which
are already known in the Latin alphabet in a less than desirable form. The best
known example may be Tchaikovsky (which is Russian into English via French) and
Tschaikowsky (Russian into English via German). Of course the Russian starts
with a character which is usually transliterated CH, moving this unfortunate
composer from near the end of the alphabet to the beginning.

The problem can even be difficult when no foreign alphabet is in-
volved: consider the usual alphabetization of the Spanish LL (double L). Many
languages alphabetize linked characters like these (not necessarily two of
the same) differently than would English. I have always been annoyed at the
problem of Danish and Swedish, because they use basically the same alphabet
but alphabetize their last three characters (which are vowels placed after Z)
in a different order. Getting back to Cyrillic (no illustrations on E-mail,
alas), cognates in Serbian and Russian are, I would suppose (being no expert
in either) going to be problematic in indexing because the scripts of these
two languages handle palatalization differently, Serbian having more digraphs.

I am sorry that I cannot solve Prof. Birnbaum's dilemma. Perhaps there
are linguist librarians who have already done so? The Library of Congress is-
sues rulings from time to time on these things, mostly on transliterations
rather than on index ordering, but even LC is not consistent. Am I wrong that
the problem is basically unsolvable, that only better and worse work-arounds
exist? In the simplest Cyrillic/Latin case, one could always print the Cyril-
lic words twice, once in the original and once transliterated, in the same
index, but that scarcely seems like an improvement, and there would always be
someone coming along to suggest that the Latin terms be printed twice too......
(3) --------------------------------------------------------------36----
Date: Fri, 20 Oct 89 15:12 EDT
From: Ivy Anderson <ANDERSON@brandeis.bitnet>
Subject: Re: Multilingual Bibliographies

At Brandeis University we grappled with a different but related problem
in developing specifications for an online library catalog which would include
both Hebrew and roman language material. Cyrillic is already romanized
in our catalog, as in many libraries, but Hebrew is not, since we have a
major research collection in this area. We haven't gotten a vendor to
undertake this development as yet, but we have formulated a set of
design criteria. On the particular question of integrating the two
alphabets within a display list, we have somewhat straddled the fence,
allowing for either a single alphabetical list (alphabetized according to
romanized equivalents) or alternatively, two separate alphabetical lists.
If we were designing a truly multilingual catalog, including Asian and
other scripts, one could imagine the problem of single alphabetical lists
incorporating numerous character sets becoming quite complicated, but
multiple lists even more so. On the other hand the card catalog provides a
perfectly good precedent for this in traditionally filing all entries
according to the romanized equivalent.

I should say that the cataloging standards for nonroman bibliographic records
in libraries require that most access points be entered in parallel fields,
one for the nonroman characters and one for the romanized equivalent. This
allows for a great deal of flexibility in how one chooses to handle the

Ivy Anderson
Brandeis University