‘Sexual Orientation’ in Swedish Preschool Policy—What Is the Problem?

The present article focuses on how ‘sexual orientation’ is represented and produced in a Swedish preschool policy document regarding discrimination and equal treatment. ‘Poststuctural policy analysis’ is employed, in line with Foucault) and Bacchi. The results show that ‘sexual orientation’ is represented as a matter for families, but for parents rather than children. In the plans for equal treatment, visualizing different families stands out as the goal of working preventively against discrimination based on ‘sexual orientation’ in preschool, and the active measures planned for are reading books and spontaneous conversations. The article argues that the discrimination perspective represented in the documents, together with discourses on childhood innocence, establish certain conditions for how ‘sexual orientation’ is produced in preschool.


Introduction
Aside from being a central part of children's life, the preschool setting is an interesting context for negotiations about discourses, norms and values concerning childhood. The present article focuses on how questions about sexuality and sexual identity interact with, and are often separated from, childhood. On the one hand, according to the official curriculum (Lpfö 18; Swedish National Agency for Education 2019), children in the Swedish preschool system are entitled to equal rights and opportunities, and the Swedish Discrimination Act (SFS 2008:567 1 ) obligates preschools to prevent discrimination based on 'sexual orientation.' On the other hand, previous research in the field has shown that heteronormativity is reproduced within early childhood education (ECE), mainly through the silencing and exclusion of non-heterosexuality, such as same-sex parents (Gunn 2008(Gunn , 2011Janmohamed 2014;Surtees 2008;Surtees and Gunn 2010). In the tension between these silences and the obligation to address 'sexual orientation' in policy, the present article aims to examine how the grounds for discrimination based on 'sexual orientation' are represented and constituted in Swedish preschools' local policies for equal treatment.
Discrimination based 'sexual orientation' is prohibited in the European Union (McGoldrick 2016), and as a member state, Sweden is obligated to work against discrimination, for example in workplaces and education institutions. The Swedish Discrimination Act (SFS 2008:567 2 ) stipulates that preschools should prevent discrimination in seven areas: Sex, Transgender identity or expression, Ethnicity, Religion or other belief, Disability, Sexual orientation and Age. The grounds for discrimination were introduced in the Swedish curriculum for the preschool with the revision of 2010 (Swedish National Agency for Education 2011). The previous wording in Lpf 98, revision 2006, prescribed that preschools should ensure that all children develop an "understanding that all persons have equal value independent of gender, social or ethnic background" (Swedish National Agency for Education 2006, p. 7). In 2010, it was replaced with the following wording: "an understanding that all persons have equal value independent of social background and regardless of gender, ethnic affiliation, religion or other belief, sexual orientation or functional impairment, [ . . . ]" (Swedish National Agency for Education 2011;2016, p. 8). The latter version is in accordance with the five grounds for discrimination that the preschools were obligated to work preventively with at that time (Sex, Ethnicity, Religion or other belief, Disability, Sexual orientation). Starting in 2017, the obligation for preschools also includes Transgender identity or expression and Age, from that point including all seven grounds for discrimination that are covered in the law (SFS 2008:567 3 ). In 2018, the following paragraph, in which seven grounds for discrimination are presented, was included in the new curriculum for preschool, Lpfö 18 (Swedish National Agency for Education 2019, p. 5): No child in the preschool should be subjected to discrimination on the grounds of the gender, transgender identity or expression, ethnic origin, religion or other belief, disability, sexual orientation or age, of the child or any person with whom the child is associated [...].
Changes in the Discrimination Act seem to have had an impact on the wording in the curriculum and 'sexual orientation' or sexuality is not mentioned elsewhere in the curriculum. I therefore argue that the introduction and presence of 'sexual orientation' in the preschool curriculum is the result of 'sexual orientation' being one ground for discrimination that the Swedish preschools are obligated to work with. In the local policy document 'plans for equal treatment,' preschools deal with the grounds for discrimination by formulating goals and active measures to prevent discrimination and support equal treatment. Given that these documents, 'preschools' plans for equal treatment,' are the most explicit representations of LGBTQ issues in Swedish preschools, their role in the constitution of these questions is central.
For more than ten years preschools are obligated by anti-discrimination law and the curriculum to work against discrimination based on sexual orientation. This has a great impact in giving this issue legitimation in the context of Swedish preschools, but the interest of the present article is how this translates into local policy and how sexual orientation is represented when goals and active measures are formulated by teachers and headmasters. For this reason, the present article aims to examine how these representations take place in local policy documents and how they interact with current discourses on childhoods and sexuality.

