Introduction

As a consequence of economic and political changes, non-profit organizations have been expected to play a greater role in welfare reforms in many nations of the western world, including Sweden, where New Public Management (private sector management styles intended to improve the public sector’s efficiency, quality and effectiveness) has dominated the public sector since the 1990s (Lundberg 2020). In Sweden’s public sector, however, having non-profit organizations provide welfare services is a contested issue that has ideological implications (Johansson et al. 2015). Voluntary social work has been simultaneously perceived as both a necessary alternative to the inefficient welfare system and as a threat to the universal welfare state (von Essen et al. 2018), and the political role of the third sector (i.e., organizations that are neither public nor private) in the Swedish welfare state is currently being negotiated. The purpose of my study is to improve our understanding of this third sector, specifically regarding how political identity in voluntary and non-profit organizations is negotiated, by analysing empirical data using linguistic theory and methods. I thereby contribute a linguistic perspective to the interdisciplinary research field of the third sector where disciplines such as political science (Lundberg 2020), social work (Johansson et al. 2015) and theology (von Essen et al. 2018) are long established. This investigation is an in-depth case study of a large dataset of texts collected from Save the Children Sweden (SCS), where I was an embedded researcher over the timespan of data collection and during the design of the study. My intention is to reveal how linguistic resources were used “in real time” in various documents and from different perspectives to negotiate that identity. Although I was employed by SCS at that time, my research funding was provided by a wholly separate source, thus creating an independent position. My location within the organization enabled me to get a deeper understanding of what texts were relevant, and also to access texts and background information on various practices via informal questions asked of employees and members.

In 2010, the main operational priorities of SCS regarding domestic issues were advocacy work directed at the state and municipalities, but the organization also performed services to a limited extent, e.g., running a psychotherapy clinic for children and youths in crisis (2010 SCS Annual Report). In 2016, SCS launched plans to expand its services substantially, in which the first initiative would have included compulsory and upper secondary schools (Strategic Plan 2017–2021). However, these plans for schools were questioned at the 2018 general assembly and were abandoned in 2020. Other services, such as youth recreation centres for afternoon activities and psychotherapy for children up to 6 years old, were realized in 2020–2022.

Texts, organizations, identity, and language

Organization studies using ethnographic methods have previously shown how important an organization’s own texts are for understanding that organization (Smith 2006; Turner 2006). I treat these texts as contextualized in the civil society framework as well as in the political, religious, and economic backgrounds of nations and geographical regions, as described by Salamon and Anheier (1997) and Anheier and Salamon (1998). In accordance with Smith (2006), I see texts as essential for many actions in organizations, where a text is an occurrence, embedded in what is happening and going forward (Smith 2006, p. 67). Furthermore, texts are vital coordinators of institutions, playing an important role in decision-making and in putting policy into action (Turner 2006).

Previous research has shown that texts are important in the creation of the ‘organizational self’, and in large organizations, texts are used to create internal images of the organization that are acceptable in various local settings and to create a coherent organization that is working for the same vision throughout (Gunnarsson 2009). Both of these tasks can be delicate and difficult. However, if one accepts that organizational identities are complex, ambiguous, imprecise, and change over time, as Albert and Whetten’s (1985) classic article on organizational identity suggested, these difficulties will not be considered necessarily problematic (which is in line with my own view on organizational identity). Similarly Sharp (2019), using grounded theory on text data, showed that rhetorical identification plays a particularly strong role for subordinate members. Eger’s (2021) analysis of ethnographic data also showed that the construction of an organizational identity involves cultural elements such as artefacts, values and language.

Linguistically, identity may be captured in various ways. The pronoun ‘we’, when it is used by organizations, provides a valuable entry point, since how ‘we’ are described, what ‘we’ do and what ‘we’ are reveal how the organization regards itself (Krizsán 2011; Seiler Brylla 2018). Like Busse (1997), I see the ‘we’—‘they’ distinction as a discourse semantic base figure. Another way to access identity may be to study metaphors. By examining the moral rhetoric as expressed in visons and mottos of two Christian voluntary organizations (Loaves and Fishes and the Salvation Army) and how that rhetoric was transferred into action and policies Allahyari (2001) revealed that metaphors and framing images have become essential to organizations. The type of metaphor that Allahyari studied differed slightly from Albert and Whetten’s (1985), as she discussed metaphors that she explicitly found in her observations, interviews, and texts, while Albert and Whetten instead addressed metaphors that could be used as overall descriptors of an organization that were not necessarily found in data, such as descriptions of university as either a church or a business (p. 283). Like Allahyari, I am looking for the explicit metaphors in my study. Departing from Lakoff and Johnson (1980), I see metaphors as prevalent and not only part of our everyday language, but part of our everyday thought. To understand abstract concepts, we use concrete ones.

