Review Article: English Influence on the Scandinavian Languages

The development of English is truly remarkable. 400 years ago it was spoken by a mere 4 to 5 million people in a limited geographical area. Now it is the native language of several hundred million people in many parts of the world. It is a second language in many countries and is studied as a foreign language in every corner of the world. English is a global language, to quote from the tide of a recent book by David Crystal (1997). The role of English is frequently debated. English has been described as a murder language, threatening the existence of local languages. The spread of English was described as linguistic imperialism in a book by Robert Phillipson (1992). The role of English is a topic frequently raised in the press. 40 years ago the readers of Dagbladet in Norway could read the following statement by a well-known publisher (quoted in translation):


Introduction
The development of English is truly remarkable.400 years ago it was spoken by a mere 4 to 5 million people in a limited geographical area.Now it is the native language of several hundred million people in many parts of the world.It is a second language in many countries and is studied as a foreign language in every corner of the world.English is a global language, to quote from the tide of a recent book by David Crystal (1997).
The role of English is frequently debated.English has been described as a murder language, threatening the existence of local languages.The spread of English was described as linguistic imperialism in a book by Robert Phillipson (1992).The role of English is a topic frequently raised in the press.40 years ago the readers of Dagbladet in Norway could read the following statement by a well-known publisher (quoted in translation): The small language communities today are in danger of being absorbed by the large ones.Perhaps in ten years English will have won the day in Iceland, in thirty years in Norway.
In Norway there have been several campaigns against unwanted English influence and for the protection of the linguistic environment (språklig miljøvern).A couple of years ago there appeared a book in Denmark with the title: Engelsk eller ikke engelsk?That is the question (Davidsen-Nielsen et al. 1999).A plan for the protection of Swedish has recently been worked out by Svenska språknämnden: handlingsplan för det svenska språket.Similar work is going on in Norway under the auspices of Norsk språkråd.
Considering the frequency and the heat of debate, there have been surprisingly few major studies that systematically survey the use of English in Scandinavia and the influence of English on the Scandinavian languages, although there has been a tendency in recent years to pay more attention to the topic.The first major study was Aasta Stene's (1945) thesis on English loanwords in Norwegian, based on material collected before the war.It took half a century before the next major works on English influence on Norwegian appeared: Anne-Line Graedler's (1998)

thesis on
Morphological, semantic and functional aspects of English lexical borrowings in Norwegian and the Anglicism dictionary by Graedler and Johansson (1997).Both focus on direct loans in written texts.The situation in Norway is discussed in a new book by Johansson & Graedler (2002).English influence on Danish has been studied particularly by Knud Sørensen, who has published two monographs on English borrowings in Danish (1973Danish ( , 1995) ) and an Anglicism dictionary (1997), which differs in important respects from its Norwegian counterpart, notably in paying a great deal of attention to translation loans and semantic loans.Like the Norwegian studies mentioned above, Sørensen's work is chiefly concerned with the written language.The most significant recent contribution is Bent Preisler's (1999) book on Danskerne og det engelske sprog, which reports on a large-scale sociolinguistic survey of the use of English and attitudes to English and also includes an in-depth study of the role of English in five subcultures where the English language was assumed to be an important defining element.(Preisler 1999: 231-232) According to Preisler, the strong position of English in Denmark is the result of a combination of forces from below {nedenfra), through the subcultures, and forces from above (ovenfra), through the educational system, business, and attitudes among the population in general.
The first comprehensive investigation of the role of English in Sweden was initiated by Magnus Ljung in the 1980s.The most important part of his "English in Sweden" project was a sociolinguistic survey of attitudes to English and to English influence on Swedish.There were a number of interesting findings: differences in reaction due to age, sex, and region; differences as regards the type of borrowing, etc. (see Ljung 1985Ljung , 1988)).Not surprisingly, age turned out to be the strongest factor; English is favoured particularly by the young.Another major study is the thesis by Judith-Ann Crystal on Engelskan i svensk dagspress (1988).This is different in many respects from Ljung's investigation.It is based on printed texts, and it deals with loanwords only, and only with direct loans; semantic loans, translation loans, and the like are excluded.
But it is a highly interesting study of the extent to which English loanwords are used and the factors which determine their integration into Swedish.As for Norwegian and Danish, there is also a dictionary of English loanwords in Swedish, produced by Bo Selten (the latest version published in 1993).
In addition to the works mentioned above, there have been more detailed investigations, such as Barbro Söderberg's (1983) monograph on the ^-plural in Swedish, and a host of minor studies.Altogether, we now have much better documentation on the role of English in Scandinavia than a couple of decades ago.There is one important gap, however.Until the publication of the new book by Harriet Sharp on English in spoken Swedish, which sparked the present review article, we had very little documentation on the influence of English in speech.

