Learning and agency for sustainability transformations: building on Bandura’s theory of human agency

Abstract How to secure a sustainable future is currently one of the key challenges across society. Bandura’s theory of human agency provides important insights into this question by recognising the role of individual, proxy and collective agency in initiating action and change. In this article, we conceptualise sustainability agency and discuss psychological features and competencies that are relevant for supporting agency in sustainability education. Agency may manifest in different forms and contexts: countless actions in individual lives, communities, and local, national and global scenes can contribute to sustainability. This range of opportunities to be engaged in sustainability transformations can be better acknowledged in sustainability education in formal, non-formal and informal learning settings by broadening and clarifying the targets and indicators of successful sustainability education. Consequently, sustainability education and related participation in advancing sustainability can become more inclusive and meaningful for actors with diverse backgrounds.


Introduction
In sustainability education tradition, there are several pedagogical approaches that emphasise action and engaging in real-life sustainability issues. Examples of these include action competence (Jensen and Schnack 1997;Breiting and Mogensen 1999;Mogensen and Schnack 2010) and place-based learning (Villanen 2014). While environmental and sustainability education initiatives have reported positive outcomes in terms of increased environmental awareness and attitude changes, there is limited empirical research on how sustainability education can support learners' capacity to take action for a more sustainable future (see systematic reviews from Ardoin et al. 2018;O'Flaherty and Liddy 2018). Conceptual clarification of targeted action is needed to guide empirical research on action-oriented sustainability education and consequently, further develop existing pedagogical models. We approach this demand by elaborating on the concept of agency, building on Bandura's (2006) theory of human agency.
There is no univocal understanding of the definition, contents and means of sustainability education (Hofman 2015;Mogensen and Schnack 2010). However, it is usually agreed on that while increasing learners' knowledge is necessary, sustainability education should also engage learners' values, emotions and behaviours (Hofman 2015;Mogensen and Schnack 2010;Ojala 2013;UNESCO, 2017). In the UN Sustainable Development Goal (SDG) on Quality Education, one of the targets is to 'by 2030, ensure all learners acquire knowledge and skills needed to promote [emphasis added] sustainable development' (UN, 2015a, n.p.). This is in line with the approach of this paper as it implies action rather than passive adjustment to a more sustainable world.
While it is not feasible to provide a universal, all-encompassing definition for sustainability education due to the contextual and complex nature of it, we have created the following definition for the purposes of this paper. Sustainability education refers to education that encourages action towards sustainability transformations, i.e. the societal changes that aim to secure the well-being of humans and the other natural world as well as just and peaceful societies within the planetary boundaries.
We approach agency primarily from the social cognitive tradition, understanding it as a property that can, as a result of an interplay of various actors, environments and contexts, manifest as different types of actions (Bandura 2006;. By acknowledging the diverse ways of enacting agency as well as the psychological features and competencies that underlie it, we want to 1) help educators to further develop pedagogical approaches that involve engaging in real-life sustainability issues as part of learning processes, and 2) guide researchers to target and structure empirical analysis on action-oriented sustainability education.
In section 2, we discuss Bandura's (2006; theory of human agency and situate it into the context of sustainability issues. We elaborate on the different modes of agency and their relevance to sustainability transformations in section 2.1. We then discuss psychological features that are linked to agency in section 2.2., which underlies our conceptualisation of sustainability agency in section 2.3. In section 3, we return to sustainability education and contemplate how agency-centred sustainability education could be implemented in formal, non-formal and informal contexts.

Connecting sustainability challenges and agency
There are several problems that endanger sustainable living on our planet. Climate change, biodiversity loss and mass extinction of species, environmental pollution, poverty, and hunger are well-known examples of sustainability issues with alarming and acute consequences. They call for rapid action yet are very difficult to solve (Waddock 2013). Peters and Wals (2013) consider the following aspects to be key characteristics that are pertinent in all sustainability issues: 'indeterminacy (the impossibility of knowing in advance what the best course of action is); value-ladenness (the crucial position of values in affecting behaviours, lifestyles, and systems); controversy (the lack of and impossibility of full agreement or consensus among all stakeholders); uncertainty (not being able to predict the exact or even near-exact impact of a chosen strategy or action); and complexity (the messy interactions between a whole range of variables operating at different intertwined scales)' (p. 88).
