Topic 4. Transcription

[Richard Noble]

At 06:07 PM 1/20/99 -0700, Bob Maxwell wrote:
>Please add option 4: Transcribe according to a regular, predictable method
>[which dcrb is NOT], such as the LCRI (1.0E) or the "solution of last
>resort" in OH. I think we need to seriously consider such an option for
>several reasons:

RCN: I have not got LCRI in reach (got to get a copy for the home office!)
to refresh my memory of 1.0E. In any case, if we're going to jettison the
current rules (fine by me) then I think the best alternative is simply to
modernize IJUV (and VV?) to the extent that they represent archaic
*typography*, rather than *orthography*. (This may be tricky sometimes in
languages in which i/j particularly are not as easily differentiable as
vowel/consonant as they are in English.)  This treatment would have to be
applied also to matter that is lower-case in the original for there to be
any kind of consistency--which is no less important a criterion than it is
in the method that we currently use, by which lower-case text is
transcribed directly and upper-case matter is lower-cased to match it.
(N.B.: The idea was not to imitate the printer's conventions just for the
sake of doing so, but to ensure that the text of any transcription should
be consistent throughout--an aesthetic motive, to be sure. One looks
further into the book only in cases where the printer's conventions are not
all exemplified in the whole of the text being transcribed.)

> 1. Patrick's point about the way these things sort in library systems.
>Especially for titles that are used frequently (either containing common
>terms or for works that have been issued in many editions), both dcrb
>transcribed titles and exact transcriptions give an entirely arbitrary sort
>order in title searches. Same titles should sort together, shouldn't they,
>independent of the vagaries of printers' preferences of u's and v's and
>catalogers' guesses about those preferences?

RCN: This opens a can of really wiggly worms. Granted, in card files it was
easy enough to disregard IJUV variants in filing, according to algorithms
that are easily enough programmed in wetware. If we were, for example, to
modernize, as I suggested above, then at least this bit of the problem
would go away when we're stuck with software.

BUT unless we are prepared to abandon the 245 as a transcribed field
altogether, or at least impose all sorts of interesting rules to suppress
all the other possible variations in titles proper, we'll have to live with
our unpredictable displays. We must beware of distorting the rules at a
fundamental level in order to solve system problems. That the collocating
function of uniform titles is not fully realized (to say the least) in many
systems doesn't mean that we'll have much better luck stuffing the 245 into
an ill-fitting uniform of its own. By this criterion we ought to change the
first sentence of 1B1, or at least redefine title proper as *not* including
anything preceding the chief title, but I have a feeling that that could be
a messy agony in Latin. We could at least strongly suggest a 246 30 for the
chief title? 

We need multi-indent displays. We're still in the early days of these
wilfully inconsistent systems. Accommodating their faults is seldom the
best basis for deciding on treatment; taking best advantage of their
possibilities and lobbying for their reform is the better angle of approach.

Basically, the producers of the unruly artifacts that we deal with were
unaware of the Paris Principles, or of any principles whatever it sometimes
seems, and we have (so far) determined that it's more useful to remain as
faithful to their products as we can, and find our own ways of getting
everything nicely organized. I don't want my vendor dictating my treatment
of research materials. I think, though I haven't had much chance to look at
it, that the IFLA paper on OPAC displays may have some ideas to offer in
all this.

> 2. Predictability. DCRB's transcription rules are almost entirely
>unpredictable to the user who does not have the book in hand (which is
>after all why he/she is searching the catalog). How is one to guess what
>the printer's practice was without having the book in hand [...]

RCN: This, to my mind, is the best argument of all against the current
method of transcription. I've always felt that I ought to provide a note to
account for the rules used for transcription--a note which could only end
up being prolix, futile, and bad PR for cataloguers in general.

> 3. DCRM for later materials. One of the points of the revision is to
>consider expanding DCRB to a DCRM that will include rules for cataloging of
>19th century and later materials. Yet here we will hit the LCRI head on,
>since up to now u/v i/j in post-1800 materials will have been transcribed
>according to the LCRI. So we will have records for the same item with
>varying title transcriptions, depending on whether they were done under
>AACR2/LCRI or under DCRM, if the old DCRB rule is retained (or more
>particularly if a rule requiring exact transcription is adopted).

RCN: Modernize the lot. Make added entries.

>Although we have not all weighed in on the principles I proposed earlier, I
>haven't heard any opposition, so I would like to cite Principle I: Does the
>rule contribute to the Paris Principle 2.1 identification function of the
>catalog? I am convinced that the current DCRB transcription rule does not
>help a user identify whether the library owns a particular item, since it
>neither calls for exact transcription nor a predictable transcription rule.
>The user unaware of DCRB 0H might well miss entirely the existence of an
>item owned by the library, particularly in systems (such as RLIN) that are
>unforgiving, in which the user has to enter the search exactly or gets no
>hits.

RCN: I think modernized typography, backed up with exact t.a.e., will give
the best chance of putting this principle into practice.

>I am not terribly convinced by the argument that we should not change the
>transcription rule because we have a lot of dcrb records out there already.
>There are probably actually many more records that have NOT been
>transcribed according to dcrb. I truly feel that the dcrb rule is a bad one
>and should not be perpetuated.

RCN: There's all sorts of stuff out there in the cataloguing mvsevm. But
Bob is right--the dcrb rules produce some of the silliest and least
predictable examples.

>[...] "exact" with the exception that we will lower-case the letters that
need to
>be lower-cased? Or should we be REALLY exact and transcribe caps as caps?
[...]

RCN: I don't think that would go down very well, and we long ago gave up on
the idea of quasi-facsimile transcription in library catalogues. But is the
Maxwellian tongue in the Maxwellian cheek? Although I suppose one could
dream briefly about a 246 for the title proper with all caps... (The one
place I sometimes do this is in 590 notes to transcribe spine titles--a
neurotic twitch.)

>To change the topic slightly, 0H and the LCRI to 1.0E both call for
>separating ligatures with the exception of oe in French; and ae in Danish
>(LCRI), or oe or ae in Scandinavian languages (DCRB). Why is this
>distinction being made? [...]

RCN: Does this reflect any sort of French or Scandinavian pressure, I
wonder? They may not be entirely happy with character deficiencies imposed
on them by anglophones. Native speakers linguae Latinae have less influence
these days.


RICHARD NOBLE : RARE BOOK CATALOGUER : JOHN HAY LIBRARY : BROWN UNIVERSITY
PROVIDENCE, RI 02912 : 401-863-1187/FAX 863-2093 : RICHARD_NOBLE@BROWN.EDU