English influence in speech
The study of speech has been greatly facilitated in recent years by the access to computer corpora of transcribed speech.For English we have a number of available corpora, but for her work on spoken Swedish Harriet Sharp had to compile her own corpora.She decided to focus on two domains where "the speakers would be prone to use a considerable number of English lexical items as their contact with the English language is intense, their English language proficiency is good, and their attitudes towards its use might be presumed to be positive" (p.35).These points should be kept in mind in evaluating the results of the investigation.

Corpora and research questions
The two discourse domains examined by Harriet Sharp are: conversation of business meetings and casual conversation of young adults.For each there is a corpus: The Cool Corpus: business meetings from Cool Carriers, where the author worked in the 1980s.Video recordings of 12 meetings, a core group of 12 speakers (including 3 women), aged 30-54.Approx.6 hours.Total number of words: approx.72,000.
The Visby Corpus: casual conversation of young adults.From a docusoap series recorded for television The Real World -Visby.34 tapes selected from 800 video tapes.7 young adults, aged 19-25 (4 men, 3 women, including one Norwegian!).Approx.10 hours (after non-verbal material had been excluded).Total number of words: approx.76,000.
The two corpora will be referred to below as the COOL corpus and the VISBY corpus, respectively.
The whole material has not been transcribed, but rather "all periods containing embedded code-switches to English" (p.44); incidentally, the term 'period' is not defined explicitly, but this does not seem to have hampered the investigation.A code-switch is defined with reference to six criteria (p.52): 1.They [the English expressions] can be found in, or are derived from, the lexical items entered in a comprehensive dictionary of the English language, 2. they have their etymological roots in the English language and are believed to have entered Swedish after 1850, 3. or they originate in other languages but have been imported from English, 4. they are used by a Swedish speaker, 5. they may be in the form of an English proper noun or name, 6. and they constitute an English code-switch of any length (number of component elements).
We notice immediately that 'code-switch' is used here in a very broad sense, and we shall come back to this point later.
The transcription is basically orthographic, with ordinary punctuation, but it includes some extra features, such as: pauses, indications of special emphasis, and notes on other features which might be important for the interpretation of the text.The transcription includes a time column which makes it possible to calculate the time between code-switches.The author has also estimated the total number of words in the two corpora.
The research questions are defined as follows in the opening chapter (p.3): 1. How frequently is English used in the corpora? 2. What are the formal and discourse functional characteristics of the English expressions used?
3. How are these expressions used from a code-switching perspective?
4. To what extent are the English expressions integrated in the Swedish discourse?
We turn now to the first question.

