Connecting the multi-level-perspective and social practice approach for sustainable transitions

Multi-level perspective
Social practice approach
Sustainability transition
Intervention design
Social change

In this section, we give a brief introduction to the main concepts and ideas of both the Multi-Level Perspective (MLP) and Social Practice Approaches (SPA) and then focus on their shared tenets and main differences. A comparison of the two approaches is shown in Table 1, which includes references that are intended to be illustrative rather than comprehensive. This exercise aims to uncover the shared foundations of the MLP and SPA to demonstrate the possibility and fruitfulness of further conceptual dialogue between the two. At the same time, it flags some differences that need to be borne in mind when conducting analysis or designing interventions that co-apply the MLP and SPA.

The MLP is a heuristic tool or middle-range theory for conceptualizing dynamic patterns that occur in socio-technical transitions. It combines key ideas from evolutionary economics, institutional theory, and various strands of science and technology studies (Grin et al., 2010; Köhler et al., 2019; see Table 1).

Transition processes lead to fundamental shifts in socio-technical systems (Geels, 2002; Geels and Schot, 2007) that consist of different elements – such as markets, infrastructures, technologies, regulations and user practices – that interact and fulfil societal needs like energy, housing, food, transportation or education. The MLP argues that a transition evolves non-linearly and results from the interplay of endogenous and exogenous dynamics at three levels (Geels, 2002; Geels and Schot 2007). The meso level consists of socio-technical regimes, which are coherent, interrelated and stabilized configurations of socio-technical systems that lead to lock-in mechanisms and path-dependencies. The micro level consists of niches, which are protective spaces where radical innovations can be nurtured to potentially challenge regimes. The macro level is the exogenous socio-technical landscape, which consists of external developments such as shocks (e.g., nuclear accidents), international politics, or incremental trends (e.g., climate change). These developments can exert pressure on a regime, potentially creating windows of opportunity for radical innovations to emerge that may eventually lead to change. MLP has been applied to a wide range of different transitions ranging from historical cases, such as the transition from horse-drawn carriages to automobiles (Geels, 2005), to contemporary and possible future transitions, such as shifting from industrial to organic agriculture (Beitz, 2004).

Social practice approaches (SPA) reflect a range of interpretations of social practice theory, which build upon earlier work in the social sciences (based on the work of authors like A. Giddens, P. Bourdieu, M. Foucault, and H. Garfinkel) and have roots in select philosophical traditions (e.g., M. Heidegger and L. Wittgenstein; see Table 1). Specifically, in defining the locus of the social, SPA seek to overcome the dichotomy between agency that operates at a structural level (for example, through culture) and agency at the level of individuals. Theorists like Schatzki (1996) and Reckwitz (2002) have proposed a contemporary conceptualisation of social practice theory, which has been widely used in sustainability studies and consumption studies in particular.

A core aspect of this understanding of social practice is a shift in focus from individuals and systems to practices as patterns of activity. Building on these insights while recognizing the heterogeneity of existing definitions, Welch and Warde (2015, p. 85) have suggested a minimal definition of social practice as “an organized, and recognizable, socially shared bundle of activities that involves the integration of a complex array of components: material, embodied, ideational and affective. Practices are sets of ‘doings and sayings’; they involve both ‘practical activity and its representations’ (Warde 2005, p. 134).” Examples of social practices are driving, cooking, cycling, showering, and decision-making. While many studies have sought to better understand a single practice or multiple inter-locked practices, some have also focused on change over time, such as explaining the emergence of Nordic walking in Europe (Shove and Pantzar, 2005) or the shift from bicycles to motorbikes in the streets of Hanoi (Hansen, 2017).

Thus, social practices are collective patterns of activity that are recognizable and reproduced over time and space, yet they constantly evolve as ‘practitioners’ enact practices in varying ways. The notion of social practice approaches (SPA) recognizes the diverse interpretations of this theoretical framework in which practices are held together by different elements, which are defined as ‘meanings, competencies, and materials’ in one interpretation (Shove and Pantzar, 2005) or ‘understandings, engagement, and procedures’ in another (Warde, 2005). For an earlier overview of different authors’ classifications of practice elements, see Gram-Hanssen, (2011).