Previous Research on 'Sexual Orientation' and Heteronormativity in Preschool
In relation to the Swedish context presented above, this section provides knowledge from international research that the present article builds upon. There are several previous studies on 'sexual orientation' and heteronormativity in preschool conducted from an adult perspective, in which caregivers or teachers are interviewed (Glass et al. 2016;Gunn 2008;Janmohamed 2014;Robinson and Ferfolja 2001;Ferfolja and Robinson 2004;Surtees 2008). In interview studies with parents and teachers who identify themselves as homosexual, bisexual or queer, it emerges that these adults believe the preschool system lacks competence as regards norm-breaking families, resulting in non-heterosexuality being excluded or silenced (Cloughessy et al. 2018;Gunn 2008Gunn , 2011Gunn et al. 2004;Robinson and Ferfolja 2001;Surtees 2008;Surtees and Gunn 2010). Directors and teachers within Australian ECE feel unprepared for same-sex parents, stating that they need more knowledge about how ECE can address heteronormativity (Cloughessy and Waniganayake 2015).
In contrast, Ferfolja and Robinson (2004) find that teacher educators question the need for anti-homophobia education for early childhood educators, given the young age of the children they teach. Additionally, Robinson (2002) finds that teachers within Australian ECE themselves believe that children in the setting are too young to understand non-heterosexuality and sexual diversity. The teachers thought that these questions were only relevant if there was a child with same-sex parents within the actual ECE setting they were working in, and not of relevance for all children, regardless of their parents' sexual orientation. Studies on how children themselves deal with heteronormativity in preschool have shown that they are active in both producing and challenging these norms, for example in their play (Blaise 2005(Blaise , 2009(Blaise , 2010; Lyttleton-Smith 2019; Removed for peer-review, Taylor and Richardson 2005). Robinson (2002Robinson ( , 2005 shows that teachers within ECE find it difficult and risky to address non-heterosexuality, given the idea that it is a sensitive and inappropriate subject to expose children to. Ullman and Ferfolja (2015) reveal that policy documents regarding LGBTQ education in Australian schools tend to focus on difficulty and marginalization of these students, and the potential controversy for teachers to address these issues, thus failing to normalize non-heterosexuality. One of the strongest discourses concerning the child and childhood is innocence, where the child is constructed as being in need of protection from knowledge of, among other things, sexuality (Bruhm and Hurley 2004;Epstein 1997;Prout 2005;Robinson 2013). According to Foucault (1990), the child is considered non-sexual, and talk about childhood and sexuality is surrounded by denial and restrictions. Sexuality, in general, and non-heterosexuality, in particular, are discursively separated from, and positioned in contrast to, children and childhood (Edelman 2004). This taboo is demonstrated, for example, in the aversion to raising these questions in preschool (Surtees 2008;Surtees and Gunn 2010;Robinson 2012Robinson , 2013Sparrman 2014). Surtees and Gunn (2010) point out that silence about alternatives to heterosexuality is the way that heteronormativity remains dominant within the ECE setting.
To summarize, this body of research finds that sexuality and sexual diversity is being silenced and excluded and, by this heteronormativity, is reproduced in the ECE setting. Given that these questions still need to be addressed in different ways, the present article focus on the presence of 'sexual orientation'. When obligated, how does Swedish preschool deal with 'sexual orientation' in the policy document? In which ways, and under what conditions, is 'sexual orientation' in the preschool setting represented and produced?