Frequencies
Estimating the frequency of English elements is far from easy.Do we include all words regardless of their age and regardless whether they are completely established and do not differ from native words (e.g.Swedish jobb)} Do we include all types of borrowings/code-switches? Do we include words which have been borrowed from English, but derive originally from other languages (e.g.Swedish jungel through English jungle from Hindi jangal)?What do we do with international words like Swedish frustration and signifikant from English frustration and significant!And, not least, what types of texts are we talking about?We may get different results depending upon how we answer these questions.
Most frequency estimates focus on direct loans and are limited to relatively new words, e.g. in the case of Chrystal (1988) words that were not recorded in Svenska Akademins ordlista (9* edition, 1950).Chrystal found that there were on average 2.5 English words per thousand words of running Swedish newspaper text.There was a marked difference between advertising and ordinary newspaper text: on average 5.9 vs. 1.3 words per thousand.
Other studies agree in finding a surprisingly low frequency of English elements.The main variable is the type of text.These are some results from studies of English elements in Norwegian, measured in words per thousand (quoted from Johansson and Graedler 2002) At the outset, one might perhaps have expected a very high frequency of English elements in the two corpora, as speech is produced off the cuff and without giving the speaker much time to choose his/her words, and particularly considering the way the two speaker groups were picked out (cf. what was said above).
Interestingly, Harriet Sharp gives frequency figures both in terms of time intervals and in relation to the number of words.The average time between code-switches was 14 seconds for the COOL group and 58 seconds for the VISBY group.The former produced Examples of such words are: jobba (55 instances), jobb (21 instances), träna (16 instances), film (13 instances).
In Table 1 I have recalculated Harriet Sharp's figures to make them more comparable with the frequency counts for written texts.As a code-switch may consist of more than a single word, the number of English words is higher than the number of codeswitches.Each lexical item is registered as a record, but the number of records is higher than the number of words, as some multiword items have been registered as a single record (e.g. of course or thank you).If we eliminate all names and all items that are 'standardised', i.e. recorded in Svenska Akademins ordlista, the English elements in the COOL corpus go down to 758 and in the VISBY corpus to 391, or 10 vs. 5 words per thousand running words.In other words, the relative proportion of English words is rather low compared with some types of written texts.
Harriet Sharp makes the point that "Proportionately speaking, English words ... play a very minor role in the spoken discourse of the investigated domains" (p.61), and her case could in fact have been made even stronger, as I have shown.She adds that "It is my contention that the reason people assume that there are more English words than there actually are has to do with the perceptual salience of English borrowings (or code-switches)" (p.62).In other words, they stand out from the surrounding text and are more easily noticed.
How do we account for the surprisingly low proportion of English elements in spontaneous speech?Partly, this has no doubt to do with the low lexical density of spontaneous speech, i.e. the low proportion of lexical words in relation to function words, and it is well-known that function words are rarely borrowed; the most typical loanwords are nouns and, much less frequendy, adjectives and verbs.Moreover, conversation has a restricted and repetitive repertoire (Biber et al. 1999(Biber et al. : 1049) ) and is typically planned in sequences, with less room for variation than in writing.
In motivating her focus on spoken material, Harriet Sharp mentions "the reported observation that linguistic change originates in speech rather than writing" (p.35), with reference to Milroy & Milroy (1999).What these authors actually say about this topic is that "Linguistic change, especially in phonology and grammar [my highlighting], originates in speech rather than writing" (Milroy & Milroy 1999: 55).The present thesis, however, deals with lexical innovations.Moreover, it is highly likely that the types of innovations due to language contact are tied to the type of contact situation.To take an example, the influence on the language of the Norwegian immigrants in America (as described by Haugen 1953) is quite different from the influence of English on contemporary Norwegian.In the latter case, the written language plays a crucial role, as suggested by the frequency figures reported above.

Analytic choices
In the second chapter of her book, Harriet Sharp discusses the problematic distinction between borrowing and code-switching and points out that "it has hithetto been impossible to draw a distinct line between them" (p.12)."It is thus probably best," she says, "to see the concepts as representing degrees on a continuum, describable in terms of a cline rather than fixed points" (ibid.).This is the background for the broad definition of code-switching referred to above.The reader has to keep in mind that it includes everything from clear instances of code-switching to just as clear instances of borrowing, such as completely established loanwords.Even words which could not by any stretch of imagination be defined as examples of code-switching are included and form a substantial part of the material, e.g.jobba, as in: [From a phone conversation between Truls and Elisabeth] T: Men ni jobbar båda två?E: [Nods] Ah.T: Näh hon jobbar inte idag.

E:
Jo [pause] -Jo hon ska.jobba idag.(p.63) Jobba is of course derived from the direct loanword jobb, but has no corresponding verb in English.
Although borrowing and code-switching are not distinguished at the outset, a distinction is made later in the analysis between mixed and unmixed utterances.This corresponds in many cases to the distinction between borrowing vs. code-switching.Compare:

Unmixed:
I do not believe id (p.101) Later in the analysis, a distinction is also made between standardised and non-standardised elements, the latter representing material that has not been recorded in Svenska Akademins ordlista.
The broad definition of the material means that all expressions which derive from English are included, provided that they conform to the criteria quoted above.An alternative choice would have been to focus on non-standardised elements, excluding names.The material would then have been greatly reduced (cf.Table 1 above), but the author would not have been able to show the consistent differences between the two speaker groups.These are well brought out in the analysis.

Differences between the two discourse domains
The detailed analysis reveals that the English elements do not just differ in frequency between the two speaker groups, but are used quite differently.The main differences are summarised in Table 2.The differences between the two speaker groups must be seen against the background of the communicative purposes in the two situations.In the case of the COOL corpus, the purpose is transactional (p.41), i.e. to get something done.Language is used to transfer information, and the referential function is in focus.Hence there is a high proportion of nouns, including a lot of names.These are mostly used in mixed utterances, as in: Jag har också tittat på quartered beef hån Queensland till Moskva [short pause] på liner terms, (p.85f.) Flagging, i.e. explicit signalling of the language shift, is comparatively rare in the COOL corpus, and where it occurs, it most typically takes the form of linguistic flagging devices, e.g.metalinguistic comments or explanations.This is consistent with the predominantly referential function.The main purpose is to convey a message.
In the VISBY corpus, the purpose is primarily interactional (p.42), aimed at the maintenance of social relationships rather than the transference of information.The relative proportion of nouns is half of that in the COOL corpus, while there are twice as many verbs (and presumably shorter and less complex clauses).Interjections and discourse markers are common, e.g. to mark back-channelling, in agreement with the interactional function.A high proportion of the English elements (40 %) is found in unmixed utterances, i.e.English language islands which are not integrated syntactically into the Swedish text.Flagging is very common and is most often paralinguistic, primarily shown by means of voice modification, laughter, and giggling.The speakers signal that they take on a different role, as in playing a game.This may take the form of what Harriet Sharp calls the Quoting Game (p.117), where the speakers play a kind of verbal pingpong using sequences from films, television programs, songs, computer games, and the like: Altogether, the code-switching behaviour is remarkably different in the two corpora.Very few words were found in both corpora, in fact only 24 different words.The English elements used by the VISBY speakers were to a large extent established loanwords of long standing or they were very basic words making up unmixed utterances, i.e. real cases of code-switching.The COOL speakers rarely resorted to unmixed utterances and used far more nonstandardised elements (cf.Table 1), typically embedded as singleword switches.Harriet Sharp interprets the difference between the two speaker groups in terms of Communication Accommodation Theory (p.113).There is a convergence in language use within the two speaker groups, shown for example through the Quoting Game in the VISBY group and among the COOL speakers in the use of English shipping jargon and geographical names.