In addition to uses of ‘we’ and metaphors, narratives are also useful for understanding organizational discourse (Cooren 2015), because narratives create meaning (Bruner 1986; Czarniawska 1998). Previous research has shown that narratives, together with intertextuality and rhetoric, can describe organizational transition (Maclean et al. 2018). Other studies have shown that managers craft historical narratives to influence workers’ identification with the organization (Aeon and Lamertz 2021). Organizations may tell more than a single version of their founding narrative, with each version reaffirming what a certain community sees as the core purpose of that organization (Foroughi 2020). Within the research field of organization studies, there have been calls for contextualizations of narratives from the past, since (i) historical claims are validated in a continuous dialogue with multiple audiences; (ii) narratives revise previously existing narratives by critiquing them or invoking earlier origins; (iii) narratives often result in rhetorical frictions that require continuous and skilful historical revisions to mitigate emerging conflicts in their reception (Lubinski 2018). A previous study on non-profit organizations that were experiencing tensions related to their governance has proposed the idea of “temporal blindness” for narratives in which certain events in the organization’s past are ignored, and “spatial blindness” for narratives in which certain parts of the organization are neglected (Bradshaw and Toubiana 2014).

Searching for ‘we’, metaphors, and narratives

The large dataset of this study requires various levels of analyses, and the strategy I use follows Krzyżanowski’s (2020) study on normalization of racism in which he separates entry-level and in-depth analyses. For my study, the entry-level thematic analysis encompassed a close reading of the data with the aim of capturing what was generally being discussed in relation to welfare services. In the dataset, two disagreeing parties within the organization became apparent—one party in favour of the expansion of welfare services on the part of the organization and one party against it. The entry-level thematic analysis enabled me to discern which individual texts contained arguments expressed by advocates of expanding welfare services, and which individual texts contained arguments expressed by opponents of expanding welfare services, where all arguments have been made on the part of the organization. In this phase, I also searched for potential linguistic differences regarding those arguments that were expressed by advocates and opponents of expansion. The view of the organization as a provider of services prevailed, resulting in youth recreation centres and psychotherapy. However, at the same time, the view of the organization as a state watchdog continued, so that SCS continued with, e.g., publishing texts on societal challenges (like child poverty reports).

The entry-level analysis indicated differences regarding the pronoun ‘we’, metaphors, and narratives, aspects that previous studies have shown to be useful for revealing rhetorical differences (cf. above). My preliminary entry-level analysis led me to operationalize my purpose into three analytical research questions, which belonged to the in-depth analyses.

RQ#1: How were ‘we’ described?

I limited the analysis of occurrences of the pronoun ‘we’ to where the word clearly referred to the organization SCS, or to individual parts of the organization. Inspired by Krizsán (2011, pp. 67–75) I analysed ‘we’ in connection to verbs, where Halliday’s Systemic Functional Grammar laid the basis for the classification of verbs (2014, p. 311). Halliday identifies six types of processes, where the verb counts as the process core. His processes are material (where the category means ‘doing’), behavioural (‘behaving’), mental (‘sensing’), verbal (‘saying’), relational (‘being’) and existential (‘existing’). However, I have applied a simplified version in which I group relational and existential processes together, and mental and verbal processes together. Material processes form one group, and I do not discern behavioural processes. This grid is coarser but still captures significant differences, and I believe the simplification makes the analysis more readily comprehensible.

I carried out a manual classification of all occurrences of ‘we’ (referring to SCS or its parts) in the dataset following the categorisations described below. I discerned between ‘we’ used about something that the organization:

  • did, as expressed by ‘we’ combined with material process cores, e.g., we perform…

  • was/had, expressed by ‘we’ combined with existence and relationship process cores, e.g., we are and we have…

  • experienced/thought/said/heard/knew/(etc.), expressed by ‘we’ combined with mental and verbal process cores, e.g., we believe…

Furthermore, following Seiler Brylla (2018)’s classification, I discerned and included occurrences of ‘we’ that were used to identify:

  • both the sender of a document and the addressee, inclusive ‘we’, e.g., we must all…

  • the sender of a document, exclusive ‘we’, e.g., we think that you need to…

RQ#2: What metaphors were used about the organization?

A linguistic metaphor is understood as a word or expression used in a way that contrasts to its more basic and concrete meaning. An example is the expression dead-end street. If it is used about a relationship like in “our relationship has hit a dead-end street” the expression forms a linguistic metaphor. In this use, it contrasts to a more basic and concrete meaning of dead-end street, namely a street that does not connect to anything else. According to Conceptual Metaphor Theory, metaphors allows us to understand a relatively abstract subject matter (in the example above, love) in terms of a more concrete subject matter (in the example above, a journey) (Lakoff 1993, p. 232). Metaphors may be very conventional, so that senders and addressees hardly notice them, or they be novel, where senders have chosen them with care and addressees become aware of them. Some studies have claimed that conventionalized metaphors can be very effective because they can influence addressees without those addressees being fully aware of the influence (Kövecses 2010, p. 12).

