University of Lausanne 
Norwegians are rather fond of thinking of the situation in which the Norwegian language finds itself as being possibly ridiculous and certainly unusual. As far as the first of these perceptions is concerned, I have argued elsewhere (see, for example, Trudgill, 1995) that, far from being ridiculous, the Norwegian language situation is very sensible indeed, especially insofar as linguistic tolerance and linguistic democracy are concerned, and in many respects one that could be learnt from and even copied in many other parts of the world. In this paper, I tackle the second of these perceptions. I argue that, although the language situation in Norway is unique in some ways, it is not really, when viewed in a world-wide perspective, particularly bizarre either. It is easy, that is, to overemphasise the extent to which the sociolinguistics of modern Norwegian is strange. I outline below some of the ways in which Norwegian is sociolinguistically a perfectly normal language, and I point out that there are many other language situations around the world which demonstrate close parallels with the Norwegian situation.
Many people in Norway appear to regard the "artificial" construction of Nynorsk as being in some way peculiar, and point to the fact that it has no native speakers. In fact, this is a perfectly normal state of affairs, and it is true of many standardised varieties around the world that they are artificial constructs developed by language planners which, often, constitute a compromise of common ground between naturally-occurring spoken dialects.
There is, however, a related and perhaps more important respect in which the development of standard Nynorsk can be regarded as perfectly normal. One way in which a clearer understanding of the Norwegian language situation and its development can be gained lies in a consideration of what I have called "Ausbau sociolinguistics" (see Trudgill 1992a;1992b). This field of study depends on the distinction initially made by Heinz Kloss (1967) between what he termed "Abstand languages" or "languages by distance" and "Ausbau languages" or "languages by extension".
All linguists are very well aware that the notion of "a language" is only partly linguistic—as we tell all our beginning students, the question of whether a variety of language should be considered "a language" or "a dialect" (or something else again) is by no means entirely a linguistic decision. Cultural, social, political and historical factors may be very heavily involved also. Kloss, however, sheds further light on this issue by pointing out that some languages can indeed be considered separate languages on purely linguistic grounds—by reason of their linguistic distance from all other varieties. In the European context, Basque is often cited as a typical example of an Abstand language. It has no close linguistic relatives, and could never reasonably be regarded as being a dialect of any other language.
Most other languages, however—the Ausbau languages—have arisen out of dialect continuum situations and are considered separate languages for reasons that are by no means entirely linguistic. The classic western European example demonstrating the nature of Ausbau languages is provided by the distinction between German and Dutch, but very many other examples could be given such as the distinctions between Polish, Slovak and Czech; or Norwegian, Swedish and Danish; or Portuguese, Spanish and Catalan. In the Dutch/German case, we can say (see Chambers and Trudgill, 1997) that there is a continuum of West Germanic dialects above which have been raised over the centuries two separate standard languages, Dutch and German. We consider dialects from this continuum that are spoken in the Netherlands and Belgium to be dialects of Dutch, and dialects from the continuum spoken in Germany, Switzerland and Austria to be dialects of German—for entirely nonlinguistic reasons.
To use Haugen’s (1968) terms, dialects of Dutch are those dialects which demonstrate heteronomy to Standard Dutch: that is, they are dependent on Standard Dutch in the sense that their speakers look to Dutch as the standard language which naturally corresponds to their vernaculars, and they learn to use Standard Dutch for the purposes of reading and writing. Similarly, German dialects are those West Germanic dialects which are heteronomous with respect to Standard German, for the same kind of reasons. (The geographical boundary between dialects of Dutch and dialects of German is thus to be located precisely on the political frontier between The Netherlands and Germany.) The two standard languages, which are codified in that they have grammar books and dictionaries devoted to them, and which have bodies of literature written in them, are correspondingly to be regarded as autonomous—they have, we can say, a form of independent existence. We are thus able to give an Ausbau definition of what a language is: a language is an autonomous standardised variety together with all the nonstandard varieties which demonstrate heteronomy towards it.
