1.2.1 Learning as Individuals

To understand learning in environmental governance, we first need a shared understanding of what learning means. As learning involves changes in cognition or behaviors, we usually start with individual-level definitions of learning. You may remember learning how to ride a bike, write a poem, or play the piano as a child. Through those memories, you might see a process to learning – such as trial and error, practice, or exposure to new ideas. But learning is not just a process. It is also a product or outcome that reflects some change in understanding or beliefs, and/or changes in behaviors or abilities. Figure 1 illustrates these ideas in a highly simplified model. If you could stay on a bike and navigate around obstacles without falling, you could say you learned how to ride a bike.

Figure 1 A simplified model of individual learning

The fields of developmental psychology, cognitive and behavioral sciences, and education have identified a wide range of individual characteristics that influence how governance actors learn (e.g., see Reference BarrouilletBarrouillet, 2015; Reference Resnick, Paris, Olson and StevensonResnick, 2017; Reference Tyng, Amin, Saad and MalikTyng et al., 2017). For instance, individual motivation, cognitive abilities, prior experiences, emotions, attention, and the capacity for self-reflection can play important roles in learning. The ways in which individuals engage in learning processes and the types of learning products that emerge can then reshape or reinforce individual characteristics that support or impede learning. Certain cognitive biases, which we discuss in detail in Section 1.3, can impede learning by limiting what type of information someone is exposed to, or whether they are open to changing their beliefs. As we know from a wealth of political psychology literature, reinforcing existing biases through one’s “information bubbles” may create learning feedback loops that focus individuals on only learning from information that supports their existing biases – a phenomenon that fosters political polarization.

1.2.2 Learning Collectively

Learning in environmental governance is not merely about the summation of individual learning in governance processes. Consider a sports analogy. All players on a basketball team can individually learn how to shoot the ball by practicing on their own. The outcome of their learning might be observed by counting their percentage of accurate shots by individual team members. Yet, having a group of good individual players does not mean that the team has learned to play well together. Take the 1969 Los Angeles Lakers, the 1993 Phoenix Suns, or the 2002 Sacramento Kings – professional basketball teams in the US known for having some of the best talent in the league, but they still could not win the national title. To learn collectively, we need the right people and the right processes and structures. A good basketball team, for instance, has players with diverse skills and experience, and usually a talented and motivating coach. The ideas that a coach introduces to the team help foster retention of knowledge and skills, and adaptability of those skills to different contexts. Such learning usually occurs through a well-designed process of practicing together, and a structure that ensures practice will happen at a set time and place. The team can then build strategies, anticipate the approach of the other team, and adapt their strategies when the unexpected happens. Sending each of the players to practice dribbling in isolation will not result in team learning.

Learning together, or collective learning in environmental governance is where groups of people acquire, translate, and disseminate innovative ideas about environmental problems or solutions (collective learning “process”), and that process leads to changes in collective understanding or beliefs (collective learning “products”) (Reference Heikkila and GerlakHeikkila and Gerlak, 2013). As we describe in more detail in Section 2, learning products in environmental governance sometimes focus on technical issues, such as how new chemicals might be contaminating drinking water. Other times, collective learning shifts our understanding of diverse human values toward environmental challenges or of the preferences of different stakeholders for adapting environmental policies (Reference RodelaRodela, 2011; Reference Raymond and ClearyRaymond and Cleary, 2013; Reference Newig, Kochskamper E. and JagerNewig et al., 2016). Collective learning products may lead to adaptation or change in governance outcomes. This concept of collective learning is illustrated in Figure 2.

Figure 2 Collective learning in a governance setting

Also shown in Figure 2, governance actors learn within a collective governance “setting” – such as a decision-making venue, an implementation program, or a coalition where actors engage. The collective setting affects both the group, and how individuals within the group learn, along with who teaches. A governance setting that is designed to bring in more diverse forms of knowledge, for instance, could create different incentives for learning than a governance process where knowledge is restricted or controlled by a few actors. Thus, learning is by no means uniform across environmental governance settings, or even across the individual actors within a given setting. Some actors may play more important roles in a learning process, based on their experience, knowledge, networks, or other social factors, which we discuss in Section 3. At the same time, structural features of governance, as we discuss in Section 4 can create differences in who has access to new knowledge, whose information is given more weight or privilege, or whose information can be shared.

1.2.3 Learning Right from Wrong

We do not want to suggest that all learning is productive, or that all learning leads to better environmental governance. History shows us that people learn all kinds of lessons that have negative or unproductive consequences. People can also learn to work only in the interests of those with political power – at the expense of the marginalized. They learn how to resist change or become fragile in the face of new crises or problems. They can learn lessons that reinforce past knowledge and support the status quo. Yet, the lessons we think are good today might look quite different in the future. Indeed, it took decades for decision-makers to learn about the detrimental impact of draining the Everglades. At the time, draining the Everglades was a learned response to flood risk more than a century ago. Similarly, in the Everglades and in many environmental governance processes, it has taken decades to “unlearn” governance practices that exclude environmental and Indigenous values at the expense of agriculture and development interests. Other times, we unlearn lessons too fast, or learning does not stick. After a crisis like an off-shore oil spill, for instance, we might see some immediate lessons about what could have been done differently to prevent the crisis, but that knowledge quickly fades into the background and old patterns persist.

We are admittedly a bit blind to whether or how the insights we offer for infusing learning in environmental governance could foster or reinforce learning the wrong lessons, or whether the learning will stick. Our hope is that we, as scholars, students, and practitioners of governance, can learn about governance processes that teach us bad lessons and correct for these. This assumes that most people are interested in learning beneficial and enduring lessons, especially the people who work diligently toward improving environmental outcomes and sustaining a healthy and livable planet.

1.3.1 Unpacking Cognitive Biases

Cognitive biases can impede learning. One of those biases is that people tend to have negative perceptions toward out-groups and positive feelings for their in-groups (Reference Tajfel, Turner, Austin and WorchelTajfel and Turner, 1979; Reference BrewerBrewer, 1999). In the governance realm where issues are often highly politicized, groups form around shared policy beliefs and political identities, which produce strong in-group affiliations (Reference MasonMason, 2015). At the same time, people may be unwilling to value information that comes from out-groups with opposing political beliefs or identities. Such “in-group/out-group bias” can also lead people to use only information that aligns with their existing beliefs when making governance choices (Reference KundaKunda, 1990; Reference Strickland, Taber and LodgeStrickland et al., 2011). When people feel their group is threatened by information that counters their beliefs they often react emotionally, which further reinforces their existing positions (Reference MasonMason, 2015; Reference Mackie and SmithMackie and Smith, 2017).

Other cognitive biases in decision-making (Reference KahnemanKahneman, 2011) can hinder learning. For example, confirmation bias leads individuals to acquire information that supports their policy beliefs and discount information that does not align with those beliefs (Reference NickersonNickerson, 1998). Likewise, desirability bias drives people to interpret information in support of the beliefs or outcomes they desire to be true (Reference Tappin, Leer and McKayTappin et al., 2017). Availability bias can also constrain the scope of information or one’s ability to learn from it when focusing on information that is closely available or simply top of mind (Reference Tversky and KahnemanTversky and Kahneman, 1973; Reference SunsteinSunstein, 2006). Finally, negativity bias can lead people to misinterpret risk, which can block learning (Reference DeCaro, Arnold, Boamah and GarmestaniDeCaro et al., 2017) because they overweigh negative information (Reference Baumeister, Bratslavsky, Finkenauer and VohsBaumeister et al., 2001). Each of these cognitive biases, and others, can impede learning in environmental governance by steering people away from relevant information that might help in understanding or addressing pressing environmental issues. Recognizing these barriers and intentionally mitigating their negative effects on learning is necessary for effective environmental governance.

1.3.2 Connecting Individual Barriers to the Governance Setting

Both the cognitive biases that create barriers to learning and the factors that can overcome such biases are shaped by governance settings. For instance, in-group/out-group biases are more likely to be triggered when a governance setting involves low levels of trust or interactions are highly adversarial (Reference Weible, Pattison and SabatierWeible et al., 2010; Reference Montpetit and LachapelleMontpetit and Lachapelle, 2017; Reference Heikkila, Weible and GerlakHeikkila et al., 2020). Multilevel settings can challenge how and where learning can occur and may create both positive and negative feedback that influence learning (Reference Di Giulio and VecchiDi Giulio and Vecchi, 2019). As size and diversity of the actors involved increase, governance actors may be exposed to a wider range of views and thus reduce the tendency toward confirmation bias. Yet, in-group/out-group biases can also increase. Likewise, the broader environmental context within which the governance setting is operating might further shape individual biases to learning. Take unexpected environmental crises, like an off-shore oil spill, which can serve as “learning trigger” because it can reveal deficiencies in governance (Reference DeverellDeverell, 2009). However, in some cases of extreme surprise and uncertainty, actors may resort to availability bias, status quo bias, or anchoring bias rather than learning (Reference Kamkhaji and RadaelliKamkhaji and Radaelli, 2017). That is, people may retreat to existing beliefs, ignore potentially relevant information, and collective learning becomes blocked – at least within the time that might be necessary to effectively address the problem.

1.4.1 Detecting Learning

In Section 2, we argue that researchers who study environmental governance and the decision-makers involved in designing and implementing environmental governance processes need to pay attention to learning directly and assess it. To help devise a strategy for this, we first distill various theories and scientific studies of learning to identify who learns, how we learn, where we learn, what we learn, and to what effect we learn in environmental governance. We also offer practical examples on how to measure learning in environmental governance.