Poststructural Policy Analysis
The methodological framework of the present article is based on Foucault's ontology of discourse and uses Bacchi's analytical approach to how problems are represented and constituted in policy. Foucault (1990) argues that sexuality is produced through discourses and regulated by power mechanisms. Sexuality discourses are based on categorizations between, for example, heterosexuals and homosexuals, women and men, and children and adults (Foucault 1990). Foucault criticizes categorizations such as these for providing the basis for a hierarchy of sexual identities, where heterosexuality is normalized to the point where it more or less disappears as an identity or subject position (see also Butler 1990). To Foucault (1990), sexual identities or orientations are not essentialist, but discourses that are being produced. Queer theorists building on Foucault have focused on normalized heterosexuality and its consequences, rather than on sexual identities and practices (Butler 1990(Butler , 1993Jagose 1996;Sullivan 2003). Discourses are defined in terms of both what is said and what is not said. But the limitation of speech is not the same thing as silence, Foucault (1972Foucault ( , 1990 argues, but shows the rules by which the discourses are produced. For example, there are a lot of limitations and conditions concerning how it is possible to speak to children about sexuality (Foucault 1990). Foucault (1972, p. 230) argues that discourses produce 'possible conditions of existence', describing how limitations and exclusions regulate what it is possible to say or do in a given context, and in what way. Bacchi (2009Bacchi ( , 2012 'What is the problem represented to be?' approach (WPR) or poststructural policy analysis (Bacchi and Goodwin 2016) draws on Foucault and offers an analytical tool that views policy as producing a problem rather than addressing or "fixing" problems (Bacchi 2009(Bacchi , 2012Bacchi and Goodwin 2016).
This analytical approach focuses on problem representations in policy, thus on how problems are produced through their proposed solutions: "what we propose to do about something indicates what we think needs to change and hence what we think the 'problem' is" (Bacchi and Goodwin 2016, p. 16). As Goodwin (2011) points out, doing a policy analysis is a matter of studying contemporary social and cultural discourses (see also Bacchi 2000). Policy does not only mirror and reproduce discourses, it also produces discourses and constitutes subject positions, for example by gendering men and women (Bacchi 2017). Bacchi (2012, p. 21) offers several questions to guide the analysis; these are summarized as: first looking at how these representations takes place in the documents and what is produced as the subject of these representations. Second, who and what is not represented in the documents. Third, in the discussion, the present article deals with how the problem representations are constructed in the way they are, and how this could be rethought.

Data
The policy documents being analyzed are 30 plans for equal treatment produced locally at preschools from different parts of Sweden. The plans for equal treatment were collected from preschools' websites during November 2016, and had been produced by the preschools between 2013 and 2016. The sample is distributed across larger and smaller municipalities from the north to the south of Sweden to represent a broad context. There are 10 documents from the three largest municipalities (Stockholm 1-4, Gothenburg 1-3 and Malmö 1-3), 10 documents from middle-sized municipalities (Halmstad, Umeå, Kristianstad, Karlstad, Eskilstuna, Örnsköldsvik, Gävle, Örebro, Norrköping and Luleå), and 10 documents from small municipalities (Upplands-Bro, Vetlanda, Landskrona, Sandviken, Sala, Lidingö, Partille, Kiruna, Sölvesborg, Trelleborg). In the analysis, no comparison between different areas or schools has been done. Like most preschools in Sweden, the selected preschools are run by the municipal authorities and accept children from one to six years of age. Another criterion for including the document in the sample was that the grounds for discrimination based on 'sexual orientation' are mentioned in the goals and active measures.
The documents are between 6 and 26 pages long and often include descriptions of the obligations that preschools have to prevent discrimination. The grounds for discrimination are often described and presented one at a time, but in some cases the goals formulated cover several grounds for discrimination. In relation to the grounds for discrimination, the preschools formulate goals and active measures to promote equal treatment as a way of preventing discrimination. For the present article, the goals and active measures formulated in relation to 'sexual orientation' were the primary data being used in the analysis. In the documents, the preschools formulated around one or two goals and one or two active measures regarding the grounds for discrimination based on 'sexual orientation.' In the following sections, the goals and active measures will be analyzed, allowing me to discuss how the representation of the problem is constituted by the proposals offered. I begin with the reoccurring goal visualize different family constellations, continuing with the reoccurring active measure reading books and spontaneous conversations. Based on the questions offered in the WPR-approach Bacchi (2012, p. 21), the following analysis deal with (a) representation of the problem in the goals and active measures, (b) categories and binaries taken for granted, (c) what is left out in the representation and (d) what is the possible effect of the representation and how it could be thought of otherwise.

Visibility for Families
The proposed solutions presented in the policy documents are related to treating all families equally, and to visualizing different family constellations. The problem representation constituted in these proposals is that all families are not equally treated and equally visible in preschool. The following example from one preschools 'plans for equal treatment' shows how families is highlighted in the description on how to work preventively against discrimination based on 'sexual orientation': Together with the children, the pedagogues have read and borrowed literature that shows on different family structures. Preschool works preventively so that everyone is respected for who you are. Everyone should feel welcome. Some group had family as thematic work where they talked about family constellations. (Gävle) This example emphasizes respecting and welcoming everyone in preschool and describes how they worked with reading and talking about families. The connection between equal treatment and families is not explicit, but reoccurring in different forms. Another document describes their preschools work regarding 'sexual orientation' as "promote understanding of different family formations and human equality" and gives three examples on how this work is done: "We talk about how families can look different. We use the word parent or guardian. We pick up the children's conversations in everyday life" (Karlstad). As in the example above, equality is, in this quotation, connected to understanding and visualizing families. In the following example, the 'goal' on sexual orientation is discussed explicitly, but in the 'measure', it is families that are central: Goal: Everyone regardless of sexual orientation should feel respected and treated equally. Measure: We work for all children to know that families can look different ways. We take advantage of and highlight situations in everyday life when we can highlight and name different family constellations. We have purchased and borrowed literature that highlights different families. (Gothenburg 2) In this example, as well as in most of the documents analyzed, the way to address working with sexual orientation in preschool is related to visualizing 'different' families. In the documents, it is not always clarified which families are excluded, thus running the risk of being discriminated against at preschool. Some document brings up homophobia in relation to sexual orientation, which signals an understanding of homosexuals being more targeted for discrimination: "Preschool has a certain responsibility to convey the common values of society to the children. It includes homophobia and the right to equal treatment regardless of sexual orientation" (Kiruna). In documents published by the equality ombudsman (DO), it is stressed that same-sex relationships are more likely to be invisible and that heterosexuality is more likely to be the norm: Another limiting and exclusive norm is heteronormativity. [ . . . ]The norm is crystallized, for example, by the fact that children are expected to have a mother and a father. This norm is very strong and causes homosexual and bisexual relationships to remain invisible. (Equality Ombudsman 2015, p. 13, my translation) In the quotation above, as an example of heteronormativity, children are expected to have a mother and a father. As shown in relation to the plans for equal treatment, it is children's parents and families that are highlighted. In the legal sense, people can claim discrimination against themselves as individuals, but not as a member of a group. In these documents, however, family is constituted as an entity that risks exclusion or discrimination. It is norms regarding families and the parents' sexualities that predominate in this problem representation.