This definition makes it clear that languages are not only cultural constructs but that they are also potentially temporary constructs. Because autonomy is a political and cultural rather than linguistic phenomenon, varieties can gain and lose autonomy. Languages can become dialects, and dialects can become languages. Many such examples present themselves. Scots, Plattdeutsch and Provencal are all examples of languages which have become dialects: they have lost their autonomy and become heteronomous with respect to other languages (in these cases English, German, and French respectively). Other languages have acquired autonomy: Afrikaans and Macedonian are both dialects (of Dutch and Bulgarian—or perhaps Serbian—respectively) which have become languages: they are Ausbau languages which have been extended or constructed or developed out of former dialects.
Two aspects of this issue are relevant to a discussion of the Norwegian situation. First, we have made much of the distinction between Ausbau and Abstand languages. However, this distinction is something of a simplification. For one thing, obviously, Abstand or linguistic distance is a phenomenon which admits of degrees of more or less. For another thing, although we have suggested that linguistic characteristics are much more important for defining Abstand languages than they are for Ausbau languages, we have to note that there is an interesting relationship between the two. Where deliberate attempts are made to destroy language status by attacking the autonomy of standard varieties—often for political reasons—linguistic distance will usually be a consideration. Thus, in the centralising fascist Spain of the 1930s, the dictator Franco attacked the autonomy of Catalan by abolishing the Chairs of Catalan Language and Literature at Barcelona University and prohibiting publication in that language, but it was also was possible for his government to argue that Catalan was "not really a language" but simply a "dialect of Spanish" because of the clear linguistic similarities between these two neighbours on the geographical dialect continuum.
Conversely, and more importantly for our purposes, we have all heard the maxim that "a language is a dialect with an army and a navy", but in actual fact an army and a navy are usually not sufficient (they may well not necessarily be essential either, even for Ausbau languages, as the example of the revival of Catalan shows, but that is another matter). Deliberate attempts to turn dialects into languages will therefore normally have to involve the maximising of what linguistic Abstand there is, as well as the development of other devices designed to accentuate individuality, such as new orthographies or even different alphabets. For example, Standard Macedonian was to a considerable extent developed, by language planning processes, out of those dialects on the South Slavic dialect continuum which were geographically, and therefore also linguistically, furthest removed from Bulgarian and Serbo-Croat—the dialects of southwestern Macedonia. This was not a coincidence. If you want to claim that your dialect is in fact a language, then there is no doubt that you have to go for the trappings of autonomy such as dictionaries, grammar books, an orthography, and, crucially, a name for your language. But there is also no doubt that the more linguistic distance you can put between your language and its neighbours on the dialect continuum the better. That way, you minimise the risk of others claiming that you language is still "really" a dialect of some other language.
Here, in the Norwegian context, we can note the special focussing by Ivar Aasen on western dialects of Norwegian. This is often presented as a concentration on those dialects which he regarded as being the most "pure" or "uncorrupted" varieties of the language. But of course one of the things that Aasen was trying to do was to secure autonomy for Norwegian, which meant independence from Danish to parallel the independence of the nation from Denmark. So, just as Macedonian has gained its independence from Serbo-Croat and Bulgarian by stressing what degree of Abstand was available to it, so Aasen selected precisely those Norwegian dialects which were furthest removed on the dialect continuum from Danish (as well as from Swedish), and were therefore linguistically most unlike it, to act as the basis for his new standardised and, eventually, autonomous Norwegian. This was only sensible, and entirely normal.
Another factor which strikes many people as being unusual about the Norwegian situation is that Aasen’s attempt to achieve autonomy for Norwegian was not the only attempt. The result of this, of course, was that Norwegian now comes in two distinct standard forms: Nynorsk, the descendant of the variety with the most Abstand; and Bokmal, which is much more like Danish but which has now acquired enough grammatical, lexical and orthographical differences from Danish to be regarded, except perhaps by certain Nynorsk supporters, as being Norwegian and not Danish.