In Table 1, we provide a comparison of the MLP and SPA and highlight the theoretical roots of the approaches discussed above.

They are presented along with the similarities and differences in their understandings of agency and social change, among other factors, as discussed further below.

The MLP and SPA share some core theoretical foundations, especially Giddens' (1984) structuration theory. Accordingly, the most evident similarities are in an understanding of agency as embedded in broader structures (what Giddens calls "duality") or elements, as well as seeing these elements (and their manifestation in systems and practices) as both constraining and enabling human action. Both theories also share a socio-material view in which both subjects (humans and various social groups) and objects (artifacts, technologies, infrastructures) intermesh, with neither taking supremacy.

The central constituting concepts of the MLP and SPA converge in important ways. Both imply a combination of elements that come together to form a practice or a system in which the elements interact with each other while also interacting with other practices or systems. Understanding how these dynamics function through the alignment or reconfiguration of elements underlies social change processes, regardless of whether they take place through system- or practice-based reproduction or transformation. For Røpke (2015), practices share similar properties with systems, but practices are "just much smaller entities" (p. 353). In this reading, socio-technical systems can be seen as made up of practices and complexes of practices (see e.g., Moore et al. 2018). Also, within this interpretation, specific practices can be related to each particular socio-technical system component, such as epistemic practices, practices inducing consumption, or governance practices (see Hölsngens et al. 2018, p. 3). For Watson (2017), practices and systems exist in a relationship of mutual co-constitution, and "practices are substantially constituted by the socio-technical systems of which they are a part, and that those socio-technical systems are constituted and sustained by the continued performance of the practices which comprise them" (p. 350). Thus, the two approaches share similar constitution mechanisms and architectures.

The concepts of rules and routines that guide and embed social action have an important place in both the MLP and SPA. Rules can be seen as coordinating and orchestrating principles that—to varying degrees, with firmness or looseness, either tacitly or explicitly—hold practices and systems together (Grin et al., 2010; Schatzki, 2002). Moreover, both the MLP and SPA acknowledge that a large part of human action has sedimented into routines that reproduce habitual ways of doing and speaking, which may result in lock-in of systems and change-resistance of practices.

The two theories have divergent approaches to agency. Agency is attributed to different actors on the niche, regime and landscape levels in the MLP, whereas agency is distributed across elements of practices, including people as carriers of practices, in SPA. However, they both encompass a non-individualist, multi-actor approach in which agents are both restricted and enabled by the habits and routines of social action. It has been claimed that SPA risk "the death of the subject" (Sovacool and Hess, 2017, p. 713) with their strong version of agentless practices. Following this line of reasoning, it can be said that there has been too much emphasis on collective agency in the MLP, leading to neglect of individual agency (Atkinson et al., 2017; Pesch, 2015). One particular sub-strand of this critique relates to the issue of overlooking consumers and "practitioners" in the transitions literature. In earlier works, this criticism was prominent (e.g., Shove and Walker 2010). However, in recent years, transition studies have included more in-depth accounts of types of users (e.g., Kanger and Schot 2016, Schot et al. 2016).

