2.1.3. Source data
As mentioned above, one of the key sources for Ethnologue data is SIL’s vast network of
field linguists. However, SIL does not have the resources to place field linguists in all
areas of the world; they concentrate their resources in areas where they currently have
projects underway. This means that SIL has very little of its own information in many
areas of the world, such as Central Asian Russia (as in the example discussed above).
Hence, the Ethnologue must rely on other sources of information to identify, locate and
enumerate speakers of languages in these areas. A number of its sources are from other
academic linguists, including important linguistic surveys such as the Language Atlas of
the Pacific Area (Wurm and Hattori 1981), and the Linguistic Survey of India (Grierson
1903-1928), general references and language family and area references such as the
Cambridge Language Surveys book series (e.g. Holm 1989; Masica 1991; Shibatani
1990; although not all available titles in the series are cited, e.g., Dixon and Aikhenvald
1999; MacAulay 1992; Mithun 1999; Posner1996). Because these surveys are
incomplete, infrequently updated and unavailable in some areas, this still leaves a
patchwork of areas that need to be covered.
Some of these areas can be covered by individual citations to academic linguistic
publications on specific languages and locales. Other gaps are filled in typically from
Christian missionary sources, such as the World Christian Database (WCD) and
Operation World. Introduction of these sources for many citations in the fifteenth edition
complicated the relationship between WCD and the Ethnologue, for the simple reason
that WCD had previously cited the Ethnologue wherever possible for information about
language populations, and the potential for circular citation made it harder for both to
check and update their sources. Moreover, the specific information that language
identifications and population estimates are based on in such sources are unlikely to be
based on the professional linguistic field assessment of the information from SIL and
other academic linguists. Rather, they are more likely based on the less formal
assessments of Christian missionaries, churches, and aid workers (collectively referred to
as “ministries”). They may even be based, directly or indirectly, on government reports,
census figures or almanacs. In the end, the provenance of this information is far less
certain than that of the census and academic sources.
Some information, including population figures, is presented without a cited
source for it. These appear to be cases where information was carried over from earlier
editions of the Ethnologue, and there did not happen to be a citation in the earlier edition.
Other times a citation year is given but no source. The Ethnologue editors have made a
decision to provide information to the extent that it is known wherever possible. In their
view, it is better to put out some form of population estimate, for example, even if it is
old, out of date or from an unreliable source, if that is all that is available. By doing so,
they are reporting as honestly as possible what they are able to ascertain about the status
of a language and its speakers. They reason that this should stimulate dialogue with the
users of the Ethnologue, who can respond either by suggesting other sources, or
providing further information of their own, that can be cross-checked and potentially
incorporated in future editions of the Ethnologue. In many cases this has had the desired
effect.
2.1.4. ISO 639-3
While the Ethnologue was conceived for SIL’s internal purposes, and is primarily used
for that today, with its publication on the web, it attained a visibility unlike what it had
previously known. This visibility brought with it, among other things, an invitation from
the International Standards Organization to participate in the standards process for ISO
639-3, a planned update for the earlier ISO 639-2 standard that libraries employ to
identify languages for cataloging and other purposes. This standard was felt to be
inadequate as it had a fixed number of languages (about 500), and no effective process
for identifying and adding the large number of new languages that might be needed.
Since the Ethnologue used a set of three-letter codes much like those of ISO 639-2, it
seemed a natural choice for developing the new ISO 639-3. The result is a new draft
standard, now undergoing the final approval process.
The development of the ISO 639-3 draft standard, and its incorporation into the
Ethnologue imposed a number of requirements on the Ethnologue system of identifying
languages. First, the internal three-letter codes that the Ethnologue had previously used
needed to be reconciled with the earlier ISO 639-2 standard. This meant changing a
number of existing codes, to avoid conflicts. The remaining Ethnologue codes were then
grafted onto ISO 639-2 to provide the additional codes needed for the ISO 639-3 draft.
This has the effect of making the Ethnologue the default catalog for the ISO 639-3
standard.
A second consequence of the standards process is that a new office needed to be
organized to maintain the standard. This office is housed inside SIL, and is staffed by
SIL, but its operation is separate from that of the Ethnologue, which submits its desired
changes to the standards office just as any other user of the standard would. Presently,
since the standard is still undergoing the approval process, there is a backlog of requests
to be processed once the standard goes into effect. Requests that would otherwise have
been made in the 15
th
edition were postponed so that the reconciliation of the
Ethnologue’s earlier code system could be accomplished. This is probably one of the
more significant changes in the Ethnologue editorial process since the 14
th
edition.
A third consequence is that a set of codes for ancient and constructed languages,
the LINGUIST codes (Aristar 2002a,b), was also affected by these changes. The
relationship between the LINGUIST codes and the Ethnologue codes significantly
predates the ISO 639-3 draft, and hence was designed to use part of the space no
occupied by the ISO 639-3 standard, of which it is not formally a part. Since there were
235 ancient language codes and 34 constructed language codes before the development of
the ISO 639-3 draft standard, there is a potential for serious maintenance issues.
2.1.5. Staff
The Ethnologue editorial staff currently has three people, Raymond Gordon, editor in
chief; Conrad Hurd, managing editor; and Paul Lewis editor; not all of whom are