It is important to appreciate that even this situation is not unique, however. First, it is not unique from a historical perspective. Norwegian was not the only European language where there were two competing solutions to the establishment of a standard upon the achievement of political independence by the nation. The development of modern Standard Greek was also characterised by lack of agreement as to what route it should follow. In Greece, too, the two competing solutions involved one variety which was based on vernacular dialects, Demotic, and another which was much more elitist, Katharevousa. Here, though (see Jahr and Trudgill, 1993), Abstand and Ausbau were not an issue, since Greek is not an Ausbau language and has no linguistic relationship at all to Turkish, the language of the former colonial power. In recent years, moreover, the competition has been resolved, in that Demotic Greek has replaced Katharevousa or puristic Greek in nearly all contexts. In highly democratic Norway, room has been made for both solutions to survive.
Secondly, moreover, we have to observe synchronically that there is nothing very unusual about having a single language with more than one standard. One way of regarding the Norwegian situation is to say that Norwegian is a single language with two standards. At this point we have to notice that autonomy, too, is a complex phenomenon which, like Abstand, admits of degrees of more or less. For modern Norwegian, we can say that there is a kind of shared autonomy, with Bokmal and Nynorsk both being autonomous varieties raised above exactly the same dialect continuum, although as we have seen it is possible to relate Nynorsk especially to western Norway (and Bokmal, perhaps, particularly to southeastern Norway).
Such complexities involving autonomy are not all that unusual. We can point out for example that the different forms of Standard English that we find in the English-speaking world, involving differences at all linguistic levels as well as in the orthography, constitute a problem for the notion of autonomy. Some of these standard varieties are clearly more equal than others, with Australian, New Zealand, South African, Caribbean, Scottish and Irish Standard English lagging some way behind English and American Standard English in the autonomy stakes. But it is clear that English, as a so-called polycentric language, has, like Norwegian, more than one standard form. Similar examples are provided by Spanish, Portuguese, Dutch and a number of other languages.
The English situation can perhaps be best described as representing cases of semi-autonomy for the different national varieties. This is characterised by somewhat different standardised varieties of what everybody regards as being the same language corresponding to different sets of nonstandard dialects in the different countries. It will perhaps be objected, however, that the strange thing about the Norwegian situation is that, unlike in the case of English, there are two standards which share autonomy in the same country. There is much strength to this argument. The situations are different, and we cannot therefore be comfortable about referring to Norwegian as being a polycentric language.
We also need clearer examples to show that Norwegian is not unique. Here are two. Hindi and Urdu are often regarded as being different languages. A more realistic way of regarding them, however, is as the same language, sometimes referred to as Hindustani, with two different autonomous standards. And although Urdu alone is the official language of Pakistan, in India both Hindi and Urdu have official status. To ask, in India, if someone is speaking Hindi or Urdu is rather like asking if someone in Norway is speaking Bokmal or Nynorsk: it is most often a question without an answer. At the colloquial, spoken level, the two "languages" are indistinguishable. It is only in writing that they differ, with Hindi being written in the Devanagari script, and Urdu in the Arabic script. At the level of vocabulary, too, they differ, to the same kind of extent as Nynorsk and Bokmal, with Hindi taking its learned lexis from Sanskrit, and Urdu from Arabic and Persian.
Serbo-Croat provides another excellent example. In the former Yugoslavia, the main national language was known by this name, with the hyphenated English form of the language showing very clearly that we are dealing with a single language with two different norms. As with Hindi/Urdu, there are differences of alphabet between the two, with the Croatian variant tending to use the Latin alphabet and the Serbian variant the Cyrillic alphabet. There are, too, phonological, grammatical and lexical differences. As with Norwegian, there is a geographical component to the differentiation, with the Croatian variant relating more to western dialects and the Serbian to eastern dialects. Again as with Norwegian, however, it was a good deal more complicated in the former Yugoslavia than that since, as with Hindi/Urdu, which variant speakers perceived themselves as using, and which variant they chose to employ in writing, might have had as much to do with ethnic and religious identity as with anything linguistic. Thus a question as to whether someone was speaking Serbian or Croatian might be answerable on linguistic grounds in the far west or far east of the country, but was more likely to be answered on ethnic grounds in the centre.