In my metaphor identification process, I relied upon the MIPVU (Metaphor Identification Procedure Vrije Universiteit) procedure proposed by Nacey et al. (2019a). MIPVU consists of manually going through the data and checking all words and expressions where a metaphorical use is possible. The procedure is: “(1) Identify the contextual meaning of the lexical unit. (…) (2) Check if there is a more basic meaning of the lexical unit. (…) (3) Determine whether the more basic meaning is sufficiently distinct from the contextual meaning. (…) (4) Examine whether the contextual meaning can be related to the more basic meaning by some form of similarity” (Steen et al. 2019, pp. 30–31). Concerning step 2, a “more basic meaning” corresponds the non-metaphorical meaning of the lexical unit. The actual checking of step 2 means looking the lexical unit up in a dictionary. The MIPVU-endorsed dictionary for Swedish data is Svensk ordbok (Nacey et al. 2019b, p. 141). How the lexical unit is presented and explained in the dictionary determines whether the two meanings (contextual meaning vs. more basic meaning) are understood as sufficiently distinct from one another (step 3). Strict MIPVU dictates that the senses should be separated by numbers or letters in the dictionary. An example is focus (Swedish fokus) which in Svensk ordbok has two senses, separated by numbers: (1) a point through a lens or on a concave mirror’s principal axis where formerly parallel light rays meet after being bent by the lens or mirror and (2) a centre for general interest. In addition, I allowed lexical units from the data to be valid as metaphors, if Svensk ordbok lists various uses under the same sense, as long as the use corresponding to the use in my data was marked as “metaphorical” (Swedish bildligt) or “abstract”. An example is end up (Swedish hamna) where the basic sense in Svensk ordbok is marked by a black circle and then described as ‘to unintendedly arrive at a certain physical place and remain there’. Under that description, a use is marked by a white circle “also regarding abstract phenomena, unintendedly find oneself in a certain situation” (Svensk ordbok). The use of end up (Swedish hamna) in my data corresponds to the description marked by the white circle and I have counted it as a metaphor. The motivation for including uses marked as “abstract” is that metaphoricity is about describing more abstract uses or senses in terms of more concrete ones (cf. citation to Lakoff 1993, p. 232 previously in this section). My modification of MIPVU for Swedish data has previously been applied by Gustafsson and Hommersberg (2018) who studied metaphors in Swedish cancer talk. To validate that I have understood the structure of senses, uses, circles, and meta-text in Svensk ordbok, I have consulted Emma Sköldberg, the dictionary’s editor-in-chief. In the metaphor analysis, I only included nouns, verbs, and adjectives from my data, and excluded all other word classes, including participles and particles. In a second stage, I analysed each identified metaphor regarding what source domain (field) it stemmed from, to understand how associations from that source domain followed from and gave meaning to the target domain in its contextual use. For dead-end street, the source domain would be journey (capitalized as domains are within Conceptual Metaphor Theory). The qualities of sharing common goals, facing impediments, and making decisions at crossroads about which direction to go give meaning to the contextual use (dead-end as an endpoint of a relationship) where the target domain is love (Lakoff 1993, p. 190).

RQ#3: What narrative(s) were told about the organization’s origin?

Finally, I searched for narrative(s) by examining whether narrative candidates exhibited certain traits. To meet the criteria, the narrative had to follow a strict disposition and include typical roles. The disposition involved:

  • Abstract—a short summary of the narrative

  • Orientation—time and place of setting

  • Complicating action—a problem

  • Evaluation—the part of the narrative that reveals the attitude of the narrator towards the situation and/or the narrative

  • Resolution/result—how the problem was solved

  • Coda—returning the reader/listener to the present time

This disposition follows Labov and Waletzky (1967) and Labov (1972). The roles, or actants, were gathered from Greimas (1966) who was inspired by analyses of folk tales. A narrative could, but need not, include all six actants. In the list below I have exemplified with roles identified in the folk tale Sleeping Beauty (modified from Martin and Ringham 2000, pp. 159–160):

  • Sender—makes the development of action possible (e.g., rumour about a sleeping beauty inside the castle)

  • Object—what the subject tries to achieve (possibly to give away) (e.g., love)

  • Recipient—the addressee who receives the object (e.g., the princess)

  • Helper—helps the subject (e.g., sword)

  • Subject—the main character who tries to achieve something (possibly in order to give it away) (e.g., the prince)

  • Antagonist—fights/competes against the subject (e.g., the thorn hedge)

Competing meanings and stories

In order to contextualize my study, I will start with a brief background of Save the Children Sweden and then give an account of it (based on related research) from two different standpoints, each of which contains competing ideas. The first standpoint is the organization’s origins as a Swedish folkrörelse, while the second standpoint can be discerned from stories about the historical origin and the chronological development of the organization.

Save the Children Sweden

Save the Children Sweden (SCS) has 58,000 members who are distributed in 147 local clubs, and these members have the option of performing volunteer work in the community (SCS Annual Report 2021). SCS can be regarded as a federation in the sense of Bradshaw and Toubiana (2014)—in other words, a member organization run democratically in several layers (local, regional, and national), each with its own board. Elected members vote at regularly held general assemblies, which make up the supreme decision-making body. Trygg (2015) showed that SCS regarded its membership as a fundamental cornerstone and an important part of the legitimacy and trustworthiness of the organization.

Competing meanings of folkrörelse

The Swedish welfare state developed throughout the 20th century, mainly under the administration of Social Democrats, who came into power in 1932. The party had close ties with Sweden’s folkrörelser, a word that is sometimes translated to ‘popular movements’, and incorporated many folkrörelser ideas into the welfare state (Hvenmark 2008). According to Johansson (1980), a folkrörelse (plural folkrörelser) is characterized as an organization that, over a long sequence of years, gathers a great number of people from various parts of the country around common ideas and interests. A folkrörelse should be democratically construed and lean on member activity, it should have idea-manifestos about humans and society, it should be independent in relation to the public sector, and membership should be optional (1980, p. 20). Wijkström et al. (in press) discerned a narrow sense of folkrörelse, which includes the Swedish Good Templar organizations, churches not associated with the Swedish state, and organizations fighting for workers’ rights. These authors further proposed a broader sense of the word which includes organizations like disability groups, sports associations, rural associations, and educational organizations targeting adults. They point out that the concept of folkrörelse sometimes includes all Swedish voluntary organizations (Wijkström et al. in press, Chap. 6, p. 7).

