Multilingualism in the North: From Baklava to Tre Kronor

Abstract

This article explores processes of place-making through the study of the linguistic landscape of a small-size town in Northern Sweden. The analysis of signs is used as a tool for examining the role and visibility of actors in the landscape. For this purpose, we examine who the authors are, what forms of multilingualism can be observed, and who has agency in the place-making of the public space. Our documentation consists of photos and fieldnotes from observations, encounters, and conversations with people during ethnographic fieldwork in 2019. Using a mixed-methods approach, all signs were first analysed quantitatively according to the categories of authors and function. Regression analysis was used to explore correlations between the categories. Secondly, multilingual signs were analysed qualitatively regarding their function and purpose in relation to their contexts. Our results illustrate a city centre with a strong presence of the Swedish language. Multilingual signs target specific groups and are intended for information, advertisement, rules and regulations; moreover, our findings indicate that the opportunities for private actors to influence the linguistic landscape are limited. The form of multilingualism in this context—visible multilingualism present mainly through English—is different from the one we can see in the socio-demographic data.

1. Introduction

Written public signs can be said to both reflect and regulate the cultural, sociolinguistic and political structure of the space they are situated in (Blommaert 2013). Languages such as they materialise, become visible or are absent in a specific geographic space, contribute to place-making since they bear and leave traces of the presence of the users of the languages, language use and attitudes to the languages. Furthermore, written language in public space has consequences for and an influence on language revitalisation (Grenoble and Whaley 2006), language ideologies (Litvinskaya 2010; Woldemariam and Lanza 2015), and language discourses and practices; these languages are, thus, part of the shared public space and at the same time part of constructing the social realities of this shared space. Linguistic Landscape Studies have an interest in documenting language used in public spaces, i.e.,

the language of public road signs, advertising billboards, street names, place names, commercial shop signs, and public signs on government buildings combines to form the linguistic landscape of a given territory, region, or urban agglomeration.

(Landry and Bourhis 1997, p. 25)

A further focus is to analyse these signs that are displayed in public places in “attempts to understand the motives, uses, ideologies, language varieties and contestations of multiple forms of ‘languages’” (Gorter 2019). A more specific definition of what constitutes a sign was developed at the initial stage of our study as follows: it should contain named languages, orthographic text either on its own or together with pictures or symbols, and the text should also be physically bounded and the font colour, shape, etc., should be cohesive. All signs identified based on this definition were photographed during fieldwork.

This paper explores processes of place-making through the study of the linguistic landscape of Kramfors, a small-sized town in northern Sweden. More specifically, this study examines which actors are visible in the urban landscape as enacted by the display of signs. For this purpose, we examine who the authors (Malinowski 2008) are, what forms of multilingualism can be observed, and who has agency in the place-making of the public space.

Previous research on Linguistic Landscapes in the Nordic countries includes Pietikäinen (2014), in which the multilingualism in a village in the rural North of Finland, how the Sami language is used, and issues around mobility are explored. Results show how “historical, political and economic processes can become embedded in the LL and, conversely, how the LL comes to inform not only these processes but also individual behaviour and the development of collective imaginations” (p. 482); analysis also shows that institutional signs made by authorities are oriented towards stability and normativity. Signs that relate to tourism, cultural activities and advertising are more oriented towards hybridity and multiplicity. The question of whether there is a direct relationship between the presence of immigrant languages used by communities and language vitality and visibility has been explored by Barni and Bagna (2010). Through statistical information, interviews and questionnaires, they have explored this issue in several urban areas in Italy, some big cities and a few smaller urban areas. Using Linguistic landscape methodology, the focus was on immigrant languages in a circumscribed area; they found that the presence of a language in an area, its vitality and visibility depend on several linguistic, extra-linguistic and contextual factors. These factors include the characteristics of the area in which the immigrants settle, the length of time the immigrants stay, the attitude of the speakers towards their own language and the linguistic policy choices of the host country and other political issues.

