Environment as mediator – a discourse analysis of policy advice on physical environment in early childhood education

ABSTRACT This article examines how environmental design is set in motion as a technique of government in Swedish policy texts issued to advise those who build and plan preschools. Drawing on Foucauldian research on governmentality and Carol Bacchis ‘What’s the problem Represented to be?’ (WPR) approach, five Swedish government policies on how to build and design preschools are examined from a critical perspective. The WPR analysis helps identify how policies produce problems in certain ways and in this case shows how the preschool environment features in policies in accordance with a certain logic. The study shows that the environment is meant to function as a mediator for disciplinary power, to shape children’s behaviors in desirable ways without coercion. The article also highlights certain silences in the material, the most prominent of these being the lack of discussion about adapting preschool environments to different needs without labeling children as disabled.


Introduction
This article focuses on the idea that the physical environment is a key factor in Swedish preschool education by providing a critical analysis of five government policies on how to build and design Swedish preschools. The Swedish preschool is included in the education system as a separate school form. Although attendance is not mandatory, approximately 85% of Swedish children between ages 1-5 attend preschool (Skolverket 2021). Preschool education is inherently linked to the idea of the Swedish welfare state and preschool is recognized as an important social, cultural, and academic preparation (Skolverket 2018). A dominant line of thought in today's Swedish preschool discourse is that the physical environment is a key factor in children's learning and development. This is a common view among early childhood education (ECE) practitioners (de Laval 2018;SI 2017) and it is suggested in government support materials for preschools that the environment and materials are crucial for how children develop, create meaning, and learn (Skolverket 2012(Skolverket , 2017. Swedish preschool pedagogy is characterized by what is called EDUCARE, a holistic understanding of education, with no clear distinctions between teaching and care (Skolverket 2018;Engdahl 2004;Sheridan and Williams 2018). According to the Swedish preschool curriculum (Skolverket 2018), education occurs as the result of planned content and happens spontaneously since children's learning is a constant and ongoing process.
The idea of the physical environment being integral to education is therefore coherent with traditional Swedish preschool pedagogy. An influence that has accentuated the focus on the physical environment in Swedish preschools over the past 20 years is the concept of the 'environment as the third educator,' which originates from the Reggio Emilia tradition (de Laval 2018;Folkman 2017;Dahlberg and Elfström 2014;Lindgren 2019). Within the Reggio Emilia tradition children and teachers are thought of as partners in learning and the teacher's role is to inspire through the environment and the setup of educational material (Strong-Wilson and Ellis 2007;Gandini 1993). Sjöstrand Öhrfelt and Lindgren (2019) describe this as a posthumanist approach in which the child is positioned in a flat ontology, since social relationships and relationships between people and things are valued equally.
This influence has over several years informed and changed how Swedish preschool environments are designed, from a homelike setting to workshops and open floor planning (Folkman 2017;Dahlberg and Elfström 2014). Research reviews and preschool audits show that the idea of the environment as integral to education can be viewed as mainstream within the Swedish ECE discourse (de Laval 2018;SI 2017).
In an international perspective, global agencies initiatives for learning environments mainly concern aspects such as infrastructure, electricity, sanitation, disability access, equality, and safety (UNESCO 2015;OECD 2018b), or overall educational quality (OECD 2018a), while the goals for learning environments in the Swedish context additionally concern the environment's inherent pedagogic properties.
Regardless of the impact of the idea of the environment as integral to education in the Swedish preschool discourse, according to de Laval (2018), there is no substantial research on the concrete effects of this idea. There is, however, a substantial body of post-structural research on how early childhood spaces are shaped as specific and different places in society (e.g. Bollig and Millei 2018;Gilliam and Gulløv 2017;Jobb 2019;Rasmussen 2004), as well as research on the specific choices of materials and environments, especially regarding various nature-inspired approaches (e.g. Harju et al. 2020;Maynard 2007;Änggård 2010).
The demarcation of education as specific places is not novel in any way. The importance of it creating specific places for children is a foundational idea of ECE. Children are understood to be at risk and the creation of specific childhood spaces can be viewed as a way to protect children from the dangers of society and as a way of imparting the values of society (Olwig and Gulløv 2013;Bollig and Millei 2018;Fenech, Sumsion, and Goodfellow 2008). Nevertheless, even if the creation of childhood spaces is understood as adults wishing to protect children it also involves a governing and shaping of children towards specific ends. Politically speaking the goal is to produce a better adult, to refine the human capital (see Gilliam and Gulløv 2017;Qvortrup 2008).