Since the break-up of Yugoslavia the situation has become more complex, with attempts by the governments of Croatia and Serbia to claim that there are in fact two separate languages. These attempts, moreover, have made life difficult in the geographical centre for Bosnian Moslems (see Trudgill, 1995). While some Croats in Sarajevo might be happy to say that they speak Croatian, and some Serbs in the same town who speak exactly the same dialect might be happy to say they speak Serbian, it is not clear what Moslems should say they speak if they are denied the use, as the Croatian and Serbian governments would wish, of the term Serbo-Croat. It is therefore not at all surprising that the government of Bosnia has now let it be known that they wish their language to be called Bosnian. It is therefore very likely that the language it still seems sensible on linguistic grounds to call Serbo-Croat will in the near future acquire three rather than two standardised variants.
Moslem users of Urdu and Hindu users of Hindi do not necessarily always take kindly to the suggestion that these are one and the same language, just as the use of the term Serbo-Croat has become absolute anathema to nationalist Croats and Serbs. Similarly, although happily in much milder form, some Norwegians object to the suggestion that Nynorsk and Bokmal are not separate languages but simply variants of the same language—important issues of group and personal identity are involved here.
But the dispassionate outside observer can, I would suggest, be entirely sanguine about arguing that what we have here, in all three cases, are single languages with shared autonomy. We can refer to such languages as multinorm languages. Recall our Ausbau definition above of a language: a language consists of an autonomous standard variety and the nonstandard dialects which are dependent upon it. For there to be two languages, therefore, we need not only two autonomous standards, but also two sets of dialects. German and Dutch are two separate autonomous standard varieties which correspond to two different parts of a dialect continuum. They therefore constitute two separate languages. It is true that in the Serbo-Croaat case, as in the Bokmal/Nynorsk and Hindi/Urdu cases, we also have two (or more) standard varieties. But, crucially, we have only one set of dialects that depend on them. We therefore have only one language. And, unlike German or Dutch speakers, speakers of these languages have a choice of standards
This means that we now have to modify our definition of a language to read "a language consists of one or more autonomous standard varieties together with those nonstandard varieties which are dependent on it or them". And this also provides us with a perfect definition of a multinorm language, for which Norwegian provides an excellent and perfectly normal example: a language where there are two or more standardised forms corresponding to a single set of nonstandard dialects.
Chambers, J.K., and Peter Trudgill 1997Dialectology Cambridge: Cambridge University Press, 2nd edition
Haugen, Einar 1966The Scandinavian languages as cultural artifacts. In J. Fishman et al. (eds) Language problems in developing nations. New York: Wiley.
Jahr, Ernst Hakon, and Peter Trudgill 1993Parallels and differences in the linguistic development of Modern Greek and Modern Norwegian. In E. H. Jahr (ed.), Language conflict and language planning. Berlin: Mouton de Gruyter. pp. 83–98.
Kloss, Heinz 1967Abstand-languages and Ausbau-languages. Anthropological linguistics 9, 29–41.
Trudgill, Peter 1992aThe Ausbau sociolinguistics of Greek as a majority and minority language. In M. Makri-Tsilipakou (ed.), Proceedings of the 6th International Symposium on the Description and/or Comparison of English and Greek. Thessaloniki: Aristotle University. pp. 213–35.
Trudgill, Peter 1992bAusbau sociolinguistics and the perception of language status in contemporary Europe. International Journal of Applied Linguistics vol. 2.2: 167–77.
Trudgill, Peter 1995Sociolinguistics: an introduction to language and society. London: Penguin, 3rd edition.
 Peter Trudgill is now Professor of English Linguistics at Norwich. We wish to thank Professor Trudgill for giving us permission to publish this article on the web.
Språkrådet | Tlf.: +47 22 54 19 50 | E–post: email@example.com