The town of Kramfors, the focus of this article, is located on the coast of the region of Västernorrland. It is a culturally and linguistically heterogeneous municipality, and is, therefore, one of the places our project group1 has identified as particularly relevant for studying how multilingualism and place-making come to expression. Given the different geographies of Norrland we have selected five different sites: these include, in addition to Kramfors, the cities of Pajala, Jokkmokk, Sorsele and Umeå. (see Cocq et al. 2020; Granstedt et al. Forthcoming). These sites reflect varying demographic and socioeconomic conditions of this part of the country: selected according to number of inhabitants, percentage of inhabitants from other countries, tourism, global industry and trade. In Kramfors, 13% (2018) of the population was born outside Sweden, i.e., they are from the Nordic countries and countries in southern Europe, East Africa, the Middle East, etc. Kramfors is also an administrative area for the national minority language Finnish (since 2019), and this implies that the municipality is responsible for promoting and supporting the language according to the Swedish Act on National Minorities and Minority Languages (Government of Sweden 2009:724); however, as this responsibility had been in place for only a few months at the time of our fieldwork, the presence of Finnish as an official minority language cannot be expected to be present in our data collection.

Multilingualism has also been a topic of great importance for the senior high school of Kramfors. Kramfors made the headlines in national newspapers a few years ago when the principal of the school allowed for a digital form of mother tongue tuition in 10–20 languages—before distance learning was supported by the School Act (Government of Sweden 2010:800)2. Despite being told off by the school inspectorate, the senior high school continued, and the principal stated that it was of great importance for students with a non-Swedish background to receive mother tongue education. The school was, thus, a participant in pushing for the change in the legislation—a change that finally took place in 2016, allowing for distance teaching in mother tongue education. Eventually, the principal received an award for this courage from the Swedish government (Lind 2016, 30 July; Sahlin 2016, 26 May).

2. Materials and Methods

Our documentation consists of photos of outdoor signs and fieldnotes from our observations, encounters and conversations with people during our fieldwork. We arrived there by train one morning in June 2019. Various languages were present in the soundscape of the city centre, as we observed, for instance, when buying coffee in Gallerian Christoffer, having lunch at Träffen, dinner at Tre Kronor, or walking through the public square. Later this day, we met two people working on an integration project (Samverket) funded by the Swedish Inheritance Fund3 who told us that Arabic and Tigrinya were two well-represented languages among children and young people. Among the food offered at the local restaurants, we found, for instance, baklava and souvlaki. A hairdresser in a street close to the city square offered “hijab” in the list of their services.

For two days, we documented the city centre (central square, streets connecting to the square), including shops and grocery stores, some national and local authorities migration office, city hall), the library and the health centre. The train and bus stations, located near the centre, were also photographed. In addition, we chose to include the Witch Museum (located 26 km from the centre) as the main attraction in the area; this was identified upon discussions with the tourist centre and the information centre at the local hotel, as the place they recommend to tourists and visitors.

Our photographic documentation resulted in 844 photos of private, public and commercial signs—here defined as everything that contains a verbal language (“named languages”) and orthographic text (i.e., not symbols, pictograms etc. without text)4. Our field notes were used in order to contextualize the material and to add an additional layer of understanding to the socio-demographic data and the photographs.

All photos were coded according to a coding scheme developed in the project, which includes placement, medium, author, function, and language/s of each sign (for a detailed account of the coding process, see Cocq et al. 2020). Our methods of analysis of ethnographic data include a quantitative analysis with descriptive statistics and logistic regression.

Table A1. Estimates of sign characteristics on the likelihood of language appearance on signs (logistic regression). Model 1—Multilingual sign. Model 2—Monolingual sign. Model 3—Monolingual sign in Swedish.