Agency refers to the capacity to initiate actions and deliberately influence the course of events (Bandura 2006). It is discussed in various research fields, e.g. in social sciences, psychology, anthropology, and gender studies (Eteläpelto et al. 2013). The roots of an agentic perspective to learning are in the age of Enlightenment (e.g. Biesta and Tedder 2007), and cultivating agency should be at the core of all education, as Bai (2020) notes.
According to Bandura (2006), 'to be an agent is to influence intentionally one's functioning and life circumstances' (p. 164). Bandura (2006) regards agency to have four core properties: intentionality, forethought, self-reactiveness, and self-reflectiveness. In other words, humans have the capacity to plan, execute and regulate their actions as well as to reflect on the reasons behind and the outcomes of them, adjusting behaviours accordingly (ibid.). Thus, agency is built on social-cognitive capacities and preparedness of individuals. However, in addition to individual agency, Bandura (2006) identifies two other modes of agency, proxy and collective, which recognise social and social-cultural contexts as a platform for agency. These will be discussed in detail in the next section.
When dealing with sustainability issues, the core properties of agency -intentionality, forethought, self-reactiveness, and self-reflectiveness (Bandura 2006) -are critical. Being able to influence one's own actions and to collaboratively work on a common goal, capacity to reflect on causes and consequences of various events, and readiness to create new courses of action are all necessary for taking up more sustainable lifestyles and furthering sustainability transformations at different scales. Bandura (2006) states that in the rapidly changing world, human agency is in a significant role to 'shape a better and sustainable future' (p. 177). Next, we elaborate on how the different modes of agency -individual, proxy, and collective agencyare all relevant to facing sustainability issues.

Individual, proxy and collective agency
Individuals' choices and actions can have an impact on sustainability issues. However, recognising the importance of acting with others or through others is crucial for larger-scale sustainability transformations. Bandura's (2006) three modes of agency are a useful approach for categorising various types of actions in sustainability education practice and research.
Individual agency is enacted when a person deliberatively does something that leads to intended consequences (Bandura 2006). For example, one might decide to reduce their consumption, switch off electronic devices when not used, start using public transportation instead of their own car, shift to a plant-based diet, or increase their own understanding of sustainability issues by taking up an online course. However, few actions fall under the influence of individuals only. Returning to the examples mentioned above, reducing one's own consumption and switching off electronic devices are rather straightforward behaviours, which nevertheless may require individual agency until they become routines. However, using public transportation is not possible if it is not offered in a city, a plant-based diet is difficult to follow if vegetarian options are rare in stores and canteens or if they are not part of the food culture, and it is more difficult to educate oneself about sustainability issues if valid information is not easily accessible. Some solutions to these challenges can be created by an individual, particularly if they are in a position of power. Nevertheless, often it is more effective or only possible to act through or with others.
Proxy agency means acting through others who are better equipped or in a better position to achieve the aspired outcomes (Bandura 2006, p. 165). For example, one can contact the city council and campaign for public transportation, striving for the desired outcome through local politicians. Collective agency, on the other hand, means acting together with others (ibid.). For example, members of an NGO can educate teachers and students about human rights, management of a company can invest in innovations that help in climate change mitigation, members of a family can jointly decide to start eating more plant-based food, and a union of countries may introduce regulations and strategies to prevent environmental pollution.
As sustainability issues are large and interconnected problems, many of them concerning the whole globe (Blanc 2015), they cannot be solved without collective agency. On a smaller scale, too, collective actions and agency are needed to create more sustainable practices and cultures (Clark 2016). Therefore, further attention should be paid to the interlinkages between individual and interpersonal scales of agency. This would support approaching sustainability from a more systemic perspective, as has been called for, among others, by the Intergovernmental Panel on Climate Change (IPCC, 2021), the Intergovernmental Panel on Biodiversity and Ecosystem Services (IPBES, 2019). Also in Agenda 2030 (UN, 2015a), it is emphasised that achieving the set goals is the shared responsibility of 'all countries and all stakeholders' (UN, 2015a, p. 1).
