John: Thanks for all of your diligent work!
In hindsight, I think there’s a fair amount of agreement that it was a mistake to encode Bosnian etc. as separate languages. This set of languages is not the only case in which ISO 639 has given multiple encodings for a single linguistic entity. One other case is Akan (ak/aka) / Fanti (fat) / Twi (tw/twi). Yet another case is Moldovan vs. Rumanian: these were coded separately in ISO 639-1 1st Edition as mo and ro, and in ISO 639-2 1st Edition as mol and rum/ron. But in 2008, the JAC decided to deprecate mo / mol. One might be tempted to consider whether we could do the same for bs/bos, hr/hrv and sr/srp, recommending instead use of sh “Serbo-Croatian”. Unfortunately, it may have worked c. 2000, but after this long I think that would be a pretty disruptive and costly change.
I read with interest the “Montengrin-European view” document. I’m curious about these statements:
“Nevertheless, for political reasons German government translators and interpreters e. g. need to use Montenegrin when interpreting/translating for a Montenegrin audience. Texts that go to all four countries must be produced in four different varieties, one for each of the four countries. Montenegrin, Serbian, Croation and Bosnian are also kept apart in translation memories or termbases.”
What does it mean to say “need to use Montenegrin when interpreting/translating for a Montenegrin audience”? Apart from any orthographic differences (which ISO 639 explicitly does not encode), will the utterances/content produced actually be different? Or will the only difference be the declaration of the target-language identity?
“The main problem is that users need a means of treating Montenegrin, Serbian, Croatian and Bosnian in a systematic and uniform way…”
This goes on to discuss the identifiers themselves. But who are the users in question and what’s the usage environment? Why would they care about the coded representations? Are they interacting directly with coded representations rather than user-interface display names? (In my industry, people that deal directly with the encoded representations would face enough other special cases that they would not have a problem with lack of uniformity.)
“If no standardised identifiers are provided to fill this gap, users will start making up their own identifiers according to their needs.”
Users already can use “local-use” identifiers in the range qaa-qtz. There’s an analogy here in the realm of ISO 3166-1 region identifiers in the case of Kosovo: it is not coded in ISO 3166-1 as distinct from Serbia, but many European agencies that have real or perceived needs for a distinct identifier have adopted the private-use identifier “XK”.
“But language selection lists with elements such as the ones below are unlikely to be widely accepted. In some cases, technical reasons might even stand against the use of a Montenegrin identifier that exceeds the typical numer of two or three characters.”
Regarding language selection lists, I again wonder: who are the users, and why would they care about the coded representation?
Btw, in order not to give any perception of preference between Serbia and Montenegro, I would have coded Serbian as sr-RS, so that the two have a uniform representation.
As for the technical limitation cited, this same argument has been used in the past for why languages coded only with an alpha-3 ID need to get an alpha-2 ID. For the same reasons as applied to those cases, I think we should not base any decisions on such technical limitations.
I want to thank Melinda for opening up this discussion. I have received quite a few requests recently from Montenegro asking that the JAC reconsider its 2010 decision to not grant Montenegrin its own language code.
The National Library of Montenegro sent me a formal request to define a new language code for Montenegrin in ISO 639-2 along with a linguistic justification document, both of which Melinda attached in her recent message to the JAC.
I sought the advice of internal (within LC) and external Slavic language experts by asking them to review the documents and asking them to consider whether there had been any further development in the Montenegrin language since 2010 to distinguish it from Serbian and to justify the definition of a new code. They all responded that, linguistically, Serbian and Montenegrin still remain virtually identical, despite Montenegro’s renewed efforts to delineate their linguistic differences in the new documents. I have provided a partial thread of the discussion I had with some of these experts (See below). It highlights the most persuasive arguments for not defining a new code for Montenegrin, based on linguistic differences alone.
The discussion led me to the following conclusion: Serbian and Montenegrin are the same language and creating a code for Montenegrin would necessarily be for political reasons. That is outside the scope of the JAC, as far as I understand it. While it is unfortunate that there are separate codes for Serbian/Bosnian/Croatian, as Ronelle Alexander points out (see below), there is more to that distinction, and getting involved further into this political thorniness could cause more complications. It is not the JAC’s responsibility to define a language because of national identity. For these reasons LC is unlikely to vote in favor of creating a new code for Montenegrin. Of course, I welcome a discussion on this topic and look forward to reading what others in the JAC have to say on the matter.
Library of Congress
Network Development & MARC Standards Office
101 Independence Ave., S.E.
Washington, DC 20540-4402
COMMENTS FROM WITHIN LC:
1) Comment from Head of the Southeast Europe Section / Germanic and Slavic Division
John and colleagues,
As a person with an MS degree in literature and linguistics and as a native speaker of Serbian, who lived in Serbia until the age of 16, and who was, from an early age, exposed to Montenegrin literature, Television Broadcasting from Podgorica (capital of Montenegro), and Montenegrins, who lived among the Serbs, I cannot, in clear conscience support this clearly politically motivated claim that Montenegrin is a separate language from Serbian. The grammatical structures are the same and the vocabularies are about 99.9% the same. A dialect yes, but a separate language no, no more than American English is a separate language from British, Canadian, or Australian English. Separate users and separate states (as indicated below) do not make them separate languages. They do not have separate orthographies, as Serbian is written in both Cyrillic and Latin alphabets, despite the recent attempts by the government of the Republic of Serbia to codify Cyrillic alphabet as the only alphabet for the Serbian language.
The passage on p. 6, of the attached document, HISTORICAL CONTEXT AND BASIC GRAMMATICAL DIFFERENCES BETWEEN MONTENEGRIN AND SERBIAN LANGUAGES, by the Institute for Standardization of Montenegro, itself, appears to argue that all the four Stokavian “languages” are the same (see the highlighted text). In fact, the whole paragraph argues that if the other three “languages” were granted independent status, then so should Montenegrin. Could this be a good enough reasoning to declare a dialect an independent language? And if yes, where does it stop?
If one would want to answer the question of whether Montenegrin language is a separate language or a standard version of the Serbian language, as it can sometimes be heard, the same question would have to be asked in the case of other three Štokavian standard languages as well – Bosnian, Croatian and Serbian. All these four standard languages are languages of the Štokavian system, standardized on the basis of separate organic speech patterns of the Štokavian system. All these four languages have separate historical development, a separate literacy, a separate literary and linguistic tradition, separate users, separate states, separate orthographies and, finally, separate language standards. The question of the existence of a separate Montenegrin standard language is therefore no different in this respect from the question of existence of the other three Štokavian standard languages, just as they do not differ among themselves. Therefore, questioning its existence cannot be linguistically motivated. In the particular case, the questioning of its existence is ideologically motivated.
The last sentence above is worth examining further, as it contradicts the argument made, rather than supporting it.
closely examining the bibliographic entries, submitted as “Evidence of sufficient number of documents to establish separate code per ISO 639-2 Annex A A.2.1,” and can assure you that most of the titles listed are in standard Serbian, with topics on a wide range of subjects, such as Montenegrin literature, politics, historiography, law, religion, etc.
A distinguished Croatian linguist, Snjezana Kordic, has written a book titled Jezik I nacionalizam (Language and Nationalism), in which she addresses the problem of subjugating languages to nationalistic ideologies, as they have manifested themselves in former Yugoslav republics. To be recognized as legitimate nations, she argues, one does not have to have its own langue. Examples are all around us, Austria and Switzerland (which still call German German). And let’s not forget Spanish and Portuguese, spoken on different continents.
In other words, let’s be rational, and recognize that a language has the right to exist as an entity independent of a current political ideology.
Thank you for giving me an opportunity to contribute to this discussion.
2) Comment from LC specialist staff in the European Division of the Collections and Services Directorate:
I also don’t think there is a *linguistic* reason to consider Montenegrin a separate language – but there is a political reason from their point of view, to separate themselves from the Serbs after independence and to follow the examples of Bosnian, Croatian, and Serbian, each now with their own codes and subject headings agreed to by LC. The old saw, “a language is a dialect with an army,” applies (although perhaps not at ISO), and isn’t Montenegro about to join NATO?
Nevertheless, in my [limited] experience, Montenegrin in its written, standard, literary form as found in books being cataloged is no more different from Serbian than American English is from British English (or a number of other Englishes), with various orthographic (center/centre; honor/honour; curb/kerb; program/programme; hodge-podge/hotchpotch, and many others) and pronunciation (mándatory/mandátory; cóntroversy/contróversy, etc.) differences. We Americans speaking English are no more different than the Montenegrins speaking Serbian.
There are some errors in the “Historical Context….” document they sent. Just a few – page 1, first paragraph: many Croatian dialects are not štokavian at all but rather kajkavian (central Croatia, including the capital of Zagreb) or čakavian (Dalmatian coast and Adriatic islands). Croatia voluntarily agreed in the 19th century to accept the Vukovian (Bosnian/Serb) standard as their literary language over their own earlier standards based on the kajkavian and čakavian dialects (LC subject headings: Kajkavian dialect; Cakavian dialect. Language code: hrv = Croatian). There is still some publishing in these dialects (poetry, folklore texts, proverbs, etc.).
Page 6, final paragraph: “Serbian language uses only the Cyrillic alphabet.” Not true—publishing in Serbia in Serbian happens in both Cyrillic and Roman alphabets – as shown two pages later in their document, which gives both alphabets as used for *Serbian* (see page 8),
COMMENT FROM EXTERNAL EXPERT:
Dear Mr. Zagas,
Thank you for writing with this interesting conundrum. I have read through the materials you provide and pondered over them for a few days. Having done so, I would advise that you remain with the previous position and not grant a new code for the Montenegrin language. Your business, after all, is clarity of classification and not the conferring of political recognition (though even if it were the latter I would still have some reservations).
There is no doubt that Montenegro is now an independent nation and that its inhabitants consider themselves Montenegrins and not Serbs. But the language is, in form and content, still essentially Serbian: adding new letters to the alphabet to mark a pronunciation distinction that is slighter than those which differentiate American and British English is not enough to make it a separate language. A further argument in my mind is that despite the energetic activities on the part of language codifiers, the fact of Montenegrin as a separate language is still not uniformly embraced by the population: many Montenegrins, while in no doubt of their Montenegrin identity, still consider the language they speak to be Serbian.
The one statement in their proposal that does give me pause is the observation that the štokavian base has been widely recognized as being more or less uniform throughout the region encompassed by Serbia, Croatia, Bosnia-Herzegovina and Montenegro, and that if international recognition has been accorded to three separate instantiations of this base, then it should be accorded to a fourth, given that each now represents a separate state. This is a well-reasoned argument and it does bring up some very basic issues. But though it looks simple on the surface, to act on this would bring one face to face with some very thorny political issues, since the proclamation of the three other languages is more or less a direct result of war; in particular, the recognition of Bosnian as a separate language is tied up also with official documents such as the Dayton accords and judgments rendered by the ICTY. The Library of Congress as an institution should not (indeed, it cannot) get involved at this level.
For all these reasons, I think you are justified in remaining with your current decision, especially since there is a precedent you can rely on (as in your example of the same code functioning for both Moldovan and Romanian).
I hope these comments are helpful to you.
Professor of Slavic Languages and Literatures
Department of Slavic Languages and Literatures
University of California, Berkeley
Initial email to the National Library of Montenegro from LC (expressing LC’s cataloging policy concerns) which led to them resubmitting their formal request for a code, along with their linguistic analysis/justification document:
(we subsequently received a response that they were not happy with the proposed solution of having Montenegrin be a variant language assigned to the Serbian code)
Dear Ms. Djurovic,
On behalf of the ISO 639 Joint Advisory Committee and as newly appointed Committee Chair, I would like to apologize for the delay in processing your request for a new code. In the intervening time since we initially received your request we have had two sudden resignations at different times due to health reasons (including the resignation of the Committee Chair at one point), and, unfortunately, the JAC was unable to conduct any business during those times of vacancy. Fortunately, with the arrival of the new year, the Committee has now been fully appointed and is ready to resume its work, with your request being its first item of business.
Before and in-between the vacancies, I had started to gather some opinions and questions for your consideration, both from JAC members and from my colleagues at the Library of Congress. As you may know, in a vote conducted in July 2010 the ISO 639 JAC decided to not create a code for the Montenegrin language because the members agreed that it was virtually identical to Serbian. Today, among my colleagues, there still seems to be a wide opinion that there is no significant linguistic difference between Serbian and Montenegrin. Before we attempt to put this out for a vote again, I was asked to get back to you so that you could explain how people in Montenegro can tell whether text is in Serbian vs. Montenegrin. Without a clear dividing principle, this request will just confuse the language code situation. In the library context, the people applying the codes need guidance on how you tell the differences between languages. It will create problems for the IT and library worlds if the assumption is "books published in Montenegro are in Montenegrin" or "text from Montenegro is Montenegrin". Determining the language of an individual work could come down to using the place of publication: if it's published in Montenegro it must be in Montenegrin, and if it's published in Serbia it must be in Serbian. That's a dangerous practice, because you could have a Serb whose work is published in Montenegro and may object to his language being identified as Montenegrin. We want to prevent disruption of cataloging practice and policy and ensure stability and continuity in the code set. Therefore, any clarification you could provide about a dividing principle or guideline you use to make a determination between Montenegrin and Serbian would be greatly appreciated.
We have received requests before that present this type of situation, where a language spoken in 2 different countries is mutually intelligible but has a different language name associated with it in each of those countries. We handle these situations by assigning variant language names but using the same identifier. An example of this in ISO 639-2 is Moldovan and Romanian, which are both coded [rum]; or Dutch and Flemish, which are both coded [dut], etc. A possible solution, if you are amenable to it, would be to add Montenegrin as a variant language to the current identifier for Serbian [srp]. It should be noted that the language codes in ISO 639-2 were developed to serve as a device to identify a language or group of languages. They were NOT intended to serve as abbreviations or short forms for languages, but rather as a code that serves as a device to identify a language name. In using the language codes, systems generally display the language name represented by the code and not the code itself to users. Therefore it becomes irrelevant whether the code is "123", "xyz", "srp", or whatever. If you think this may be a viable solution, please let me know. Such a solution may have the benefit of being adopted quickly by the JAC. Unfortunately, I cannot guarantee that a second vote to establish a code for Montenegrin will pass in the JAC, given the fairly recent outcome of the last vote in 2010.
I very much welcome your comments and eagerly await your reply.
Please do not hesitate to ask me any questions.
Library of Congress
Network Development & MARC Standards Office
101 Independence Ave., S.E.
Washington, DC 20540-4402
E-Mail: [log in to unmask]