1.4.2 The Social Foundation for Learning

Section 3 explains how to build the social foundation for learning together in environmental governance. In general, when diverse participants interact regularly to define the problems, share knowledge, and set and implement solutions, learning is more likely. But creating an environment that is conducive to learning is not always easy. Also needed are groups that represent a diverse range of interests, who engage in genuine and face-to-face dialogue, and who have skillful learning leaders acting as brokers between people with competing interests or goals.

1.4.3 Structures to Support Learning

The social elements of governance identified in Section 3 need structures to guide and support learning. Structural features include formal feedback loops to institutionalize learning, dedicated resources to sustain learning, and well-connected venues and organizations that can hard-wire learning across a network.

Sections 3 and 4 discuss several caveats when implementing the social and structural governance foundations we identify. These caveats acknowledge the importance of paying attention to the larger contexts surrounding environmental governance. These contexts include political, socio-economic, and institutional conditions that can influence the feasibility of implementing different social or structural features to support learning into environmental governance.

1.4.4 Infusing Learning into Environmental Governance

Sections 3 and 4 uncover a set of foundational governance features that can guide research and practice, but they are not meant to be a step-by-step recipe. No magic formula exists that works across all organizations or decision-making venues of environmental governance, or that works across all governance questions. Instead, it may be important to infuse learning in a way that allows for continuous adaptation of environmental governance systems and processes. In doing so, attention should be given to how the social and structural features of governance interrelate and reinforce the capacity for learning. This requires ongoing diagnoses – from academics and practitioners – of how well the individuals and organizations involved in environmental governance are learning and what pieces of the learning puzzle are missing. In the last section of this Element, we encourage environmental governance scholars to infuse learning intentionally into environmental governance by developing a diagnostic approach and building learning systems. In doing so, we offer strategies for how scholars and practitioners can collaborate in advancing intentional learning.

2.1.1 Who Is Learning (and Teaching): Power and Empowerment?

In environmental governance, who learns is a key question. This decision begins with who is invited to the proverbial governance “table,” who has power at the table, and who is not at the table. It also begs the question “who is teaching,” which influences whose knowledge informs decision-making, along with “what is learned” and “to what effect” as we discuss later. Diverse forms of power can be at play in learning processes, including economic power, informational power, operational power, and social power (e.g., derived from personal skills and knowledge), as well as formal authority and capacity (Reference Roome and WijenRoome and Wijen, 2006).

Learning scholars often describe “experts” as both learners and teachers. For instance, policy scholars emphasize the role of policy brokers or policy entrepreneurs to initiate learning processes and facilitate policy change (Reference Howlett, Mukherjee and KoppenjanHowlett et al., 2017; Reference Valin and HuitemaValin and Huitema, 2023). Similarly epistemic communities, or networks of professionals with expertise, through their knowledge generation and dissemination at both global and local scales in environmental governance, are sources of learning and teaching (e.g., Reference HaasHaas, 1992; Reference Mabon, Shih, Kondo, Kanekiyo and HayabuchiMabon et al., 2019). In addition, coalitions and networks of elite policy actors (e.g., interest groups, industry, environmental nonprofits, universities, and donors), are both teachers (i.e., by sharing their preferences with decision-makers) and learners who can affect policy change (Reference Moyson, Scholten and WeibleMoyson et al., 2017; Reference PattisonPattison, 2018).

Experts and policy elites are not the only actors who learn (and teach) in environmental governance. The roles of citizens or communities (Reference Pormon and LejanoPormon and Lejano, 2023) and local or Indigenous knowledge in environmental governance (Reference Bohensky and MaruBohensky and Maru, 2011) are widely recognized as providing relevant governance knowledge, often grounded in daily experience and practices (Reference Eshuis and StuiverEshuis and Stuiver, 2005). Moreover, by embracing interactions with a broader array of governance actors, the capacity to learn and understand complex issues increases (Reference Pescaroli and AlexanderPescaroli and Alexander, 2018). In other words, who learns, and who teaches, in environmental governance should be examined from a broad lens that acknowledges the roles of diverse types of actors, particularly those who traditionally do not have power in government.

One of the ways local and Indigenous actors might have opportunities to both learn and teach in environmental governance is through collaborative knowledge production, or co-production processes (e.g., Reference Baker, Ekstrom, Meagher, Preston and BedsworthBaker et al., 2020; Reference Turnhout, Metze, Wyborn, Klenk and LouderTurnhout et al., 2020). In these processes, knowledge users and producers interact to develop tailored information fit for purpose. The extent of that learning, and whether certain individuals play more central roles in collective learning (e.g., leaders, knowledge brokers) relative to traditionally marginalized groups, however, can depend on various governance characteristics, such as the quality of interactions between groups (Reference Baker, Ekstrom, Meagher, Preston and BedsworthBaker et al., 2020; Reference Norström, Cvitanovic, Löf, West, Wyborn, Balvanera, Bednarek, Bennett, Biggs, de Bremond and CampbellNorström et al., 2020) – which we describe in more detail in Section 3. It may also depend on how much agency marginalized groups are afforded to self-define their roles, responsibilities, objectives, and priorities (Reference Turnhout, Metze, Wyborn, Klenk and LouderTurnhout et al., 2020) – a consideration for learning structures we touch upon in Section 4. Ideally, learning processes under any approach would empower community members in ways that help build broader resilience and sustainability (Reference Pormon and LejanoPormon and Lejano, 2023).

2.1.2 How We Learn Together: Acquiring, Translating, and Disseminating

Collective learning unfolds through a process of acquisition, translation, and dissemination of knowledge, information, and experience (Reference Gerlak and HeikkilaGerlak and Heikkila, 2011). The acquisition phase involves the collection or receipt of information. Information can be acquired experientially, through searching and noticing, practice and conditioning, and through trial and error (Reference Fazey, Fazey and FazeyFazey et al., 2005). Organizations may also engage in self-appraisals, experiments, or assessments (Reference Schneider and IngramSchneider and Ingram, 1990; Reference Walters and HollingWalters and Holling, 1990) to gain information that can allow actors to learn about errors or problems or how processes have been unfolding over time.

Through translation, which involves interpreting the meaning of new information, information is understood and becomes shared knowledge. Individuals, groups, and networks may seek out sources of expert knowledge or rely on trusted advisors to help interpret information (Reference Heikkila and GerlakHeikkila and Gerlak, 2013: 489). Subconsciously, mental framing, individual habits, professional training, and behavioral patterns may create “ways of knowing” that influence translation of ideas across diverse members of a group (Reference Lejano and IngramLejano and Ingram, 2009).

Finally, individual-level knowledge within a collective group develops into shared knowledge through dissemination, or the transference and embedding of information across a group (Reference ArgoteArgote, 2011). Dissemination can occur in a variety of ways, such as through collective routines and shared communication processes (Reference Feldman and RafaeliFeldman and Rafaeli, 2002). Storytelling can also help promote dissemination because it can help individuals within a group relate to new information about problems or new approaches or solutions (Reference Crossan, Lane and WhiteCrossan et al., 1999).

The phases of collective learning are not always differentiated or directly observed when studying learning in environmental governance. For instance, researchers of social learning focus on interactions between people as the primary way individuals learn (Reference Armitage, Marschke and PlummerArmitage et al., 2008; Reference Pahl-WostlPahl-Wostl, 2009). Reference Lejano, Haque and BerkesLejano and colleagues (2021) emphasize the process of mutual development and sharing of knowledge through iterative reflection on observation and experience.

Others pay attention to specific mechanisms that feed into multiple stages of learning. By “mechanism” we are referring to a pathway that brings an outcome, such as learning, into being, or a process that triggers the multiple stages of learning. For instance, storytelling can be a powerful process not only for disseminating information, but also helping people acquire knowledge about marginalized groups (Reference Adelle, Black and KrollAdelle et al., 2022), interpret information more intuitively (Reference Goldstein, Wessells, Lejano and ButlerGoldstein et al., 2015; Reference BorisBoris, 2017), or make sense of nonroutine, uncertain, or novel situations (Reference Bietti, Tilston and BangerterBietti et al., 2019). Similarly, scenarios are commonly used to help people acquire information about possible alternative futures, and then interpret how that information connects to present actions and decision-making around risk and uncertainty (e.g., Reference Marchau, Walker, Bloemen and PopperMarchau et al., 2019; Reference Gerlak, Jacobs, McCoy, Martin, Rivera-Torres, Murveit, Leinberger and ThomureGerlak et al., 2021a).

2.2.1 The “What” in Learning: Not All Learning Is the Same

Learning researchers commonly discuss what individuals, groups, and organizations learn in governance processes, often as “types” of learning. Reference MayMay’s (1992) classic typology (e.g., see Reference BennettBennett, 1997; Reference MeseguerMeseguer, 2005) includes: political learning, which is learning about the strategies or tools to influence policy; instrumental learning, or learning about the feasibility and effectiveness of policy implementation; and social policy learning, which is about the definition or social construction of a policy issue. Other typologies refer to conceptual versus technical (Reference NilssonNilsson, 2006) or more cognitive versus relational kinds of learning (Reference Armitage, Dzyundzyak, Baird, Bodin, Plummer and SchultzArmitage et al., 2018; Reference Huitema, Cornelisse and OttowHuitema et al., 2010). Others differentiate between non-learning, where no information is processed, and blocked learning, where learning occurs but is not translated or disseminated into group knowledge or action (Reference Vagionaki, Dunlop, Radaelli and TreinVagionaki, 2018).

Some typologies of learning blend how learning occurs with what types of learning arise. Reference RietigRietig (2019), for example, suggests three basic archetypes of learning: factual learning, experiential learning, and constructivist learning. While factual learning focuses on what knowledge is gained, experiential learning is about how engagement in a governance process produces a different type of learning. Likewise, constructivist learning is about what changes to beliefs result from processes that are “more than a reflection on the input and a resulting increase in knowledge or experience” (Reference RietigRietig, 2021: 165). This concept of constructivist learning is like reflexive learning (e.g., see Reference Dunlop and RadaelliDunlop and Radaelli, 2013), which is about a reflexive mode – or how learning occurs – and resulting changes in preferences – or what learning occurs.

