Recent Focused Series »

Indo-European Origins
Siberia
Northern California
The Caucasus
Imaginary Geography
Home » Cartography, Linguistic Geography, World

Misleading Language Maps on the Internet

Submitted by on July 10, 2012 – 4:06 pm 31 Comments |  
Although the internet allows easy access to manifold cartographic treasures, it provides even more rapid access to misleading, poorly constructed, and laughably inaccurate maps. Consider, for example, language maps at the global scale. A simple Google image search of “world language map” yields over 600 million results, although only the top hits, and by no means all of them, actually show linguistic maps of the world. Those that do can in general be divided into two categories: maps that depict language families, and maps focused on the most widely spoken individual languages. Today’s post considers the latter category, analyzing Google’s eight most highly ranked “world language maps” that portray the distribution of specific languages.

All of these maps are actually best described as “political linguistic maps,” as they organize their depiction of language distribution in accordance with the territories of internationally recognized states. As a result, multilingual states—which constitute the majority of the world’s countries—tend to be mapped as monolingual. Canada is divided into English- and French-speaking zones in roughly half of these maps, but few other countries are treated in such a manner. In almost all cases, official languages are highlighted regardless of whether they are actually spoken by the majority of the population; as a result, Mali appears to be as much a French-speaking country as France. The criteria for language selection generally go unmentioned and in most cases seem inconsistent, but it does appear in general that the “number of native speakers” outweighs the “total number of speakers.”  As a result, one of the world’s most widely spoken and politically significant languages, Indonesian/Malaysian (“Malay”), is usually ignored, often in favor of much less widely used European languages.

Such problems, however, are difficult to avoid. Multilingualism alone—both on the individual and on communal level—makes language mapping a frustrating exercise. But in all of the maps analyzed, the flaws run much deeper. Most are riddled with errors, many at a quite elementary level. As a result, the use of such readily accessible maps risks undermining knowledge of the world by delivering misinformation. To substantiate such harsh allegations, the remainder of the post will examine in some detail each of top eight world language maps that depict individual languages.

The first map has relatively few obvious blunders, although portraying Namibia, Lesotho, and Swaziland as French-speaking is a howler. Mapping Djibouti as an Arabic-speaking country is also problematic; although Arabic has official status—as does French—relatively few Djiboutians speak it, as Somali and Afar are the main vehicles of communication. Map 1 does divide a few countries into separate languages, and does so with a degree of accuracy. Not only is Canada split at the provincial level into Anglophone and Francophone areas, but so too is Cameroon, while Chad is divided into Arabic- and French-speaking zones. Nowhere else, however, do language boundaries deviate from those of states. Only Kenya and Tanzania are portrayed (through diagonal stippling) as containing multiple languages in the same locations, but the effort is flubbed ; presumably the intention was to show English intermixed with Swahili, but Swahili does not appear in the key. As is true of almost all maps of this kind, official languages of European origin in sub-Saharan Africa are exaggerated; curiously, however, Botswana, Malawi, and South Africa—where English has official status and is widely used—are not included in the English-speaking set. The portrayal of India as uniformly Hindi-speaking is also problematic, as is the mapping of China as completely Chinese-speaking—especially considering the fact that  “Chinese” is not exactly a spoken language, but rather a group of related languages that are, with the exception of Mandarin, locally conceptualized as mere dialects.

Map 2 is a far more comprehensive effort, depicting 23 separate languages. Most are limited to a single country, sometimes incorrectly so (Austria, for example, is not depicted as German-speaking). Outrageous errors here include the depiction of Sakhalin as Japanese speaking, Mali, Cyprus, and Azerbaijan as Arabic speaking, and Belgium as speaking some uncertain language (the color used for Belgium does not appear in the key). The criteria for inclusion in this map seem particularly odd; why, for example, are relatively major languages such as Vietnamese, Bahasa Indonesia, Italian, and Polish, ignored while Finnish and Norwegian are mapped? India is depicted accurately here as “multilingual,” but it is the only country so classified!  The text-box labeled “other major languages spoken in the world” is confusing; how can “French and English” be classified here as “other languages” when both are extensively mapped? In actuality, it appears that the numbers in the box were designed to have been placed on specific countries: “1: French and Sango,” for example, pertains to Central African Republic. Unfortunately, that step was neglected.

Map 3 at least attempts to show areas of language overlap and multilingualism, although it does so in a crude manner. Note, for example, the extension of the North American French-speaking zone well out of Quebec into Newfoundland, Nova Scotia, and Maine. The map’s reduction of the vast majority of languages in the key to “local dialects, misc.” is risible. One also finds the Inuit language (actually, “languages”), spoken by fewer than 100,000 people, depicted as more significant than languages spoken by more than 100 million, such as Hindi, Bengali, and Indonesian. Yet the map simultaneously puts one of the main Inuit-speaking areas, the coastal strip of southern Greenland, in the “local dialects and miscellaneous” category.

Map 4—probably the worst of the lot—strictly depicts all sovereign states as linguistically uniform—except Canada. Switzerland and Belgium are simply mapped as French speaking—as, absurdly, are Romania, Vietnam, and even Albania. Equally egregious is the depiction of Thailand, Laos, Burma, Cambodia, and Malaysia as Mandarin speaking.  The portrayal of the entire former Soviet Union as Russian speaking is also misleading, as is the mapping of Somalia and Eritrea as Arabic speaking (although Arabic is a co-official language of both countries). Note that Israel is also mapped as Arabic speaking. The Portuguese language is oddly ignored, and Portuguese-using Guinea Bissau has been colored as a Francophone state. Francophone Burundi and Rwanda* have in contrast been depicted Anglophone, whereas Anglophone Malawi and Swaziland have been excluded from the same category.

On first glance, Map 5 appears to be comprehensive and sophisticated than the others—but it is not.  This map violates basic protocols by placing individual languages and language families at the same level of analysis. Here, for example, one finds “German” rather than “Germanic” but at the same time “Turkic” rather than “Turkish.” Yet the Turkic language family is severely misconstrued, as Turkic-speaking Azerbaijan is placed in the non-existent “Caucasian” language family, whereas non-Turkic-speaking Mongolia is included. Dividing the Slavic family on the basis of script rather than linguistic relationship is inexcusable,** as is the use of imaginary language categories, such as “West African.” The depiction of Ethiopia as entirely Amharic speaking is problematic enough, but the placing of Somalia in the same category is indefensible. Belize, Haiti, Guyana, Suriname, and French Guiana are all incorrectly mapped as Spanish speaking. For India, Bangladesh, Bhutan, and Sri Lanka—as well as Mainland Southeast Asia—cop-out categories of geographical rather than linguistic reference are employed. Bizarrely, the “languages of [Mainland] SE Asia” class is extended to Madagascar and Melanesia. And in Europe, while the relatively minor language of Albanian is mapped, Albanian-speaking Kosovo is incorrectly depicted as Slavic speaking.


One might think that the Wikipedia map (#6) of world languages would be reasonably accurate—but one would again be mistaken. Although it is difficult to see, the map severely misconstrues the Caribbean, where neither Guadeloupe not Martinique are depicted as French speaking, but Dominica—where the official language is English—is. (In the Pacific, Fiji and Samoa are also mapped as French speaking.) Dutch- and Papiamento-speaking Curaçao and Aruba, however, are portrayed as English speaking. All of Timor is depicted as Portuguese speaking, even the Indonesian half of the island.  For both India and Pakistan, the archaic term “Hindustani” is employed, which is depicted as uniformly extending across India. Kyrgyzstan is shown as Russian speaking; although Russian is an official language, it is by no means the country’s major tongue. As with many other maps of this type, the extent of Arabic is exaggerated by including Eritrea, Somalia and South Sudan. By the same token, the extent of French in Africa is overplayed, yet that of English in the same region is ignored altogether. And while Swahili is indeed an official language of Uganda, the country can hardly be regarded as Swahili speaking; English also has official status, and is more widely used. The mapping of Afghanistan as Persian speaking is justifiable, but the exclusion of Tajikistan from the same category is not. One might also ask why Italian merits depiction, but not Japanese, Turkish, Korean, Vietnamese, and Indonesian/Malaysian.

Map 7 does a better job than the others in depicting multilingualism. Yet it oddly depicts Guinea, Gabon, and Senegal as entirely French speaking, unlike the other Francophone countries of sub-Saharan Africa; compare also the divergent mapping of Mozambique and Angola in Lusophone (Portuguese speaking) Africa, and note the depiction of southern Africa as entirely English speaking. This map also mixes individual languages with language families (Turkic and Slavic), yet it manages to misconstrue its own categories.  Note that Azerbaijan is incorrectly mapped as non-Turkic, just as Bosnia and Macedonia are incorrectly mapped as non-Slavic. Finnic-speaking Estonia, however, is put in the Slavic category.