For my purposes, it is noteworthy that, in the first decades of 20th century, the workers’ movement and the Social Democrats resented notions of social charity and philanthropy, as such support was tied to an unequal class structure and to ideals and values that traditionally can be characterized as conservative and/or liberal. Even so, charity organizations had a prominent function at the time when Sweden formed its modern social politics. Many activities that were initially operated by charity organizations were subsequently transferred to the public sector, sometimes through public reforms, sometimes by an organization’s own initiative. However, representatives from folkrörelser (in Wijkström et al.’s narrow sense of the word) and other ideological associations had more radical approaches as they wanted to change the whole of society more thoroughly, e.g., through various national insurances (such as health insurance for all citizens). This approach challenged the contemporary thinking about charity and philanthropy. More specifically, the pressure from folkrörelser and other ideological associations forced the charity organizations to adapt to the new political climate, and as welfare reforms and higher standards of living in Sweden developed, charity organizations found new areas in which to operate. Charity organizations could now function as innovators (two examples are home help and elder care, which were later incorporated into the services provided by municipalities), as a complement, or as an interest group organization (Wijkström et al. in press, Chap. 6, pp. 20–28). My crucial question is how SCS relates itself to the term folkrörelse.

Competing stories Footnote 1 about SCS’s origin and identity

Earlier research showed that SCS has rewritten its history regarding its origin and identity (Sturfelt 2018). The first history version or story, which Sturfelt has discerned from texts published in the member magazine during SCS’s 90th anniversary year 2009, claims a transition from a charity organization into a rights organization, and that the 1989 Convention on the Rights of the Child marks the transition point between the two phases. The second story, which Sturfelt has distinguished from a jubilee book published 2018 by SCS in anticipation of its 100th anniversary, instead claims that SCS started in 1919 as a pure rights organization, which it has remained since then, where the declarations on child’s rights in 1924 (League of Nations), 1959 (United Nations), and 1989’s Convention on the Rights of the Child (UN) are milestones. In the contextualization of both stories, SCS claims that 20th century humanitarianism globally was subject to a sharp paradigm shift, from charity to rights (Sturfelt 2018). However, Sturfelt (2018) also argued that such a sharp 20th century humanitarianism paradigm shift never actually occurred, and that such an understanding obscures what the breaking points in the story of human rights are and when they happened. Sturfelt instead suggested that our understanding of the history of human rights would benefit from admitting that there are contradictory and competing stories. Lindkvist’s (2018) study supported the first version of SCS’s story, since her results showed that, around 1980, SCS transformed from a charity organization into a rights organization. My interest at present is whether these or possibly other stories can be connected to narratives in the data (cf. Section RQ#3: What narratives were told about the organization’s origin?).

Internal texts 2010–2020

The data were gathered from SCS and comprised relevant and representative internal texts written from 2010 to 2020. I collected texts from SCS to create a one-million-word corpus, the sections of which are listed in Table 1. The time span was set 2010–2020 to capture the change regarding the proposed service expansion plan, where the change seemed to occur around 2016. I chose central, internal texts where the negotiation about welfare services was likely to be seen in terms of suggestions (motions, bills, strategic plans, and business orientations) and decisions (minutes), but also in terms of alterations over time (charters, platform of values, annual reports from both national and local levels, member magazine, and newsletters). I also included agendas from arenas where members and employees met (dialogue forums and activity conferences). These texts formed the large dataset.

Table 1 Large dataset/Narrow dataset
Full size table

From the large dataset, I gathered texts that touched upon domestic accountability and welfare services in Sweden, which formed a narrow dataset, indicated in bold in Table 1. The analyses were performed on the Swedish original documents and have been translated in the examples given in the Results sections (The linguistic negotiation, How ‘we’ was used, How metaphors were used, and Narrative told). After this point, when the word “data” is used, it should be understood as referring to the narrow dataset. A complete list with a direct public link to all documents (as PDFs) is provided (SCS Sources, large data set).

The linguistic negotiation

My empirical findings, generated from analysis of how the pronoun ‘we’ was used about the organization, metaphor analysis, and narrative analysis, demonstrate how a part of the third sector is changing its approach to welfare. This example is captured “live” within the sector, through negotiation about political identity involving two parties in a voluntary organization. One party in the organization advocates for the position that the third sector itself should provide welfare services, while the other party in the organization opposes such provisions and instead stresses that the role of the third sector is to demand accountability from the public sector. The close reading of the entry-level thematic analysis revealed in what texts the various arguments are articulated. Advocates elaborate their arguments mostly in strategic plans, e.g., We provide third sector welfare services where they are needed (Strategic Plan 2017–2021) and in the board’s answer to motions, e.g., The fundamental purpose of testing school activities is, through social innovation, to develop schoolwork for the children in question (General Assembly 2018, Answer to motion 25). Opponents tend to explain their arguments in motions, e.g., There may be a risk that the child rights issues are neglected as SCS hastens away towards SCS3.0, third sector welfare provision and other direct support (General Assembly 2018, Motion 19) and member magazines, e.g., We within SCS don’t want to take over the responsibility of the public sector (Member magazine #2, 2011). Advocates in favour of the organization providing services used linguistic resources to convey the message about the voluntary organization as deed-crafty, united, prioritizing, goal-oriented, intervening, and as a self-evident historical welfare service provider. Opponents of this view, who preferred the organization to be more of a public sector watchdog, used linguistic resources to send a message about the organization as analytical, as part of the greater society, and as a forum for discussion (Fig. 1).

Fig. 1
figure 1

Internal negotiation on the organization’s political identity

Full size image

As mentioned in the introduction, this negotiation in SCS resulted in a substantial expansion of its welfare provision. Even though the plans for schools were abandoned in 2020, as of 2022, other welfare services were being provided. The outcome enables me to suggest theoretical claims about which linguistic resources were advantageous in the negotiations about the organization’s political role (see final section).

I will now give a detailed account of the findings emerging from this in-depth linguistic analysis. Examples from the dataset will be written in italics, and the word/expression in focus (‘we’ and metaphors) will also be bolded.

How ‘we’ was used to negotiate the organization’s identity

The findings show that the group in SCS that advocated for the organization to expand its voluntary welfare services (hereafter “advocates”) produced far more and longer texts than did the group that opposed such expansion (hereafter “opponents”) (Figs. 2, 3 and 4). As will soon become apparent, advocates and opponents used ‘we’ very differently, both in terms of the connection of ‘we’ to process cores (material process cores, existence and relationship process cores, or mental and verbal process cores) and in terms of inclusivity/exclusivity.

Fig. 2
figure 2

Welfare service advocates’ use of ‘we’ in connection with three categories of process cores

Full size image
Fig. 3
figure 3

Welfare service advocates’ use of inclusive and exclusive ‘we’

Full size image
Fig. 4
figure 4

Welfare service opponents’ use of ‘we’ in connection with three categories of process cores

Full size image

The advocates clearly connected ‘we’ to action verbs in a majority of the cases (60%, Fig. 2), such as We will implement relevant activities for and with children in the most deprived and marginalized areas of Sweden, through third sector welfare (Strategic Plan 2017–2021). I interpret this use of ‘we’ combined with material process cores as stressing the potency of the organization. ‘We’ used in connection with material process cores carried the organization’s identity and brand.

Advocates used ‘we’ in its inclusive way, i.e., referring both to sender and addressee, in nearly all cases (96%, Fig. 3), e.g., SCS’s main strength is that we are a popular movement [Swedish folkrörelse] (Strategic Plan 2017–2021).

I see this usage as a means to fortify SCS, portray it as united, and legitimize its welfare provision. The specific example SCS’s main strength is that we are a popular movement resonates with Trygg’s (2015) results, which showed that SCS regarded its members as an important part of the legitimacy and trustworthiness of the organization.

At first glance, it seems as if the opponents had a more balanced use of ‘we’ in terms of the three categories, each of which was used more or less a third of the time (Fig. 4). However, closer scrutiny revealed that the opponents in two cases connected ‘we’ to material process cores combined with ‘don’t/not’ to indicate what the organization should not do: We within SCS don’t want to take over the responsibility of the public sector (Member magazine #2, 2011), and once connected ‘we’ to material process core in a question: But sometimes when the social responsibility of enterprises is discussed, the motive of the enterprises is questioned. Do they want to mitigate their guilty consciences? Buy themselves free? And if that is the case, should we as third sector organizations really help them with that? (Member magazine #2, 2011). Even so, ‘we’ was used positively with material process cores regarding domestic issues such as welfare services in Sweden nine cases, corresponding to 28% of all cases.

Compared to how advocates used ‘we’ combined with material process cores (60% of all cases, see Fig. 2) I interpret this pattern as indicating that the opponents were downplaying material processes in relation to the identity of the organization. When opponents combined ‘we’ with existence and relationship process cores, they only referred to SCS itself in two cases out of ten, but in eight cases out of ten, they referred to SCS in terms of something bigger, such as the whole society, e.g., society as something we are all part of (Member magazine #2, 2011), which qualifies as an inclusive use. Alternatively, ‘we’ could refer to civil society in relation to other actors, e.g., depending on what types of organizations we represent, we have different missions and play different roles (Member magazine #2, 2011). My interpretation is supported by a sentence in the same member magazine piece: Today, an active discussion is taking place in Sweden about the relations between public and civil society (Member magazine #2, 2011). The ‘we’ referring to civil society qualifies as an exclusive use in relation to the member magazine’s readership, because not only members but also private persons in general (as well as professional officials such as teachers, nurses, etc.) were targeted by these texts. The opponents’ inclusive use of ‘we’ was slightly less frequent (41%) than their exclusive use of ‘we’ (59%) (Fig. 5).