In our own work, we categorize what is learned into both cognitive and behavioral forms of learning (Reference Heikkila and GerlakHeikkila and Gerlak, 2013). Cognitive learning is about what changes occur in the minds of individuals. They can include changes in factual understanding, knowledge of political strategies or even changes to core values and beliefs. The behavioral manifestations of learning involve changes to strategies, rules, or policies in response to new information.

These diverse typologies of what is learned, and their intersection with how learning occurs, have led to messiness in research design and conceptualization. Still, typologies of learning can be useful for developing more nuanced theories and hypotheses of how or why different types of learning may emerge. They also can allow researchers to develop insightful comparisons of learning products across governance settings and contexts. Doing so might then enable us to trace whether who learns, where governance actors learn, and how they learn can shape what they learn.

2.2.2 “To Which Effect”: Depth and Breadth of Learning

What is learned, and often how learning happens, are often characterized – whether explicitly or implicitly – in terms of their depth. The concept of “loop” learning is perhaps the most common categorization of learning depth. Single-loop learning involves the detection and correction of errors, which allow organizations to implement their objectives and norms, while double-loop learning goes deeper by modifying those norms and objectives (Argyris and Schön, Reference Argyris and Schön1996). Certain learning processes, such as knowledge co-production, are more often linked to double-loop learning because they enable governance actors to gain insights into varying perspectives and understandings of policy solutions. Triple-loop learning goes even deeper by examining how we learn about and eventually change underlying governance norms (Reference Johannessen, Swartling, Wamsler, Andersson, Arran, Vivas and StenströmJohannessen et al., 2019). Researchers studying adaptive environmental governance often see triple-loop learning as key to improving governance processes and building capacity to better respond to uncertain environmental changes (Reference Armitage, Marschke and PlummerArmitage et al., 2008).

From an individual learning perspective, deep forms of learning occur when people overcome prior experience – or experience that could result in inappropriate choices when facing complex or dynamic situations – and adapt their decisions to information about new circumstances (Reference OhlssonOhlsson, 2011). Deeper forms of learning can eventually transform groups and social processes (Reference MezirowMezirow, 1994). Such “transformative learning” in environmental governance allows participants to construct new cause–effect relationships, which can lead to more integrated, sustainable solutions to difficult environmental problems (Reference Sims and SinclairSims and Sinclair, 2008). Others describe these deeper forms where participants change preferences of learning as “thick” learning, as opposed to “thin” learning, where participants simply learn to cope with a problem without changing preferences (Reference RadaelliRadaelli, 2008: 244). Similarly, deliberative learning (Reference Dunlop and RadaelliDunlop and Radaelli, 2013) where actors co-create knowledge and governance solutions, can lead to new ways of knowing (Reference Lejano and IngramLejano and Ingram, 2009) that would not be feasible without deliberation and reflection.

In addition to the depth of learning, the breadth of learning can be an important consideration in environmental governance. The breadth of learning raises questions such as how many actors involved in governance are learning together, to what extent learning is shared across individuals or organizations in a network, or to what extent the coordination of actors in a polycentric governance system allows for learning across venues (Reference BodinBodin, 2017). Although the scholarship on learning in environmental governance tends to focus on exploring the depth of learning, potentially important considerations about the quality of learning, such as how power struggles play out in learning processes (e.g., Reference Betsill and BulkeleyBetsill and Bulkeley, 2004), can be assessed when accounting for questions around breadth. Empirical studies in fact suggest that learning among only a few key policy actors may be necessary to effectively push lessons into policy action (Reference Rietig and PerkinsRietig and Perkins, 2018). Thus, considering the breadth of learning, when examining “who” learns, can open important theoretical and practical questions.

2.2.3 “Where” Learning Happens: Diverse Governance Settings

Learning in environmental governance takes place within a wide variety of institutional locations and places, decision processes, or forums and venues (Reference Gerlak, Heikkila, Smolinski, Huitema and ArmitageGerlak et al., 2018), which we described in Section 1 as the governance setting. Sometimes these settings are more formal and institutionalized like a regulatory venue or a commission for an international environmental treaty. They can also be informal and ad hoc such as a group of citizen scientists who share data on air quality monitoring with a government agency or nonprofit group, or a task force of people brought together to provide advice on an emerging environmental issue. Alternatively, learning might occur within subgroups of venues, such as a leadership committee, or across a network of local government leaders working on climate change. Whether formal or informal, governance settings can range in jurisdictional scale – from local to regional, national, and international – and in size, geographic scope, and institutional complexity.

The characteristics of governance settings where learning can occur are important because they establish both the social foundation (see Section 3) and the structure (see Section 4) that shape collective learning. For instance, the structure dictates who learns or who is empowered to learn, and what types of social interactions or processes of learning are supported and allowed. As Reference Dunlop and RadaelliDunlop and Radaelli (2013) recognize, different modes of learning are triggered by institutional contexts, such as hierarchies versus epistemic communities. Some settings might have design features that intentionally enable learning – such as policy labs that engage in experiments. Venues also can create different barriers to learning or exacerbate individual learning barriers (Reference Heikkila, Gerlak and SmithHeikkila et al., 2023). In their study of large, diverse forums for national-level climate deliberation, Reference Wagner and Ylä-AnttilaWagner and Ylä-Antilla (2020) find that actors opt to interact with those already in their network who share their beliefs. In doing so, climate policy actors narrow the scope of learning.

Likewise, when trying to observe and diagnose where learning occurs in environmental governance, the complexity of the governance structure can be important to consider. For instance, in highly polycentric governance systems with multiple overlapping venues, the structure of these systems can have critical implications for how learning occurs. Moreover, as the boundaries or scope of the governance system become more complex, certain characteristics of the system, such as its connectivity, may become more important for learning.

2.3 Advancing Measures of Learning

Researchers of learning have adopted a diverse set of approaches to detect and measure learning in environmental governance. Document analysis, interviews, and surveys are common tools (Reference Crow and AlbrightCrow and Albright, 2019; Reference De Voogt and PattersonDe Voogt and Patterson, 2019; Reference KoebeleKoebele, 2019), Others use process tracing and ethnographic approaches (Reference Mukhtarov, Dieperink, Driessen and RileyMukhtarov et al., 2019), participant observation (Reference RietigRietig, 2019; Reference Ricco and SchultzRicco and Schultz, 2019), or cognitive mapping (Reference Huitema, Cornelisse and OttowHuitema et al., 2010). How scholars measure and study learning depend upon whether they are interested in a particular piece of a learning process or a learning product and which level of analysis within a governance context (e.g., individual actors, policy venues, collaborative groups, or networks) is most relevant for their learning questions at hand.

Researchers have struggled with measurement and operationalization challenges, especially around the connection between learning processes and products, and between learning and environmental governance outcomes (Reference Siebenhüner, Rodela and EckerSiebenhüner et al., 2016; Reference Armitage, Dzyundzyak, Baird, Bodin, Plummer and SchultzArmitage et al., 2018). Part of this challenge is due to the latent nature of learning – and our inability to see it directly. We often use proxy indicators, such as changes in behaviors or actions that might arise through processes other than learning (like force). Other times we use self-reported perceptions of learning, which can be biased by social desirability to report learning. Additionally assessing collective learning using measures based on the responses of individuals remains a perennial challenge (e.g., see Reference Montpetit and LachapelleMontpetit and LaChapelle, 2015). For instance, how many individuals must learn to say that the collective has learned? In addition, learning as a process can easily be conflated with changes in policies or outcomes (Reference Muro and JeffreyMuro and Jeffrey, 2012). Challenges also arise around the temporal dimension of learning. Learning sometimes takes more time than researchers have available to study learning. So, we might be able to detect learning processes, but fail at detecting the products or outcomes.

Despite these limitations, we have observed significant improvement in the measurement and conceptualization of learning in recent years. We see more attention to explicit conceptual definitions and operationalization, more sophisticated measurement approaches using multiple methods, and more rigorous analytical tools – both using qualitative and quantitative methods – and advancements in observing learning over multiple time periods. Reference Gronow, Brockhaus, Di Gregorio, Karimo and Ylä-AnttilaGronow and colleagues (2021), for example, measure learning in policy networks that are involved in reducing emissions from deforestation. They provide direct evidence of how belief change, as a form of learning, happens and how different network structures shape learning. Others have used experiments (e.g., Reference Yu, Shin, Perez, Anderies and JanssenYu et al., 2016) to measure learning processes, and outcomes. We have also seen signs of more engaged scholarship approaches. Reference Ricco and SchultzRicco and Schultz’s (2019) study, for example, engaged with US Forest Service employees during discussions of forest plan revisions, with follow-up interviews, to inform their study of learning.

Developing clear measures and approaches for assessing learning processes and products is not simply an academic exercise. From a practical perspective, the who, how, what and to which effect of learning gives us some diagnostic tools that we discuss further in Section 5 for assessing learning in practice. Identifying different types of learning might be useful for thinking more critically about what type of learning may be needed to solve shared problems. Focusing on instrumental learning or more technical forms of learning, for instance, may not improve governance outcomes when the underlying source of an environmental problems is that people simply hold different values about the importance or relevance of the technical issues.

Disentangling the characteristics of the governance setting (where) from the process (who and how) and products (what and to which effect) further is important for identifying and testing the learning hypothesis. The hypothesis assumes that the environmental governance setting (where) can influence how learning happens and what results from it. It also helps separate some of the bundled concepts in learning – like relational learning, which assumes both a process of dialogue and a learning outcome of improved understanding of relations. By unpacking these elements, we might discover that a governance process emphasizing dialogue and communication among diverse stakeholders triggers an unwillingness or inability to learn about one other’s values, rather than triggering relational learning. What might matter in such a process is not just that diverse stakeholders are present, but whether the dialogue is facilitated and what resources are in place to support engagement. We draw out what types of social and structural foundations of environmental governance are more likely to support learning in the next two sections.