Map 8 is the odd one out in this series, as it does not actually map languages, but rather merely provides information on the percentage of people who speak certain languages over five continent-like divisions of the world. I cannot imagine how this information could be useful to anyone in any circumstance. Note also that it also makes errors in categorization, as it lists individual languages along with a language family of uncertain coherence (“Kwa”) and a certain type of language (“Creole”).

It is of course one thing to harshly criticize such maps and another to produce something better. Stay tuned for a GeoCurrents map of the world’s main languages later this summer.

 

*Rwanda is admittedly tending in an Anglophone direction—much to the consternation of France—but French is still more important than English.

**This distinction is also incorrectly applied, as Montenegro actually favors the Latin rather than the Cyrillic alphabet.

 

Previous Post
«
Next Post
»

Subscribe For Updates

It would be a pleasure to have you back on GeoCurrents in the future. You can sign up for email updates or follow our RSS Feed, Facebook, or Twitter for notifications of each new post:
        

Commenting Guidelines: GeoCurrents is a forum for the respectful exchange of ideas, and loaded political commentary can detract from that. We ask that you as a reader keep this in mind when sharing your thoughts in the comments below.

  • http://profile.yahoo.com/A3QWSXGGMUKHBJHMQ5XHSPTY44 Peter

    Although French competes with local languages in much of Africa, I nonetheless have heard that there are more people who speak French as a first language in Africa than in France.

    • Randy McDonald

      As a second language, yes, but not as a first. The only Francophone African countries I know of where there is a shift towards French as a common first language, on the model of Portuguese in Angola, are Cote d’Ivoire and Gabon. In the first country, a very ethnically complex country opted for French as a neutral common language in urban areas, especially Abidjan; in the second, a long and extensive French presence led to the displacement of African languages, notably Fang, from the public sphere.

      • http://geocurrents.info Martin W. Lewis

        Manny thanks the additional information. Abidjan was long the core city of Francophone Africa, and the language politics of Cote d’Ivoire are extremely complicated.

        Do you know what the situation is in Dakar, Senegal?  The city has a large French expat community, as well as a large number of Lebanese. 

        • Randy McDonald

          From what I know, Dakar is basically like Senegal writ large, a country with a Wolof-speaking majority that uses French in certain stable niches.

          I strongly recommend Jacques Leclerc’s site “L’aménagement linguistique dans le monde”, “Language Management in the World”.

          http://www.tlfq.ulaval.ca/AXL/

          It has detailed profiles of language situations in all the countries of the world: facts on the ground, official policies, and more.

      • jeronimo constantina

        I agree. I may be mistaken, but heard that the only two  (?) places in Africa where French is a first language are Abidjan, Cote d’Ivoire and Libreville, Gabon.

    • http://geocurrents.info Martin W. Lewis

      I think that Randy McDonald is right: as a second language, but not as a first.  

      • http://www.pereltsvaig.com Asya Pereltsvaig

        Indeed. And the degree to which people actually speak French as a second (or more commonly, third) language varies greatly from place to place…

  • http://twitter.com/joinacult Brandon

    ” Note, for example, the extension of the North American French-speaking zone well out of Quebec into Newfoundland, Nova Scotia, and Maine.” The map is fairly accurate in this respect. The map actually highlights New Brunswick, Canada’s only officially bilingual province, as well. Acadian French is endemic to Quebec, Nova Scotia, New Brunswick, and Maine, despite its being in a different country. They do err on Newfoundland, however. Although it was settled by Irish and French colonists, it’s almost strictly English-speaking.

    • Randy McDonald

      Agreed. While language shift to English is ongoing, there is a “bilingual belt” straddling Quebec’s frontiers that includes large Francophone populations in northeastern Ontario and New Brunswick, just under a million people between the two provinces.

      • http://geocurrents.info Martin W. Lewis

        Good points by Brandon and Randy McDonald, but I still think that the map exaggerates the extent of French in the Maritime Provinces and in northern New England, as these areas are all shown as more blue than red. 