Highlighting Difference
The reoccurring goal in the plans for equal treatment is to visualize different family constellations. For example, these preschools formulated a goal regarding children's knowledge about families: We visualize and affirm different family constellations. (Sandviken) We inform children about different family constellations. (Halmstad) In the example above, family/families are described as different, which stands out as a keyword in the discourse on sexual orientation in preschool represented in the plans for equal treatment. The sexual orientations defined in the Swedish Discrimination Act (SFS 2008:567 4 )-heterosexual, homosexual and bisexual-are never mentioned in the goals or active measures. They seem to have been replaced with the expression 'different sexual orientations' or 'different family constellations'. The description of different families assumes that there is a difference between families that is based on the parents' sexual orientation. The difference between families is thereby constituted by the gender position of the parents, and how they are defined as being "same-sex" or "different-sex" parents/families. The notion that these families are different from each other is related to the discourse in which homosexuality and heterosexuality constitute different categories and subject positions (Foucault 1990).

Highlighting Parents
When 'sexual orientation' is mentioned in the goals or active measures, it is in most cases in relation to parents or legal guardians and in some cases in relation to the pedagogues. The following examples describe how children are at risk of being targeted for indirect discrimination based on their parents' sexual orientation: Pedagogues should be observant that children are not discriminated against based on their parents' sexual orientation. It is also not acceptable for staff to be discriminated against based on their sexual orientation. (Stockholm 2) We should treat all families the same and think about how we use concepts such as family, parents and legal guardian. (Umeå) Sexual orientation is connected to the adults in the preschool. Children are mentioned, as in the citation above, in relation to their parents' sexual orientation. Highlighting the family and primarily the parents constitutes a strong focus on adults in relation to preschool, rather than on the children attending preschool. The effect is that, when the "problem" is situated in relation to children's homes and families-outside preschool, the grounds for discrimination seem less relevant to children's interaction and play when at preschool. In the next section the analyses focus on how children are (not) produced as subjected to this problem's representation.

Invisibility for Children?
The problem is represented as concerning the family, and does not address children in preschool to the same extent. Aspects such as how (hetero)sexuality is negotiated in children's interaction and play are largely left unproblematized. In this section, the reoccurring active measures of the documents are presented in relation to what are proposed as a 'solution to the problem' and how this mirrors the discourses that these documents are produced within (Bacchi and Goodwin 2016). Given that the goal in the problem representation is to visualize different families, these active measures in the documents answer to the question of how preschools should visualize different families.

Reading Books
One reoccurring active measure formulated in the documents-reading books-proposes an action in which pedagogues make texts and pictures (created by someone else) available to children. In relation to highlighting different families, the pedagogues are described performing activities such as borrowing and reading books: Presenting reading books as the main action makes the books seem to be artefacts that have the agency to present knowledge and enable visibility. To carry out this action, writers, publishers and libraries have to provide books with non-heteronormative content. Even if the pedagogues read the books at preschool, some of the responsibly for availability and content is placed outside the preschool setting. In this action, it is not often mentioned what kinds of families they intend to read books about; instead the focus is on broader representation. One exception is a preschool that more explicitly describes how fairytales could be used to introduce same-sex parents-two dads-to children in preschool: We use fairy tales and songs, for example, play with words, change the fairytale about Goldilocks, two "papa bears" (Gothenburg 1) In this example, the pedagogues seem to be more active when they describe playing with words, which indicates that they do not only read but also change and add to the stories. These queerings of fairytales are sometimes deployed in school to introduce new themes using old stories, for example, 'lesbian Cinderella' (Allan et al. 2008).

Spontaneous Conversations
The second recurring action described-the spontaneous conversation-is more reactive; it is based on answering (children's) questions. These active measures are formulated around dealing with situations that occur which have not been planned by the pedagogues: We educators think about how we express ourselves and also talk to the children when opportunities arise spontaneously. (Halmstad) In these types of active measures, the initiative to bring these questions to the table is given to the children: We deal with children's questions when they arise. Here, an interesting paradox regarding children's and adult's responsibility and agency can be found in the goals and active measures presented. On the one hand, children are represented as a less relevant category in relation to 'sexual orientation,' in that they are primarily affected indirectly by their families and their parents' sexual orientation. In that case, children's agency in producing normalization and/or queerness is not taken into to account. On the other hand, children are represented as central actors with a great amount of agency, in that they are expected to initiate conversations with the pedagogues at preschool. As Surtees points out, one potential weakness associated with teachers leaving it up to children to raise questions about sexuality is that "teachers may simply not notice and recognize children's interests" (Surtees 2008, p. 14). Moreover, teachers may not support and encourage children's talk about "inappropriate" subjects.