Variable	Model 1 Multilingual Sign			Model 2 Monolingual Sign			Model 3 Monolingual Sign in Swedish
	Coeff.	Sign.	Odds Ratio	Coeff.	Sign.	Odds Ratio	Coeff.	Sign.	Odds Ratio
Placement/On building	0.553	0.329	1.704	0.230	0.636	1.258	0.085	0.863	1.089
Placement/On pole	−1.288	0.118	0.276	2.165	0.006 ***	8.717	1.839	0.013 **	6.291
Placement/Message board	0.785	0.224	2.193	0.003	0.995	1.003	−0.175	0.775	0.840
Placement/Sidewalk	0.885	0.386	2.422	0.441	0.653	1.555	0.485	0.624	1.624
Placement/Parking machine	0.164	0.813	1.178	0.642	0.323	1.901	0.575	0.386	1.776
Placement/Semi Movable	0.716	0.284	2.045	0.196	0.750	1.216	−0.192	0.753	0.825
Medium/No medium	0.045	0.936	1.046	0.708	0.144	2.030	0.746	0.117	2.108
Medium/Poster	1.847	0.001 ***	6.340	−1.037	0.036 **	0.355	−1.076	0.027 **	0.341
Medium/Sticker	1.123	0.056 *	3.073	−0.357	0.489	0.700	−0.293	0.564	0.746
Medium/Piece of paper	1.058	0.060 *	2.879	−0.259	0.597	0.772	−0.231	0.632	0.794
Medium/Sandwich board	1.167	0.233	3.212	−1.011	0.298	0.364	−1.273	0.190	0.280
Medium/Product	1.093	0.142	2.984	−0.361	0.601	0.697	−1.510	0.025 **	0.221
Author/Company	−0.660	0.017 **	0.517	0.859	0.002 ***	2.360	0.893	0.001 ***	2.441
Author/Authority	−0.452	0.156	0.636	0.608	0.054 *	1.836	0.877	0.006 ***	2.405
Author/Non-profit organization	−1.151	0.001 ***	0.316	1.318	0.000 ***	3.737	1.476	0.000 ***	4.373
Content/Company name	−0.256	0.720	0.774	1.619	0.014 **	5.048	0.832	0.200	2.299
Content/Information	0.559	0.378	1.749	0.813	0.158	2.256	1.004	0.093 *	2.730
Content/Advertisement	0.724	0.294	2.063	0.654	0.309	1.923	0.857	0.196	2.356
Content/Rules	−0.076	0.921	0.926	1.411	0.050 **	4.100	1.607	0.029 **	4.987
Content/Art	−0.114	0.902	0.893	1.499	0.089 *	4.476	1.544	0.085 *	4.683
Constant	−1.958	0.011		−1.098	0.080	0.333	−1.306	0.045	0.271
LOG-L	915.3			923.1			954.5
N	858			858			858

* p < 0.10, ** p < 0.05, *** p < 0.01.

in Appendix A provides the full modelling results (logistic regression) for the estimation of the likelihood of language appearance on signs. We have estimated three models focussing on multilingual signs, monolingual signs and monolingual signs in Swedish, respectively. The modelling results show that only a subset of the identified covariates (sign characteristics) has a significant association with the goal variables. Briefly, signs belonging to the Author category show many significant associations with languages on signs. Almost all estimates in this category are significant. There is also a clear difference between multilingual and monolingual signs in that estimates are either positive or negative. Signs belonging to the Author category are very likely to be monolingual, usually in Swedish. On the other hand, it is very unlikely to have multilingual signs from Companies and Non-profit Organizations. Generally, multilingual signs are mostly found on Medium/Poster, whereas monolingual signs are found on Placement/pole; Author/Company; Author/Interest Organizations; Content/Company Name, and Content/Rules.

In a second step, a qualitative analysis of the multilingual signs was conducted in order to discuss visible multilingualism and forms of multilingualism that occur in the data. Following the model of analysis of types of multilingual information developed by Reh (2004), the contents of all multilingual signs collected during our fieldwork were closely examined in order to identify instances of duplicating, fragmentary, overlapping or complementary multilingualism.

3. Results

3.1. Descriptives

A majority of the signs in the centre of Kramfors were issued by commercial companies (53%), such as food chains, restaurants, local shops or banks (see

Table 1. Authors of signs (N = 850).

Author	Number of Signs	% of Signs
Authorities	208	24.47
Companies	454	53.41
Organisations	96	11.29
Private actors	6	0.71
Multiple	36	4.24
Other	50	5.88

). Around 25% of the signs were issued by authorities, in particular by the municipality of Kramfors or the Swedish Migration Agency, and 11% by interest organisations, for example, churches or sports clubs. Less than 1% of the signs were authored by private persons.

As illustrated in

Table 2. Function of signs (N = 850).