Fig. 5
figure 5

Welfare service opponents’ use of inclusive and exclusive ‘we’

Full size image

I interpret these proportions as reflecting the fact that the opponents were significantly more inclined to separate themselves from the addressees of the documents than were the advocates. Furthermore, the relatively high degree of mental and verbal process cores (31%, see Fig. 4) helped the opponents convey SCS’s analytical capacities: we have seen in other areas, for example in the women’s shelter work, that the public sector has tried to put too much [work load] on the voluntary organizations (Member magazine #4, 2011). What one experiences and how one analyses it are important tools in advocacy work, so these uses can be understood as strengthening the advocacy role of SCS.

How metaphors were used to negotiate the identity of the organization

My analysis shows that advocates for and opponents of expanded services sometimes used metaphors about the organization in similar ways, and sometimes in different ways. Both groups used metaphors from the source domain sight. Among the advocates’ metaphors, one particularly common type was about focus, e.g., a child rights organization with engaged members, with a focus on activities (General Assembly 2020, Bill).

The literal meaning of ‘focus’ (Swedish fokus) is a point through a lens or on a concave mirror’s principal axis where formerly parallel light rays meet after being bent by the lens or mirror. In its metaphorical meaning, focus is used to indicate ‘a centre for general interest’ (Svensk ordbok). Clearly the metaphorical sense was used in the quotation.

In contrast, the opponents did not use metaphors based on focus. Instead, they used metaphors from the source domain sight about their points on various issues, e.g. she [chair of SCS] sees a risk that municipalities will try to relocate their responsibility to the voluntary organizations (Member magazine #4, 2022). ‘Sees’ (Swedish ser) has a literal meaning, ‘to perceive from sight’, and at least two metaphorical meanings, ‘to notice something’ and ‘to judge something’ (Svensk ordbok). It is not easy to determine which of these two metaphorical meanings was used in this example (possibly a mix of both).

Both groups also used metaphors relating to the source domain source-path-goal. Interestingly, only the advocates had metaphors highlighting the goal, e.g. An important goal is that more children should pass compulsory school so that they can carry on to upper secondary school (General Assembly 2018, Answer to motion 25). ‘Goal’ (Swedish mål) has a literal meaning, ‘the endpoint of a motion path’, and a metaphorical meaning, ‘the intended result from an activity’ (Svensk ordbok), but in the quotation, clearly the metaphorical meaning was used. The opponents, on the other hand, instead used these metaphors to convey lack of intention, such as ‘ending up’ somewhere: In my circumstances, I often end up in discussions about who is in fact responsible for our common welfare (Member magazine #2, 2011) ‘End up’ (Swedish hamna) in its literal sense means to unintendedly arrive at a certain physical place and remain there (Svensk ordbok). It may also be used in relation to abstract phenomena, in which case it means to find oneself in a certain situation (Svensk ordbok). In the quotation, the discussion is compared to such a situation, and thus ‘end up’ is regarded as metaphorical. I interpret this use as the opponents wanting to show more open-mindedness as opposed to have an orientation strictly toward achieving a goal.

Finally, the advocates used metaphors relating to the source domain in/out, where the target domain consisted of what interventions, activities, and services the organization would do, such as: SCS will scale up the use of more direct interventions for children, including running third sector welfare provision (e.g. recreation centres, schools, refugee housing etc.) (Strategic Plan 2017–2021).

The Swedish word used here, insats (plural insatser, translated to ‘interventions’ above), refers to a loose part of an object that may be inserted into a main object—for example, the filter of a tea pot, or the removable cot of a pram. But in its metaphorical sense, insats means well-performed work that contributes to a certain result (Svensk ordbok). The word insats occurred frequently in the data, which makes it it the most common metaphor used by the advocates.

The opponents did not use metaphors from the source domain in/out to the same extent; instead, they used metaphors from the source domain sound, in terms of metaphors about raising one’s voice to address an issue, e.g., to ensure that the General Assembly/the member motion has an opportunity to provide a clear view on the issue [Swedish fråga] (General Assembly 2018, Motion 25). The noun ‘fråga’ has two uses, the first of which corresponds to literally ask for information (Svensk ordbok) in speech but also in text, and which would correspond to English ‘question’, while the second use means a relevant subject that needs to be discussed (Svensk ordbok), and so corresponds to English ‘issue’. In the example sentence, the latter, metaphorical use was being used. Other metaphors from the target domain sound were used about voice, e.g., SCS is a strong child rights organization and a strong voice for the rights of the children (General Assembly 2018, Motion 19). ‘Voice’ (Swedish röst) has a literal sense, ‘to use the human organ that produces sound, mostly for speech or song’, and at least two metaphorical senses, the first meaning conscience and the second one opinion (Svensk ordbok). In the quotation, clearly neither speech nor song were intended, but rather conscience or opinion (or possibly both). I interpret the differences between how advocates and opponents used metaphors as each party trying to effectively convey their respective messages: the advocates about action and interventions, both dynamic aspects, the opponents about discussion and advocacy work, both more static aspects. In all, the metaphors were conventional, and as such, it can be argued that they probably passed “under the radar” for the readers, which, according to Kövecses (2010, p. 12) may be an effective way to convey one’s message.

Narrative told about the organization’s origin

My analysis reveals that advocates put forward a strong narrative about the origin of the organization to legitimize the expansion of welfare services. I found no traces of narratives about SCS’s origin that could legitimize the opponents’ view that SCS shouldn’t expand its welfare services.

