On The National Interest, John Allen Gay has written an interesting piece about what Google’s Ngram Viewer can tell us about the use of place names that have been changing over time. You should read his text first, so I don’t have to explain again what it is all about. Gay’s article focuses on English place names. This made me wonder if German toponyms, often similar in form to the English ones, show the same usage patterns. Let’s have a look at it. I will discuss the examples in the same order as Gay and, whenever possible, stick with the time ranges he selected.
Allow me one cautionary remark before presenting the data: What Google Ngram Viewer gives us is the proportion of a certain term relative to all words in publications of a certain year that are part of the Google Books catalogue. However, the Ngram Viewer does not tell us anything about the context in which our search term appears. If we find an outdated form to be as frequent as a more recent form, this does not necessarily imply that writers have used the outdated form in reference to the modern state or city. It could just as well be a historical reference, but we can’t tell that apart. I would therefore try to resist drawing any firm conclusions from the data, unless there is a prolonged and robust frequency advantage of one term over the other. Gay basically adheres to this principle, but it can’t hurt to remind oneself and the readers of such potentials confounds in the data. Let me also make you aware of one linguistic feature of German that makes this kind of analyses more cumbersome: Unlike English, German distinguishes four grammatical cases that are associated with different suffixes. In most cases, I have included all possible forms of the toponyms in the analyses (or, when this proved impossible due to length restrictions in the search term, at least included the same inflected forms for all toponyms).
But now to the data: For English, Gay shows that ‘Burma’ still is preferred over ‘Myanmar’, 25 years after the country has officially been renamed. The German data are less conclusive: They show fairly similar frequencies for ‘Myanmar’, ‘Birma’ and ‘Burma’ since the mid-90s. In the last years, ‘Myanmar’ seems to fall in between the other two names in terms of frequency.
The German data for ‘Thailand’, formerly called ‘Siam’, show somewhat sharper rises and falls than the English data. It can clearly be seen that ‘Thailand’, the official name of the country from 1939 to 1945, was more frequent than ‘Siam’ in exactly those years (whereas it just spiked, but remained less popular in English). When ‘Siam’ became the official name again (1945–1949), ‘Thailand’ fell behind in German and, just like in English, only became the more frequent form from 1953 on.
According to Gay, the English data for ‘Thailand’ and ‘Siam’ show “how hard it can be to change names”. However, the transitions in the German data seem fairly straightforward and swift to me. The case of ‘Burkina Faso’ and ‘Obervolta’ (Upper Volta), by contrast, is considered a “successful switchover” by Gay although the lag between the official renaming of the country and the first evidence of a higher frequency of the new name is only one year shorter than for ‘Thailand’ and ‘Siam’ (three years instead of four). In any case, the German data for ‘Burkina Faso’ and ‘Obervolta’ are very similar to the English ones, but add another year of usage lag. The country was renamed to ‘Burkina Faso’ in 1984, but the new name only becomes more frequent than the old one in 1988.
Even more quickly than in those cases, ‘Goldküste’ (Gold Coast) was replaced by the country’s new name: Ghana. After no more than a year (in 1956), the latter had surpassed the former in both languages.
The naming situation is more complex in the case of what is now called ‘Zimbabwe’ (or, more frequently, ‘Simbabwe’ in German). When the territory fell into colonial hands at the end of the 19th century, it was called ‘Rhodesia’ after the British politician Cecil Rhodes. After a couple of years, Rhodesia was split up into a Northern part (now Zambia) and a Southern part (now Zimbabwe). ‘Südrhodesien’ (Southern Rhodesia) was never more frequent in German texts than ‘Rhodesien’ although the former was the official name of the country between 1911 and 1953. Between 1953 and 1963, Northern Rhodesia, Southern Rhodesia and Nyasaland (now Malawi) formed to the ‘Föderation von Rhodesien und Njassaland’ (Federation of Rhodesia and Nyasaland), inofficially also referred to as ‘Zentralafrikanische Föderation’ (Central African Federation). Even when combining the two terms, as in the diagram below, their frequency remains low with a modest peak around 1960. ‘Rhodesia’ remained the more popular term and rose to new heights when it became the official name of nowadays Zimbabwe again after the dissolution of the Federation of Rhodesia and Nyasaland in 1963. When the country’s name eventually changed to ‘Zimbabwe’ in 1980, it was in the same year that, frequency-wise, the new name overtook the old one in German. In English, it took another four years for this to happen.
Have a look at the graphs for ‘Malawi’ and ‘Zambia’ (or ‘Sambia’) as well.
John Allen Gay mentions the usage of ‘Congo’ and ‘Zaire’ as a counterexample to a clean break after independence from colonisation. This might be the case, but I am convinced that Google’s Ngram Viewer won’t be able to show us. After all, ‘Congo’ (and ‘Kongo’ in German) is as much the name of a river as it now is (part of) the name of two countries. You might want to assume that the frequency of the river should have remained stable over the decades, but that is just speculation. When we search for ‘Congo’ or ‘Kongo’, we simply cannot know what the words refer to. It does therefore not make much sense to put their frequency in relation to, say, ‘Zaire’. When trying to compare the frequencies of the names of what is now called ‘Republik Kongo’ (Republic of the Congo), things aren’t any clearer: Our frequency waters are muddied by the fact that between 1960 and 1966, ‘Republik Kongo’ was the official name of what is now called ‘Demokratische Republik Kongo’ (Democratic Republic of the Congo). I think we just have to conclude that thorough research instead of a simple frequency check is needed to learn more about how any country that has or had ‘Congo’ in its name has been referred to.