3.1.1 Find the Right Mix of People for the Problem

Environmental problems are complex. Learning to address this complexity requires insights from people who bring different perspectives to bear on these problems. Not everyone affected by an environmental governance action, decision, or program, knows in advance that they need to be at the table. Furthermore, the definition of stakeholders or community is often unclear in environmental governance processes. Yet, how we define stakeholders or community determines who gets to be engaged (Reference Bendtsen, Clausen and HansenBendtsen et al., 2021), and therefore who learns and who governance actors learn from.

Over a decade ago, the threat of high valued fish species collapsing in the United Kingdom (UK) triggered a process where diverse stakeholders engaged in dialogue and learning. The threat of the banning the sale of skates and rays “caused shock waves throughout the fish supply chain,” according to researchers at England’s University of Hull (Reference Garrett, MacMullen and SymesGarrett et al., 2012: 184). A series of workshops brought together retail companies, scientists, managers, and working fishers to better understand the scope of the problem and develop potential solutions. The outcomes of these processes included an enhanced understanding of the different situations facing these species, a database to provide timely trade information, and an identification guide to assist fishers. Around the same time, the UK government was threatening to prevent access to parts of the most productive UK scallop fishing grounds through the designation of marine protected areas. Voluntary closed areas were introduced in a few bays across Scotland, Wales, and England but breaches in the agreements resulted in statutory closures in a few bays. Fearful of these closures, the scallop industry brought together a diverse group of stakeholders to counter the negative publicity surrounding the industry and to “moderate proposals to prevent access to scallop grounds” (Reference Garrett, MacMullen and SymesGarrett et al., 2012: 185). Stakeholders included conservation agencies, voluntary wildlife trusts, processors, retailers, and scallop fishers. These stakeholders devised a plan to pilot new fishing gear aimed at reducing habitat damage, improving catch quality, and cutting fuel consumption.

In both cases, the diversity of actors across industry, government, nongovernmental organizations (NGOs), and scientists brought essential, but distinct perspectives on the problem of overfishing. The conversations were not limited to fisheries scientists and policymakers. By engaging more diverse actors and interests, the process allowed for learning about the science-informed tools that industry could use in collaboration with policymakers to create more feasible solutions.

3.1.2 Go Beyond the Usual Suspects

When diverse types of actors are at the table, sometimes those actors are the usual suspects. The usual suspects are often the people and organizations who regularly voice their positions on environmental governance issues, or who are well-known in a specific governance community. A typical regulatory hearing on an environmental issue for instance might include environmental nonprofits, regulated industry, advocacy organizations, and other related government agencies. Although diverse in terms of the organizations they represent, such a mix of actors may lack creative insights or lessons from outside their system that can foster learning and innovation in governance. Going beyond the usual suspects might involve bringing in more local communities when developing projects or environmental solutions targeted at local issues and problems (Reference Folke, Hahn, Olsson and NorbergFolke et al., 2005). Similarly, Indigenous knowledge, or traditional ecological knowledge (TEK), brings essential historical understanding of environmental issues and experience on how to manage natural resources (Reference Bohensky and MaruBohensky and Maru, 2011).

Since western science is often privileged in environmental governance over other forms of knowledge, creative approaches that start early in decision-making processes often are necessary to incorporate alternative types of knowledge. The Glen Canyon Adaptive Management Program in the US follows a mix of strategies to access diverse forms of knowledge (Reference Gerlak, Karambelkar and FergusonGerlak et al., 2021c). Created in 1992, this federal program coordinates dam operations, downstream resource protection, and monitoring in the Glen Canyon region of the Colorado River. In this program, stakeholders can raise emerging issues and invite experts in to present to the larger group, allowing for an influx of outside ideas. Efforts have not always been ideal, however. In response to growing tribal dissatisfaction with the Glen Canyon process, in 2012, the Cultural Resources Ad Hoc Group proposed the inclusion of TEK as a pilot project. The group suggested that the culture of science failed to represent humanistic values and acknowledge a role for Indigenous communities in the program. To rectify this, in 2015 the first-ever Tribal Consultation Plan outlined a process to seek, discuss, and consider the views of Tribes and exchange ideas, not simply providing information.

Another example of creative approaches for bringing diverse stakeholders into the mix is highlighted in an urban forestry program in Melbourne, Australia (Reference Gulsrud, Hertzog and ShearsGulsrud et al., 2018). The city developed several in-person and online forums to intentionally create opportunities for learning around urban forestry. Through their efforts, they brought in voices less frequently heard in community engagement forums. For example, they hosted local in-person forums in neighborhoods across the city at various times of the day and week. They engaged youth in envisioning their local urban landscapes through an art competition. They designed review councils composed of demographically representative citizens to engage in discussions around urban renaturing. Finally, the city created an online platform where citizens can learn about specific trees across the city, which included individual email address for each tree species so citizens could send their thoughts and feelings, or questions about specific trees.

Through these approaches, citizens learned about how proposed strategies could accommodate citizen goals for the neighborhood. They also learned how these goals would fit alongside the city’s indicators of urban forest health like aesthetic preferences and tree planting and removal strategies. Adding to the fun, children actively contributed by drawing neighborhood maps with their favorite trees and plants. In the end, Melbourne’s citizens helped establish 10 urban forest precinct plans through a process involving a more diverse set of people than the city had in past processes. Based on citizen input, the city implemented a long-term strategy to invest in urban green infrastructure with the aim of promoting landscape resilience and community wellbeing. Ironically, when they first initiated the process, City of Melbourne officials had thought the effort would lead to a tree removal as a drought response strategy in the face of climate change. However, the lessons learned through the various community engagement activities were more transformative than what the city had imagined – leading to novel management actions.

3.1.3 Watch Out for Pitfalls When Engaging Diverse Interests

Not all processes that include diverse actors and interests will foster learning. Power differentials or priorities of dominant actors can be challenging. In a study of municipal flood risk governance in the Netherlands, local priorities and understandings of risk directly affected the degree to which municipalities were likely to use information provided by external sources (Reference De Voogt and PattersonDe Voogt and Patterson, 2019). Similarly, researchers studying municipal climate change governance in Durban, South Africa found that strong personalities served as gatekeepers and barriers to external expertise, as experts were perceived as being out of touch with local realities (Reference Leck and RobertsLeck and Roberts, 2015). As a result, department officials narrowly focused on developing and strengthening internal knowledge sources. Sometimes social features reinforce in-group/out-group biases. Participants may gravitate toward those who are most like them, which can reinforce negativity biases toward dissimilar groups (Reference BrewerBrewer, 1999). Engagement approaches that encourage diverse participants to see themselves as the “other,” and not as the center of expertise can help overcome these challenges (Reference PalmerPalmer, 2011).

A fundamental challenge of bringing together diverse people is that they may not speak the same language, including technical languages often adopted in governance processes. They also may not be capable of translating ideas into lessons that are directly applicable to the context of the governance system. It can take considerable time to build a shared or common language to help translate ideas in similar ways and come to agreement on these ideas (Reference Richerson and BoydRicherson and Boyd, 2005). But, when diverse groups of people build a common language, they engage in deeper forms of learning and become more capable of accurately understanding those individuals or groups with competing interests and beliefs. They also are more likely to overcome some of the individual cognitive biases that can impede collective learning. In the case of Melbourne, Australia’s urban forestry efforts, city officials came to see how their intentional efforts to promote learning helped to develop a common language for the community and even served to promote deep learning and transformation among the most vocal and skeptical citizen groups (Reference Gulsrud, Hertzog and ShearsGulsrud et al., 2018).

Despite some of the difficulties that can arise in more inclusive processes, the benefits of bringing diverse interests together likely outweigh the alternative. Limiting participants to a narrow band of interests can lead to poorly informed decisions. Take an example from community wildfire protection planning in the eastern US. Researchers from the University of Minnesota and the US Forest Service found that when participation is limited to certain individuals – in this case, state and local foresters and fire chiefs – a narrow type of institutionalized knowledge of wildfires was present (Reference Brummel, Nelson, Souter, Jakes and WilliamsBrummel et al., 2010). Because the groups primarily consisted of traditional agency partners, they focused on existing management goals of fire suppression and fuels mitigation. This resulted in a less innovative approach to designing new practices and constrained learning.

3.2.1 Create Mutual Respect Through Face-to-Face Interaction

Face-to-face interaction, built on respect and equity, is a key feature of genuine dialogue (Reference Pahl-Wostl, Craps, Dewulf, Mostert, Tabara and TaillieuPahl-Wostl et al., 2007a; Reference Gerlak, Heikkila, Smolinski, Huitema and ArmitageGerlak et al., 2018). Respectful, face-to-face interactions where mutual criticism is tolerated can support learning and greater group creativity (Reference OhlssonOhlsson, 2011). This can be particularly important where people have different mental frames or ways of knowing based on their experience or connection to an environmental issue. Diverse perspectives and opinions, expressed in a mutually respectful manner, can allow people to question their own assumptions and envision new or different futures (Reference Innes and BooherInnes and Booher, 1999).

Ensuring a climate of respect is especially important for vulnerable or disenfranchised groups. When Thai farmers were treated respectfully by public irrigation staff in participatory irrigation management meetings, they were more open to diverse points of views and willing to learn different perspectives from fellow farmers (Reference KumnerdpetKumnerdpet, 2011). At the same time disrupting unequal power relationships and investing in trust-building are often essential to fostering learning (Reference Davidson-Hunt and O’FlahertyDavidson-Hunt and O’Flaherty, 2007).

3.2.2 Use Creative Approaches

Opportunities for dialogue that involve more creative, and playful approaches, such as simulation gaming approaches (Reference Haug, Huitema and WenzlerHaug et al., 2011), scenario building (Reference Innes and BooherInnes and Booher, 1999), field trips (Reference Gerlak and HeikkilaGerlak and Heikkila, 2011), brainstorming, participatory modeling, and cognitive mapping can support learning. Researchers studying adult learning in activist organizations similarly find that using humor and other creative approaches, such as reading circles or knitting groups, can spur learning about activist goals and strategies (Reference OllisOllis, 2020). Through such approaches, governance actors can examine their beliefs in relation to others and learn about the dynamics of social-ecological systems (Reference Özesmi and ÖzesmiÖzesmi and Özesmi, 2003). Such approaches also allow for informal discussion and developing trust (Reference Sinclair, Collins and SpalingSinclair et al., 2011), which can support learning.

One example is “mental modeling,” which elicits people’s internal beliefs about an external reality – such as environmental conditions – and makes those models explicit. Mental modeling exercises, through a series of participatory workshops, helped community members learn about and devise natural disaster strategies in Hawaii’s North Shore of Oahu (Reference Henly-Shepard, Gray and CoxHenly-Shepard et al., 2015). The mental model exercises drew out community members’ beliefs about the natural hazards they were facing, and their perceptions of trade-offs in different adaptation responses. These activities allowed participants to first learn about each other’s understanding of the community system, build trust and empower participants. They also developed strategies to reduce hazard risks to the community. Researchers documented changes in participants’ mental models of natural disasters, finding that people progressed from more superficial to deeper forms of learning over the course of the workshops.

“Serious games” can also provide interactive ways to engage diverse participants in environmental governance. Serious games are useful learning processes when incomplete knowledge, diverse cultural norms, or concerns with power asymmetries exist (Reference Rodela, Ligtenberg and BosmaRodela et al., 2019). In the Cedar River Watershed in eastern Iowa, disciplinary experts, policymakers, and local stakeholders played the multi-hazard tournament, developed by Agriculture and Agri-food Canada (Reference Hill, Hadarits, Rieger, Strickert, Davies and StrobbeHill et al., 2014). It educated and trained participants in decision-making skills around drought and water (Reference Bathke, Haigh, Bernadt, Wall, Hill and CarsonBathke et al., 2019). Participants were organized into teams and charged with crafting solutions for reducing flood, drought, and water quality impacts under climate scenarios. In addition to the game players, they had referees dressed in costume, team facilitators, fans, and an announcer. The game format consisted of (1) a scenario presentation, (2) facilitated discussion of the scenarios, (3) adaptation option selection, (4) crafting of a press release, and (5) scoring. Each of the seven teams had a $1.6 billion budget in the first scenario for adopting water management strategies for a 20-year planning period. Subsequent rounds decreased budgets and incorporated maintenance and operating costs and introduced hazards like flood and drought. Players engaged with new individuals across sectors and worked collaboratively through the scenarios. They gained knowledge about water management options and strengthened their ability to evaluate these options critically.

3.2.3 Sustain Dialogue

Learning is more likely to be cultivated when dialogue is ongoing. Continuous interaction helps actors develop deeper connections to the cultural, historical, and institutional contexts that drive environmental outcomes and better understand the underlying drivers of environmental issues (Reference NyeNye, 1987). This process of deepening our understanding of complex issues is a form of “sensemaking.” Sensemaking involves a continuous process of reorganizing, interpreting, and integrating different forms of information to view issues with greater complexity and nuance (Reference Garrett, MacMullen and SymesGarrett et al., 2012).

A study comparing water management groups in Germany and Ireland found that when German water management groups had more opportunities and time to connect, participants learned more about each other than the Irish groups (Reference Muro and JeffreyMuro and Jeffrey, 2012). Similarly, in the UK, fisheries management groups with longer lifespans proved more likely to develop mutual trust, expand networks of collaboration, and produced higher problem-solving capabilities and more complex, reflexive learning (Reference Garrett, MacMullen and SymesGarrett et al., 2012). Conversely, in the fisheries management groups characterized by weaker engagement and shorter lifespans, learning was more technical.

3.2.4 Pay Attention to Conflicts

Face-to-face dialogue can, at times, aggravate or create conflicts in the governance process, especially if not managed well. Conflict is not necessarily bad. It can be an opportunity for change and learning (Reference Folke, Hahn, Olsson and NorbergFolke et al., 2005) especially when it heightens a sense of urgency for action. Conflicts also can lead people to reinforce or deepen their existing beliefs and perspectives and limit their learning from people with conflicting beliefs. For instance, a survey of people involved in climate and energy policy governance in the state of Colorado in the US found that those individuals with more extreme policy beliefs on climate and energy issues were more likely to reinforce their beliefs, rather than change them (Reference PattisonPattison, 2018).

Creating strategies to avoid the escalation of conflict, such as mediation or facilitation, can minimize the negative effects of conflict on learning in several ways. For instance, mediation can mitigate against intolerance, suspicion, and blame – all of which constrain receptiveness to different ideas across a group (Reference Edmondson, Woolley, Easterby-Smith and LylesEdmondson and Woolley, 2006) and discourage willingness to share or respect others’ ideas. Even with good facilitation, productive dialogue with a large group of participants can be difficult. Building a governance process that can effectively manage conflict is not easy; it often requires establishing best practices and norms around dialogue at the outset (Reference Stöhr and ChabayStöhr and Chabay, 2014).

3.3.1 Spur Innovation

Learning leaders encourage creativity, challenge institutionalized knowledge, and actively disseminate innovative ideas from individuals to groups and organizations (Reference SiebenhünerSiebenhüner, 2008). Their leadership approach is often “transformative” – where they encourage others to question assumptions and look at old problems in new ways. Doing so enables people to adopt more generative and exploratory thinking processes and be more innovative and creative in problem solving (Reference García-Morales, Jiménez-Barrionuevo and Gutiérrez-GutiérrezGarcía-Morales et al., 2012).

Deeper forms of learning, which include both cognitive and behavioral change, demand leaders who are not afraid to build in processes within their organizations to promote learning (Reference SiebenhünerSiebenhüner, 2008). These processes encourage exploration, which involves risk taking and experimentation and requires feed-forward and feedback learning that challenges current institutionalized learning (Reference Jansen, Vera and CrossanJansen et al., 2009). Leaders can initiate, promote, and support innovative approaches and strategies in a variety of ways. For example, an environmental agency director could bring in outside voices or plan activities outside the regular meeting space.

Community art is one emerging approach that leaders can support and guide to bring in more diverse interests and perspectives, and to communicate diverse narratives and knowledge in environmental governance spaces. Through “poetry routes” around river management, local communities in the Netherlands and Indonesia have been engaging in a facilitated arts-based process to produce and combine images and poems in a series of printed banners (Reference Witteveen, van Arensbergen and FliervoetWitteveen et al., 2022). The resulting poetry route then travels and creates spaces for dialogue with other stakeholders similarly concerned or engaged with the issue. Similarly, visual problem appraisal (VPA), a film-based learning strategy, involves sharing various perspectives through a series of filmed interviews. The diversity of filmed stories builds and unfolds through workshops, which requires leaders who can facilitate the storytelling and enable stakeholders to hear other perspectives and examine conflicting interests. Photography and storytelling also have been used to facilitate sharing of Indigenous and local knowledge of marine issues in South Africa (Reference Strand, Rivers, Baasch and SnowStrand et al., 2022). With photovoice, participants were asked to take photographs using smartphones and then share photographs to engage discussions and dialogue between diverse stakeholders to help inform ocean governance processes. Leaders who can incorporate arts-based platforms into governance processes help others access and share alternative ways of understanding complex and dynamic systems and inspire people toward more creative ways to adapt to these changing conditions (Reference AdlerAdler, 2006).

3.3.2 Facilitate

Learning leaders are skilled at facilitating and managing group processes in environmental governance (Reference Imperial, Ospina, Johnston, O’Leary, Thomsen, Williams and JohnsonImperial et al., 2016). Facilitation requires a wide range of techniques, such as asking questions, promoting inclusion, constructive discussions, practicing empathetic listening, recognizing power imbalances, managing tensions, and building better mutual understanding of interests and values (Reference Fritz and FritzFritz, 2021). Fundamentally, facilitation is “the art of surfacing, stimulating, and honing the creative tension within a group to help it move where it wants to go, needs to go, and how it wants to go” (Reference FierroFierro, 2016: 33).

Facilitation skills matter for navigating through conflicts and communication challenges that inevitably arise when diverse people come together to solve environmental issues. Facilitation speeds up innovation and learning (Reference Lubell, Blomquist and BeutlerLubell et al., 2020). In various environmental governance processes, the presence of people with strong coordination and facilitation skills is often associated with better environmental outcomes (Reference Olvera‐Garcia and NeilOlvera‐Garcia and Neil, 2020).

In their study of moose management in Sweden, researchers highlighted the convening and facilitation role of leaders in the governance process (Reference Dressel, Sjölander-Lindqvist, Johansson, Ericsson and SandströmDressel et al., 2021). Group members noted how facilitation assisted in providing a more balanced view from the various actors, including hunters and landowners, and helped to collect and share information key to learning. Participants spoke of the importance of functional meeting agendas to guide their work and contribute to more focused and meaningful discussions. Delegation of tasks helped harness the different strengths and expertise of the group members, cultivating an environment of inclusiveness. Participants acknowledged the role of leaders in championing the process and valued their flexibility and creativity in communication strategies and management tools.