English -a threat to the survival of the Scandinavian languages?
The analysis and interpretation of the differences between the two speaker groups are the main strengths of Harriet Sharp's book (and there is more to say than the points taken up above).Additionally, it is significant that we have now for the first time reliable documentation on English influence in two domains of spoken discourse.This is a considerable achievement.
To what extent can we generalise from Harriet Sharp's investigation?Does English pose a threat to the survival of Swedish?The author is very clear on this point: It is my contention that the influence on Swedish syntax and lexis of the young adults' use of English expressions is negligible.Swedish is, and will remain, the language spoken in Sweden.
It is my firm belief that English does not pose a threat to the survival of Swedish.Although English is present in many different contexts, it is in principle used as an auxiliary language for specific purposes in Swedish discourse domains.In this capacity it enriches our lexical stock, enables stylistic variation, adds expressivity and signals certain interpersonal relations and values.English words are thus an asset rather than a liability for Swedish speakers, (p. 199) Here the author jumps to conclusions.These are some issues that could profitably have been taken up in the concluding discussion: • To what extent might the results have been different if other discourse domains had been selected?
• Why is the influence stronger in some types of written texts than in the two spoken discourse domains (if the frequencies are recalculated in the way I have shown above)?
• As the younger generation has previously been shown to be most positive as regards the use of English and attitudes to English in Sweden (cf.Ljung 1985Ljung , 1988)), why is the frequency of English elements lower in the VISBY group (age: 19-25) than in the COOL group (age: 30-54)?
It would also have been interesting if Harriet Sharp had related her results to the findings in Preisler's (1999) study and to his theory of influence from above and below.
Last but not least, borrowing and code-switching are not the main threat to the Scandinavian languages, but rather the fact that English seems to be taking over in some domains, e.g. in business and research.Cf. the following comment in a recent article from Aftenposten: Norsk skriftspråk (saerleg bokmål) vil truleg tapa område for område -naeringsliv, reiseliv, forsking, marknadsføring, børs, bank.Lengst vil norsk skrift halda seg i grunnskole, massemedia, politikk, litteratur.

Where do we go now?
Harriet Sharp's study shows very clearly that the influence of English may be quite different depending upon the particular domain.If we want to understand the role of English in Scandinavia (and, as a result, perhaps take some action in support of the Scandinavian languages), we need more studies of this kind.Much remains to be done.There is a need to study more domains in depth, and to investigate both use and attitudes.Some important areas are: computing, business, advertising, pop music, fashion and beauty, film and TV, sports.Needless to say, any linguistic study of this kind is incomplete without considering the wider cultural context.
Most of what has been done so far has focused on direct loans {importord).There is a need for more work on semantic loans and on possible areas of syntactic influence; for some examples of syntactic influence, see the opening chapter in Johansson & Graedler (2002).One channel that deserves to be investigated more fully is the influence on the Scandinavian languages through texts translated from English; see e.g.Gellerstam (1986) and Johansson (2001).
The major studies of English in relation to other languages have generally, though not exclusively, been carried out by English language specialists (Aasta Stene, Magnus Ljung, Knud Sørensen, etc.; Judith-Ann Chrystal is a notable exception).A significant recent publication is A dictionary of European Anglicisms edited by the German Anglicist Manfred Görlach (2001), where there is a comparison of Anglicisms across sixteen European languages, including Norwegian (contributor: Anne-Line Graedler).Large-scale comparative projects involving all the Nordic languages are currently under way, initiated by researchers specialising on the Nordic languages.Unlike the studies carried out by English language specialists, which have had a descriptive slant, these new projects are more concerned with the maintenance and preservation of the Nordic languages, both issues to do with borrowing and the loss of domains.With the works that have appeared in the last couple of decades and the results the ongoing projects can be expected to yield, we will be in a much better position than before to evaluate the role of English in Scandinavia and possible effects on the Scandinavian languages.

[
Ken's girlfriend sees the camera crew and wonders what is going on.Gorm and Ken offer an explanation] G: Du har hamnat på en porrfilmsinspelning [Laughs] Okay då kör vi igång![Starts taking his shirt off] K: Congratulations this is dolda kameran![Laughs] G: Smile!You are on Candid Camera! (p.118)

Table 1 The frequency of English elements in two corpora of spontaneous speech
Chrystal (1988)er thousand, the latter 22 words per thousand.The overall frequency is approximately ten times the overall figure reported byChrystal (1988)for Swedish newspaper text.There is a need for a great deal of caution, however; the figures are not comparable, because: