In work on language revitalization, a number of scales have been developed to assist in operationalising the status of endangered (and non-endangered) languages. One of the most influential scales is Fishman’s Graded Intergenerational Disruption Scale (henceforth GIDS, Fishman, 1991), which classifies the world’s languages on a scale from 1 (vital, national language) to 9 (dormant). The GIDS has remained a central instrument of language policy research, arguably due to its wide applicability across language scenarios, and its effectiveness in highlighting the complex factors that influence language revitalization. In Fishman’s (2004) discussion of the GIDS, he argues that the success of language planning greatly depends on whether or not the measures in question are appropriate for the stage at which the language finds itself. Though the GIDS is not to be understood as a strict hierarchy, Fishman (2004) urges language planners to lend particular importance to Stage 6, which is concerned with intergenerational language transmission and use within the family. If high-level measures—such as those intended to secure greater power or prestige amongst the majority society—are introduced before Stage 6 (and intergenerational mother-tongue transmission is secured), the success of those measures will be short-lived or limited in scope. Fishman gives the example of the Irish language, which received extensive support at the nation-wide level for education programmes, but which today is often first learned as a school language, rather than as a mother tongue (p. 429). At worst, pushing for high-level measures before their time can lead to diminishing trust, and waste of resources that are already in low supply, such as money and time (p. 435). What exacerbates the problem is that, according to Fishman, language planning is difficult precisely at Stage 6, since language behaviour at home, with one’s family and with one’s own children, takes place in a realm of ‘spontaneity’ and ‘intimacy’ that is hard to reach with rationally-planned measures (p. 431). Nonetheless, if mother-tongue revitalisation is the goal, then ensuring language use in the family realm is essential.

The GIDS is also useful for our particular case, because it enables a comparison of languages across individual countries. The EGIDS, which is an expanded version of Fishman’s GIDS (Graded Intergenerational Disruption Scale), developed by the Ethnologue, places North Sámi at Stage 2 (‘provincial’) in Norway, meaning that language is used in education, work, mass media and government within administrative subdivisions. The EGIDS is a scale that operationalizes 11 different levels of language engdangerement, with ‘international’ (0) at the top, representing the most vital languages, and 10 at the bottom (extinct). In Sweden, North Sámi is placed much further down the scale at Stage 6 (threatened), meaning that the language is used for face-to-face communication, but that it is losing users’ (Eberhard et al., 2022). The implications of Stage 6 are, as mentioned above, that the focus should be on intergenerational transmission (Eberhard et al., 2022). If we were to draw implications from this classification, it would be that higher-level revitalization measures for North Sámi should be possible in Norway as compared to Sweden.

The exact number of ethnic Sámi is unknown, since information on language and ethnicity is not collected by Sweden or Norway in the national census. Estimates range between 65,000 and 100,000, though only a minority is still able to speak a Sámi language, with estimates ranging between 18,000 and 30,000 speakers, mostly in Norway (Aikio-Puoskari, 2009). While all nine remaining Sámi languages are endangered (UNESCO), North Sámi is the most vital and has the largest estimated number of speakers (~20,000). The Endangered Languages Project (ELP) considers North Sámi vulnerable, and the other Sámi languages endangered to varying degrees (Todal, 2013; Hettema and Outakoski, 2020; Vangsnes, 2022). Geographically, about two thirds of the North Sámi speakers live in northern Norway, and a smaller share lives in northern Sweden (~5,000–7,000; Rasmussen and Nolan, 2011) and Finland (~2,000; Rasmussen and Nolan, 2011). Besides North Sámi, the other Sámi languages spoken in Sweden and Norway include South Sámi, spoken by 500–1,000 people (Hettema and Outakoski, 2020), and Pite and Lule Sámi, which are spoken by only a small handful of individuals (Hettema and Outakoski, 2020). The Sámi languages enjoy the highest levels of vitality in the Sámi administrative areas, which exist in Norway, Sweden, and Finland. These territorially-confined areas consist of municipalities in which extended language rights apply (described in section 2.3.).⁴ The purpose of the administrative areas is to provide resources and opportunities beneficial for the survival and advancement of the Sámi languages. For example, state-run schooling through the medium of Sámi exists in all Sámi administrative areas (Huss and Lindgren, 2022), as does the right to communicate with the authorities in a Sámi language (Svonni, 2008; Aikio-Puoskari, 2009). However, there are no monolingual speakers remaining; Sámi speakers are bilingual with the majority language and a Sámi language (Aikio-Puoskari, 2009).

Today, the Sámi people are largely assimilated and do not differ from the majority in terms of economic measures, unlike most other Indigenous populations world-wide (Yasar et al., in preparation).⁵ Nonetheless, the Sámi consistently report experiencing higher levels of discrimination than the majority population (see Hansen et al., 2008, 2016, for Norway, and Yasar et al., in preparation, see footnote 5, for Sweden). Therefore, despite socio-economic equality, inequalities seem to persist. Language is one area where this becomes apparent: A recent study by Yasar et al. (in preparation, see footnote 5) showed that speaking Sámi outside the family context was related to a higher proportion of perceived discrimination. Such findings may reflect a higher sensitivity to missing opportunities for engagement with Sámi issues. For example, not all Sámi children have access to education through the medium of Sámi (see section 2.3.2). For those who do, it is common to discontinue education in Sámi and to pursue secondary education in the majority language (Vangsnes, 2021). Based on studies on small speaker communities, it would seem that language shift from Sámi to the majority languages is still ongoing. In a study on reindeer-herding populations in Sweden, Svonni (2008) evidenced North Sámi proficiency in 60% of the older generation (over 60 years), dropping to 45% in the 30–59 year-olds, and 20% in the younger generation. However, no quantitative data on today’s parents and their language use patterns with their children exist.

In the following, we outline the differences in national-level policies across the domains of language, education, and budgetary spending to provide a picture of the overall language policy context. Although the implementation at local or municipality level might vary, we chose to focus on national-level policies because they define the general scope of which language measures can be implemented. Further, our data would not be suited to a municipality-level comparison since we do not have an even distribution of participants across municipalities, and many of the participants no longer live in the areas they grew up in.

While designing the survey, we actively sought feedback from local stakeholders (such as government representatives) and scholars working in the field. Participation was voluntary and participants could withdraw from the survey at any time. All data was fully anonymized, and will be made freely available to the community in the near future.

Data collection took place in two stages in 15 northern municipalities in Norway and 5 in Sweden in summer 2021. Stage I of data collection consisted of short, computer-assisted telephone interviews; Stage II was an online survey, described in more detail below. The 20 municipalities were selected based on the relatively high share of residents on the electoral role for the Sámi Parliament elections in 2017, which was substantially higher in these municipalities than in others. In the 15 Norwegian municipalities we selected, the mean share of residents on the Sámi electoral role was 20.9% (min = 8.5% in Alta, max = 68.6% in Kautokeino). In the five Swedish municipalities the mean was 9.1% of residents (min = 7.5% in Kiruna, max = 15.5% in Jokkmokk).

Stage I of data collection consisted of a short, representative survey carried out by local survey companies in the form of computer-assisted telephone interviews (CATI), and contained questions on demographics, ethnic self-identification, and language use. The companies employed trained interviewers who conducted the interviews in Norwegian and Swedish, respectively. Participants were offered the opportunity to schedule an interview in North Sámi, though no one chose this offer. No sampling of respondents took place. Instead, Stage I was a full population survey in which the companies tried to reach out to every adult with a phone number registered in the selected municipalities. In Norway, our local cooperation partner had access to 17,096 eligible adults with a telephone number registered in the selected municipalities. They managed to reach 11,153 people, of whom 21.5% (N = 2,396) agreed to participate in the CATI. The Swedish survey company had access to 22,073 people. They managed to reach 6,265 people on the phone, of whom 48.2% (N = 3,020) participated in Stage I.⁷ As shown in Table 3, in Norway, a total of 2,396 telephone interviews were carried out, of which 1,072 (44.7%) were conducted with ethnic Sámi respondents. In Sweden, 3,020 people participated, 847 of whom (28.1%) were ethnic Sámi. Though there are many approaches to defining ethnicity, we defined having a Sámi ethnicity as fulfilling one or more of the following criteria: (i) self-identification as Sámi, (ii) one parent or grandparent with Sámi ethnicity, and/or (iii) a Sámi language was spoken at home. Others were defined as having non-Sámi ethnicity.

The maps in Figures 1, 2 display the number of responses from the 20 municipalities in Norway and Sweden. The vast majority of respondents came from Alta (Norway) and Kiruna (Sweden), as indicated by the shaded areas in dark blue. These large shares reflect the fact that these two municipalities have the largest populations; in fact, the relative share of the population that participated in the survey is reasonably evenly-distributed across the municipalities (see Supplementary materials for the participation shares).

Stage II of data collection consisted of an online survey⁸ with participants who were recruited during Stage I. The questionnaire was available in three languages: the majority language (Swedish or Norwegian), North Sámi, and in English. A total of n = 502 (including 239 ethnic Sámi) participated in Stage II in Norway, and n = 918 (including 263 ethnic Sámi) in Sweden. All participants regardless of ethnicity answered questions on demographics, self-assessed Sámi proficiency, and Sámi language learning at school. Ethnic Sámi respondents (identified as such in Stage I) received a longer version of the questionnaire that contained questions about Sámi language use across the lifespan and ended with two languages tasks, including a vocabulary test and a speech recording. Participants received a voucher for their participation, and a second one for completing the language tasks.

To examine how Sámi language use has changed across the generations (RQ1), we draw on data from the Stage I CATI interviews. In the telephone interviews, participants were asked whether they themselves, their parents, or grandparents used a Sámi language at home. As shown in Table 4, the proportion of respondents’ grandparents who used Sámi is high, but higher in Norway (75.60%) than in Sweden (56.54%). A Chi-squared test showed this difference was highly significant (χ² = 75.1307, p < .001). The number of respondents who report using Sámi today is much lower, at 21.80% in Norway and 20.60% in Sweden (this difference was not significant; χ² = 0.4051). These results suggest that the large gap that was apparent between the two countries for Sámi use amongst grandparents has disappeared for Sámi speakers today. This cannot be due to the age of the respondents, which was similar in Norway (mean = 53.16, sd = 14.37) and Sweden (mean = 52.13, sd = 16.31).

A potential explanation might be that many of the respondents are from Alta (Norway) and Kiruna (Sweden). These are relatively urban areas, where mixed-ethnicity households are probably more prevalent, making the use of Sámi at home less likely (cf. Trosterud, 2008). Further, there has been considerable migration to Alta from rural parts of the so-called ‘core area’ in Norway (Melhus et al., 2020). These areas have been less affected by the assimilation and modernization policies (e.g., Trosterud, 2008; Rasmussen and Nolan, 2011; Hermansen and Olsen, 2020). Hence, if the Norwegian parents or grandparents were from areas such as Tana, Karasjok or Kautokeino, they would have been more likely to use a Sámi language. In Sweden, by contrast, language shift seems to have been more sweeping, and even grandparents in the rural inland might have been less likely to use Sámi at home. Further, this data does not show how frequently Sámi is used by the individual. More refined measures of language use from the Stage II survey are reported in what follows.

In addition to the self-reported measures used in the survey, a small subset of participants who specifically indicated that they spoke North Sámi completed a vocabulary task in North Sámi (n = 44 in Sweden, and n = 38 in Norway). Since no easy-to-implement measure of North Sámi proficiency was available, we developed the North Sámi Vocabulary Task (NSVT) specifically for this purpose (see Gyllstad et al., in preparation, see footnote 3). The test was modelled on the DIALANG placement test (Alderson, 2005), and contained 50 real words and 25 pseudowords (see Supplementary materials).¹⁰ The items appeared on the screen one at a time, and participants were instructed to click ‘Yes’ if they thought the item existed, or ‘No’ if they did not. One point for each correctly-identified item was given, rendering a maximum score of 75.

Table 6 shows that the mean score on the NSVT is higher in Norway (55.97) than in Sweden (47.29). A simple linear regression indicated that the effect of ‘country’ on ‘vocabulary score’ is significant (β = −8.67, SE = 2.659, t = −3.263, p = .001). Assuming vocabulary knowledge can be taken as a proxy for general proficiency (e.g., Alderson, 2005), we can carefully conclude that North Sámi proficiency was higher in Norway. This might reflect the better educational policies in Norway, whereby children have better opportunities to learn the (written¹¹) language at school.

In an attempt to better understand which aspects of language experience were predictive of the vocabulary test scores, we looked at the effect of ‘age of first exposure to Sámi.’ We found that participants with a very early age of onset (i.e., between 0 and 1 years) achieved substantially higher scores on the test than participants who began acquiring Sámi as older children or adults. A basic linear regression showed that the difference between the ‘0-1 age group’ and all other age groups was significant (p < .01 or lower). This likely reflects the higher amount of input received by the ‘0-1 group’ as compared to the other groups (Figure 8).

RQ1 was concerned with changes in Sámi language use across generations. In the data from the Stage I CATI interviews, we observed a drastic drop in self-reported Sámi use between the generation of the grandparents (Norway = 75% Sámi users, Sweden = 53% Sámi users), the parents, and the participants themselves (23% vs. 20% Sámi users). Here we see the systematic effects of the older generation, i.e., the grandparents of today, having been forced to give up their language at school, and not passing their language on to their children, the parents of today (Olthuis et al., 2013). However, more-detailed results from the Stage II survey give us reason to believe that there is a pushback against this trend in Sámi-speaking parents today. Namely, the rate of participants who reported having used ‘mostly Sámi’ with their parents as preschool children is approximately the same as the proportion of participants who today use ‘mostly Sámi’ with their own pre-school children (around 5% of Sámi respondents in Sweden and 10% in Norway). This suggests that those who received high amounts of linguistic input as children themselves, i.e., today’s fluent Sámi-speaking parents, generally pass their knowledge of Sámi on to the next generation. One possible explanation is that attitudes towards the Sámi languages have improved in today’s parents as compared to previous generations, the majority of whom did not pass the language on to their children.

RQ2 addressed the levels of self-assessed language proficiency in non-Sámi respondents, who have not yet been the subject of investigation on Sámi matters to date. We showed that only 24% of Swedish non-Sámi and 36% of Norwegian non-Sámi participants reported being able to understand at least a ‘few words’ of a Sámi language. Considering that our respondents live in Sámi municipalities or in areas with a large share of Sámi residents, we think these numbers show that more work is needed to increase proficiency of Sámi in the majority population. Though self-assessed proficiency has to be taken with a grain of salt, we think these data are interesting, and they might reflect a higher aware of non-Sámi Norwegians in Sámi matters. Support for this idea can be found in (anecdotal) data we obtained in open comments fields, which we will report on briefly. We received a total of 19 comments from Swedish non-Sámi respondents, who indicated the question of Sámi at school was not relevant to them because they were not Sámi. However, we did not receive such comments from the non-Sámi Norwegian participants. This might suggest that Sámi learning is more widely seen in Norway as being relevant to everyone. For Sweden, this could be achieved by raising awareness of the Sámi culture in school education. Hettema and Outakoski (2020) points out that textbooks in Sweden do not contain ample information on national minorities, and also argues that more awareness building is needed in this area.