The contribution of the present study is to explore how the discourses on physical environments and education arrange, assume, and reproduce subjectivities, by examining the ideas about the environment in Swedish government policy and how these may impact the subjectification of children through education. By drawing on Foucault's (2010) later writings on governmentality and the tradition of governmentality studies, this article examines five Swedish policy texts on how to build and organize environments in preschool education. To operationalize the governmentality analysis the article makes use of the 'What's the Problem Represented to be? (WPR)' approach developed by Carol Bacchi (2009). The analysis is structured around five questions drawn from the WPR framework, which will be outlined in some detail below. From a post-structural, critical perspective, this article asks what problematizations are produced in policy on educational environment, and how these problematizations form the prerequisites for preschool children. Put in other words, my main question can be phrased as follows: How do ideas about the environment as a part of education function as a technology of government aimed at shaping the preschool child?

Governmentality
The concept of governmentality can in a wider perspective be understood as the governing of life, i.e. a type of governing that aims for security, welfare, and freedom for the population as a whole.
This kind of governing is not suppressive or authoritarian in character, but, rather, can be viewed as a refining of the human capital into what is considered a successful and desired end for the state and the population in its entirety (Cruikshank 1999;Foucault 1991aFoucault , 2010Foucault , 2014. Educational institutions in particular stand out as a main arena for governing both children and families. The task of education from this perspective is not only to impart knowledge, but also to ensure children's 'normal' development and mold them into desirable citizen subjects (Rose 1999). The end goal is to govern the population towards a desired end and children are in this respect understood as a biopolitical resource (Cliff and Millei 2011). To govern means, in this sense, to act upon the acts of others, which can apply to any 'discourse, program or strategy that attempts to alter or shape the actions of others or oneself' (Cruikshank 1999, 4; see also Foucault 1991aFoucault , 1991b. It is a way to bring individuals' wills and desires in line with notions of the common good (Cruikshank 1999;Foucault 2010Foucault , 2014. Many childhood scholars with a Foucauldian approach, particularly within governmentality studies, have studied childhood education as a site for disciplinary power aiming at normalization (e.g. Cliff and Millei 2011;Cohen 2008;Franck and Nilsen 2015;Mac Naughton 2005;Millei 2005), and how risk discourses have made possible various techniques for governing and regulating children (e.g. Fenech, Sumsion, and Goodfellow 2008;Maynard 2007).
In the context of a governmentality analysis, the present study examines how educational policy 'arranges' the ECE environment and design to achieve certain ends. The purpose is not to condemn educational strategies for shaping children in specific ways, but to highlight the privileging of certain knowledges over others and discuss the possible unexamined effects these have in children's lives in early childhood education (see Mac Naughton 2005). An analysis from this perspective views the organizing of space not as a coincidence or a natural state but, rather, as strategies, ideas, and calculated efforts to achieve certain outcomes (see Foucault 1991a). Thus, this study does not examine the lived life within learning environments but the discourses on learning environments as they appear in policy. Since policies aim at change, they by their very nature contain implicit or explicit problematizations that allow an examination of what assumptions and knowledges underlie the discourses on learning environments (see Bacchi 2009)

Analytical tools
To analyze the policy proposals on the Swedish preschool environment, I have used Carol Bacchi's WPR analysis. The abbreviation 'WPR' stands for 'What's the Problem Represented to be?' and is a Foucault-inspired, post-structural approach for policy analysis. The aim of the WPR analysis is to challenge the idea that policies address problems that already exist. Instead, the analysis considers that policies produce problems in certain ways, and that the way that problems are produced shapes lives and worlds (Bacchi and Goodwin 2016). The term problematizations in this context refer to governmental practices: how issues, things, or persons are defined as problematic in relation to any type of moral, knowledge, or in relation to a specific practice, to render them as governable objects (Rose 1999). The key point in a WPR analysis is to identify proposed solutions in policy to be able to track and examine the implicit problematizations within it and to identify the presuppositions, assumptions and discourses that need to be in place for the produced problematizations to make sense (Bacchi and Goodwin 2016).
The WPR analysis is structured around seven questions. The questions are not a static framework, since they include a measure of overlap and repetition; rather, they constitute a tool that is adaptable to the specific research aim (Bacchi and Goodwin 2016). In the present study, I have primarily used five of the WPR questions (Bacchi and Goodwin 2016, 20), which I will describe shortly below: (1) What's the problem represented to be in a specific policy or policies?
(2) What presuppositions and assumptions underlie this representation of the 'problem'?
(3) How has this representation of the 'problem' come about? (4) What is left unproblematic in this problem representation? Where are the silences? Can the 'problem' be thought about differently? (5) What effects are produced by this representation of the 'problem'?
