APG-L Archives

Archiver > APG > 2009-01 > 1232218042

From: <>
Subject: Re: [APG] "Fun" with ancestry.com
Date: Sat, 17 Jan 2009 12:47:22 -0600
References: <589E02758F15403EA07D5BD1F1DED59D@YOUR58BA15CF1B><213220.51692.qm@web111211.mail.gq1.yahoo.com>
In-Reply-To: <213220.51692.qm@web111211.mail.gq1.yahoo.com>

Langdon wrote:
> Personally, I don?t trust Ancestry.com data bases anyway,

Langdon, no database can be "trusted." Not any, by anybody.

Take, for example, the highly acclaimed, scholarly database by the renowned
Gwendolyn Midlo Hall, Research Fellow, Tulane University; Professor Emerita
of History, Rutgers University; and member of the International Advisory
Board of the Harriet Tubman Resource Institute on the African Diaspora (York
University, Toronto, Canada).

That database, "Afro-Louisiana History and Genealogy, 1699-1860," was
published by Louisiana State University Press. It's now online at various
sites, including Ancestry. I use it regularly and I value it for what it is:
a finding aid--not a collection of "records."

In the 9 years since this database was released by LSU Press, I have
compared hundreds of its db entries against the original records, and ***in
more than half of those*** there are errors. Names are regularly "misread."
(Even though Hall's assistants are considered scholars familiar with
Louisiana culture, the language, and the penmanship of the era, they do
not--and cannot be expected to--know the correct identity of every settler
and every slave whose names could be variously read.) Beyond that,
assumptions of kinship are made without evidence to support those
assumptions. Some documents were totally missed within this-or-that record
set. For other documents, identities of multiple individuals are collapsed
into one.

Shall we now roundly castigate LSU Press for "sloppiness" in publishing that
database without verifying the accuracy of every entry? Shall we castigate
Tulane or Rutgers or the Harriet Tubman Resource Institute for its
association with someone whose work is not perfect? Or shall we simply say:
"Thanks for providing this finding aid. It helps me mine the original
records more expeditiously and more economically. And, by the way, in
Document xxxx, the name Whatever should be Whoever. Attached is my evidence
to support the correction."

Which approach is the most realistic and the most helpful to all concerned?

Last week, you posted your frustration with a different record set from the
South Carolina State Archives. Or, not so much frustration with the records
itself as with the staff who (a) make copies for researchers without
precisely identifying what they are copying; and (b) can't seem to 'get
their act together on how to cite those records,' to the point of actually
telling you, "It doesn't matter. Whatever citation you use, we will find

As researchers, we have a different need from that of the learned staff at
the Archives. To us it *does* matter, because we're viewing the situation
from a different standpoint. Archivists, in their academic training,
typically learn the standard explanation for citations ("cite your sources
so others can find where your information comes from"). We, as researchers
know there is a far more important reason for identifying our sources
thoroughly: the need to identify the *characteristics* of a record set so we
can make a sound judgment as to the reliability of the source.

The reality is that the academic world is far more trusting than the careful
genealogist is. Most scholars work with the "big picture"--an approach
wherein random errors on small points rarely affect their overall
conclusions about society at large and its patterns. Good genealogists are
microhistorians; for us, an error on a small point can be fatal.

So, no. We cannot trust any database--academic, archival, or commercial. We
use them, we're delighted to have them, but we don't trust them.


Elizabeth Shown Mills, CG, CGL, FASG
APG member, Tennessee

This thread: