From "Servia" to "Serbia"

« previous post | next post »

[The first part of this post is from an anonymous contributor.]

The Serbian legation in London complains to the media about the spelling Servia, which is 'highly offensive to our people'.

(It is true that there is a place in Greece called 'Servia', whose name 'derives from the Latin verb servo, meaning "to watch over"'.)

The Chinese for 'Serbia', 'Sai'erweiya' 塞爾維亞 is obviously derived from 'Servia', not 'Serbia' (which latter would have been 'Sai'erbiya'). But, where did English get 'Servia' anyway?

Background to the 'Servia/Sai'erweiya' transcription:

The pronunciation of standard Chinese, as mediated by the hanographs anyway,

1) has no [v]; [w] is the closest thing

2) has no [wi]; [wei] is the closest thing

3) has no monophthong [ɛ]; [aj] is the closest thing

4) requires [ə] before [ɹ]

'Servia' is a 'historical English term, taken from Greek language, used in relation with Serbia, Serbs or the Serbian language', says Wikipedia.

But since when does English go and transcribe old words from Greek with beta as <v> instead of just transliterating as <b>?

In Koine beta was already [v]:

The consonants [of Koine Greek] also preserved their ancient pronunciations to a great extent, except β, γ, δ, φ, θ, χ and ζ. Β, Γ, Δ, which were originally pronounced /b ɡ d/, became the fricatives /v/ (via [β]), /ɣ/, /ð/, which they still are today, except when preceded by a nasal consonant (μ, ν); in that case, they retain their ancient pronunciations (e.g. γαμβρός [ɣambros], άνδρας [andras], άγγελος [aŋɡelos]). The latter three (Φ, Θ, Χ), which were initially pronounced as aspirates (/pʰ tʰ kʰ/ respectively), developed into the fricatives /f/ (via [ɸ]), /θ/, and /x/. Finally ζ, which is still metrically categorised as a double consonant with ξ and ψ because it was initially pronounced as σδ (sd), later acquired its modern-day value of /z/.[9]

But English has 'Bartholomew' and 'Bosphorus', not 'Vartholomew' and 'Vosphorus' .

Wikipedia also says of names of Serbia (historical renderings in other languages):

  • Servii, Latin rendering.[24]
    • Serviani/Servians, medieval French and English rendering of the Serbs.
 So, that's the English source, French via Latin?


The following is by VHM:

The pronunciation with "v" was not just in English, but also in modern French.  See the first page of Journal de l'Orne (pdf) (August 22, 1914), under the column titled "Memento de la Guerre"), July 21.  It is curious, however, that for July 23, 24, 26, and 28, the name is spelled "Serbie".

In a recent post, I cited a valuable discussion on the pronunciation and spelling of "asterisk" in English Language & Usage Stack Exchange.

The same forum also took up the question about "Servia" and "Serbia" raised above in a discussion initiated on 8/27/15:

"Why did Servia become Serbia? "

This discussion begins with the observation that, at the start of the First World War, the nation in the Balkans was referred to as Servia, but in "numbers" [sic] published after the second half of 1916, it became Serbia.  I suspect that this dramatic change (as shown in an accompanying Google NGrams chart) was the result of the 1915 initiative of the Serbian government reported in this article from the New Zealand North Otago Times, Volume CI, Issue 13235, 5 March 1915, Page 7:


If my suspicion is true, this shows that government intervention can radically influence language usage on a global scale.

Those interested in further investigating the origins and etymology of the names of the Serbs and Serbia may consult this Wikpedia article.


  1. Anschel Schaffer-Cohen said,

    December 21, 2015 @ 10:06 pm

    "If my suspicion is true, this shows that government intervention can radically influence language usage on a global scale."

    Even if your suspicion is false, I think the cases of Thailand and Iran are pretty strong evidence for this assertion.

  2. Moacir said,

    December 21, 2015 @ 10:12 pm

    "this shows that government intervention can radically influence language usage on a global scale."

    I understood that this was the case in the move from "The Ukraine" to "Ukraine" as well. Wiki says:

    Since the Declaration of Independence of Ukraine the English-speaking world has changed its usage from "the Ukraine" to "Ukraine".[40][41][4][42] From November 1991, several American journalists began to refer to Ukraine as Ukraine instead of the Ukraine.[4] The Associated Press dropped the article 'the' on 3 December 1991.[4] This approach has become established in journalism and diplomacy since (other examples are the style guides of The Guardian[43] and The Times[44]). In 1993 the Ukrainian government requested that the article be dropped, which has a Russian analog in the use of "в" or "на" as the appropriate pronoun for "in" (I once accidentally said "на Литве" in Russian class and was scolded for making a gross political error).

  3. Moacir said,

    December 21, 2015 @ 10:14 pm

    Er, I forgot I was quoting wiki… and started telling a story.

  4. Victor Mair said,

    December 21, 2015 @ 10:43 pm

    @Anschel Schaffer-Cohen

    So you think that my suspicion is true? If not, what is the referent of "this assertion" in your statement?

  5. Anschel Schaffer-Cohen said,

    December 21, 2015 @ 10:58 pm

    The assertion that "government intervention can radically influence language usage on a global scale."

  6. Victor Mair said,

    December 21, 2015 @ 11:03 pm

    So you don't think my suspicion is true?

  7. Anschel Schaffer-Cohen said,

    December 21, 2015 @ 11:07 pm

    I don't know any more about the history of the word "Serbia" than I just read here, so I don't have enough evidence to have an opinion on your suspicion. My point was that the bit about governments being able to change what their countries are called is independently true.

  8. Michael Watts said,

    December 22, 2015 @ 3:23 am

    The pronunciation of standard Chinese […]

    1) has no [v]; [w] is the closest thing

    This is certainly correct as a description of standard Chinese. Is it correct in any absolute sense? Modern standard Mandarin has phonemic /w/ and phonemic /f/, but [v] is an allophone of /w/. (Actually, it's not quite correct to say that MSM has no [v], as the sound occurs all the time. But everyone knows that it's really /w/.) I believe Old English had phonemic /f/ and phonemic /w/ as well, but [v] was an allophone of /f/ (cf heofon -> heaven).

    What I'd really like to know is, if I had to study Chinese without access to a native speaker who I could have attempt to pronounce foreign words for me, could I predict that [v] was a variety of /w/ rather than of /f/ or some other sound? What's controlling that?

    (Or, since [v] actually does occur in Chinese, I could pose the same question with a sound that never does: how could I, studying Chinese from the future when no living speakers exist, predict that [θ] would be interpreted in that language as /s/ rather than /f/ or /t/?)

  9. John Walden said,

    December 22, 2015 @ 3:33 am

    Cities are a bit of a red herring, what with Munich and so on, but isn't Beijing an example of where official influence has had some effect? I do know that the story is more complicated than just that.

    Côte Ivoire seems intent on not being (The) Ivory Coast,

    Wikipedia: Therefore, in April 1986, the government declared Côte d'Ivoire (or, more fully, République de Côte d'Ivoire to be its formal name for the purposes of diplomatic protocol, and officially refuses to recognize or accept any translation from French to another language in its international dealings. Despite the Ivorian government's request, the English translation "Ivory Coast" (often "the Ivory Coast") is still frequently used in English, by various media outlets and publications.

    The "the" in (The) Lebanon and (The) Congo and (The) Gambia is another change in usage. I'm not even sure about the capitalisation of the "the". I suspect that the disappearance is a result of moving away from German and French usages and not any particular preference of the countries.

    I can vaguely remember "The Argentine" being used.

    Is there a general tendency to get closer to what a country's inhabitants call it? I write this from Spain though, where we cheerfully accept that Spain gets called whatever it is in other languages while also using Spanish for more than just countries: the whole British royal family gets transposed into Spanish (Carlos, Isabela, Andres, Guillermo and so on) while English seems to have stopped changing first names, for the most part.

  10. Bart said,

    December 22, 2015 @ 3:49 am

    I wish the government of 'the Czech republic' would tell the world to standardise on some sensible name for their country in English: Czechia for example.

  11. Bob Ladd said,

    December 22, 2015 @ 3:55 am

    I recall an article in the Economist some years ago about the pressure they were under to use Beijing, Myanmar, Chennai, etc. etc. if they wanted to continue trouble-free selling of their "newspaper" in the respective countries. So government pressure certainly affects these things. But governments care a lot more about what they get called in English than in other languages. The Dutch still call Serbia Servië, and the Italians still call Beijing Pechino, and I don't think it matters much to anyone official in Serbia and China.

    Independently or not, there's also a definite tendency for English speakers to care more about using the "right" name than speakers of many other languages, as John Walden notes. There was a long discussion about this on Language Log some years ago – can't find it now – in which a number of English speakers implied that it is somehow morally reprehensible to use an English name for a place that calls itself something else. I don't have Liberman-style data to back this up, but even without moral arguments it certainly seems that some anglicised place names are on the way out even without pressure from anyone official. Leghorn for Livorno is an obvious example.

    John Walden's example of royal names is certainly true in a number of European languages besides Spanish, but I think the only context in which all European languages follow the practice of nativising names involves popes (John Paul, Benedict, Francis, etc.).

  12. John Walden said,

    December 22, 2015 @ 4:18 am

    This may be special pleading but it's not really the pope's name is it? What is his name in the original language? And more to the point, what is the original language? Latin?

    Being whimsical, I suppose something is not your name unless you turn your head when someone says it out loud. I remember as a child thinking, when I was told that London was Londres in French, "But that's not its Name". It is an odd idea to say "Your name is Spain. Didn't you know?"

    But then I don't like being told what to do either. So Ivory Coast and Bombay they stay.

  13. неко said,

    December 22, 2015 @ 4:42 am

    This is probably off-topic, but this reminded me of something…

    In Serbo-Croatian, there's the word "Šiptar" which is an extremely derogatory word for an Albanian. The source Albanian word is "shqiptar" (it also means Albanian). However, some Serbs love to use it for Albanians even though the standard word is "Albanac", and when someone gets told to stop using "Šiptar" oftentimes the response is "but that's how Albanians call themselves", implying that no harm was intended. Which is about as true as saying that "Servi" is how Serbs call themselves. Altho Servi isn't a word in Serbo-Croatian, some non-Serbs might try to insult a Serb with that word (the similarity to the Latin root serv- is clearly apparent to both the person dishing out the insult and the one taking it).

  14. Victor Mair said,

    December 22, 2015 @ 5:14 am

    @Anschel Schaffer-Cohen

    What's the difference between my suspicion and "this assertion" in your statement? And where did you get "this assertion" anyway?

  15. J.W. Brewer said,

    December 22, 2015 @ 8:57 am

    If you look at the google books n-gram viewer, "Serbia" becomes more common than "Servia" in exactly 1914. Wartime was probably an especially propitious time for Anglophone publications and copy editors to be inclined to accede to the expressed wishes of an allied government. I don't know what role government preference had in the more complicated transition Roumania -> Rumania -> Romania.

  16. J.W. Brewer said,

    December 22, 2015 @ 9:04 am

    Anglophone publications flirted briefly after 1918 with the spellings "Jugoslavia" and "Jugo-Slavia" before settling down to Yugoslavia. Since the "indigenous" spelling is "Jugoslavija" (for Slovene and the Latin-scripted version of The Language Formerly Known As Serbo-Croatian) this would seem the opposite of the Servia/Serbia transition, unless for some reason the Yugoslav authorities affirmatively wanted the English spelling to be one that cued the correct pronunciation under English orthographic conventions.

  17. Mr Punch said,

    December 22, 2015 @ 9:52 am

    Back in 1976, an acquaintance of mine proposed that in honor of its bicentennial the United States of America should adopt an actual name — he suggested "Wyoming."

  18. James Wimberley said,

    December 22, 2015 @ 9:57 am

    Does the current government of Serbia object as I do to the adjective "Serbian" instead of the traditonal and more concise "Serb"?

  19. Adrian Bailey said,

    December 22, 2015 @ 10:27 am

    @James That's a discussion that's been had before on this blog. The two words suggest different things. Roughly speaking, "Serb" suggests ethnic group, "Serbian" suggests nationality.

    Of course in many situations there aren't two words available to use to make this distinction, so the country name is often used as an adjective. e.g. "Belgium footballer" rather than Belgian because you don't have to be Belgian to be Belgian or to play for Belgium.

  20. J. W. Brewer said,

    December 22, 2015 @ 10:37 am

    I would imagine these days most Flemings and most Walloons do not think of "Belgian" as an ethnic identity.

  21. Robert Coren said,

    December 22, 2015 @ 10:53 am

    @Bart: Yes, I too have hated the name "The Czech Republic" since it's invention. By the same (or a similar) token, if it had been up to me I would have called the country to Russia's southwest "Ukrainia".

    @Bob Ladd: I'm unaware of any tendency for English-speakers to say "Venezia", "Roma", "Napoli", "München", "Moskva", etc.

  22. Bob Ladd said,

    December 22, 2015 @ 11:30 am

    @Robert Coren: No, of course not. I just said that some of these forms are losing ground. Leghorn and Ratisbon are completely obsolete, as far as I know; Lyons and Marseilles with the superfluous s are losing ground (Google Maps says Lyon but still keeps Marseilles); Bombay and Peking have all but disappeared (and NB those were not name changes, like, say, Leningrad vs. St. Petersburg – Bombay was always Mumbai in Gujarati and Marathi, and Peking was always Beijing in Mandarin); and so on. (OK, OK, not "always"; but Beijing was the Mandarin version of that place name for many years before English stopped using Peking.)

    About the Czech Republic: the Germans have happily called it Tschechien (i.e. Czechia) since it separated from Slovakia. I don't know why English speakers didn't decide to do the same thing. English speakers who are actually familiar with the Czech Republic seem to refer to the place just as Czech, which I think mirrors a usage in Czech itself.

  23. D.O. said,

    December 22, 2015 @ 11:33 am


    It is absolutely certain that governments of relatively obscure countries (I define "obscure" as a country whose name is used overwhelmingly by expert class, not general population) can dictate how their name is rendered in English. I would imagine that an attempt by governments of Germany or India to change English names of their countries would be summarily laughed at.

  24. Michael Watts said,

    December 22, 2015 @ 11:47 am

    English speakers who are actually familiar with the Czech Republic seem to refer to the place just as Czech, which I think mirrors a usage in Czech itself.

    When I was in a Chinese class with a Czech guy, it worked out that English references to that country took the form "Czech". I had assumed that this was influenced by the Chinese usage, which is just jie-ke, but this would definitely explain it.

  25. David Marjanović said,

    December 22, 2015 @ 1:38 pm

    What I'd really like to know is, if I had to study Chinese without access to a native speaker who I could have attempt to pronounce foreign words for me, could I predict that [v] was a variety of /w/ rather than of /f/ or some other sound? What's controlling that?

    (Or, since [v] actually does occur in Chinese, I could pose the same question with a sound that never does: how could I, studying Chinese from the future when no living speakers exist, predict that [θ] would be interpreted in that language as /s/ rather than /f/ or /t/?)

    Such things may be altogether impossible to predict. All kinds of factors are involved, including (mis)interpretations of spellings and pronunciation traditions which may even be taken from other languages.

    Close to everyone with a heavy German accent will turn English [θ] into [s] (or [z] in most environments if they have that sound in their native inventory). Personally, and I'm not the only one, I find [f] much more similar, and German of course has a /f/. But I belong to the last generation that was at some point taught to pronounce it "like s, but with the tongue between your teeth" – and if you stick the "corners" of your tongue between your teeth, rather than the tip, the outcome will sound almost exactly like [s]. I soon learned better from hearing more English and from having better teachers later; earlier generations generally didn't have an opportunity for that.

    Or take /æ/. A French accent turns it into [a], obviously influenced by the spelling. A German accent never does that, even though the German /a/ and/or /aː/ is just about as front as the French one in most accents. Instead, people equate it with the letter ä, whose name is [æː] even outside the one geographically limited accent that retains an /æː/. The pronunciation of short ä is [ɛ], and that's what the English /æ/ tends to end up as. Hypercorrectivisms abound among those with a less strong German accent, so bed and bad can merge either way.

  26. oulenz said,

    December 22, 2015 @ 3:18 pm

    @Victor Mair: Your 'suspicion' was that English Servia > Serbia was due to the efforts by the Serbian government. The 'assertion' that @Anschel Schaffer-Cohen references was that "government intervention can radically influence language usage on a global scale". I believe all they meant to do was to contribute two independent examples to substantiate this assertion, given that your choice of words ('suspicion') left open the possibility of some other explanation in the particular case of Serbia.

  27. J. W. Brewer said,

    December 22, 2015 @ 3:59 pm

    Note also fwiw that there doesn't seem to be much movement toward using "Beograd" in English instead of the traditional "Belgrade." (The "L" that's missing in the local version of the name seems ubiquitous in other languages, including the closely-related Bulgarian/Macedonian, at least according to; maybe the Serbs themselves had it in an earlier version of the name but then mislaid it due to phonological change?)

  28. Andrew (not the same one) said,

    December 22, 2015 @ 5:00 pm

    D.O.: I would have thought that Persia was quite widely talked about before it required us to change its name to Iran.

  29. Michael Watts said,

    December 22, 2015 @ 5:58 pm

    David Marjanović:

    What you're talking about isn't exactly the same as what I meant to ask about. So, a hypothetical German person has formally learned that english /æ/ is german "ä", however the latter might be pronounced. That definitely happens, but there are other ways to get sound equivalences —

    A hypothetical Irishman might pronounce [θ] as [t]. I'd be willing to assume that's a learned equivalence similar to what you've described in German.

    A very young child still learning to speak English will typically pronounce [θ] as [f]. That's not because they don't know the difference, and it's also not because of any learned equivalence — they know perfectly well that there's supposed to be a difference, and they can hear it, but they haven't learned to pronounce it yet.

    And similarly, a hypothetical "Platonic Chinese"-speaking person with no knowledge of english will hear [θ] as /s/, and repeat it back as [s] if asked to repeat an auditory cue. I had a friend in China who was quite conversant in english, and obviously well aware that we thought the difference between "th" and "s" was important for some reason, but she couldn't hear that difference as pronounced by me (a native speaker of US english).

    That should be predictable, given sufficient knowledge of Platonic Chinese, because it's replicable between speakers who haven't been contaminated by a common cultural tradition, whereas the german/english examples you mention seem to be confounded that way.

    (As a possibly-related observation, I have trouble distinguishing /m/, /n/, and /ŋ/ as pronounced by Chinese people, and I find this particularly irritating because all of those sounds are separate phonemes in my own native language. Some of that is because a lot of Chinese people themselves don't distinguish syllable-final /n/ from /ŋ/, but I've experienced this with syllable-initial /m/ – /n/ too…)

  30. D.O. said,

    December 22, 2015 @ 7:31 pm

    Re: Iran. Apparently it is not working very well, because if Wikipedia is to be believed, now a sizable part of Iranian/Persian society wants its Western name to be changed back.

  31. Michael Watts said,

    December 23, 2015 @ 1:06 am

    What is the story behind the "true name" of Iran/Persia? I tried to look into it on wikipedia a while back and the only conclusion I felt I could reach was "wikipedia is not self-consistent".

    So… in we see

    In 1935 Reza Shah requested that foreign nations use the name Iran rather than Persia in official correspondence. The name of the country had internally been Iran since the time of the Sassanid Empire (224–651), whereas the name Persia is descended from Greek Persis (Περσίς), referring to a single province.

    One Greek-ish guy who had a lot of experience with Persia, and called it Persia, was Alexander the Great. He died in 323 BC, 500 years before the Sassanid Empire began. Were Greeks at the time just confusing the name of a province of some nameless foreign empire with the empire itself?

    Well… displays the Achaemenid Empire's name for itself as four cuneiform symbols, with the transliteration "Pārsa". It has since been the subject of an edit war, actually being named in Arabic — which was definitely not in use at the time — for a period, and the current page has turned bashful and declines to offer any such information at all. But it's certainly hard to trust. Successive edit wars have seen the claim that Parsa was the name of the empire, that the name of the empire was actually امپراطوری بزرگ هخامنشیان (!), that Parsa referred to "the area of the Persians", who were called parshu and who were in charge of the empire, but of course didn't refer to the empire itself, and, in the current version, that Parsa actually referred to the Persians themselves.

  32. Thor said,

    December 23, 2015 @ 1:14 am

    As a technical editor for various UN entities for three decades I have had to adapt to what has seemed at times to be mere whimsy on the part of representatives. Thus no "USA" — always spell out. Myanmar, not Burma; Cote d'Ivoire was an early change. Persia to Islamic Republic of Iran came later. The perceived symbolism behind official names has been intriguing, but one simply adds it to the stop-list of words to be confirmed before use, and moves on.

  33. Michael Watts said,

    December 23, 2015 @ 1:21 am

    We can also see, in :

    (1) I am Darius [Dâryavuš], the great king, king of kings, the king of Persia [Pârsa], the king of countries, the son of Hystaspes, the grandson of Arsames, the Achaemenid.

  34. неко said,

    December 23, 2015 @ 3:27 am

    @J. W. Brewer

    Yes, the old Serbian name was "Beligrad". The standard dialect went through L-vocalization and "beli" turned to "beo" in Serbian (> Beograd). Names like "Belgrade" (en), "Belgrad" (de) or "Belgrado" (it) are just remnants of the old name.

  35. Jon said,

    December 23, 2015 @ 5:24 am

    The modern (UK) world atlas that I have clearly has a policy that country names will be spelt as traditionally in English, and anything else will be spelt as in the local language. So no Copenhagen or Munich, except as a cross-reference in the index.
    I've also seen an EU map that went the whole hog, giving country names in local languages. I had to work out what Shqiperia was. Greece was given in Greek, with 'Hellas' underneath.

  36. mira said,

    December 23, 2015 @ 10:42 am

    @Bob Ladd: I think that you're thinking of the word "Čechy". This word is occasionally used by Czechs to refer to the whole country, but it technically just means Bohemia and as such only refers to one part of the country. "Česko" is used much more frequently, particularly in the media, but some people hate it — Havel said that whenever he heard "Česko" he felt like slugs were crawling on his skin. Various Czech politicians have been trying to promote "Czechia" — I've heard they're going to put it on the uniforms at the next Olympics. I support it in theory, but I think "Czechia" is hideous.

    I rather prefer "Czechland".

  37. Rolig said,

    December 23, 2015 @ 2:20 pm

    In the fall of 1990, I was working as a copyeditor at the Baltimore Sun newspaper when we were given the order that henceforth "Moldavia" would be "Moldova" and "Belarussia" "Belarus". While I was sympathetic with the latter change, since I appreciated that Belarusians woudn't want their country viewed as an extension of Russia, the Moldova change struck me as unnecessary: Moldavia was just the country's English name. We don't say Sverige or Italia or even Rossi(y)a, so why should we have to say Moldova? I suppose the change has something to do with the fact that part of Romania is also known as Moldavia, although even that part is called in Romanian "Moldova".

  38. David B Solnit said,

    December 23, 2015 @ 2:23 pm

    We will not mention Sevastopol/Sebastopol…

  39. ngage92 said,

    December 23, 2015 @ 5:05 pm

    The funny part about this is that Serbian has a V in places where you might expect a B, and where Croatian and Bosnian indeed have the B.

    Babylon -> Vavilon/Babilon
    Byzantium -> Vizantija/Bizantija

  40. J. W. Brewer said,

    December 23, 2015 @ 5:44 pm

    Since both "Bohemia" and "Moravia" are perfectly good toponyms of longstanding pedigree in English-language usage, a second-stage Velvet Divorce dismantling the country-in-search-of-a-plausible-name and making each of the Czech Lands sovereign (well, as sovereign as they can be within the EU) would solve the problem quite nicely. (Reunion of Moldavia/Moldova/Bessarabia with Roumania/Rumania/Romania in keeping with the post-WWI would eliminate another naming problem although perhaps it would cause other problems of an economic and/or geopolitical nature.)

  41. неко said,

    December 24, 2015 @ 8:00 am


    That's because of the different source for those loanwords. Western parts of Yugoslavia got the loanwords from Latin, eastern parts got those loanwords from Greek. The "v" is expected in Serbian.

    A similar split between western and eastern influence is in respect to cases like "chemistry", it's "hemija" ("хемија") in Serbian, but "kemija" in Croatian,

  42. FM said,

    December 27, 2015 @ 2:01 am

    So the Serbian national anthem (mentioned in the letter linked at the beginning of the post) strings the syllable "srp" over several notes. To sing this, one has to either trill the r for a long time, or add an epenthetic schwa, or maybe something else. What is the done thing in this case?

RSS feed for comments on this post