While we argue that individual, proxy and collective agency are all needed to advance sustainability, it is important to note that not all human behaviours and choices in everyday decision-making are deliberate. Instead, they often occur unconsciously as habits and routines (Klöckner 2013) or based on cognitive shortcuts (heuristics) that facilitate processing the information load people are surrounded with (Morewedge and Kahneman 2010). Contextual factors and norms can sometimes guide people into making sustainable choices without intentional thought processes, which is taken advantage of in behaviour change initiatives that alter the choices that people have (e.g. Arbuthnott 2009;Lehner, Mont, and Heiskanen 2016). Moreover, motivations for executing behaviours that can be regarded as sustainable vary. For example, cycling or reducing consumption may not be taken up with the intention of advancing sustainability (Gifford and Nilsson 2014;Klöckner 2013). Thus, exerting the core properties of agency is not necessary in all daily sustainable behaviours (Klöckner 2013).
However, before sustainability becomes a 'default option' in all the functions of societies, living in a sustainable manner cannot rely only on unconscious behaviours. Also, agency is not needed only to influence one's own daily behaviours. Realities are continuously re-constructed in interaction of agency, environment and current structures (Bandura 2006, p. 167). Importantly, to ensure that social, political, economic and technological systems support sustainable living, active change agents are needed to intentionally promote sustainable alternatives in different roles in local and national contexts, organisations, workplaces and communities, as well as in the global community.

Psychological features and competencies linked to sustainability agency
There are several psychological features which are linked to the emergence and continuity of agency. First, enacting agency is intentional (Bandura 2006), which implies that the actor wants to perform it. Therefore, motivations that drive sustainability actions are needed. Furthermore, believing that one can bring about desired outcomes, which is essential for agency, is inherently linked to efficacy perceptions (see Bandura 2006, p. 170). Being motivated and finding actions meaningful connect to values and beliefs about what is true, good, desirable and possible (see Bandura 2001, p. 2). Furthermore, taking any actions demands competencies to do so. Next, we briefly discuss these partly overlapping features -motivations, efficacy perceptions, values, beliefs, and competencies -and their connections to agency.
First, the concept of motivation is elaborated on, which means the underlying reasons for any actions. Motivations can be intrinsic or extrinsic (Ryan and Deci 2000), which refers to the source of motivation. When motivation is intrinsic, an action is done for 'its inherent satisfaction' but in the case of an extrinsic motivation, the driver is 'some separable consequence' (Ryan and Deci 2000, p. 56). Motivations can be also classified as autonomous or controlled, which refers to the degree of volition versus coercion as motivational factors (Vansteenkiste, Lens, and Deci 2006).
According to self-determination theory, there are three basic psychological needs, which, when being met, bring about intrinsic and autonomous motivation (Ryan, Bradshaw, and Deci 2019). These are the needs for autonomy (the sense of acting and making decisions volitionally), competence (mastering skills that are needed to follow these aspirations), and relatedness (feeling connected to others and a sense of belonging to a group) (ibid.). Self-determination theory is applied in several studies that aim to understand commitment to sustainable behaviours (e.g. De Groot and Steg 2010; Lavergne and Pelletier 2016). Empirical studies suggest that intrinsic motivations are strong among those who actively contribute to sustainability (e.g. Koistinen et al. 2020). Therefore, we argue that the core components of self-determination theory are of practical value when developing sustainability education that supports learners' agency.
Second, efficacy perceptions, which are beliefs about whether actions can lead to aspired outcomes or not, are discussed (Bandura 2006;Bandura 1994). Efficacy perceptions operate on two scales: personal (self-efficacy) and collective (collective efficacy) (Bandura 1994). High efficacy beliefs increase motivations to take up courses of action and persevere when facing difficulties (Bandura 2006;Bandura 1994). This perseverance is crucial for remaining agentic regardless of the indeterminacy, uncertainty and complexity of sustainability issues (Peters and Wals 2013).
In two meta-analyses of psycho-social determinants of pro-environmental behaviour, a significant correlation was found between efficacy perceptions and pro-environmental behaviour (Bamberg and Möser 2007;Hines, Hungerford, and Tomera 1987). A sense of collective efficacy is also conducive to acting in a sustainable manner (Chen 2015;Jugert et al. 2016). Without collective efficacy and agency, moving towards a more sustainable future may seem utopian. Psychological defence mechanisms, such as denial or apathy, can be activated as a response to disturbing issues or problems that seem too difficult to solve, and this can decrease active problem-solving (Coelho et al. 2017;Kollmuss and Agyeman 2002). To avoid this, working together with others and creating and maintaining hope and positive visions for the future are essential (Chawla and Cushing 2007;Hicks 2014, p. 109;Ojala 2017). The importance of collective efficacy also links with the relatedness aspect of self-determination theory and the need to feel unity with others (Ryan, Bradshaw, and Deci 2019). Therefore, creating experiences that allow learners to gain experiences of self-and particularly collective efficacy is relevant for supporting their emerging agency.