3.3.3 Broker

Learning leaders are not always senior administrators or heads of organizations. Often, they are the individuals and organizations that make key connections in an environmental governance process – such as brokers or boundary spanners. They also include people who can formally facilitate the exchange of ideas in contentious contexts. To promote learning in environmental governance processes, such leaders often manage divergent stakeholder interests and navigate power differentials (Reference Gerlak, Heikkila and NewigGerlak et al., 2020).

Leaders with brokering skills can support learning by facilitating communication across different groups and building shared meaning (Reference CarlileCarlile, 2004). Brokers can aid decision-makers in acquiring and considering information that they would otherwise not obtain or incorporate in their decision-making (Reference MichaelsMichaels, 2009). Similarly, boundary-spanners can draw critical connections between disconnected groups in environmental governance. For instance, research on aquaculture partnerships found that when government actors play a leading role in the partnership, individuals in the partnership are more likely to learn (Reference Resh, Siddiki and McConnellResh et al., 2014).

Arizona State University’s Decision Center for a Desert City (DCDC), which has brought scientists and policymakers together to collectively learn about and address specific issues for water governance, engaged in a variety of leadership practices (Reference Crona and ParkerCrona and Parker, 2012). Their approach to reconciling competing demands required strategically timing engagement among different stakeholder groups within modeling workshops. When leaders felt a growing distrust in the decision model and the data on which it was based, DCDC leaders worked quickly to engage the workshop participants in a new round of discussion around the model and offered to customize and down-scale it to a level where it could be more useful to practitioners. A key player that they brought in – with credentials in both academia and in the water policy community – helped broker between different social spheres around the decision model. Another senior policy actor who served as a bridge between the water policy and the urban development community, provided critical linkages for water management in the region.

3.3.4 Avoid Over-Reliance on Traditional Leaders

Over-reliance on traditional leaders – or those established by their hierarchical position or power – can create barriers to learning (Reference Pahl-Wostl, Becker, Knieper and SendzimirPahl-Wostl et al., 2013). Leaders are not a panacea (Reference Imperial, Ospina, Johnston, O’Leary, Thomsen, Williams and JohnsonImperial et al., 2016), nor does leadership reside exclusively in one individual. By recognizing groups as sources of leadership (Reference Uhl-Bien and OspinaUhl-Bien and Ospina, 2012) and using distributed leadership at various times in governance processes (Reference Fitzsimons, James and DenyerFitzsimons et al., 2011), an arsenal of strategies, skills, and approaches can be leveraged to promote learning (Reference Imperial, Ospina, Johnston, O’Leary, Thomsen, Williams and JohnsonImperial et al., 2016: 128–129). At the same time, the types of leadership needed for learning will depend on context-specific demands or skillsets of involved participants (Reference Dressel, Sjölander-Lindqvist, Johansson, Ericsson and SandströmDressel et al., 2021).

Community members can also serve key roles in leading learning processes alongside professionals (Reference Van de Kerkhof and WieczorekVan de Kerkhof and Wieczorek, 2005). Researchers in Australia studied how community members participated in collaborative water planning processes in a water-scarce town in Victoria, Australia (Reference Lindsay, Rogers, Church, Gunn, Hammer, Dean and FieldingLindsay et al., 2019). Through an iterative process of workshops and focus groups over an eight-month period, project leaders brought together industry, government partners, and community members to develop a 50-year vision for the town. They found that community champions were able to challenge groupthink and question water professionals’ assumptions and preconceived solutions.

4.1.1 Structure Venues for Learning

Formalizing learning means ensuring that venues for information sharing, knowledge building, and information dissemination are available. Learning-oriented venues can be found in policy labs that engage in experimentation and analysis (Reference Lee and MaLee and Ma, 2020), workshops aimed at building knowledge around uncertain or complex issues, collaborative working groups and teams, or task forces that bring together diverse groups or knowledge sources. For example, the South Florida Restoration Task Force, which oversees the Florida Everglades restoration planning and coordination that we introduced in Section 1 was organized to bring multiple federal and state agencies together with tribes. The goal was to learn about the complexities and challenges associated with the implementation of a multibillion-dollar restoration effort (Reference Gerlak and HeikkilaGerlak and Heikkila, 2011). In other ecosystem restoration programs, collaborative working groups, such as California’s Sacramento-San Joaquin Delta Interagency Ecological Program, similarly are designed as venues for intentional learning through collaborative monitoring and data synthesis by multiple agencies, NGOs, and scientists (Interagency Ecological Program, 2023).

Institutionalizing rules for diverse stakeholder participation or nontraditional forms of knowledge, especially in forums that traditionally are seen as “expert based” (Reference Esguerra and Van der HelEsguerra and van der Hel, 2021) can further support learning. Of course, formalizing or institutionalizing such participation directly influences the social characteristics of governance, described in Section 3, that are critical for learning. That formalization is not just about who has a seat at the table. Rules that mitigate power imbalances in information exchange, for instance, can be helpful for learning when diverse interests come together (Reference Riche, Aubin and MoysonRiche et al., 2020). Balancing formal rules with informal norms can further support the types of creative approaches we expect from learning leaders (Reference Riche, Aubin and MoysonRiche et al., 2020).

A project known as “Enhancing Livelihoods and Food Security from Agroforestry and Community Forestry in Nepal” (EnLiFT) offered a policy lab-like venue for integrating knowledge into forestry governance (Reference Ojha, Regmi, Shrestha, Paudel, Amatya, Zwi, Nuberg, Cedamon and BanjadeOjha et al., 2020). The project was established by the Nepalese government, Australian universities, and nongovernmental researchers in Nepal. The policy lab created a space to host events that focused on specific policy-related topics (e.g., clarifying the government-community contract for community forestry, scientific forest management, regulatory hurdles, land-use, and food security, among others). Each event brought together relevant participants knowledgeable about the issues – such as researchers, government agencies, various forestry-sector participants, donors, and politicians. Their learning outcomes included increased recognition of policy-related problems, identification of shared interests and concerns around private forestry, enhanced capacity for listening among conflicting policy actors, and policy change around timber harvesting. Researchers studying the project argue: “There is merit in going beyond the conventional forms of research dissemination to structure an iterative framework with which actors of science and policy can interact” (Reference Ojha, Regmi, Shrestha, Paudel, Amatya, Zwi, Nuberg, Cedamon and BanjadeOjha et al., 2020: 101997).

4.1.2 Establish Feedback Loops

As the previous example acknowledges, formalizing learning also requires learning feedback loops, which are common in adaptive management and planning. Adaptive management has been embraced in numerous environmental governance contexts, from the Glen Canyon Adaptive Management Program (Reference Gerlak, Karambelkar and FergusonGerlak et al., 2021c) to governance of the Florida Everglades (Reference Gunderson and LightGunderson and Light, 2006) to the implementation of strategies for addressing water quality challenges under the European Union’s Water Framework Directive (Reference Kochskämper, Koontz and NewigKochskämper et al., 2021).

Adaptive management can be formalized through policy – as required under California’s Delta Plan – to foster learning processes of planning, implementing, monitoring, and assessing management strategies (Delta Stewardship Council, 2023a). This approach is codified in state legislation (Delta Reform Act, 2009), which establishes adaptive management as “a framework and flexible decision-making process for ongoing knowledge acquisition, monitoring, and evaluation leading to continuous improvements in management planning and implementation” (California Water Code section 85052). Of course, the promise of adaptive management is not always realized in California’s governance of water resources and ecosystems within the Delta (Delta Independent Science Board, 2016). Still, under this more formalized approach, the state has institutionalized an ongoing review of adaptive management (with advice from an Independent Science Board) and organizes regular adaptive management forums to support shared learning across stakeholders involved in or affected by adaptive management in the Delta (Delta Stewardship Council, 2023b).

At a more local level, projects for improving water quality in northern Germany have demonstrated that formalizing adaptive management can result in different types of learning – including both reducing ecological uncertainty and reducing social uncertainty about how stakeholders will respond to interventions (Reference Kochskämper, Koontz and NewigKochskämper et al., 2021). More importantly, these adaptive management cases in Germany demonstrated that the “dual-feedback cycle” between ecological and social learning can be institutionalized. Doing so enables deeper forms of learning and facilitates the spread of lessons beyond a project to the broader governance system. Creating learning feedback loops across society and governance institutions can further support transitions toward more sustainable behaviors (Reference Newell, Daley and TwenaNewell et al., 2022).

4.1.3 Enable Knowledge Co-production

Formalizing knowledge co-production that is inclusive of multiple types of actors who have direct experience with, or knowledge of an environmental governance issue can also support learning (Reference WybornWyborn, 2015; Reference Frantzeskaki and KabischFrantzeskaki & Kabisch, 2016; Reference Zurba, Petriello, Madge, McCarney, Bishop, McBeth, Denniston, Bodwitch and BaileyZurba et al., 2022). Although knowledge co-production can occur in informal ways, more formalized approaches, such as through co-management structures or formal agreements are often key (Reference Dale and ArmitageDale and Armitage, 2011). When historically marginalized governance actors are engaged in knowledge co-production, however, it can be important for such groups to have the agency to define their roles in the process and objectives (Reference Turnhout, Metze, Wyborn, Klenk and LouderTurnhout et al., 2020).

In the Canadian Arctic, Indigenous groups, such as the Innuit and Innuvialuit, alongside different Canadian land and fisheries bodies, have established formal agreements to jointly manage fisheries and other resources (Reference Armitage, Berkes, Dale, Kocho-Schellenberg and PattonArmitage et al., 2011). One such case involved management of beluga whales that occasionally became trapped in ice. Initially, in the late 1980s, the Innuvialuit were treated as implementing parties and government officials were the decision-makers when dealing with this issue. This structure created barriers to knowledge sharing and learning. By the early 2000s stronger communication processes were institutionalized for understanding beluga entrapment problems. They encouraged Indigenous communities to share knowledge alongside government actors and build shared strategies for managing entrapment. The approach created long-term learning and enhanced capacity building for future beluga management.

4.1.4 Minimize Rigidity in Formal Structures

Formalizing learning can produce certain types of rigidity or blinders. For instance, creating formal feedback loops runs the risk of applying past lessons to future contexts that may not align with past experiences. This is called the overgeneralization trap (Reference Beach and SmeetsBeach and Smeets, 2022), or learning the wrong lessons.

At other times, we may build formalized processes that appear on the surface as mechanisms for learning, but in practice they are tools for reinforcing existing decisions or path dependencies. Consider environmental impact assessments and reviews as an example. These are tools that government agencies around the globe use to comply with national and subnational environmental impact laws when actions are proposed that could negatively harm the environment – like building new infrastructure. In theory these processes conduct extensive research and assessments on the potential risks, and sometimes benefits, of actions to issues like water quality, species habitat, air quality, and recreation access. These processes build in formal requirements for affected communities and other interest groups to comment on the draft impact statements to ensure that relevant impacts were assessed, and the analytical approach was reasonable. Yet, such tools at their worst often are rubber stamps for proposed actions. At their best, they can lead to changes in the design of actions and sometimes stop actions – which can be indicative of learning. But without strong stakeholder engagement, transparency, and dialogue around the process – such learning is less likely (Reference Nita, Fineran and RozylowiczNita et al., 2022).

At the same time, informal exchanges offer the potential for more personal and meaningful engagement that fosters learning (Reference Sattler and SchroterSattler and Schroter, 2022). In other words, the formal structures need to align with the principles that we described in Section 3. As Reference Riche, Aubin and MoysonRiche and colleagues (2020: 12) argue, learning is more effective in governance processes when informal norms allow for creativity, but also when formal rules allow for adequate information exchange. Formalizing learning may be necessary to overcome some of the risks of relying too much on processes that could be subject to the whims of the next leader, or the personalities of people involved. But building in too much rigidity into the process through formal procedures can hinder opportunities for social learning and genuine engagement (Reference Pahl-Wostl, Craps, Dewulf, Mostert, Tabara and TaillieuPahl-Wostl et al., 2007a).

4.2.1 Invest in Novel Sources of Information

Often one of the barriers to learning stems from our tendency toward confirmation biases, status quo biases, and desirability biases in decision-making, as we discussed in Section 1. However, governance can be structured to help overcome these biases and think creatively. For instance, learning can be enhanced in governance structures that facilitate the active pursuit of novel sources of information (Reference Bodin, Crona and ErnstsonBodin et al., 2006; Reference BodinBodin, 2017). These tools, such as scenario planning, can mitigate cognitive biases and build openness to innovative ideas (Reference Haasnoot, van ‘t Klooster and Van AlphenHaasnoot et al., 2018; Reference Ward, Forrow, Kirk, Worthington, Paling, Stacey and BruntWard et al., 2024). They also can include methods that explicitly assess the disparity between what is known and what needs to be known for a responsible decision (Reference Ben-Haim, Marchau, Walker, Bloemen and PopperBen-Haim, 2019).

The Global Governance Project studied learning in eight international organizations with a focus on international environmental policy over more than a 10-year period. It focused on the role of learning mechanisms, such as independent evaluation reports, expert workshops, and project reviews. Researchers found these mechanisms helped acquire new knowledge and implement solutions to organizational problems (Reference SiebenhünerSiebenhüner, 2008). When international organizations, like the World Bank and UNEP, scanned both the internal and external environment, they were better prepared to trigger more reflexive learning than those that closed off their organization against such knowledge. Dedicating resources to novel information also enabled foundational social features, which we described in the previous section that can support learning – namely diverse forms of engagement between stakeholders, staff members, and external experts. And their learning resulted in rethinking the environmental impacts of the Bank’s operations and supported modifications to lending policies and the internal culture of the World Bank.

4.2.2 Borrow from Your Neighbors

Another way that resources can be allocated to support learning, especially in resource constrained environments, is by borrowing knowledge or lessons from similar governance venues or processes. When we first studied learning in the Everglades, we heard that leaders of the restoration effort were hungry to learn how other types of large-scale ecosystem restoration efforts across the US were grappling with similar challenges. Learning about governance strategies in similar contexts can be supported by strong networks, as we discuss in more detail below. Learning from neighbors may require dedicated staff or technology that can help track and identify which types of governance actors are engaged in or facing similar challenges, while also putting resources into creating the space and time for conversations and lesson sharing.

For instance, in a study of hurricane recovery in Gulf of Mexico communities, task teams with outside experts with recent memory of major storms were important for planning, securing funding, and implementing recovery (Reference DunningDunning, 2020). These experts helped fill knowledge gaps and enabled successful planning and competitive applications for federal aid. Similarly, in their study of noise pollution at airports, researchers in the Netherlands highlight the notion of “transnational communication” where countries learn from the environmental policies of other countries (Reference Veenman and LiefferinkVeenman and Liefferink, 2014: 156). Countries learned through informal mechanisms like phone calls but also from other countries working to mitigate noise pollution through more formalized meetings. In Germany, different federal states learned from each other about how best to design participatory flood risk planning (Reference Newig, Kochskamper E. and JagerNewig et al., 2016). Lessons sourced from other jurisdictions or policy fields can spur such learning by providing inspiration.

4.2.3 Dedicate Capacity, at the Right Time and Place

Fostering and maintaining learning resources also requires dedicated capacity to strategically deploy those resources. Examples of this capacity include experts and technical resources for information management and analysis (Reference LagadecLagadec, 1997), legal and policy guidance, and resources for convening and sustaining the type of dialogue we identified in Section 3. Such resources are particularly critical if decision-makers or leaders of governance venues and programs are serious about connecting with actors who bring Indigenous knowledge and TEK or environmental justice perspectives. As we know from experts who study public participation processes (e.g., Reference FungFung, 2015), simply expecting that nongovernmental actors will show up and have time to engage in dialogue can be naïve. And worse, such expectations can reinforce past inequities if it is costly for historically marginalized actors to engage. Participating in governance takes time and compensating people who engage outside of their regular professional roles can be critical to ensuring that their capacity exists to be part of learning processes. This capacity can be structured within the strategic goals of governance actors and in their approaches to crafting, implementing, monitoring, and enforcing policies and programs.

Research on environmental disasters and crises has reinforced the importance of dedicating key resources in support of learning. Reference Crow and AlbrightCrow and Albright’s (2019) study of communities that were severely impacted by massive flooding in Colorado’s Front Range found that deeper forms and more deliberative forms of learning in recovery efforts appeared in communities that had both strong resource flows – from financial, technical, and human resources – and strong collaborative relationships to higher levels of government. Although learning may be constrained when financial resources are limited (Reference Taylor, Jeschke and ZarbTaylor et al., 2023), there are ways to allocate nonfinancial resources for those critical times when learning capacity needs to ramp up (Reference Ricco and SchultzRicco and Schultz, 2019).

4.2.4 Sidestep Resource Constraints

There is no question that resources for learning will be constrained in every environmental governance context. As a result, leaders of governance organizations or programs might be more interested in investing in resources targeted for mission-critical activities rather than toward those that support learning. Regardless of the perception or the reality of resource constraints, targeted action can be directed to support learning. Recognizing how learning is key to supporting mission-critical activities or outcomes could be the first step. Utilizing tools for learning that might be less resource intensive, borrowing resources from partners in networks (as we discuss in more detail below), and targeting resources towards key actors or actions that are known to foster learning can further help overcome resource constraints (Reference Gerlak, Heikkila and NewigGerlak et al., 2020: 8).

In line with experience from the COVID-19 pandemic, using lower-cost structures such as online platforms, to support the type of dialogue described in the last section is one approach to mitigate resource constraints. Reference Andrade, van Riper, Goodson, Johnson, Stewart, López-Rodríguez, Cebrián-Piqueras, Horcea-Milcu, Lo and RaymondAndrade et al. (2023) facilitated a series of online deliberative engagements for local stakeholders from Interior Alaska to discuss landscape change and associated management practices. Using the online forums helped identify shared values and spark value shifts across stakeholder groups. Despite some potential limitations with online forums, their use lowered participation costs for people who might have to travel long-distances and allowed for more diverse viewpoints to be brought into environmental decision-making. Virtual spaces, however, may not be appropriate for more reflexive learning, which requires thoughtful structuring of communication norms. Other approaches might also help overcome resource constraints. For instance, we might be able to use AI to compile lessons on how other decision makers are enabling learning. This could include looking for publicly available information on forums that incorporate diverse stakeholder dialogue, or searching for examples of creative stakeholder engagement, or finding insights on leadership training programs that incorporate learning skills.

4.3.1 Create Multilevel Scaffolding

Governance networks need to be well-connected across multiple levels of actors and venues to support learning (Reference Crona and ParkerCrona and Parker, 2012; Reference Resh, Siddiki and McConnellResh et al., 2014; Reference Gerlak, Heikkila and NewigGerlak et al., 2020: 7). Broad reaching networks that cross diverse types of actors and positions can facilitate information sharing and knowledge access and production (Reference Newig, Günther and Pahl-WostlNewig et al., 2010; Reference Heikkila and GerlakHeikkila and Gerlak, 2013: 496). Learning can also be fostered when multilevel networks align around shared goals (Reference Armitage, Berkes, Dale, Kocho-Schellenberg and PattonArmitage et al., 2011) and establish communication pathways across levels, from individuals to organizations (Reference Crossan, Lane and WhiteCrossan et al., 1999). Such connections further need to ensure that actors have adequate access to novel sources of information and knowledge (Reference Bodin, Crona and ErnstsonBodin et al., 2006; Reference BodinBodin, 2017). Such networks cannot be developed haphazardly. Tailoring them to the functional purpose of a governance system can be important for learning (Reference Rijke, Brown, Zevenbergen, Ashley, Farrelly, Morison and van HerkRijke et al., 2012).