Iran is another interesting (and, luckily, clearer) case: ‘Iran’, the oldest name for this country, has not been the most common term for centuries. ‘Persien’ (Persia) was used instead. In 1935, Rezā Shāh tried to establish ‘Iran’ as the official name in international contexts—a suggestion that engendered much opposition. In 1969, Rezā Shāh’s successor, Mohammad Reza Pahlavi, approved the proposal to use ‘Iran’ and ‘Persia’ interchangeably. Nevertheless, ‘Iran’ has been more frequent in German since 1955, a switch that only took place eight years later in English sources. It also seems to me as if ‘Iran’ had a slightly stronger frequency advantage over ‘Persien’ in German than over ‘Persia’ in English.
Ethiopia, a country with only a brief period of (Italian) colonisation, has formerly been referred to as ‘Abyssinia’. The frequency distribution of ‘Äthiopien’ (Ethopia) and ‘Abessinien’ (Abyssinia) in German differs markedly from that of its English counterparts. 1939 was the year in which ‘Ethiopia’ passed ‘Abyssinia’ in English, two years before the country became independent again. In German, ‘Äthiopien’ had been much less frequent throughout the 19th and the first half of the 20th century than ‘Ethiopia’ in English. This might be the reason why the switch in German was delayed until 1956.
Considerable differences between English and German usage also surface when it comes to the frequency of ‘Netherlands’, as opposed to ‘Holland’. Strictly speaking, the latter term only refers to a region in the Western part of the Netherlands, but it has been used metonymically for the entire country. In German, both terms have been similarly frequent for centuries with ‘Holland’ traditionally taking the lead. This could not only be caused by a usage preference for ‘Holland’, but also by the fact that ‘Holland’, as an ambiguous term, can be used to refer either to the Holland region within the Netherlands or to the country as a whole. ‘Niederlande’ cannot be used that way. The latter became the more frequent term in 1956, whereas in English, ‘Holland’ only dipped below ‘Netherlands’ a remarkable 25 years later (in 1981). The adjective switch from ‘holländisch’ to ‘niederländisch’ in German happens in 1964 (unfortunately, you can’t test that in English with the adjective being ‘Dutch’ in any case).
There might also be differences between English and German for the frequency of ‘Tschechoslowakei’ (Czechoslovakia) and the names of its successor states, ‘Tschechische Republik’ (colloquially: Tschechien; Czech Republic) and ‘Slowakische Republik’ (Slowakei; Slovak Republic). Gay’s diagram, however, does not seem to give an accurate representation of the linguistic state of affairs: He compares ‘Czech’, which can also be used as an adjective outside the country name, to ‘Czechoslovakia’ (noun only, no adjective form) and ‘Slovakia+Slovak’ (noun + adjective). This reveals ‘Czechoslovakia’ falling behind ‘Czech’ and ‘Slovakia+Slovak’ in 1994 and 1996, respectively. The German data are much less clear-cut with all forms (except for ‘Mähren’, more on that soon) being about equally frequent since the mid-90s:
When we do not take the adjectives into consideration and only compare country names, the English diagram suddenly looks much more similar to the German one:
However, one difference remains: As Gay rightly notes, “the German attempt to brand the Czech lands “Bohemia” (and Moravia) during their occupation was largely unsuccessful” – in English. In German, ‘Böhmen’ and ‘Mähren’ show a sizeable spike between 1933 and 1945.
When it comes to the names of what we now call ‘Russia’, Gay again draws a comparison that does not seem to accurately represent the linguistic situation: He contrasts the frequency of ‘Soviet’—which was part of the term ‘Soviet Union’, but could also be used as an adjective independently—to that of ‘Russia’ without including the adjectival form ‘Russian’, finding ‘Soviet’ to remain more frequent until the end of the data set (in 2008). Comparing ‘Soviet’ to ‘Russia+Russian’ or ‘Soviet Union’ to ‘Russia’, however, shows a frequency advantage for ‘Russia(n)’ since 1994 or 1992, respectively. In German, the frequencies of the current and the former term are much closer to one another, both in the country name (shown below) and in the adjective forms. The frequency advantage of ‘Russland(s)’ is more marked when comparing it to the German equivalent of ‘Union of Soviet Socialist Republics (USSR)’.
Following the English text, we now turn from country names to city names.
Gay discusses the frequency of ‘Istanbul’ and ‘Constantinople’, concluding that in English, the former is still less common than the latter, “more than five hundred years after the the city fell to the Muslims”. Technically, this statement is true for the data the Google Ngram Viewer shows us. However, I strongly doubt that ‘Constantinople’ (or ‘Konstantinopel’ in German) is really widely used to refer to present-day Istanbul. It seems more likely to me that this term, especially in academic writing, relates to the 1,500 years of history that the city has seen under the name of ‘Constantinople’. The German data at least show converging proportions for the old and the new name from the late 90s on.
