For those of us living in homogeneous Central Europe, the patchwork of minorities, ethnicities, and languages in some of the regions that TOL covers can seem incomprehensible or, at least, extremely foreign to our own experiences (even if I am originally from Brooklyn, which could probably compete with many of those areas in diversity).

Mapping tends to bring these issues home like few other tools, and I was thus pleased to recently come across a blog post called “Re-Mapping Languages of the Caucasus.” The author, Asya Pereltsvaig, is a linguist at Stanford and the post appeared on a site that she helps run, Geocurrents, which describes itself as a “a map-illustrated forum dedicated to exploring global geography.” The site appears to be particularly strong on TOL’s coverage region, having run a series of posts earlier this year on the Caucasus, and, over the past few weeks, on Siberia.

As part of the Caucasus series, the site worked with a cartographer, Jake Coolidge, also at Stanford, to map “the ethno-linguistic mosaic of the region.” As Ms. Pereltsvaig describes:

It quickly became clear to us that while many ethnic groups are defined by language, in some cases ethnic and linguistic affiliations do not coincide: for example, Jews speak different languages, depending on where they live, and most Greeks in the Caucasus—to the extent that these communities still exist—speak Urum (a Turkic language) rather than Greek (an Indo-European language). We therefore decided to focus on mapping the diverse languages of the region, a task that proved more complicated than we had expected.

In the end, they used a variety of ethnic and linguistic maps, other demographic data, and census data to construct a map of the entire Caucasus and then one focusing just on Dagestan. The result is a strikingly colorful map that must have been an immense amount of work (a segment is above, and the full map can be accessed on the site).

As comments below the post indicated, not everyone understood what was being depicted. The goal here was to show where languages are spoken, not to depict dominant languages or ethnicities.

Given that many of the languages depicted on this map are quite small in the number of speakers, they would not show up at all if we depicted the majority language in each area. So we decided to go with “where are speakers of X located” question.

Some people also seemed itching to read more into the depiction of this or that language. For example, little of anything other than Armenian is, of course, now spoken in Karabakh; the map’s creators illustrated that, but they were quick to indicate that they weren’t saying the region was now part of Armenia:

This map is meant to represent languages, not geopolitics. The color of Karabakh represents the language spoken there, Armenian. According to the latest census data (2005), Armenians constitute 99.7% of the population in Nagorno-Karabakh. BTW, note the two Armenian areas in Abkhazia: they are not meant to make any geopolitical claims either!

I was also fascinated (but not surprised) to hear that many other maps have exaggerated or understated the territory held by certain groups:

As we researched these posts, it became obvious that most available maps of the ethno-linguistic groups of the Caucasus are inadequate to the task. Many maps over-represent the extent of the territory occupied by certain groups, especially those speaking Northwest Caucasian languages: Abazians, Adygheis, Kabardians, Cherkess. Partially, this has to do with extending their territories to uninhabited areas in the high mountains, as is the case with this map. In other instances, certain groups are shown as occupying more extensive territories than they really do, often at the expense of neighboring peoples. The range of other groups is exaggerated by the use of out-of-date data. Even new maps often fail to capture the wholesale migrations, episodes of ethnic cleansing, and population exchanges have changed the situation on the ground.

Sometimes those “errors” are probably unintentional, as the authors point out (old data), but one can only imagine that some mapmakers have been motivated by a desire to address historical wrongs or to gerrymander certain regions in favor of one group or another. That would be an interesting investigation.

As for the Geocurrents maps, yes, as the authors admit, “mistakes and imprecisions have crept into these maps.” One commenter noted, for example, that Azeri speakers supposedly inhabit an area of southwestern Azerbaijan located between the borders of Armenia and Nagorno-Karabakh, but few, if any, have lived there since the end of hostilities. The mapmakers have, however, been admirably quick to acknowledge such constructive criticism, and have even already incorporated earlier comments into the newer versions that were posted earlier this month.

So I would encourage all readers of East of Center and TOL to take a look at the maps and send along your comments. The mapmakers are listening!

Jeremy Druker

Jeremy Druker is the executive director and editor in chief of Transitions Online. Twitter: @JeremyDruker Email:

More Posts