Function	Number of Signs	% of Signs
Art	24	2.82
Commercials	66	7.76
Company name	100	11.76
Graffiti	1	0.12
Information	592	69.65
Menu	8	0.94
Opinion	1	0.12
Rules & regulations	58	6.82

, the primary function of the signs was to convey information (69.65%), followed by company names (11.76%), commercials (7.76%) and rules and regulations (6.82%). A few signs conveyed artistic messages or menus.

The majority (70%) of the signs were monolingual and in Swedish (see

Table 3. Languages on monolingual signs (N = 596).

Language	Number of Signs	% of Signs
Swedish	564	94.63
English	27	4.53
German	0	0
French	0	0
Hybrid	3	0.50
Unclear	1	0.17
Other	1	0.17

). Around 4.5% of the monolingual signs were in English and small numbers in other languages.

Swedish was also the dominant language on the 30% of multilingual signs (see

Table 4. Languages on multilingual signs. N = 255 for the main language and N = 329 for additional language as some signs included more than one additional language.

	Main Language		Additional Languages
	Number of Signs	%	Number of Occurrences	%
Swedish	216	84.1	16	4.86
English	30	11.76	177	53.8
German	1	0.39	0	0
French	0	0	11	3.34
Hybrid	1	0.39	39	11.85
Unclear	2	0.78	35	10.64
Other	5	1.96	51	15.50

). Similar to the monolingual signs, English was the second most common language on multilingual signs. “Other” languages in our coding included Finnish, Arabic, Tigrinya, Armenian, Bengali, Arabic, Portuguese, Spanish and Mandarin. A few signs included perfect cognates, i.e., words you can find in several languages with a common origin, similar spelling, pronunciation and meaning, such as Ariel. Words that combine 2 or more languages (almost exclusively names of companies and products, for instance, Nordhydraulic) were coded as “hybrid” (see Granstedt et al. Forthcoming). Languages that could not be identified (for instance in the name of a company) were coded as “unclear”.

In order to get an overview of how sign characteristics, as defined by our coding methodology, correlate with languages appearing on signs, we have estimated logistic regression models; this modelling approach makes it possible to determine how monolingual and multilingual signs as well as signs written in Swedish, English or other languages are linked to sign characteristics such as where they are placed (indoors or outdoors), on what kind of object they are placed (building, post, bulletin board, etc.), medium—object on which the sign is attached (clothes, bags, stickers, posters etc.), issuer (authority, companies, organisations, private persons, etc.), and content (graffiti, name of enterprise, menu, information, rules, opinion formation, advertising, art, etc.).

The results indicate that posters, stickers and notes are more likely to be connected to multilingual signs; however, this is not the case regarding signs issued by companies and interest organizations, which are more inclined to convey their messages in Swedish. Monolingual signs tend to be found on posts and issued by companies and interest organizations; this type of sign is more likely to include company names, rules and art.

Most monolingual signs carry messages written in Swedish. It can be concluded that the Swedish language dominates the linguistic landscape of Kramfors. English also appears on signs, but makes up only a small proportion of the signs documented within the city. All other languages are almost completely absent, which is quite surprising given that about 13% of the municipality’s population originates from many other countries. The non-Nordic population amounts to 8%, representing countries from all over the world; this gives evidence of an abundance of language skills in the local context. Many languages are most likely spoken in private domains and at home, but not much of it can be seen in the public space.

3.2. What Forms of Multilingualism Do We Find?

In a second step, a qualitative analysis was conducted by a close reading of the signs that contained more than one language. These signs were examined for the purpose of distinguishing what forms of multilingualism—based on the model proposed by Reh (2004)—could be observed in the city centre of Kramfors.

First, we observe that visible multilingualism is present mainly through English. As we illustrate here with a number of examples, English is used mainly in the names of products and for branding in commercial contexts, or, in some instances, as a lingua franca. Other contextual factors and addressees can be identified in the case of the use of other languages such as Arabic or Tigrinya, among others.

