At 02:28 PM 7/5/2005, Nancy Sack wrote:
>Are you saying that catalogers have been unneccessarily adding death dates
>all along? I'm shocked. SHOCKED!

I'm thinking of sending something like the following.  Any comments?

How refreshing it is to have extended discussion on a matter of policy on
the PCC list! It will be interesting to see whether any of our comments has
any effect on the final policy.

While I think it is possible for us to come to consensus on the small point
raised by the proposed policy change, I think it would be better if we
looked at the larger question first (and at all kinds of headings, not just
personal names), and then see how this particular issue fits in.  I think
we do better when we have a set of general principles that we should
normally follow, rather than only a series of policies addressing
individual situations.

I would take as the starting point the assumption that adequate resources
(in-house personnel--regarded in terms of both quantity and ability, local
system features, money for outsourcing, outsourcing vendor capabilities,
and perhaps other considerations) are not available at most institutions
for the reliable and efficient performance of database changes. If true,
this would for me imply that we should reduce as much as possible the need
for heading maintenance: whatever maintenance we are able to perform must
be maintenance that makes a difference for catalog users (of any type,
including library staff). To put it another way: headings should not be
changed unless change is necessary.

So for me, the general question might be restated like this: When is it
necessary that an existing heading be changed?  Several clearly-delimited
categories come immediately to mind:
         1) When an existing heading contains errors of fact (typographical
errors, for example)
         2) When an existing heading conflicts with another existing heading
         3) When an existing heading conflicts with a heading being
constructed, and it is not possible to adjust the new heading to avoid the
conflict

Viewed in this way, a change to an existing non-conflicting heading simply
because more information has become available is an unnecessary change and
should not be allowed. So if for example a person's date of birth or the
full form for a name represented in a personal name heading by an initial
becomes available but the person's heading is not in conflict, the
birthdate or other information should be recorded in a proper 670 field
(not a 678!) for future use, and the heading should stand unchanged. (I
would extend this principle also to headings constructed for CIP materials.
Once the authority record has been distributed, the heading is established;
the heading should not be changed simply because more information such as
date of birth or full form of a name has become available--even if the
additional information comes from the published form of the very item for
which the heaidng was established. I would similarly extend this principle
to cases in which the person who established the heading missed information
in the item for which the heading was established: add dates, full names
etc. to the 670 field for the item, but leave the heading alone.)

* * * * Digression

To move for the moment from the abstract to the particular, assume we have
this set of established headings:

         Smith, Flash
         Smith, Flash, $d 1952-
         Smith, Flash, $d 1960-

If we discover the birthdate for the first Flash after the heading has been
established, we may add a new 670 field to Flash's authority record (or
modify an existing 670 field if an existing 670 field is incomplete).
Because there are no conflicting headings, we make no change to any of the
headings.

If we are cataloging a new item for a Flash that we determine is none of
the existing Flashes and we have the date of birth for the new Flash, we
include the date of birth in the new heading, and leave the existing
headings alone.

         Smith, Flash
         Smith, Flash, $d 1952-
         Smith, Flash, $d 1960-
         Smith, Flash, $d 1965- <a new heading>

But if we are cataloging a new item for a Flash that we determine is none
of the existing Flashes and we do not have the date of birth for the new
Flash but we have in the authority record a date of birth for the existing
no-date-in-heading Flash (July 2, 1954, say), then we change the existing
heading to include the date of birth, to make way for the new heading:

         Smith, Flash <a new heading>
         Smith, Flash, $d 1952-
         Smith, Flash, $d 1954- <formerly "Smith, Flash">
         Smith, Flash, $d 1960-

Or consider the following set of established headings:

         Smith, Flash, $d d. 1792
         Smith, Flash, $d 1781-1824

If we discover the birthdate for the first Flash, we add a new 670 field to
the authority record for that heading, and leave the heading itself alone,
because there is no conflict.  (I would not bat an eye if someone added to
the authority record a reference tracing that included the birth and death
dates. But there are many changes to policy for reference tracings that
might be suggested and I'd like not to stray too far from the main point.
Save that for another time.)

* * * * End of digression

We might choose to define one exception to the general principle ("make no
change unless it is necessary") and clearly label it as an exception: If an
existing personal name heading contains a date of birth and it is
discovered that the person is no longer living, the heading may be changed
to include the date of death if known or to insert "b." before the date of
birth if the year of death is not known. The reason for allowing this
exception has been stated by others several times: it makes us look stupid
to indicate in our catalogs that Sartre is not yet dead. We'll abandon
principle in this case for cosmetic or public relations reasons. In this
kind of change (when the former personal name heading contains a date) the
automatic flipping of headings by program without operator review will not
result in incorrect reassignment except in the most unusual cases (once
every 20 years, I'm thinking). Note that this exception applies only to
headings for persons that include an "open" date; under the general
principle, I would not have us add a death date to an existing
non-conflicting heading that contained no dates at all.

Another exception we might recognize comes immediately to mind: If a
corporate body is constructed provisionally because the form of name in the
appropriate language is not available and the proper language form later
becomes available, then it is correct to change the heading to reflect the
proper language. (So a heading for an Algerian government agency
constructed in English because that's all that was available at the time
would be changed to reflect the Arabic (or whatever) form when that form
was encountered.)

For me, the need to keep maintenance to a minimum also means that headings
should be constructed from the beginning with an eye on future upkeep. So I
think there is a great deal of merit in one of the counter-proposals made
by Alva Stone. As I would restate the proposal, we would retitle the
"Headings that do not conflict" section from LCRI 22.17 "Headings for
persons known to be deceased that do not conflict:" we would no longer add
dates to non-conflicting headings for persons assumed to be living at the
time of heading creation. (One small and unscientific sample indicates that
3/4 of new personal name headings with dates don't actually need the dates
to resolve conflicts.) In the authority records for persons assumed to be
living we would include birth dates along with other information as a
matter of course in 670 fields; if information such as dates were
discovered at a later time, they could be added to the authority record
(again, in 670 fields).Should headings come into conflict, information in
authority records (and other sources as necessary) would be consulted and
dates added here or there as policy and need dictate. Maintenance would be
performed only when needed, and headings would also not contain "open"
dates for patrons to complain about.


Gary L. Strawn, Authorities Librarian, etc.
Northwestern University, 1970 Campus Drive, Evanston IL 60208-2300
e-mail: [log in to unmask]   voice: 847/491-2788   fax: 847/491-8306
Forsan et haec olim meminisse iuvabit.