[DCRM-L] Reconsidering digraphs

[Laurence Creider]

I find Brian's post best represents my beliefs on this rather difficult
problem.  The issues of display can be considerable.  I was hoping not to
have to say anything about this topic, mostly because I do not have the
knowledge of early printing practices that so many of the respondents do.  
However, I have a visceral belief that 1) ae and oe in Latin should not be
transcribed as digraphs and 2) that Anglo-Saxon ae should be.  When I look
at the reasons for this (even visceral reactions have reasons, however
inadequate), I think that the difference is that the Latin ae/oe digraph
in printing (or ms) is not a matter of orthography or meaning, but a
matter of combining the letters.  The reading (either grammatical reading
or pronunciation) would not vary on the basis of the print representation.  
In this fact the situation is not much different from the ct or st
ligatures.  In Old English, however, the ae digraph is a separate letter.
While the digraph or the two separate letters would, I believe, have been
pronounced the same, AS ae is not a dipthong but a single sound.

In general, I am not a fan of basing cataloging decisions on the
(in)capabilities of local systems.  However, two things about the
ligatured forms of ae and oe give me pause.  First is the memory that when
Penn migrated to Voyager back in the mid 1990's we lost the ligatured
forms completely.  We had to change a lot of instances of the word
"uvres!"  This leads me to reflect that since diacritics in general are
not in the front of the mind of most of those who do system conversion and
since the special character sets vary from system to system, there is more
chance for problems when a record goes from RLIN to OCLC or vice-versa or
from one local system to another.  Presumably, when the world is Unicode
compliant, this will cease to be a factor.  The other problem is that when
a digraph is included, some systems will require an added entry
representing the separate components of the digraph so that the searcher
without the capability to represent the digraph or knowledge that one
might be present in the title will be able to find that title.  In other
systems, such an added entry will be redundant.  It is not hard to imagine
a situation where one's bibliographic utility and local system are
different in this regard.  It seems to me that a great many redundancies
will be eliminated if we simply generally transcribe these ligatures or
digraphs as two separate letters except in the cases in which they are
generally done.

One last question, how often is the difference between one edition or even
issues signalled soley or even primarily by the presence of a digraph?  
If often, then we may need to transcribe the digraph as such for purposes
of identification.  If this is seldom the case, then we lose little by
routinely separating the letters.
	Larry

Laurence S. Creider
Head, General Cataloging Unit
New Mexico State University
Las Cruces, NM  88003
Work: 505-646-4707
Fax: 505-646-7477
lcreider at lib.nmsu.edu

On Wed, 23 Feb 2005, Hillyard, Brian wrote:

> As somebody who served on (even moderated) the relevant Working Group
> (WG 2) at the DCRM Revision Conference, Yale, in March 2003, my memory
> is that our concern was exclusively with the exceptions (French,
> Scandinavian, etc.) which we did agonise over; I don't remember that
> we seriously questioned the separation of the ae and oe ligatures in
> (e.g.) Latin.  I've read some of the other postings on this, and I
> cannot see that this is a question of content as opposed to form.  I
> suspect the ESTC thinking was influenced by the possibility that
> transcribing ae and oe ligatures as such in records for (e.g.)
> editions of Caesar's works would sometimes help distinguish between
> editions.  One could point out that to the same end it would be useful
> to transcribe long S (lowercase) as a different character -- I don't
> think any descriptive cataloguing rules have tried to prescribe this,
> but because ae and oe ligatures had traditionally been retained in the
> printing trade, it seemed acceptable and possible to retain those in
> transcription (when I began cataloguing in 1977, typewriters used by
> NLS cataloguers had these ligatures).  It does seem to me to be an
> arbitrary line we're drawing here and not a distinction between
> content and form.  Where some modern languages do now use ligatures,
> there may be a case for using those ligatures, but as we don't use ae
> and oe ligatures now in printing Latin, there may be a case for not
> using them -- even though we could.  In the vocabulary of DCRM(B)
> principles, are these cases of how we balance representation and
> standardization?
> 
>  
> 
> Here in NLS I remember we did transcribe ae and oe ligatures when we
> began MARC cataloguing and I also remember the problems of OPAC
> display (e.g. instances of "Casar", I think) and retrieval that caused
> us to revise the practice.
> 
>  
> This is an entirely personal view: no consultation at all.
>  
> Brian
>  
> 
> ******************************************** 
> Dr Brian Hillyard 
> Rare Book Collections Manager, National Library of Scotland 
> George IV Bridge, Edinburgh, EH1 1EW 
> b.hillyard at nls.uk: 0131-226 4531 (voice): 0131-466 2807 (fax) 
> 
> -----
> 
>