The sign in Figure 1 displays a short text in Swedish, “[Get] out and enjoy” and the name of a Swedish chain of food stores. The names of the products are in other languages; however, this linguistic dimension does not contain any information or knowledge necessary for the understanding of the sign. We can assume that people passing by the sign perceive it as an advertisement for ice cream, regardless of their understanding of the meaning of the name of the product.

This sign illustrates visible multilingualism as complementary multilingual writing (compare Reh 2004); this form of multilingualism can be observed in many instances in the linguistic landscape of the city centre of Kramfors.

This poster (Figure 2) announcing the “City party of Härnösand”, and other similar announcements for concerts and festivals, can be encountered frequently in the linguistic landscape of Kramfors; this is not surprising in the early summer of a city in Northern Sweden (in pre-corona times). The occurrence of these posters resulted in a rather high representation of English in our quantitative analysis. What we see here is that English occurs through the names of artists (“Bryant”, “Damn”).

Methodologically, it might appear problematic to consider these occurrences in terms of multilingualism; however, in line with Backhaus, we find that “it is problematic to ignore English names and slogans in total because the general proclivity towards English in itself has something to say” (2006, p. 62). Our data show that this form of visible multilingualism, together with complementary multilingual writing in, for instance, advertisements has such a strong presence in the city centre of Kramfors that it would be misrepresentative to exclude these from the materials collected and from our analysis.

Figure 3 illustrates globalization rather than multilingualism. As e.g., Coulmas and Heinrich (2005) have shown, the power of global market forces can be observed in urban environments. Not least, in this photo, as in other examples in our data, the high prestige value of English, in relation to globalization, motivates the use of signs in another language than the ones spoken by the population. The city centre provides many examples of the impact of globalization on “local linguistic ecologies” (Backhaus 2006, p. 56).

We find other linguistic influences that can be interpreted in relation to globalization and its impact. The sign in Figure 4 reads “Chinese restaurant” in Mandarin, China (English) and restaurant (in Swedish). The use of the three languages illustrates practice of overlapping multilingual writing (Reh 2004, p. 12): its information is repeated, but parts of the text are in one language only. It is, however, worth noticing that knowledge in all three languages is not required for a reader to understand the message displayed on the sign.

Figure 5 is an illustration of duplicating multilingual writing, i.e., the same text is presented in more than one language (Reh 2004, p. 8). Here, only one part of the text is translated. In the hierarchy of languages, Swedish is the dominant language and translations in Tigrinya and Amharic are provided for only one part of the text, conveying to the language groups the information assumed relevant to them. Other information on the poster addresses exclusively Swedish speakers (sometimes informing in Swedish that the event will take place in other languages). While the examples above have commercial actors as authors, here, an interest group (a church) is the author.

Other examples show instances of multilingual writing with national authorities as authors. Figure 6, for instance, illustrates a similar form of practice as Figure 5. Here, this sign placed on the door of the migration office provides information in Swedish and English. The use of English here also reflects the position of this language as a lingua franca, since we know that the variety of countries of origin among asylum seekers includes many more languages than English.

Figure 7 displays the same text in Swedish, English, Arabic and Tigrinya; this example of duplicating multilingual writing acknowledges the existence of societal multilingualism, i.e., “the existence of more than one language in the target community” (Reh 2004, p. 8); this sign was placed in the window of premises organizing activities and hosting offices. The target groups might here be more specific than the examples above (the migration office and a shop), something reflected in the choice of the languages included on the signs.

4. Discussion

Our first day of fieldwork started in a small mall where we stopped for a cup of coffee. The café offered warm and soft drinks, sandwiches and baklavas. The menu was in Swedish, but other languages were present in the soundscape, for instance when we heard a conversation between the personnel in the café. Later during our fieldwork, we end up at the restaurant Tre Kronor for dinner. Tre kronor, “Three Crowns”, is the national emblem of Sweden, composed of three yellow or golden crowns on a blue background. The emblem is used as a symbol of official State authority, for instance by the Government and Swedish embassies around the world, but also as the symbol for Sweden’s men’s national ice hockey team (hence being called “Three Crowns”); it is also a common name for pizza restaurants in Sweden. The restaurant Tre Kronor in Kramfors, from its appearance, seems to fit in the genre of a Swedish restaurant that offers standard meals at an affordable price. The signs (see, for instance, Figure 8) and menu are in Swedish. While we were documenting the surroundings of the restaurant with our camera, the owner came out, interested and curious about what we might be doing. He encouraged us to come by for dinner later, which we eventually chose to do. It turned out that the owner (and also chef), of Greek origin and who seemed to have lived in Kramfors for many years, offers a couple of traditional meals from his own country that he is particularly proud about. In other words, the linguistic landscape when passing by the street displays a Swedish monolingual place, whereas the interaction with the owner also displays the presence of Greek language and cuisine. The restaurant Tre Kronor as a commercial actor was visible, while the owner as a private actor was not; this discrepancy between the landscape and its inhabitants appears in many instances in our study.