        • http://www.facebook.com/people/James-T-Wilson/682045086 James T. Wilson

          I’m not familiar with the Canadian situation, but New England reminds me of the recent posting on French heritage in the Midwest. I have never come across a town where I heard a variety of French spoken casually, as one does in southeastern Louisiana. If one looks, one can find a great deal of evidence of previous French culture in place names and in some localities in family names, but it is not an ongoing presence.

      • David A.

        there are more than a million French speakers in Canada, 20% speaks French. And there is no “shift” to English

  • Ixn

    Given the degree of overlap and local variation, surely there’s only one solution here – a series of maps, one for each language, demonstrating the approximate number of speakers (per capita, rather than net total) in each set geographic region. It would be easy to use flash/similar to overlay these as desired,or cluster results to form language families etc. Different series could use political boundaries or set spatial grids (ie 500x500km cells); or political ‘official language’ statuses vs actual prevalence in daily use.  

    It would be particularly fascinating, if data exists, to show these maps over time – ie colonisation, or more recent Anglicisation, etc.

    Is that the map Geocurrents has in mind?

    • http://geocurrents.info Martin W. Lewis

      Yes, something like that — certainly multiple maps, and GIS maps with multiple layers, will be necessary.  We also hope to use “crowd editing” again, as we did with the map of the Caucasus. 

      • http://www.pereltsvaig.com Asya Pereltsvaig

        The overall idea is great, but I am afraid that getting reliable (and in some cases, meaningful) data would be a questionable proposition. Also, when it comes to non-native speakers there are varying degrees of “speaking language X”, and I am not sure how we can go about representing that…

  • James

    In view of the complex multilingualism which is the worldwide norm these days, isn’t it something of an anachronism to even try to relate language to geographical spaces and places? Vive le mélange des langues!

    • http://geocurrents.info Martin W. Lewis

      I would’t go that far, as language use is still highly geographically structured, both in terms of what language people actually speak and what languages have official status. But you point is well taken. 

      • http://www.pereltsvaig.com Asya Pereltsvaig

        I agree with Martin here. Especially if we take away the big, international languages, 99% of the world’s languages are spoken in fairly limited geographical locations…

    • http://www.pereltsvaig.com Asya Pereltsvaig

      Vive the mix языков! :)

  • http://blog.zolnai.ca/ Andrew Zolnai

    Hilarious yet sad. Have you seen ethnologue.com and Steve Huffman’s worldgeodatasets.com? Or on arcgis.com
    http://www.arcgis.com/home/webmap/viewer.html?services=990baa47865a42859bab116358a53971

    UN Data has a dearth of info here, covering only 84 countries!
    http://data.un.org/Data.aspx?q=language&d=POP&f=tableCode%3a27 

    As in my previous blogs its complexity can be mapped and aggregated by country: https://www.google.com/fusiontables/DataSource?snapid=S571825OPJzor

    or by language: https://www.google.com/fusiontables/DataSource?snapid=S571834H_d4

    Look for a blogpost to illustrate the complexities and limitations.

    • http://geocurrents.info Martin W. Lewis

      Many thanks, Andrew, for bringing these resources to our attention!

  • Pingback: Mapping International Languages « Geonotes « GeoCurrents

  • http://blog.zolnai.ca/ Andrew Zolnai

    More links at the risk of repetition:
    WALS online http://wals.info http://2008.wals.info/ (Oxford Uni./Max Planck Dig. Lib.) 
    World mapper http://www.worldmapper.org/display_languages.php?selected=583 (Uni. Suffolk and its famous cartograms)

  • Jeronimoconstantina

    The Philippines, as you will readily see, is one of the gross victims of these misleading maps. While maps 1, 2, 6 and 7 can plead innocence, with the country either labeled as “other,” or has a color that does not appear in the legend. Map 3 commits the sin of calling Philippine languages “local dialects, misc.” (in fact, it is depicted in two colors, representing English and “local dialects, misc.”).   Map 4 unequivocally depicts the Philippines as English-speaking. While English is an official language, it is inferior in this respect to Tagalog aka Filipino, since the Constitution says that the official languages are Filipino, and English, unless otherwise provided by law (in other words, the official status of English can be repealed by legislation, while that of Filipino-Tagalog cannot, unless the Constitution itself were suspended or repealed. Map 5 is similarly misleading, as it depicts the entire Philippines as speaking Filipino, the government name for Tagalog. In fact, only about a third or less of Filipinos are native speakers of Tagalog. They have committed an offense similar to Google, Blogger, Blogspot, and other Tagalog defaulters, which, under the notion that all Filipinos are native speakers of Tagalog, they have made Manila’s language the default medium from Batanes to Tawi-Tawi. OTOH, map 8 is also guilty, although in a different way – the Philippines is entirely by a pie chart, so we do not know how it is mapped.