The first question is intended to identify the implicit problematizations in policy documents. The question can be thought of as a lever used for opening up the policy for analysis. The second question concerns the deep-seated presuppositions or assumptions that underlie the representation of the problem at hand. It seeks to identify the conceptual premises that make the specific problematization make sense. The third question focuses on the discursive practices that give authority to certain knowledges over others. The fourth question examines apparent silences in the material and opens up for thinking about alternative ways to understand the represented problem. The fifth question considers three specific types of effects: discursive effects that constitute the terms of reference that are established by a specific problem representation; subjective effects regarding how a specific kind of subject is implicated in the problem representation; and the 'lived effects,' which term refers to how these discursive and subjective effects form the way people can live their lives (Bacchi and Goodwin 2016, 20-23).

The governing of Swedish preschools
Swedish preschool education is regulated by law (SFS 2010, 800), and by the Swedish preschool curriculum (Skolverket 2018). There are two government agencies assigned to ensure that all Swedish educational institutions fulfill the decisions and demands that are placed upon them by laws and regulations. These are, the Department of Education (Skolverket), a managing authority with responsibility to govern and support the Swedish education system, and the Swedish Schools Inspectorate (SI) (Skolinspektionen), an authority that is responsible for inspection and evaluation of education (SI 2020, https://www.skolverket.se). In addition to these, there are several other state agencies involved in the governance and support of the Swedish education system depending on the specific area or issue.
The agency that is of greatest interest for this study is the Swedish National Board of Housing, Building and Planning (Boverket). Boverket is the state agency in charge of all issues concerning built environments, physical planning, management, and community planning in Sweden. Its assignments include establishing regulations and guidelines and spreading knowledge about its area of expertise. As such, the policies distributed by Boverket are state-issued suggestions, regulations, and advice, and the agency is a governmental authority for planning and building. The guidelines provided by Boverket are based on Swedish legislation, public prescriptions, and national aims, including the Swedish Education Act, the preschool curriculum and other reports and policies issued by the authorities of education. Apart from the materials included in this analysis, Boverket's website also contains legal documents, building regulations, information about municipal rules, inspirational materials, and links to a list of recommended sites and publications (https://www. boverket.se).

Material
The scope of this study is government policies that address the educational aspect of preschool environments. Thus, policies that concern other aspects of education, or, policies that concern general building regulations are not included in the analysis. Four of the five policies selected in this study are guidelines for planning and designing children's physical environments issued by the above-described agency Boverket. The fifth policy document is an assessment report issued by the SI.
The Department of Education (Skolverket), does not currently offer policies on how to build and design preschool environments. It does offer an online class for educators, called 'Preschool's learning and environment' that is said to address how preschool environments can contribute to play and how digitalization and equality can become an integral part of developing environments (https://www.skolverket.se). This material was excluded from the study since it is not a policy text. In the following, the selected policy texts will be described in more detail.
Two of the selected documents are published on Boverket's website as a part of what the agency calls an 'inspirational example collection.' The purpose of the example collection is described as offering inspiration and guidance to those who plan, organize, manage, design, or build preschool and school environments. The site was first published in 2019 on behalf of the government's proposition 'Politik för gestaltad livsmiljö' (Politics for a designed life environment) (Prop. 2017/ 18:110). The two texts are 'Hedlunda preschoolwith the child in the center' (Boverket 2020a) and 'Kometen preschoolwith the forest as a play environment' (Boverket 2020b). These two texts are descriptions of the physical environments at two different preschools and how these have been developed in collaboration with architectural firms.
The other two selected policies published by the above-described agency are addressed to everyone who designs and plans environments for children and youth, and they are not only directed at preschools. One is a guide for planning and designing children's outdoor environments, titled, 'Make room for children and youths! Guidance for planning, designing and management of schools' and preschools' outdoor environment' (Boverket 2015a), the other is an impact assessment: 'Impact assessment of the National Board of Housing, Building and Planning's general advice on free space for play and outdoor activities at leisure centers, preschools, schools, or similar organizations' (Boverket 2015b). The fifth selected material is one out of three audit reports published by the SI based on a 3-year audit of the Swedish preschools conducted by the SI. Of the three, I selected this specific report because it contains the SI's audit results concerning the physical environment (SI 2017).
Moreover, I selected these policies because they are texts that are written with the purpose of being prescriptive, offering opinions and advice on how to organize, design, and arrange the educational environment in Swedish ECE (see Bacchi and Goodwin 2016). The selected materials were triangulated with the Swedish preschool curriculum (Skolverket 2018).