Third, in addition to beliefs about self-and collective efficacy, other types of beliefs also influence sustainability agency. According to Heimlich, Mony, and Yocco (2013), 'all people hold beliefs about the environment and specific environmental issues, their relationship to the issue and the environment, and toward the efficacy of their own actions on the issue' (p. 263). For example, a belief in climate change is one of the key factors influencing whether one takes climate actions or not (e.g. Vainio and Paloniemi 2013). Values are also central in explaining sustainable actions (Schwartz 1992;Swami et al. 2010;De Groot and Steg 2007). For example, post-material values, such as self-expression and quality of life (e.g. Paloniemi and Vainio 2011), and nature connectedness (Dale et al. 2020;Martin and Czellar 2017) seem to encourage sustainable behaviours. However, sustainability issues are multifaceted, thus values that encourage agency may vary between people and depend on what is relevant in the specific context. For example, in a study of 85 climate educators, social justice issues were more often mentioned as a source of motivation for climate action than biospheric themes (Howell and Allen 2019).
Fourth, practising agency is not possible without having the necessary skills and competencies. Also, as mentioned earlier, feeling competent is one of the main drivers of motivation (Ryan, Bradshaw, and Deci 2019). Several academics have sought to shed light on competencies that are relevant in the context of advancing sustainability (e.g. De Haan 2010; Rieckmann 2012; Wiek, Withycombe, and Redman 2011). According to a synthesis conducted by UNESCO (2017), these key sustainability competencies include behavioural and strategic competencies that facilitate taking action and finding viable solutions to sustainability problems in collaboration with diverse people; cognitive competencies with a particular focus on critical, systems and futures thinking; ethical and reflective competencies, which include capacities for critical contemplation of values, norms and personal choices; and affective competencies, e.g. understanding and processing own feelings as well as developing empathy and sensitivity towards others (UNESCO, 2017, p. 10). These competencies partly overlap with Bandura's (2006) perception of core properties of agency -intentionality, forethought, self-reactiveness, and self-reflectiveness -as they require both reflective and anticipatory thinking and the ability to direct one's actions to the desired direction.
While these sustainability competencies are context-independent (Rieckmann 2012), skills needed for advancing sustainability in specific circumstances vary. For example, maintaining a garden in a sustainable manner requires knowledge and practical skills of pollinator-friendly native plants and organic fertilisers and pesticides, whereas online campaigning for increasing the proportion of vegetarian meals sold in restaurants nation-wide demands good communication skills and a solid understanding of how to reach wide audiences. Thus, we argue that in addition to key competencies for sustainability, context-dependent competencies are highly relevant for advancing sustainability in practice. Also, due to the long-term, iterative nature of sustainability pursuits and continuously emerging new information, we highlight that life-long learning skills and motivation to continuously learn more are necessary for enduring agency (see Sterling 2010;Vare and Scott 2007).
The features that support sustainability agency -motivations, efficacy perceptions, values, beliefs, and competencies -are interlinked and partly overlapping. While efficacy perceptions, values, beliefs, and competencies are essential elements for building agency directly, they assumably also encourage it indirectly through increased motivation to act based on one's own interpretations on sustainability transformations. Also, practising agency can support and strengthen these key features; thus the interactions are not linear but rather cyclic. However, the interpretation of these elements and their interactions remains hypothetical and calls for further empirical investigation and endorsement. Also, it is likely that in empirical studies other relevant elements may emerge.

Defining sustainability agency
Based on Bandura's (2006) work and what has been discussed in this paper, we propose the following definition for sustainability agency. It refers to a. the desire and associated competencies to take individual, proxy or collective actions in personal life, in a community or in society, to contribute to sustainability transformations locally, regionally, nationally or globally b. the intentions to continuously motivate personal and collective learning for sustainability, aiming to adjust actions and behaviours in the light of previous experiences, the information available, and changing circumstances.
With this definition, we want to highlight the diversity of ways to participate in building a more sustainable future. People, contexts in which they function, and the locally most acute sustainability problems are different (Miller 2013). Not everyone has to be interested in or competent to promote the same aspects of sustainability with similar means. This connects with the need for autonomy, which is one of the central components of motivation in the self-determination theory (Ryan, Bradshaw, and Deci 2019).
Similarly to agency in any life sphere, societal, cultural, political, economic and technological systems as well as social and physical environments influence the emergence of agency and create prerequisites, opportunities and barriers for different modes and scales of sustainability agency (Bandura 2006). We acknowledge the uniqueness of and diversity in people's lives, living circumstances and opportunities, which can be drastically different within and between families, communities, and countries. For example, the society one lives in (e.g. authoritarian vs. democratic), origin, financial position, available resources, social status, personal qualities and interests, age, gender, and health all influence the opportunities at hand and how much agency and effort a certain action may require. Therefore, sustainability agency should not be too narrowly defined or exclusive, and agency-centred sustainability education should be sensitive to the individual, social and contextual factors that create prerequisites for learning and taking action.
It is important that agency endures in the long term because the problems of unsustainability and its root causes are not easily nor quickly fixed and new problems are likely to emerge (see Wals 2010). Therefore, interest and motivation to continuously learn more and persevere when facing difficulties is an inherent quality of sustainability agency. Furthermore, knowledge concerning sustainability issues and how to tackle them is intertwined with uncertainty (Miller 2013;Peters and Wals 2013). Therefore, what is sustainable needs to be continuously re-discovered (ibid.). Hence, continuous reflection and learning are needed to iteratively choose meaningful actions in changing and uncertain circumstances, which calls for the self-reactiveness property of agency (Bandura 2006, p. 165). This continuous reflection and learning do not need to remain the responsibility of individuals as collective endeavours are crucial also here. To read more about collective actions and social learning theory, see for example Clark (2016) or Sol, Beers, and Wals (2013).
Behaviour change, which is sometimes regarded as the goal of sustainability education (Hungerford and Volk 1990;Kollmuss and Agyeman 2002), is not a synonym for agency. The term behaviour change implies that a certain behaviour, e.g. energy (Vainio et al. 2020) or eating behaviour (Peltola, Kaljonen, and Kettunen 2020), is aimed to be changed to a predetermined direction. On the other hand, sustainability agency as defined above refers to the capacity to be attentive and proactive in finding possibilities to either act in a more sustainable manner or to advance changes towards sustainability individually, through others, or with others.
Agency can lead to changing one's own behaviour (e.g. starting to compost biodegredable waste) or to planning interventions that support other people's behaviour change (e.g. city planners working to find solutions that encourage sustainable transportation). In these types of situations, knowledge of behaviour change theories can be useful. For example, some famous models have been developed to describe determinants for sustainable behaviour (Bamberg and Möser 2007;Hungerford and Volk 1990;Hines, Hungerford, and Tomera 1987;Kollmuss and Agyeman 2002). A more recent evaluation and synthesis of various behaviour change theories, the behaviour change wheel, can also serve as a useful tool for designing interventions that target specific behaviours (Michie, van Stralen, and West 2011). However, agency extends beyond behaviour change and while behaviour changes are certainly needed for reaching the ambitious goals set for humankind, such as Agenda 2030 (UN, 2015a) or the Paris Climate Agreement (UN, 2015b), encouraging and enabling behaviour changes can require agency from, for example, educators, communication professionals, researchers, policy-makers, or various communities and organisations. Therefore, we want to emphasise that sustainability agency is needed and can be enacted in various roles.

Encouraging sustainability agency through education
Human agency is not a fixed attribute but it evolves and can be practised throughout the lifespan (Biesta and Tedder 2007). Seeking to strengthen everyone's sustainability agency is critical for two reasons. First, people have a right to be as prepared as possible for the uncertain today and future, to be able to function and maintain their well-being within planetary boundaries in evidently changing circumstances (Ojala 2007;Sterling 2010). Second, as sustainability agency may be realised in different forms and contexts, people's individual and collective efforts combined should lead to extensive enough transformations that can secure a viable future for the current and future generations, which likely requires some form of agency from everyone (see Kaaronen and Strelkovskii 2020). Therefore, embedding agency in (sustainability) education is critical. In this section, we discuss agency and sustainability education and propose means to implement agency-centred sustainability education in formal, non-formal and informal learning settings.
The basic orientations in sustainability education can be roughly grouped into two approaches, 'intrinsic' and 'instrumental' (Sterling 2010;see Poeck and Vandenabeele 2012;Vare and Scott 2007;Wals 2011). This distinction refers to the main aims and means of sustainability education. In the intrinsic view, the primary focus is on quality learning processes, including supporting learners' competencies and resilience to act and make decisions in an informed and independent manner (Sterling 2010). In turn, the instrumental approach refers to education that emphasises learning outcomes that are set beforehand, e.g. acting in a certain pro-environmental manner (ibid.). Our perception of sustainability agency and how to foster it in education leans towards the intrinsic view of education.
However, following Barth and Michelsen's (2013) remark about sustainability education being too focused on individuals and their personal sense of responsibility to bring about change, we emphasise collective learning for sustainability. Often in sustainability education, individual actions are emphasised over collective actions and actions with lower impact, e.g. switching off lights, are promoted instead of more influential actions, such as writing political petitions (Aarnio-Linnanvuori 2018, p. 67; Barth and Michelsen 2013;Chawla and Cushing 2007). When acting together, the potential impacts are larger and the strengths emerging from social-psychological mechanisms, such as social approval and cohesion, can encourage further action (Arbuthnott 2009;Wals 2011).
Acknowledging the diversity of actions that can advance sustainability as well as learners' autonomy to work on matters they consider as meaningful is important to avoid indoctrination (Chawla and Cushing 2007;Katsenou, Flogaitis, and Liarakou 2013). Taking this into account can help to co-create projects and action plans that enable learners to develop their self-and collective efficacy (Chawla and Cushing 2007;Katsenou, Flogaitis, and Liarakou 2013). Simultaneously, supporting learners' knowledge production of causes of different sustainability issues and influential ways of tackling them is essential (Hungerford and Volk 1990).
There is plenty of research and practical examples of environmental and sustainability education with children and young people (see reviews from Ardoin et al. 2018;Hedefalk, Almqvist, and Östman 2015;O'Flaherty and Liddy 2018;Stern, Powell, and Hill 2014). This is sensible because the foundations for values, attitudes, knowledge construction and behaviour patterns start to develop early (Chawla 1999;Daries et al. 2009;Evans, Otto, and Kaiser 2018;Samuelsson 2011). Youth is also a fruitful phase for sustainability education because cognitive and moral development are at stages that enable young people to delve more into the reasons behind and consequences of complex sustainability issues (Ojala and Lakew 2017;Smetana, Campione-Barr, and Metzger 2006). While autonomy over one's own choices and actions increases in youth, simultaneously do opportunities for political and other activity (Smetana, Campione-Barr, and Metzger 2006). Moreover, as children and young people will live in the world more affected by global environmental changes and other global trends, they have a right to participate in constructing it (Corner et al. 2015;O'Brien et al., 2018;Ojala 2007). Sustainability education can create circumstances for the needed participation (Chawla and Cushing 2007).
However, as Walker (2017) notes, children are sometimes too idealistically positioned as agents of change in their homes and in the wider society in sustainability education narratives. Given that rapid and systemic changes are needed to respond to the acute sustainability problems (e.g. IPBES, 2018;IPCC 2021;UN, 2019), the importance of the decisions made and actions taken by current adult populations must not be underestimated. Therefore, it is critical to engage all societal sectors and stakeholders in learning for sustainability in formal, informal and non-formal contexts (Dlouhá et al. 2013;Koskela 2019;Sol, Beers, and Wals 2013;Zachariou and Symeou 2009). In order to reach diverse groups of people and provide structures for meaningful learning and participation, creative and extensive research and practical experiments are needed.
Applying transformative learning theory could be useful for approaching agency-centred sustainability education among diverse groups. Originally an adult learning theory, transformative learning refers to learning that profoundly shifts learners' thought patterns and influences their actions, too (Mezirow 1997;O'Sullivan 2012, p. 164). According to Mezirow (1978), the phases of transformative learning include a disorienting dilemma, which can lead to critical reflection of fixed thought and behaviour patterns and planning and trying new courses of action. While this kind of learning process demands agency from the learner, it can also lead to increased agency.
Exerting the core properties of agency -intentionality, forethought, self-reactiveness, and self-reflectiveness -is a continuous process (Bandura 2006). Similarly, strengthening sustainability agency is iterative. Learning and acting for sustainability can occur in parallel and the three modes of agency can be used flexibly depending on the issue, competencies, and resources at hand (see Figure 1 for a summary of sustainability agency). Gradually, competencies develop and new knowledge is produced, which enables applying agency to new spheres of action. Ideally, this could lead to a cycle of life-long learning and life-long activity for sustainability.
In Table 1, we propose six steps for how to support learners' agency in sustainability education. In this table, we draw on what has been discussed earlier in this article and bring together agency, sustainability education, and psychological features and competencies that support agency. The theoretical background behind each of the suggestions is presented in the table. Also, the sustainability competencies that a particular step links with are included. As discussed earlier, it is necessary that agency-centred sustainability education is sensitive to the individual, social and contextual factors that create prerequisites for learning and taking action. These steps can serve as guidance and a starting point when designing agency-centred learning experiences.

Discussion
In this paper, we have conceptualised sustainability agency, which refers to the desire and associated competencies to take individual, proxy or collective actions in personal life, in a community or in society, to contribute to sustainability transformations locally, regionally, nationally or globally; and the intentions to continuously motivate personal and collective learning for sustainability, aiming to adjust actions and behaviours in the light of previous experiences, the information available, and changing circumstances. Supportive features for such agency include but are not necessarily limited to motivations, efficacy perceptions, values, beliefs, and competencies, the latter three of which can further fuel motivations to act for change. Taking into account these supportive features, we have proposed six steps for how to strengthen learners' agency in sustainability education.
Longitudinal research on how agency evolves over time is important to enlighten what kinds of phases, events and practices can either increase or decrease the sense of individual, proxy or collective agency among people from different cultures and diverse socio-economic backgrounds. How individuals and communities experience and strengthen their agency and what kinds of actions they view as feasible, legitimate, effective, and transformative in different political systems is another relevant research area. Furthermore, empirical research that explores agency and transformative learning together could provide useful information on how to support particularly adults' sustainability agency.
As progress and detectable sustainability transformations can occur rather slowly, it is particularly relevant to find means and practices that help to sustain agency and hope in the long term. Therefore, it is important to empirically study whether and how developing sustainability  (Bandura 2006), which can be supported by creating structures for formal, informal and non-formal learning.
agency and related competencies as well as flexible application of the three modes of agency can help to manage setbacks, negative emotions, and even anxiety (Pihkala 2020) related to complex sustainability issues. Finally, while there is strong theoretical and empirical evidence that supports the relevance of agency for achieving the desired sustainability goals, the impacts that collectively increased agency has not only on learners but also on actual sustainability transformations is a relevant question. This aspect requires evaluating and monitoring of the impacts of different measures the learning and participation experiences lead to in the short or long term, which is often neglected when studying sustainability and environmental education (Heimlich 2010;Keene and Blumstein 2010).
To find solutions that maximise synergies and minimise trade-offs between different scales and aspects of sustainability, the core properties and all modes of agency -individual, proxy,  (Bandura 1994) • self-determination theory (e.g. ryan, Bradshaw, and deci 2019) • systems thinking competency • anticipatory competency 6) involve elements that enable learners to experience joy of doing meaningful things together and recognise what has already been achieved.
• self-efficacy theory (Bandura 1994) • self-determination theory, particularly relatadness (e.g. ryan, Bradshaw, and deci 2019) • Emotional awareness, hope and anticipation (ojala 2013; ojala 2017) • collaborative competency • anticipatory competency • self-awareness competency six steps for designing sustainability education with theoretical background and related sustainability competencies. and collective -are essential. Deliberatively and continuously creating cross-sectoral opportunities for people from various backgrounds to participate and engage in decision-and change-making processes, to learn, and to develop and enact agency is a necessity for moving towards more sustainable societies (Barth and Michelsen 2013). Hence, we argue for systematically embedding agency-centred sustainability education -sensitive to psychological, social and contextual factors -in formal, non-formal and informal learning settings.