    • http://geocurrents.info Martin W. Lewis

      Excellent points — many thanks. Cebuano long had more native speakers than Tagalog, but recent reports suggest that Tagalog/Filipino has surpassed it.  Ilocano certainly has its partisans; when I lived in northern Luzon in the 1980s, most people, other than the young, resisted the national language, favoring a combination of Ilocano and English in addition to their native tongues. I lived in a Kankana-ey speaking village, close to the linguistic border with Ibaloi. Most people were “quatrilingual” — learning Filipino entailed becoming “quintilingual.” 

      In classroom use, I love to show the language map of the Philippines, as it demonstrates  how the seas historically connect and the mountainous interiors of islands divide language communities. 

  • jeronimo constantina

    The Philippines, as you will readily see, is one of the gross victims of these misleading maps. While maps 1, 2, 6 and 7 can plead innocence, with the country labeled as “other,” or appearing in a color that is not in the legend. Map 3 commits the sin of calling Philippine languages “local dialects, misc.” (in fact, it is depicted in two colors, representing English and “local dialects, misc.”).  Map 4 unequivocally labels the Philippines as English-speaking. While English is an official language, it is inferior in this respect to Tagalog aka Filipino, since the Constitution says that the official languages are Filipino, and English, unless otherwise provided by law (in other words, the official status of English can be repealed by legislation, while that of Filipino-Tagalog cannot, unless the Constitution itself were suspended or repealed). Map 5 is similarly misleading, as it depicts the entire Philippines as speaking Filipino, the government name for Tagalog. In fact, only about a third or less of Filipinos are native speakers of Tagalog. They have committed an offense similar to Google, Blogger, Blogspot, and other Tagalog defaulters, which, under the notion that all Filipinos are native speakers of Tagalog, they have made Manila’s language the default medium from Batanes to Tawi-Tawi. OTOH, map 8 is also guilty, although in a different way – the Philippines is hidden entirely by a pie chart, so we do not know how it is mapped.

  • Ikerfoot

    Any news on when the new, accurate map can be expected, now that we’re at the end of July? Thank you! (I’d like to use it in my class this fall.)

    • http://www.pereltsvaig.com Asya Pereltsvaig

      Thank you for your continuing interest, Ikerfoot! Unfortunately, several other projects have taken priority (see below) and this being a summer, things are bit slower too. So the map project has been rather on the back burner lately, but we hope to get back to it before long. However, we are hoping to release the GeoCurrents Map Archive, with numerous maps of various kind, classified and searchable, before too long. Also, we are running a contest among our readers, with a personalized travel map as the prize:

      http://geocurrents.info/contests/mapping-your-world-travels-personalized-world-traveler-map-giveaway

      • ikerfoot

        Thank you so much! Please keep posting updates when any more info becomes available.

        • http://www.pereltsvaig.com Asya Pereltsvaig

          I will. Our most immediate plans include a mini-series on Hawaii, another one on the Olympics and the results, and we are also hoping to address a recent paper by Atkinson et al. in Science about the origins and spread of the Indo-European language family (discussed today in the New York Times)…

  • Pingback: Link love: language (45) « Sentence first

  • Brian

    Sideways, 3-D pie charts are always wrong.

  • Oodles

    “I cannot imagine how this information could be useful to anyone in any circumstance.”

    I’m sorry to say, but I think your imagination might be at fault here. Having percentages of speakers, in combination with geographical area, can give one a rough idea of the likelihood of being able to communicate with someone in that area (relative to your own language set).

    Granted, this map is not specific enough with the areas it defines. But I can easily see how such a map, better defined, might be very useful for tourism, advertising or missionary purposes. Can’t you?

    At the end of the day, all maps are simplifications of reality and induce errors. However, a percentage approach is interesting, because it at least allows for the possibility of multilingualism.