RQ3 targeted language use and proficiency amongst the ethnic Sámi respondents. For self-assessed proficiency, just over half in Sweden (57%) and just over two thirds in Norway (70%) said they could understand at least ‘a few words’ of Sámi. This result lines up with the most positive estimates in the field, e.g., those given by the Ethnologue, who estimate that 25,000 of 30,000 Ethnic Sámi people are language users (Eberhard et al., 2022). However, our results suggest these estimates are not ‘users,’ but more accurately people with loose contacts to the language, who understand a few words. In contrast, less than one third (27%) of Sámi in Norway and less than one fifth (18%) in Sweden assessed their comprehension as moderate or better, i.e., as potentially good enough to follow a conversation. This finding is closer to estimates by Rasmussen and Nolan (2011) and Hettema and Outakoski (2020), who gauge that only around one third of the Sámi population are Sámi users. Native-like proficiency was specified by only 11.3% of participants in Norway, and 4.18% in Sweden.

We also asked about language use in various contexts, which has not yet been done in previous studies. Based on a battery of questions about language use with family members across the lifespan, we showed that in both countries, around 20% of the ethnic Sámi population uses Sámi at least occasionally. However, the proportion who used ‘mostly’ or ‘only’ Sámi was extremely low, at 10% in Norway and at less than 5% in Sweden. We believe that this is the proportion of ethnic Sámi who can be considered as fluent speakers of a Sámi language.

Further, we noted a surprisingly high amount of receptive Sámi use in both countries. Over a third of Swedish Sámi and over half of Norwegian Sámi listened to the radio, watched TV, or used streaming services at least occasionally. This finding highlights the important role that is played by such channels in providing sources of language input. At the same time, we also know that passive exposure in itself is insufficient, and needs to be complemented by active use. This is because receptive bilinguals are less likely to be able to pass their language on to the next generation.

We further think that the higher rates of receptive engagement with the Sámi languages in Norway is due to the better support that is given for cultural measures, as evidenced by the higher budget for the Norwegian Sámi Parliament. By its own account, Samefolket is the only Sámi newspaper in Sweden, and publishes the majority of its texts in Swedish. In contrast, there are several Sámi-language newspapers in Norway, such as SÁMi magasiidna¹² and Ávvir. They receive considerable funding from the government. In fact, in 2013 Ávvir received the highest government subsidies of all newspapers in northern Norway.

To measure language proficiency directly and to test the reliability of our indirect, self-reported measures, we implemented a vocabulary task (NSVT) with a subset of participants who spoke North Sámi. Two findings are of interest: First, high test scores were generally only achieved by early bilinguals, who had acquired Sámi from birth. This was true for Sámi speakers in both countries, and reflects what we know from bilingualism research, namely that an early age of onset is crucial (Johnson and Newport, 1989; Abrahamsson and Hyltenstam, 2009), of course paired with high quality input (Chondrogianni and Marinis, 2011). Given the observation that only 10% of ethnic Sámi participants in Norway and less than 5% in Sweden use Sámi regularly with their children, we can conclude that current levels of language transmission are likely to be insufficient for the language to survive. In both countries, there is a need for a larger number of L2 learners—both Sámi and non-Sámi—and the need to work on raising the perceived importance of the Sámi languages in majority society. Second, we once again found significantly higher scores in Norway than in Sweden. Of course, this might be a reflection of the larger North-Sámi community in Norway than in Sweden, and it is likely that we would receive better results in Sweden for a test of South or Lule Sámi.

With RQ4 we examined whether potential differences between the two countries could be explained by their respective policies. Obviously, one-shot observational data like ours do not allow for a verifiable analysis of policy effects. Instead, the aim of this paper is to present novel data on Sámi language use and proficiency, and to elaborate on how these might be associated with themes found in the national Sámi policies.

Indeed, it seems the higher language use and proficiency levels we found in Norway may reflect educational policies. In Norway, there is a larger number of schools that offer Sámi lessons for a longer period of time and our data confirm that there is also a larger proportion of Sámi who learn a Sámi language at school (a total of 22.17% of Norwegian Sámi reported having had some form of Sámi lessons at school, but only 13.98% of Swedish Sámi respondents did so). As a result, Sámi-medium education should be extended to the high school years in Sweden (Belančić and Lindgren, 2020), as is already the case in Norway. At the same time, even in Norway existing policies may be insufficient. Vangsnes (2022) recently pointed out that, even if all school learners of Sámi became ‘bearers’ of the language and carry the language forward in time, the total number of Sámi speakers is likely to drop unless additional measures are taken, such as increasing the number of L2 learners, and making improvements to the curriculum. Further, as suggested by Albury (2015), Sámi learning at school needs to be made relevant to all citizens, including non-Sámi. Finally, it is conceivable that immersive teaching (i.e., a Sámi track in which all teaching is in Sámi) would be needed to create new speakers, as done for example in the Māori language in New Zealand (Hill and May, 2014).

Our analysis of budgetary spending showed that the Sámi parliament in Norway allocates a larger proportion of their budget to language and culture than in Sweden. Budgetary policies have a both symbolic and tangible effect on language experience. Although the total budget of the Sámi Parliament in Sweden is increasingly approaching that of Norway, the share of the budget that is related to the promotion of Sami languages seems to be declining slightly. The amount of financing available for the promotion of the Sámi languages conveys a message about how much the languages are valued, and spending a higher proportion of public resources on the Sámi languages signals a higher societal appreciation of the languages. More practically, a higher budget means more funding for producing radio and TV-shows, publishing books and newspapers in Sámi, and for other language undertakings. Taken together, the differences we found between the two countries in data seem to reflect the two countries’ Sámi language policies. Norway’s national Sámi language policies provide more comprehensive Sámi education opportunities and more monetary resources for Sámi language work, thus conveying a more affirmative status message about the languages to the overall public. As a result, we also see more frequent use of Sámi languages and higher proficiency levels (self-assessed and objective) in Norway.

It also seems likely that Norway’s move to make Sámi the co-official language has positively impacted on the awareness and acceptance of the Sámi languages amongst the non-Sámi population, who were shown in our study to be more likely to learn a Sámi language at school, and to have a higher level of Sámi language proficiency. Although some measures have been undertaken in Sweden, it seems paramount that awareness of the Sámi languages is increased in the majority population. One strategy could be higher budgetary allocations to language and a higher legal status of the Sámi languages. As mentioned earlier on, the EGIDS places the North Sámi language in Sweden on Stage 6, meaning that intergenerational transmission is taking part, but is not necessarily secure. Therefore, such measures need to be taken in parallel to strengthening Sámi language use in families. Otherwise, top-down measures introduced in isolation could threaten to institutionalise the language, rather than elevating the position of the Sámi languages to languages that are acquired as L1s by young children.

Admittedly, there are several limitations to these conclusions. Based on our analysis it is not possible to make a statement about the causal nature of these effects. The significant differences we found between Norway and Sweden for almost all measures might be due to the larger speaker community in the former country. At the same time, it may be that the larger speaker community is there due to the political support. Further, regardless of how good a policy may be, its ultimate success on the ground depends on complex societal conditions. For example, Sámi revitalization policies are challenged by increased globalisation and speaker mobility, whereby people who would benefit from these policies no longer live in the traditional Sámi areas, and in which multicultural family constellations make language transmission increasingly difficult. A further point is that, even if policies towards a minority group are improved, the negative attitudes towards these languages (and, by implication, towards the people who speak them) may be entrenched. Such persisting attitudes may prevent people from using and developing their minority language. Therefore, policy measures must go hand in hand with measures to increase the visibility and appreciation, also in non-Sámi areas. Ideally, Sámi and non-Sámi school students should be proud of learning a Sámi language, and they should be rewarded for doing so. Learning Sámi should become as attractive as learning any (foreign) language. Finally, it is worth pointing out that our study was carried out in the Northern provinces where there is generally more awareness of Sámi matters. For future studies, it would be interesting to look at awareness in the Southern provinces, where we would expect even more drastic differences.