In the long quotation from the data below, the one instance of narrative supporting the SCS’s intervention and provision of welfare services is given, both in its English translation and the Swedish original. I have separated the sentences into sections and labelled each section.

[Abstract:] Swedish society is undergoing a shift in the view on civil society’s role and responsibility. The situation has changed for civil society, with clearer demands concerning relevance, effectiveness and feedback.

Swedish original: Det svenska samhället genomgår ett skifte i synen på civilsamhällets roll och ansvar. Förutsättningarna för civilsamhället har förändrats med tydligare krav på relevans, effektivitet och återkoppling.

[Orientation:] In the early 20th century, the Swedish popular movements were beacons of social innovation. Activities such as public dental care, home help and preschool were initially founded by popular movements and organizations and then later transferred into the Swedish welfare system. Over the past few decades, we have seen incredible progress for the vast majority of children in Sweden,

Swedish original: Under tidigt 1900-tal stod de svenska folkrörelserna för social innovation. Verksamheter som folktandvården, hemtjänsten och förskolan grundades av folkrörelser och organisationer och övertogs och förflyttades sedan in i den svenska offentliga välfärden. Under de senaste decennierna har vi sett en fantastisk utveckling för de allra flesta barn i Sverige

[Complicating actions:] but groups of children have nevertheless been left behind.

Swedish original: men grupper av barn har samtidigt lämnats utanför.

[Evaluation:] SCS’s most recent child poverty report (2015) shows that 12 percent of children in Sweden live a life of social and economic exclusion. (…) We meet children whose prospects for completing school with good grades are non-existent and we meet children whose life chances are inhibited by violence, in the home and in the public sphere. (…) The crisis of confidence between the public service and the excluded citizens remains large and continues to deepen. (…)

Swedish original: Rädda Barnens senaste fattigdomsrapport (2015) visar att 12 procent av barn i Sverige lever i socialt och ekonomiskt utanförskap. (…) Vi möter barn vars utsikter att klara skolan med godkända resultat är obefintliga, och vi möter barn vars livschanser beskärs av våld, både i hemmet och det offentliga rummet. (…) Förtroendekrisen mellan den offentliga servicen och medborgare i utanförskap är fortsatt stor och fördjupas. (…)

[Resolution/Result:] The conclusion is that SCS, like other actors in the third sector, must once again take up the role of social innovator in a drive to promote a socially sustainable society that safeguards all child rights.

Swedish original: Slutsatsen är att Rädda Barnen, likt andra aktörer i den idéburna sektorn åter igen måste gå in i rollen som samhällsbyggare i strävan att verka för ett socialt hållbart samhälle som säkrar alla barns rättigheter.

[Coda:] The focus will be on “Socioeconomically deprived children”, “Children in migration and displacement” and “Children subjected to violence”. Activities are expected to be run by members and active volunteers, as well as employees with particular skills.

Swedish original: Fokus kommer vara på “Barn i socioekonomisk utsatthet”, “Barn i migration” och “Barn utsatta för våld”. Verksamhet förväntas bedrivas av såväl medlemmar och ideellt engagerade som anställd personal med särskilt kompetens.

(Strategic Plan 2017–2021, pp. 9–10, Swedish original, p. 12)

The narrative includes an orientation that established the time (early 20th century) and place (Sweden) where folkrörelser developed innovations that were later transferred to the welfare system. However, over the past few decades, this development has been interrupted by complicating actions that resulted in groups of children being left behind. The evaluation provided by SCS is that these children live without opportunities, their lives are restricted by violence, and a crisis of confidence is growing between the public services and excluded citizens. The resolution/result consists of SCS once again accepting the role of social innovator, and the coda takes the reader back to the present time by stating who would be targeted and by whom the work would be done.

The actant analysis shows that SCS is the subject, i.e., the actant that would perform the action, since it is claimed in the quotation that SCS, like other actors in the third sector, must once again take up the role of social innovator in a drive to promote a socially sustainable society. This statement matches well with the advocates’ emphasis on action, as previously shown in the analysis of ‘we’ in connection to action-oriented process cores, and also with the advocates’ use of metaphors with the source domain source-path-goal, where goal is emphasized. The groups of children left behind are the recipients, i.e., those who would benefit from SCS’s action. These children are specified at the end of the narrative as socioeconomically deprived children, children in migration and displacement and children subjected to violence. The object, i.e., what should be handed over to the recipients, is rather vague in the quotation, but is later specified as activities such as integrated social enterprises, case management centres, recreation centres, compulsory schools and upper secondary schools (p. 12), in other words, welfare services. Again, this narrative goes well together with the advocate’s action-orientation, revealed in the previous analysis of ‘we’. No helper can be discerned from the quotation, but later it is stated that the commitment of the members and their will to contribute actively form an indisputable added value which gives SCS a strong local anchoring (General Assembly 2018, Answer to motion 25), which indicates that the members function as helper. However, it is not evident what actant the public sector corresponds to in the narrative. From one perspective, the public sector qualifies as the actant sender: New operational forms such as third sector welfare … open up new opportunities for financing from municipalities, county councils and the state (p. 18), but from another perspective, the public sector qualifies for the actant antagonist, at least an antagonist against which to measure one’s strength: Society’s inability to support these children compels SCS to operate in places where others have been unable to deliver results (p. 10). The latter perspective is supported by the metaphors from the source domain sight (cf. above) that point out the inadequacies of the public sector.