The presence of several languages in Kramfors can, as we could expect based on the population statistics, be observed in various places and in many instances; however, we find that multilingualism is not represented to a large extent in the linguistic landscape. We find no direct relation between the linguistic knowledge of the population, as indicated in the socio-demographic data and the publicly visible written language use. Rather, our results illustrate a city centre with a strong presence of the Swedish language, in particular on permanent signs. Multilingual signs, in contrast, target specific groups, are temporary and intended for information, advertisement, rules and regulations.

Our findings also indicate that the opportunities for private actors to influence the linguistic landscape is limited. Even though we do not know to what extent graffiti, stickers etc might occur, or if they are cleaned away quickly when they do5, the absence of such vernacular expressions in the centre indicates that this specific place is not the one where common citizens can express themselves through orthographic signs as private persons. The degree of agency of these private actors is rather low, whereas it is considerably higher for commercial actors. A few non-profit organizations add some variety to the linguistic landscape, but to a limited extent; this distribution of agency in a public place does not support, promote or give visibility to multilingualism as it exists in private domains in the same municipality.

The forms of multilingualism that can be observed are complementary, overlapping and duplicating forms of multilingual writing. Societal multilingualism appears in a few instances and instances of the influence of globalization are many. The presence of the multinational company Air Liquide and the food chain Subway are examples of how globalization materializes in language use. But also, we see that English is preferred, as it has positive associations in a Nordic context. Even local actors in the region of Kramfors, or national companies, choose sometimes to take an English name (for instance “Active Sport Club”, a small sports centre; “Nordic snus”, a Swedish tobacco company (with no export)). Another finding is that, when multilingualism is present, it is often through the occurrence of Swedish and English on the same sign. Posters and notes are positive significant explanatory variables for the occurrence of English in multilingual signs.

5. Conclusions: Whose Languages?

Previous studies have emphasized the role of commercial discourse in the linguistic landscape (see, for instance, Vandenbroucke 2016). Our study illustrates the consequences of such a role: in Kramfors, the presence of signs by private actors is rather low, and commercial actors have priority in this specific public space; it is thus their language use that is studied. These languages and this form of multilingualism are different from the one we can see in the socio-demographic data. Where commercial actors dominate, private actors have less opportunity to influence. While official signs reveal who is addressed, included and visible (or not) in the city, vernacular signs would indicate who claims, uses and interacts with the urban space. The absence of such data in our materials indicate that such forms of expressions and interactions do not take place in the city centre of Kramfors.

The material collected indicates that there are almost no private authors, but a high percentage of companies behind the signs displayed in the city centre of Kramfors. As a public space, it is shaped by commercial actors, whereas other possible authors of linguistic signs (interest/non-profit groups, private persons etc.) are not present. The linguistic landscape is thereby formed by the politics and dynamics of commercial actors. The presence of English, for instance, is surprisingly high if approached in relation to the socio-demographic data, but does not astonish in the context of our globalized Nordic urban cultures.

In contrast with other studies (e.g., Moriarty 2014), processes of mobility (migration and diaspora) are visible only to a very limited extent. The fact that only traces of languages represented in the socio-demographic background are visible in the documented area has several implications, for instance in regard to societal aspects such as integration, inclusion and attitudes to foreign languages. The languages are present elsewhere (in the soundscape of the city, for instance) but the linguistic norm, however, is dominantly monolingual and the presence of multilingualism is rarely a reflection of the population of Kramfors, but rather of the presence of commercial actors.