One example of how multilevel networks facilitate collective learning is the Communities Committee of the Seventh American Forest Congress, where forestry leaders and stakeholders come together to discuss US forest policy issues (Reference Cheng, Danks and AllredCheng et al., 2011). The Communities Committee started in 1996, although the Forest Congress has been in place since 1882. The Communities Committee is an open membership group focused on fostering forest stewardship and has worked with other nongovernmental forestry groups to inform federal forestry policy. They also help community forestry practitioners learn about various tactics and strategies for working with the federal government and identify opportunities for community-based forestry, while providing learning opportunities for federal elected officials about community forestry.

4.3.2 Include Boundary Spanning Partners

Many boundaries are often at play in environmental governance including land ownership, functional management tasks, organizational missions and cultures, and even different individual and organizational conceptions of and knowledge of environmental matters (Reference Davis, Huber-Stearns, Cheng and JacobsonDavis et al., 2021: 4). Spanning these boundaries is important in fostering resilience and adaptive governance within social-ecological systems (Reference Quick and FeldmanQuick and Feldman, 2014; Reference Newell, Daley and TwenaNewell et al., 2022). Particularly in governance settings with a wide variety of fragmented actors and venues, boundary organizations can create strategic bridges between actors who do not have opportunities to interact or who work in vastly different domains (Reference Sattler and SchroterSattler and Schroter, 2022). These bridges foster learning when new relationships and shared knowledge are built (Reference O’Mahony and BechkyO’Mahony and Bechky, 2008; Reference FischerFischer, 2015). Boundary organizations also can drive knowledge co-production and integration of diverse knowledge into policymaking and practice (Reference GustonGuston, 2001; Reference Crona and ParkerCrona and Parker, 2012; Reference Feldman and IngramFeldman and Ingram, 2009). To do this, boundary organizations may need to create and utilize boundary objects, which take up concrete or actionable forms like maps, models, scenarios, or reports, agreements, and shared data (Reference Davis, Huber-Stearns, Cheng and JacobsonDavis et al., 2021: 8).

In the early 2000s, a Field School in the Waorani and Achuar communities in the Ecuadorian Amazon was created to provide employment and to capitalize on opportunities to generate revenue from the educational tourism industry (Reference Buzinde, Manuel-Navarrete and SwansonBuzinde et al., 2020). The Field School serves as a boundary organization comprised of members representing academic institutions and members of Indigenous communities. The role of this boundary organization is to work cross-culturally to build diverse networks that are based on trust and respect and can enhance adaptive capacity through knowledge exchange and learning. Through the program, locals teach the local language and bio culture to Western students who pay to get immersed in local knowledge systems related to sustainability. Boundary objects are used to acquaint non-Indigenous people with local Indigenous perspectives and include videos that capture key aspects of Indigenous knowledge. Games, as discussed in Section 3, can also serve as boundary objects helping to facilitate shared learning and collaboration between stakeholders.

4.3.3 Understand the Tradeoffs of Different Network Structures

Networks vary in terms of their structure and actor composition, which impose different strengths and weaknesses for learning (Reference Rijke, Brown, Zevenbergen, Ashley, Farrelly, Morison and van HerkRijke, 2012). Especially in complex governance settings, network structures that limit interactions between different subgroups can impede learning (Reference BodinBodin, 2017). In these situations, learning may be restricted to the concerns and ideas of those who are most powerful (Reference Schusler, Decker and PfefferSchusler et al., 2003; Reference Roome and WijenRoome and Wijen, 2006). Although more connectivity or density across a network may allow more opportunities to acquire information and learn quickly, excessively dense networks can also produce more homogenous information and impede learning (Reference Rijke, Brown, Zevenbergen, Ashley, Farrelly, Morison and van HerkRijke et al., 2012).

Similarly, networks with highly centralized actors can enhance information flows, but they may also constrain diversity of input and deeper levels of learning (Reference Bodin, Crona and ErnstsonBodin et al., 2006). In the case of the Venice canal system, a highly hierarchical and centralized governance system that lacked connectivity resulted in closed coalitions, and limited communication and knowledge sharing, which made policy change difficult (Reference Munaretto and HuitemaMunaretto and Huitema, 2012). Conversely, networks with less centrality may facilitate broader engagement and shared learning (Reference HenryHenry, 2009), but may increase the transaction costs of learning processes (Reference Gerlak and HeikkilaGerlak and Heikkila, 2011).

Furthermore, there is a tendency for people to form networks with people they share beliefs with, or bonding ties, and the boundaries of such a network create constraints on learning (Reference Henry and VollanHenry and Vollan, 2014). While people may be more likely to trust information coming from people who share similar beliefs, homogeneity of actors in networks may constrain the diversity of ideas and impede learning (Reference Newig, Günther and Pahl-WostlNewig et al., 2010). This is because more homogeneous governance structures – that is, intra-sectoral networks – may not allow for the type of “bridging” ties that are needed for knowledge acquisition (Reference Dow, Haywood, Kettle and LackstromDow et al., 2013). When people build governance networks around shared beliefs, learning may be focused on reinforcing existing beliefs (Reference Weible, Olofsson and HeikkilaWeible et al., 2023). A “dynamic interplay” of bridging and bonding links, which serve to build more internal trust and norms, may be needed to promote learning to support environmental governance (Reference Newman and DaleNewman and Dale, 2005).

5.2.1 Ask Learning Questions

Researchers and practitioners can start by using the social and structural features and mechanisms from Sections 3 and 4 to diagnose where learning in governance settings can be enhanced. Through a diagnostic approach, we can better assess how to adapt the social and structural elements of governance needed for learning, particularly as organizational and political contexts change. This involves introducing the concept of learning and then examining the processes used in a specific governance venue, decision-making process, or organization. You can begin by asking if any of the social or structural foundational features of learning exist.

• Which of the social and structural features are present;

• What is missing; and

• How are foundational features interacting?

Furthermore, guiding questions outlined in Section 2 can be posed about who learns, how learning occurs, where people learn, what they learn, and to which effect. Doing so can make learning processes and products more visible (e.g., as highlighted in Figure 6). From there, it is easier to draw connections between the indicators of learning and the social and structural features of governance we would expect to support learning (also shown in Figure 6).

In diagnosing how learning occurs, consider whether different collective learning stages are activated (i.e., acquisition, translation, and dissemination of information). As we described in Section 2, even if individuals are learning on their own, learning for environmental governance cannot be triggered without a collective learning process. That means governance actors intentionally seek out new ideas, interpret, and disseminate that information across groups, organizations, and networks. Asking whether and how learning processes are enabled will also point out where, or which governance processes, might need investments in the foundational features described above. In some cases, learning processes can be enhanced by a stronger social foundation, like engaging diverse actors and interests, fostering genuine dialogue, and nurturing skilled learning leaders. In other cases, deeper structural change might be needed to spark learning. Key questions to ask might be:

• Is a new collaborative body needed to bring diverse actors together;

• Where could resources be directed to build greater capacity; and

• What kind of networks are desirable to achieve the types of learning outcomes that will help advance governance goals?

5.2.2 Set Learning Triggers

A second pathway to infuse learning into environmental governance is to establish triggers for when and how we diagnose learning. Paying attention to notable moments of learning or non-learning from past actions can help diagnose what governance changes could support learning in the future. This might mean looking at timelines of planned programs or policy implementation efforts and devising a “learning summit” at key points in the process. Through dialogue and reflection, actors in a particular governance setting can then undertake a formal diagnosis of past actions. Again, using the tools from Section 2 to analyze the who, how, what, to which effect, and where of learning is key to the diagnosis. Such assessments can allow actors to identify where learning blockages might be occurring (e.g., who is holding up the flow of information and why), or where the social foundation may be incomplete or lacking. Conversely, it can reveal the presence of mechanisms, such as learning leaders or networks, which supported learning or contributed to blocked learning or non-learning in past actions.

We can also look to examples where environmental policies or programs have failed to meet key goals and use such examples to trigger learning processes. Failures are often instructive because they focus attention on problems that can build new knowledge about what to do differently in the future (Reference McConnellMcConnell, 2015; Reference Sokołowski and HeffronSokolowski and Heffron, 2022). However, due to our negativity bias, we can over-learn or focus too intently on our failures as triggers for learning, potentially ignoring important lessons from cases of success (Reference Beach and SmeetsBeach and Smeets, 2022). Considering policy successes – or at least cases that are not deemed as failures – can further infuse learning into governance. For example, an appreciative inquiry approach encourages organizations to identify what is working well in their organization and envision how to make positive outcomes occur more frequently (Reference Fam, Mitchell, Abeysuriyaa and MeekFam et al., 2013). With this approach, problems are reframed to focus on strengths and successes by asking participants to identify what works well, why it works well, and what they want to see more of in their organization (Reference Coghlan, Preskill and Tzavaras-CatsambasCoghlan et al., 2003). Another approach is a premortem, which occurs before a project or effort is launched, and through a hypothetical asks the participants to explain the reasons why the effort might fail to anticipate and head off future problems (Reference KleinKlein, 2007: Reference SmerekSmerek, 2018).

Like the tools of program evaluation and policy analysis, being able to assess different governance outcomes, and understand the drivers of those outcomes creates a formalized process for learning. It embeds regularized acquisition of information into governance, ensuring that processes are in place to formally assess and interpret that information. Whether the lessons from that process result in changes into the design of future policy or governance depends on how well the social and structural foundations of learning are established.