The analysis process I started my analysis by looking at the two texts from Boverket's (2020aBoverket's ( , 2020b inspirational example collection. Based on the first question in the WPR analysis, the analytical focus was first directed at what solutions the preschool environment is thought to provide according to the policies. To understand what the preschool environment is suggested to solve, attention was directed at how it is described 'to be' and what it is described 'to do.' The purpose was to find the problematizations in the policies by way of studying the solutions (see Bacchi and Goodwin 2016). For example, one solution that the preschool environment is described to provide is to inspire children to be physically active. The implied problem dealt with here then, based on what is presented as a solution, is children's activity levels.
The scope of the empirical materials was then strategically widened by working backwards, continuing with material referenced either as inspiration, or as the scientific, or political reason for specific recommendations on the main site, which resulted in the extended scope described in the previous section. The additional three policies were read with focus on the parts of the policies that offered environment advice directed at preschools. Because of the overlapping character of the WPR questions, the questions were used repeatedly and to some extent simultaneously during the analysis process.
The results of the analysis were grouped according to central themes of problematizations, to show how the idea of the environment features in the policies in accordance with a specific logic, giving shape to specific problematizations. Accordingly, in the following, the results of the analysis will be presented by describing how physical environment and materials are mobilized in policy as a strategy and a technique of governing. The analysis will show how the technology of governing through the physical environment is shaped in policy directed at ECE.

The shaping of the independent child
The issue that I want to address in this first section of the analysis is what the represented problem is in these policies on a more overarching scale. Based on the first question of the WPR analysis, focus was put on how the environment is described 'to be' and what the environment is described 'to do', in order to understand what problematizations the policies are aiming to solve (see Bacchi and Goodwin 2016). In one of the two inspirational texts, the Reggio Emilia approach and the concept of 'the environment as the third educator' is specifically used as an explanation as to why the environment is important, indicating that education is incorporated in the environment and to some extent carried out by the environment.
The physical environment is called 'the third educator' and we care about adapting architecture and interior design to the size of the children to promote their independence and self-perception. The view of the Reggio Emilia pedagogy is clearly reflected in how Hedlunda preschool is designed, above all through the heart of the buildingthe creative studio. (Boverket 2020a) In the above quote, the environment is called 'the third educator' and the Reggio Emilia pedagogy is said to be reflected in the preschool's interior design. The specific pedagogic approach is therefore both used as a reason for the environment design and viewed as inscribed in the environment; consequently, one thing that the environment is described to do is to educate. The concept of 'the environment as the third educator' is not explicitly defined in the policy, but what is emphasized in relation to the concept is that the architecture and furnishings are designed to promote children's independence and self-perception, indicating that one problematization in the policy is children's lack of these desired qualities. Another emphasized part of the pedagogic design is the aspect of creativity. The same document includes the statement that: 'The Reggio Emiliapedagogy highly values children's creations and creativity.' (Boverket 2020a). In the following passage, it is further described how the educational approach operates through architecture and the building itself, as the preschool is said to have-… distinctively translated Reggio Emilia pedagogy into architecture. Space for learning, play, and creation is satisfied through studios, exhibition spaces, playful scale, and spaciousness. (Boverket 2020a) It is noticeable in both texts published as a part of Boverket's inspirational collection that there are specifically designed places where certain activities are meant to be stimulated. In the above example, it is 'studios' and 'exhibition spaces,' but there are several other examples like these throughout the material.
To further examine how, and what, the environment is proposed to educate, I focused on the words used to describe the educational environment. Words specifically connected to creative activities are extensively used and it is described as important that the environment should offer space for creativity and that it should feel inspiring: The main theme in rooms and the design of the learning environment, with place and space for creativity, discovery, and play, makes the environment feel inspiring and lively. (Boverket 2020a) The outdoor design, which is the focus of the second text published in the inspirational collection, is also described as a trigger and inspiration for specific activities, but these are different from indoor activities: In nature, play and imagination get started in a totally different way. The children are keeping themselves busy all the time and we, the personnel, can be there as support (Boverket 2020b) The above quote shows that the outside environment is meant to function as a trigger for specific activities but also to enable children to keep themselves busy, which corresponds with the previous emphasis on independence. Nature is described as having a different effect, which will be discussed in a later section.
Since the environment is described as so important for how children act in relation to specific spaces, the function of the physical environment emerges as a way to divide and arrange the children within the preschool spaces. This can be analyzed in relation to how Foucault (2017) described the disciplinary function of the traditional classroom, where the teacher's division and placing of students in the classroom functions as a type of disciplinary power and a way of gaining control over bodies. The difference here is that the guidance material suggests that it is the environment itself that organizes the children and inspires activity in the educational space. Thus, this can be understood as a way to assign disciplinary power to the environment, rather than assigning it to the teacher. The environment then functions as a mediator of disciplinary power.