Discussion
The present article has argued that the goals and active measures for equal treatment regarding the act of discriminating based on 'sexual orientation' focus on adults in the family, rather than on children in the preschool setting. In the Swedish Discrimination Act, the understanding of sexual orientation as categories (homosexual, bisexual and heterosexual) puts certain limitations on addressing these issues from a child perspective. Preschool children do not often position themselves in relation to categories relating to sexual identity. Instead, children are positioned as part of a family, and it is suggested they may be indirectly discriminated against through the association with their parents. Although categories such as homosexual and heterosexual are crucial in the Swedish Discrimination Act on which the plans for equal treatment build, these categories-identities such as lesbian and gay-or explicit talk about what makes the families 'different' does not occur. According to Robinson (2002), talking about families and parents instead of about homosexuals is a strategy employed when sexuality is perceived as a sensitive topic; one part of this is to remove terms such as 'lesbian' and 'gay.' Talking about families is, in this sense, perceived as a neutral and non-provocative way of addressing 'sexual orientation' in preschool: "The legitimation of lesbian and gays within the context of 'the family' can be viewed as part of a normalization process, where sexuality and sexual differences can be eclipsed, albeit temporarily, by more familiar, acceptable and comfortable of family diversity and mothering and fathering" (Robinson 2002, p. 428). The results of the present article show that family is the way to address 'sexual orientation' in the local policy documents of Swedish preschools.
When representing the 'problem' with 'sexual orientation' as invisibility for different families, or the differences between families, it does reproduce 'difference' between sexualities. Brown (2006, p. 50) critics the tolerance discourse of the discrimination perspective: "As it essentializes difference and reifies sexuality, race, and ethnicity at the level of ideas and practices, contemporary tolerance discourse covers over the workings of power and the importance of history in producing the differences called sexuality, race, and ethnicity". If we are to talk about difference, there is a need to separate the difference in hierarchy orders between subject positions such as heterosexual and homosexual, and the difference that reproduces the categorization that precedes the hierarchies.
Based on these results of how sexual orientation is represented in preschools' 'plans for equal treatment' this concluding part turns to the questions in the WPR-analysis that deals with how this representation turned out this way, and how it could be thought otherwise (Bacchi 2012). The notion that sexuality and queerness are sensitive issues in relation to children (Epstein 1997;Prout 2005;Robinson 2013) could provide a possible explanation for why sexuality and norms surrounding sexuality do not emerge in preschool policy and curricula, other than in relation to discrimination. For this reason, the present article suggests that both the discrimination discourse and the innocent child discourse condition the way in which sexuality is represented in Swedish preschool policy. Thinking otherwise regarding the representation of 'sexual orientation' in preschool, there is no need to include children in the categorization of sexualities to make them subjectified by the Swedish Discrimination Act. Rather, a queer theoretical move would amount to questioning the predominance of the juridicification of these questions within preschool and offer new perspectives on children, sexuality and queerness. For example, Stockton (2009) argues that, even if children are seen as aspiring to heterosexuality-not yet straight, the discourse of the child as non-sexual or pre-sexual (see also Foucault 1990) enables childhood to be a potential queer space. Categorization being the condition for hierarchy between sexualities (Foucault 1990) not being categorized is a possibility rather than a shortcoming. Instead of bringing children into the juridificated classification of different sexual orientations, there is a need to bring other perspectives on children into dealing with heteronormativity at preschool. These may not emerge in preschools' plans for equal treatment, as these documents are bound to discrimination as a concept and definitions of the grounds for discrimination. Instead, the issues need to be addressed in other ways, with a focus on children's perspective and the potential queerness existing inside preschool, in play, conversation and other activities during the day.