Gay has already accurately described the German usage patterns of Zagreb (also called Agram), Ljubljana (Laibach) and Praha (Prague/Prag), so it seems unnecessary to discuss these here again. Also, I am skipping ‘Wien’: Comparing it to ‘Vienna’ is interesting in English, but less so in German.
Next up is ‘Sankt Petersburg’ (Saint Petersburg), formerly called Petrograd (during the First World War) and Leningrad (between 1924 and 1991). Again, I have some doubts about whether the Ngram Viewer can paint a realistic picture of the name usage in this case. Why that? In English, German and Russian, Saint Petersburg is frequently referred to as ‘Peter(s)burg’, so this form should be included in the search term. However, there are plenty of towns called ‘Petersburg’ in the United States and elsewhere. Is searching for ‘Petersburg’ going to artificially inflate the frequency due to mentions of places outside Russia? The differences between searching for ‘Saint/Sankt/St. Petersburg’ on the one hand and ‘Petersburg’ on the other are huge in any case: In English, ‘Saint/St. Petersburg’ remains behind ‘Leningrad’ until the end of the data set, even passing ‘Petrograd’ only in 1992, whereas ‘Petersburg’ has almost always been more frequent than both other names. The German data are similar for comparisons using ‘Sankt/St. Petersburg’ or ‘Petersburg’ as search terms. The Russian data, shown below, are unlikely to be influenced by American place names. They resemble English and German plots that use just ‘Petersburg’, so these comparisons probably have to be considered most informative:
Let’s look even farther East. Bombay, Calcutta (or ‘Kalkutta’ in German) and Saigon are probably all worth a visit, but there’s just one thing we can say about usage changes with respect to their names: There aren’t any changes. In German, like in English, hardly anybody calls these cities ‘Mumbai’, ‘Kolkata’ or ‘Ho-Chi-Minh-Stadt’, respectively. The situation is similar for Yangon in Myanmar: The city was founded under the name of ‘Dagon’ (but it’s no use searching for that in the Ngrams Viewer because ‘Dagon’ is also the name of a Mesopotamian fertility god) and renamed to ‘Yangon’ in 1755. The city was under British occupation between 1824 and 1826. It was probably at that time that the name ‘Rangoon’ came into existence, an attempt at anglicising the local pronunciation of ‘Yangon’. In English sources, ‘Rangoon’ is still used overwhelmingly. In German, ‘Rangun’ is also more popular than ‘Yangon’, but the share of the old/new name seems to be somewhat larger than in English.
Not only Iran as a country, but also its capital, Tehran, has seen interesting usage changes in its names—or, in the latter case, in the English rendering of the Persian (or Farsi) pronunciation of the same name. Called تهران [tehˈɾɒn] locally, the city was referred to in English as ‘Tehran’ or ‘Teheran’ interchangeably throughout the 19th century. ‘Tehran’ then fell out of popularity for decades, started rising again after the Second World War and has been more frequent than ‘Teheran’ since 1976. In German, by contrast, ‘Tehran’ never enjoyed any popularity. I have to admit that, while I was aware of the Persian pronunciation of the name, I didn’t even know that the city wasn’t called ‘Teheran’ in English:
Coming back to the title of this post, I have one more intriguing example of differences between English and German: ‘Beijing’ is what the capital of China is now called in the vast majority of English texts in the Google Books catalogue. That’s a fairly recent development. Before 1975, hardly anybody used that form. ‘Peking’ (or, in the 19th century, ‘Pekin’) had been the preferred English rendering of the name for more than two centuries. It seems to have had a strong impact that in 1979, ‘Beijing’ became the obligatory spelling in official Chinese publications. In German, the impact was much smaller. ‘Beijing’ has become more frequent since then, but ‘Peking’ remains the most frequent form.
The frequency advantage is even stronger in non-academic usage: ‘Peking’ occurs about 20 times as often as ‘Beijing’ in the German IDS corpus, mainly consisting of newspaper sources. Why that? The letter/sound correspondences of the Pinyin transliteration are not very intuitive for speakers of German. Every native speaker of English can probably guess that the diphthong written ‘ei’ should start with an [e] instead of an [a] and that ‘j’ does not symbolize [j], but [d͡ʒ]. Many Germans would get this wrong, which might be one of the reasons why German authors have been slower in adopting ‘Beijing’. The same does not hold true for ‘Nanking’, the ‘Capital of the North’: The Pinyin rendering ‘Nanjing’ is about as frequent as the older transcription in German (and English). Again: Why? Because in all probability, hardly any non-sinophile German knows of the existence of ‘Nanjing’. Therefore it’s mentioned mostly in specialist publications where naming innovations catch on faster than with the general public.
Is that all? No, of course not. John Allen Gay’s article ends at that point, but there are plenty more interesting name changes and usage changes to look at. I’ll get to that as soon as possible.