These findings corresponded with the SI's audit report. The report discusses the inspirational capacity of preschool environments and provides a more in-depth example of how the material should function according to the authority, by including examples of what is deemed as bad environmental designs: There are examples from the audited preschools where materials are in boxes, which makes the content nonvisible, or where material is placed on high shelves and therefore not reachable for children. If the material is not visible and available the children are not inspired and their possibility for variation is limited. (SI 2017, 21) Availability and exposure are not granted, which in turn requires personnel ready to help children to take out material and provide guidance on how to use it (SI 2017, 21) Two problematizations can be discerned in the above quotes: one is that a 'non-visible' environment will not inspire children and give them the possibility of variation. The other problematization is that, by not exposing materials, personnel is required to be ready to help and guide children. The logic seems to be that if environments are not exposed and available, children will not be inspired to do different types of activities and they will also be unable to act on their own, i.e. independently.
From a governmentality perspective, the aim to enable children to become independent can be understood as a form of governing that operates within a discourse of empowerment (Cruikshank 1999). The logic of governing through the environment can consequently be understood as a technique for shaping children towards independence in two ways: firstly, by inspiring them to act in a variety of ways, and secondly, by arranging the environments in such a way that the educators do not need to assist the children. The shape of the problematization in the policies can, then, be viewed as a problematization of children's dependency on adults working in preschool and the solution to this is to arrange the environment in a way that will render children, and education, less dependent on adult interaction.
As is obvious from the quotes above, there is a strong recommendation that environments ought to be stimulating, visible, and rich in content, and the SI report criticizes preschools that keep material in boxes or on high shelves. By comparing these texts to Boverket's guide for planning outdoor environments, a gap in the material is exposed. The guide for planning outdoor environments offer several examples of how to adapt environments for children with disability, for example by limiting sensory impressions.
Many children with disabilities have difficulties sorting sensory impressions which leads to fatigue and concentration difficulties. The outdoor environment therefore needs to include retreat sites with restorative qualities offering break and recovery of concentration. (Boverket 2015a, 64) Children with perception and concentration difficulties need a clear division of space for different activities in the outdoor environment. Distinct rooms, with "walls" of e.g. shrubs and fences delimit sensory impressions and enable quiet moments in play. (Boverket 2015a, 68) To further explore this gap, I wish to use another quote from the SI report. The report states that educators at some of the audited preschools had made an active choice to keep a stripped-down environment so as not to overload the children with impressions. This stance is implicitly criticized in the SI report: In some preschools the personnel have made an active choice to keep a stripped-down environment with the purpose of not overloading the children with impressions that can give them difficulties to concentrate. It does not appear that the preschools in question have evaluated such orientation. (SI 2017, 21) The gap between these examples indicates that adapting environments to children's different needs is recommended if the children in question are considered disabled, but it is questioned by the SI as a general educational approach. The SI's criticism can be viewed as power exerted by experts to silence the situated and specific knowledge of the educators, making the teachers both the object and the subject of expert knowledge (see Cohen 2008;Fenech, Sumsion, and Goodfellow 2008;Millei 2005).
This gap in the policy texts can be viewed as an example of how language is used as a dividing practice (see Foucault 1982). The effect of such language is that the only child subjects made possible here are either the 'normal' child who responds to the chosen environmental design in a desired way, or, the disabled child, positioned outside the norm, who needs an adaption of the environment. An alternative way to understand and address diversity is to acknowledge that there are multiple possibilities to respond to children's diverse potential, including both competencies and needs (see Cohen 2008).

Problematizing children's future health and wellbeing
In this section, I explore the ways in which a discourse of health and wellbeing functions both as an aim for governing and as a central knowledge for shaping the governmental logic regarding the educational environment (see Bacchi and Goodwin 2016).
The inspirational material contains several examples of how the preschools have been designed and built to inspire children to become more physically active. For example, the texts contain headlines such as 'Environments rich in movement' (Boverket 2020a(Boverket , 2020b. These headings indicate that the buildings are designed with the purpose of creating activity and movement. The following quotes are examples of how specific places, both indoors and outdoors, are constructed to promote physical activity: The large play area incites running, play, discovery and meeting with the wildlife that comes for visits from the nearby forest. (Boverket 2020b) There are several built-in activity elements in the form of small steps, stairs, a slide, a place for dancing, a climbing frame, and huts, which gives movement and physical activity indoors. (Boverket 2020a) The above-described strategies to create environments to stimulate physical activity can be understood as examples of solutions to a produced problematization of children's activity levels. The problematization leading to these solutions is more explicitly stated in Boverket's advisory material, as seen in the following quote: There are many indicators that suggest that children's physical activity has declined, and that sedentary activities have increased. This could lead to an increase in chronical illnesses in the future. (Boverket 2015a, 14) One problem the policies are meant to counter is consequently the risk of ill health in children when they are adults, and one solution is to make the children more physically active during preschool. The underlying assumption seems to be that the reason for the decreasing health of the Swedish population is that children's physical activity has declined. As shown in previous studies (Åkerblom and Fejes 2017;Åkerblom 2020), the Swedish 'health problem' refers to the risks of, so-called, 'welfare diseases' and 'lifestyle-related health issues.' In particular obesity and mental health issues have increasingly become viewed as a problem for the future nation, and for this reason, it has become an educational aim to educate children about healthy lifestyles by primarily endorsing physical activity.
Further exploring the health discourse in the policies in line with the third question in the WPR analysis shows that the aim for increasing health appears to be related to a wider spectrum of problematizations in addition to the risk of increasing chronic illness. The health discourse is also strongly connected to academic achievement and social adjustment, as is shown in the following quote: The physical environment is of great importance for children's and youth's health and learning. Trends of increasing sedentary lifestyles, overweight, and growing numbers of illness are concerning; simultaneously, goal achievement in school work is declining, according to international comparisons. (Boverket 2015b, 5) The quote illustrates how overweight specifically is viewed as a health risk, but also, how assumptions about academic achievement and future health are intertwined. According to Åkerblom and Fejes (2017), there is an interconnecting logic between the discourse on health and that on education because education is viewed as a means to achieve a healthier life, while illness is considered to be a cause for declining goal achievement. What is at stake in the problematization of health is consequently not only children's health, but also children's academic future. The underlying assumption regarding what health consequently involves much more than simply weighing the risks of obesity; good health is considered to be connected to academic achievement and the overall idea of a good life. The links that are made between obesity, sedentary lifestyles, illness, and declining goal achievement are also indicative of an underlying stigma surrounding overweight and ill health, a stigma that may divert attention from other factors that affect academic achievement (see Bacchi and Goodwin 2016).
What is distinctive about the policies is that health as a technology of governing entails several diverse underlying assumptions about what healthy behaviors are and makes recommendations on how to achieve these through different uses of the physical environment. The aim to increase the population's health and wellbeing can be viewed as a logic that makes governing of a variety of behaviors possiblebehaviors that are understood as healthy in a wider context: physiologically, psychologically, and sociallyto ensure that the children as future adults will fit into society (see Åkerblom 2020).
I wish now to turn the discussion to how health and wellbeing appear to be connected to specific assumptions about nature based on the different descriptions of what nature and natural materials are thought to 'do' according to the policy texts. The notion of what is understood as healthy behaviours appears strongly connected to time and place, making health and wellbeing dependent on when and where. The following quote shows how the desired level of activity differs depending on place: The children stay outdoors for half of the day and the personnel states that they are calm and balanced indoors because of this. (Boverket 2020b) In the above quote, staying outdoors is not described as desirable per se; rather, staying outdoors is seen as desirable because it makes children 'calm and balanced' when they return indoors. The wording used in the quote, 'calm and balanced,' suggests mental wellbeing, which is connected to the health discourse. Hence, the promotion of health does not only concern activation of children, but it may also concern decreasing children's activity levels in a certain space and at a certain time. When analyzed as a technique of governing, the way the outdoors is designed, as described in the quote, can be understood to function as a strategy for making the children calmer indoors. These contrasted spaces suggest that different types of desirable acts have clear spatial boundaries.
What is distinctive about the policies is that the outdoors, and nature, are specifically described as having elevated significance for wellbeing, almost as though there is magic in the effect they have: The health-improving effects of outdoor activities are a consequence of how they combine playful movement and contact with nature. (Boverket 2015a, 15) There is an underlying assumption that is consistent throughout the policies, namely, that nature and natural elements have a specific and desirable effect on children's actions. The following quote highlights how natural play materials, such as sand and water, are chosen for their supposed effect on children, inducing them to be active in a desirable way. On the other hand, human-made play materials, such as bicycles and swings, are described to lead to negative behaviors and therefore need to be taken away: … the personnel then choose to remove the swings and replace them with sand play and water play, as a complement to the natural land. [A quote from a member of the teaching staff:] 'It is so nice that there aren't any swings or bicycles! They only lead to competition and waiting. Those kinds of play equipment create passivity and conflicts.' (Boverket 2020b) As suggested in the two above quotes, the use of natural materials is described in the policies to promote and initiate certain 'good' activities, but also as preventing specific undesired activities such as competition, conflicts, and passivity. These undesired behaviors are described as created by certain play material rather than as children's own expressions of real emotions or desires.
One conclusion that can be drawn by looking at how the policies recommend the use of natural materials is that nature and natural materials are to be strategically used to shape desirable behaviors. This can be put into context with social emotional learning programs commonly used in preschools to organize their value-based education to teach children to become socially competent (see Bartholdsson, Gustafsson-Lundberg, and Hultin 2014). It is common that children's expressions of certain emotions, in particular anger, are viewed as something that needs to be controlled and disciplined. The difference compared to a social emotional learning program is that the policies promote the use of nature as a technique to shape the correct behavior. Hence, nature and natural materials are presented in the policies as a way to prevent future pathologies such as violence and promote socially desirable traits (e.g. Rose 1999).
This valuation of nature and natural materials can be derived from two different ideas about nature and children. One is the idea of nature as something innately good and pure, which is related to a romanticized and idealized idea about nature and about the child as nature (Harju et al. 2020). The other is the logic of the binary nature and culture divide, where nature represents a neutral place where children can live out their true selves, while culture is by contrast understood as artificial and unnatural. In the educational discourse, this logic connects to the centuries-old idea that children need to be protected from culture (e.g. Harju et al. 2020;Änggård 2010).
There are several possible alternative understandings of the 'problem' of human-made play materials. In the previous quote, waiting is interpreted as 'passivity,' a negative behavior caused by the play equipment. An alternative understanding could be that the children choose to wait for equipment that they enjoy. Furthermore, 'being able to wait', could be understood as an important skill for children to acquire, as opposed to as an undesired or unhealthy behavior.

Shaping the self-reflective citizen
The policy texts describe the aims for the preschool design as solutions to the overarching produced problematizations, such as the aims to improve the population's future health and the shaping of the independent child, described above. However, there are also examples of more explicit problematizations that appear as solvable through the environment. One such problematization, presented in the policy documents, which appears to be strongly connected to the physical environment is that of stereotypical gender norms. In the inspirational examples and also in the report from the SI, gender understanding and, to some extent, gender expression is shaped as a problem that needs to be addressed (SI 2017;Boverket 2020a). One of the inspirational texts describes 'gender thinking' as a guide for the design: The preschool is a good example of a learning environment where sustainability and gender thinking have guided the design (Boverket 2020a). The quote is an example of advice that suggests that gender thinking could and should be incorporated into the physical environment. The unarticulated problem seems to be the risk of a 'wrong' kind of gender thinking. The concept of 'gender thinking' that is used in the text is not formally defined but, rather, used as an everyday expression when referring to adopting a gender perspective in workplaces, schools, and such. It is also left unexplained in the policy and implicitly this lack of explanation could signal that the reader should already understand what the 'correct' gender thinking is. Another assumption that is unexplained in the inspirational example text is how gender is to be a part of the design. A more explicit articulation of how materials and spaces can induce a specific 'gender learning' can be found in the reports of the SI, which suggest that in the absence of rich, exposed, available, and variable materials, … boys' and girls' possibilities to be challenged in existing gender roles and mindsets about what girls and boys can do and how they are allowed to act, are affected. (SI 2017, 21) The logic here seems to be that if children do not experience exposed, available, and variable materials in rich supply they will not challenge the existing gender roles. Furthermore, the report states that the environment should be designed to counteract stereotypical gender patterns and that preschools should avoid design choices that are gender-coded (SI 2017). These findings correspond with the preschool curriculum that states that: 'The preschool environment should inspire and challenge children to widen their abilities and interests without being limited by gender stereotypes.' (Skolverket 2018, 8).
There are several underlying assumptions in the material about what gender is and what gender learning might be. The two most prominent assumptions are, firstly, that children's understanding of gender is connected to visual and material cues, and, secondly, that a child's interaction with the right materials changes the child's 'gender thinking' in a desirable way. The type of governing that is put into play in the policies is therefore an activation of children's self-governing. According to Qvarsebo (2019), the kind of governing at play within the framework of what is called 'norm-critical pedagogics' does not present itself as moral practices; rather, it presents itself as techniques that operate through the self-reflection of the subjects, aiming to make them aware of the social norms embedded in everyday practices, and thereby making more desirable ways of thinking and acting possible. The way that the physical environment is described to function in relation to gender learning in the analyzed policy texts can be understood in accordance with a similar logic as that of norm-critical pedagogy. Incorporating a specific 'gender thinking' into the design is thought to encourage children to engage in self-reflection and, as a result, to think about and 'do' gender in a more desirable way (see Qvarsebo 2019).
In the Nordic countries, the childhood institutions are strongly connected to the welfare state, and are thought to civilize children into society's norms, not through discipline but, rather, through a soft standardization of what is perceived as acceptable behavior and the boundaries for such (Gilliam and Gulløv 2014). From this perspective, if we look at the function of environment design as a promoter of certain ways of thinking, it can be understood as a technique for governing from a distance. And governing, as a normative shaping of the thoughts of the future citizens without having to resort to what appears as disciplinary action, is turned into self-governing (see, Cruikshank 1999;Rose 1999).

Discussion of the environment as mediator and technology of government
By analyzing policy materials on preschool environment, this article has illustrated a number of problematizations represented in policy texts, problematizations that are thought to be solved through the environment. The preschool environment is presented in the policies in accordance with a certain logic, a logic where the environments and materials function as a technique to shape children's acts in desirable ways. The proposed function of the environment can be understood to foster children's independence and to shape specific behaviors in specific spaces. Based on the analysis, I argue that the technology of power put to play in the policies is the use of the environment as a mediator for disciplinary power. This notion constitutes the child as an active constructor of knowledge and mobilizes mechanisms of power that utilize children's self-regulation (see, e.g. Cruikshank 1999; see also Cliff and Millei 2011).
One effect of repurposing the environment design to shape children into independent learners is that the role of the teacher becomes repositioned. I draw this conclusion on two grounds: firstly, because the physical environments are not described as places where teachers educate but rather as environments that stimulate education on their own; and secondly, because the implications are that environments that require adult guidance are considered weak environments. According to Jobb (2019), an educational discourse where children are positioned as active and independent learners, risks creating a binary view of education where child-centered education is put in opposition to teacher-led education, leading to the view of teacher-led pedagogics as authoritarian whereas child-oriented pedagogics are viewed as more responsive to children's interests. However, this binary construction does not provide an answer to how power reinscribes itself in these shifting educational approaches.
In the light of my analysis, the technique of governing through the environment can be understood as a way for the state to distance itself, and the teacher, from the governing practice. Governing is then conducted not by coercion, but through the enabling of free choice (e.g. Cruikshank 1999;Fejes 2010;Rose 1999). The environment functions as a mediator to enable a distance between government and the governed subject. The child is regulated through the physical environment, a governing practice that is dependent on children's self-regulation, enabling a soft governing towards desired behaviors within preschool spaces (see Cliff and Millei 2011). This shift in forms of governing entails not only a shift in the power relation, but also a shift in responsibility. This shift is closely related to the development of education politics, where the concept of learning has been disconnected from the concept of education and teaching, a mechanism that renders the learner herself responsible (e.g. Simons and Masschelein 2008).
Another possible effect of governing education through the environment is that other factors affecting education can be overlooked as the environment tend to take over also social and relational aspects of education, such as, for example, children's understandings of gender, how to manage conflicts, or, academic attainment.
One of the most crucial gaps in the material concerns whether, and how, the various environments can be adapted to children's different needs. One reason for this gap may be that it is the result of several contradicting but interwoven discourses on children. One of these is the influential discourse of the competent child that is very pronounced in the Swedish, and Nordic, educational contexts (Sjöstrand Öhrfelt and Lindgren 2015;Månsson 2007;Plum 2014). Also, discourses on ECE as a medium for early intervention, which are strongly tied to the authority of developmental psychology in education (see Cohen 2008;Franck and Nilsen 2015;Mac Naughton 2005). The discourse on the competent child regards children as competent per se. Competence is then understood as compliance to the early childhood institutions' rules and norms, which for instance entails children being independent and self-regulating, but the discourse also constructs a norm of what is not a competent child (see Franck and Nilsen 2015;Plum 2014). Since the discourse emphasizes children as competent, this may also serve to disqualify talk about children's needs and make children's need of support invisible (Månsson 2007). However, in educational policies the importance of discovering and addressing special needs is stressed, which entails a need for identifying and measuring children against a standardized norm (Cohen 2008;Franck and Nilsen 2015;Mac Naughton 2005;Millei 2005). These contradicting discourses particularly affect children who are not formally diagnosed, yet not attaining the norm of competent children (Franck and Nilsen 2015).
The logic in the policy texts seems to be that the 'normal' child subject is constructed as a child that operates independently and appropriately within specifically designed educational environments; a logic that also has the effect of creating boundaries for the 'undesired' child subject, that is, any child who does not respond in a desirable way to the environmental and material cues. The effect of such dividing practice risks positioning children that do not do well in a specifically designed environment as 'deficient'.

Disclosure statement
No potential conflict of interest was reported by the author(s).