The narrative functions as a strong legitimation that the organization, thanks to its origin, should take up welfare services.

The outcome of the negotiation, and implications for third sector research

The fact that the advocates of welfare services pictured the voluntary organization as deed-crafty, united, prioritizing, goal-oriented, intervening, and as an evident welfare service provider, thanks to history, may have contributed to their successful result, i.e., that the organization indeed expanded its welfare services substantially. The opponents of welfare services used linguistic resources to convey an image of the organization as analytical, as part of the greater society, and as a forum for discussion, and they were defeated. The advocates showed that they were willing to play a greater part in welfare reforms, as they were expected to do, in the wake of the introduction of New Public Management (cf. Lundberg 2020), and the advocates took a clear side in the greater societal debate on voluntary social work, namely that welfare services provided by voluntary organizations form a necessary alternative to inefficient public welfare (cf. von Essen et al. 2018).

The most interesting difference between the advocates and the opponents is that only the advocates used a strong narrative (with one manifestation in the data), which can be interpreted as support for Bruner’s (1986) and Czarniawska’s (1998) claims that narratives create meaning. The narrative about SCS, published in 2016 in the Strategic Plan 2017–2021, then becomes an alternative to the two competing stories revealed by Sturfelt (2018): the first one portraying SCS as a charity organization which later transformed into a rights organization (supported by member magazine data from 2009 in Sturfelt’s study), and the second one portraying SCS as a rights organization from the beginning (supported by jubilee book data 2018 in Sturfelt’s study). The strong link between rights and advocacy could qualify either story, or even both stories, as supporting the opponents’ idea of SCS. However, such support is just a hypothesis that would need further testing, as the two stories revealed by Sturfelt did not appear in my data (our data not being identical). The narrative I found, which was published 2016 in the Strategic Plan 2017–2021, portrays SCS as a folkrörelse that, since its very beginning, was (and still is) an innovative welfare service provider: In the early 20th century, the Swedish folkrörelser were beacons of social innovation. Activities such as public dental care, home help and preschool were initially found by folkrörelser and organizations and then later transferred into the Swedish welfare system. (…) The conclusion is that SCS, like other actors in the third sector, must once again take up the role of social innovator in a drive to promote a socially sustainable society that safeguards all child rights. The narrative then qualifies as one of several versions of SCS’s foundational stories (cf. Foroughi 2020) and it revises the previously existing story by a skilful history revision (cf. Lubinski 2018) without contextualizing the organization’s evolution during the 20th century (cf. Lubinski 2018; Sturfelt 2018). This historical revision is accomplished by using the term folkrörelse in its broadest sense, in which all voluntary organizations may be included, and by erasing the conflict between, on the one hand, representatives for folkrörelse in its narrow sense, and on the other hand, charity organizations (cf. Wijkström et al. in press). It does not seem to matter that SCS did not in fact contribute to the foundation of public dental care, home help and preschool—documentation can only confirm that SCS donated toys to preschool-like settings in Swedish refugee camps in 1945 (Fredricson 2020, p. 36). Promoting such a pragmatic narrative could count as what Bradshaw and Toubiana (2014) refer to as temporal blindness, but I would prefer to suggest that those promoting this narrative are using ‘temporal blinkers’, since parts of the history can be confirmed by documentation. The pragmatic narrative supports Albert and Whetten’s (1985) view on organizational identity as complex, ambiguous, imprecise and changing over time. My results further support Smith’s claim (2006) that texts are essential for many actions in organizations. In addition, these results give support to Turner’s point (2006) that texts are vital coordinators of institutions, as texts play an important role in decision-making and putting policy into action. My findings are also in harmony with Gunnarsson (2009), who showed that texts are very important in the creation of ‘organizational self’.

My theoretical contribution to third sector research is the suggestion that, in negotiation about political roles in the welfare state, it is advantageous for advocates in favour of change to:

  • Use ‘we’ in a way that includes both senders and addressees of the text in order to appear united (e.g., SCS’s main strength is that we are a popular movement in an internal document where the intended audience includes members and employees of SCS)

  • Combine ‘we’ with material process cores (e.g., we will implement) to appear deed-crafty

  • Use metaphors grounded in activity (e.g., an important goal is that …) to appear potent

  • Provide a more or less accurate historical narrative, possibly using ‘temporal blinkers’ to appear legitimate in one’s “continued” activities (e.g., The conclusion is that SCS, like other actors in the third sector, must once again take up the role of social innovator)

Through my empirical and critical linguistic study of a non-profit organization’s texts, I have shown how the pronoun ‘we’, metaphors, and narrative may be used as important linguistic resources for a third sector organization as it negotiates its place in the welfare state. My analysis complements and expands upon Eger’s (2021) ethnographic work on the co-construction of non-profit organizational identity, and Sharp’s (2019) text-based, grounded theory work on organizational change. In contrast to Allahyari (2001), who revealed framing images and metaphors of organizations that had been “sedimented into structure” (2001 p. 200), my study shows that “live” negotiation takes place among competing framing images and metaphors. It is my hope that my consideration of a text data collection and critical linguistic approach will inspire and facilitate subsequent studies.