The MLP and SPA share important assumptions in their views on social change, even though notable differences exist (cf. Erdoğan Öztekin and Gazihusuloy, 2020). Both address the tension between reproducing normality and planting the seeds of transition. Much of the literature engaging with SPA in consumption and environmental studies has focused on explaining routines and reproduction and normalization processes (for household appliances, see e.g., Aro 2017, Hand and Shove 2007, Sahakian 2019). However, recent studies have applied SPA to explore how social change processes can be co-designed or imagined in the future. These studies have envisaged strategies like interventions that would seek to change networks of non-desirable practices (Vihalemm et al., 2015) or target change points in practices that might lead to less resource-intensive ways of living (Hoolohan and Browne, 2018). Some scholars claim that SPA do not sufficiently elucidate variation in space, whilst they are much better at explaining variation in time (Sovacool and Hess, 2017, p. 713). In contrast, the stasis versus dynamism critique states that the main emphasis of the MLP tends to be on transition and change with a relative disregard for the embeddedness that results in stasis (Tyfield and Zuev, 2018). At the same time, both frameworks are cognizant of and potentially complementary in analysing how change processes result in both systemic lock-in mechanisms (for a typology of different lock-in mechanisms of existing socio-technical regimes, see Klitkou et al. 2015) and obduracy of social practices, which in their turn, impede further change. This has been convincingly shown in a study of Maastricht underground parking by Stanković et al. (2020).

The MLP and SPA share a processual orientation that homes in on struggles between new and old configurations that emerge as co-evolutions of heterogeneous elements manifesting the inseparability of society and technology (Avelino et al., 2017; Geels et al., 2015). This leads to an understanding of change as a fundamentally processual, adaptive and reflexive phenomenon in which the need for extensive social learning is acknowledged (Hargreaves et al., 2013; Hielscher et al., 2013).

Finally, both MLP and SPA have a shared substantive interest in sustainability and transitions (even though they are not their only focus). Both recognize the need for systemic change rather than incremental tinkering or individual behavioural change alone. However, a large body of SPA studies remains focused on the micro-level of domestic everyday practices. This has led to some criticism regarding its inability to keep sight of 'the big picture' of systems and structures. The alleged inadequacy of SPA to illuminate systemic forces or developments (Huber, 2017, p. 65; Sovacool and Hess, 2017, p. 713) or 'large-scale'/'supra-practice' social phenomena has been recently acknowledged and discussed in the practice theory community (see Nicolini 2016, Welch 2020). This includes Schatzki's (2019) own analysis of how social change occurs over wider constellations or bundles of practices that form complex nexuses of activities and processes.

The MLP and SPA tend to conceptualize change in different ways. In the MLP, socio-technical regimes might change because of landscape pressures that create opportunities for mature niche innovations. In SPA, each practice contains the seeds of change (Warde, 2005), and change can occur either through reconfiguration of elements of practices (Sahakian and Wilhite, 2014; Vihalemm et al., 2015) or through new relations between practices (Spurling et al., 2013). Thus, these different ways of conceptualising change point towards an ontological distinction between the MLP and SPA—the MLP stems from a vertical ontology with hierarchies between three levels, while SPA are based on a flat ontology with no dominance a priori of any element of a practice or between practices.

Nevertheless, some authors have argued that the MLP and SPA are less incompatible than they may initially appear, especially from a methodological and analytical perspective. This is highlighted in Geels' (2011, p. 37) response to a critique: "Although people often summarize the MLP as 'micro-meso-macro', the levels are defined as referring to different degrees of structuration of local practices, which relate to differences in scale and the number of actors that reproduce regimes (and niches). Levels thus refer to different degrees of stability, which are not necessarily hierarchical. This is a deviation from earlier MLP-work, which used the notion of 'nested hierarchy'". This updated understanding of the MLP could contribute to resolving the incompatibility issue according to Welch and Warde (2015, p. 86): "The recent stress of MLP proponents that their model primarily distinguishes degrees of institutionalization, defined as the stability and consistency of interactions and power relations over time and space, goes some way to address these concerns."

As a result, many authors have identified numerous similarities and shared notions between the two theories, which explains why both approaches are widely used side-by-side in socio-technical analysis (see e.g., El Bilali 2018, Geels et al. 2015).

To fulfil our research objective, we first sought to map the extent and types of studies in which the MLP and SPA have been co-employed. For this, a systematic literature review was carried out. The process consisted of four steps. First, we defined a fixed search string of keywords that combined social practice theory, the multi-level perspective, and sustainability transitions: "Practice theory" [OR] "theory of social practice" [OR] "social practice approach" [AND] "multi-level perspective" [OR] "multilevel perspective" [AND] "sustainability transition".¹ These keywords were used to identify articles in Scopus in February 2021. To ensure that we did not miss an important article, we also carried out a search on Science Direct and Web of Science. Second, we verified whether the articles met the eligibility criteria by verifying that the abstracts and full texts engaged with social practice theory, the multi-level perspective, and sustainability transitions and did not simply refer to them. Articles identified through a snowballing strategy were manually added. This entailed identifying relevant articles within the reference list of known articles, as well as going back to key contributions in the field to make sure important papers were included. In total, 118 peer-reviewed articles and book chapters were used for the present study.

In the third step, we coded the identified articles to yield quantifiable results for analysis. The coding was initially carried out by one of the authors, and the codebook was then transferred to a second author who checked that the criteria was systematically and consistently applied.

We identified (a) the type of each study, i.e., whether it was empirical and/or conceptual; (b) the conceptual focus of each study, i.e., how exactly the MLP and SPA were combined or treated in the article or chapter; and (c) the empirical focus of each text. During this third phase, we read 51 articles in greater depth. These texts were selected based on the criterion that they explicitly attempted to link the MLP and SPA or discussed the possibility and opportunities of doing so, rather than simply mentioning both approaches. That is, the selected texts employed the main terms and notions from both approaches systematically and explained the implications of such a co-application exercise. We developed an initial analysis based on this review. This informed our understanding of the two conceptual approaches and their methodological implications. Specifically, for each of these articles that attempted to co-articulate the MLP and SPA, the main practical implications, possible research questions that arise, and potential challenges were identified based on the empirical and conceptual findings. These implications, research questions and challenges were discussed among all the authors and involved several iterations to identify the most salient themes and insights.

Finally, as a fourth step, we used this analysis of 51 studies to record insights into trends, opportunities and suggestions for linking the MLP and SPA, as well as the heterogeneity of cases in which both these approaches were linked. Specifically, we delved into the theoretical explanations with which the authors justified their attempts to link the MLP and SPA. Based on the approach described above, we teased out the practical implications of their efforts and identified useful insights for future research. This provided us with a set of assumptions about opportunities that arise from articulating MLP and SPA for further application in studies informing social change, which we outline in Section 4.

Our results demonstrate a surge in the number of publications that link the MLP theoretical framework and social practice theory since 2010, and in particular between 2017 and 2019. Note that although only two months are included in 2021, they account for five papers, suggesting an upward trend (despite what is presented graphically in Fig. 1). Overall, the articles in our sample were published in many different journals, but six journals in particular published three or more: Environmental Innovation and Societal Transition ($N\; =\; 13$), Research Policy ($N\; =\; 9$), Journal of Cleaner Production ($N\; =\; 9$), Technological Forecasting and Social Change ($N\; =\; 8$), Technology Analysis and Strategic Management ($N\; =\; 5$), and Global Environmental Change ($N\; =\; 3$).

The identified papers varied in terms of their type, empirical focus and methods (Table 2). Excluding review papers, 33 papers were exclusively conceptual, whilst 45 were conceptual but also contained at least one empirical illustration. 40 papers were primarily empirical. 51 explicitly articulated or co-developed the MLP and SPA; these were read in more depth and used for the co-articulation and bridging efforts in Section 4, as mentioned above. The studies also differed in terms of their empirical focus (Table 2). Most (i.e., 60%) focused on three main empirical settings—grassroot movements and niche innovations, energy, and food and agriculture, with many papers combining several topics. The category 'other' contains examples like marketing initiatives, taxation systems, and institutional entrepreneurship. This suggests that a rather limited number of empirical settings have been explored in this field of research. Finally, the research methods employed in the identified studies varied greatly, and many of the papers used more than one type of methodology. In Table 2, we also present some of the main methods that were applied, which include interviews and case studies (most recurrent), but also desk research, observations, document analysis, surveys, focus groups and workshops.

These quantitative results indicate that bridging the MLP and SPA offers many different perspectives and provides specific insights into transitions phenomena and social change for sustainability in multi-scalar and multi-dimensional ways. These results also shed light on the limited, yet diverse, fields of investigation and methodologies employed in these studies. Again, this highlights the potential for linking both approaches to tackle social change for sustainability and develop understanding of transition phenomena in various contexts and fields of study. It also shows the diversity of methodologies that can be applied depending on the empirical focus and the research questions at hand.

To expand the combined use of SPA and the MLP in a more structured and consistent way, these results reiterate the need to develop a co-articulated framework to provide the means to further understand and foster social change for sustainability and identify specific intervention points. This framework could then be applied to different domains of study using various context-specific methodological approaches. The next section synthesizes the analysis of the 51 papers into seven bridge-building insights that demonstrate the value of articulating both approaches to develop a more complex and holistic picture of transition phenomena and opportunities for social change.

Given the focus of SPA on everyday life dynamics and the inclusion of regime and landscape dynamics in the MLP, it seems evident that the two approaches are most often combined in sustainability and consumption studies in order to zoom in to the level of mundane daily activities through SPA and zoom out to analyse the bigger picture with a bird's eye view through the MLP. A frequently cited MLP figure showing levels with niches, regimes and landscapes inspires zooming out and gives less attention to the micro-dynamics at various levels (see e.g., Hirt et al. 2020, van Sluisveld et al. 2020). In SPA, understanding practices and their elements (such as meanings, skills and materials in Shove et al. (2012) interpretation) provides a handy and sufficiently granular framework for investigation of unfolding everyday action that helps to break it up into analysable categories. Thus, the MLP and SPA standpoints complement each other as frameworks for investigating social change and intervention programmes, as the rich evidence from published studies clearly demonstrates.

We use the visual metaphor proposed by Nicolini (2010) in which zooming in and out with SPT and MLP, respectively, means switching theoretical lenses and “sequential selective re-positioning” (2010, p.1396). Our use of the term “zooming out” allows for a broader range of analysis, however. For Nicolini (2010), zooming out relates to the study of how “translocal phenomena” come into being and how practices connect with each other in constellations. Based on the articles we studied, zooming out may also include attention to socio-technical regimes and systems, which are criss-crossed by practices (without treating practices as confined to particular systems). For Nicolini, zooming in means using (micro)ethnographic observation to examine in detail how practices unfold as performances in real-time. We suggest that the level of detail and the exact method varies according to the requirements of a particular analytical enquiry and need not begin with a close-up of practice-making and real-time activity.

Thus, when studying change processes and co-applying MLP and SPA, one point of departure is to consciously decide on a research strategy of zooming in and out. This entails choosing a starting point, whether it is a regime-level socio-technical system with competing niches and landscapes as backdrop or a particular configuration of social practices that are embedded in and span across socio-technical systems. Laakso et al. (2021) explain the zooming in and out approach in their study. They use the notion of “reconfiguration” as a heuristic device to explain how changes can be analysed through both the MLP and SPA. They provide a way to focus on regime reconfiguration by understanding practices within regimes, and they use MLP to consider how practice elements play out at a broader scale. The latter also embraces cultural and normative factors of how various actors (including intermediaries and professionals) get recruited (or fail to do so) into novel practices. Gismondi et al. (2015) provide another example of zooming in and

Seven insights for combining MLP and SPA in research and practice: exploring possible research questions and challenges.

(continued on next page)

zooming out. In their study, they use the MLP to observe how Alberta Community and Cooperative Association intermediaries interacted with financial systems at a regime level to enable cooperative local investment niche-innovation to emerge. The authors then drew on SPA and their three components of practices (i.e., images, skills, and practices) to distinguish between mainstream and alternative borrowing and investing. Through this combination of zooming in and out, the authors identified “friction points”, pinpointing where local borrowing and investing are locked into broader practices that relate to the existing regime.

Related to the first proposal is the observation that “vertical” and “horizontal” analysis are often enabled by combining the MLP and SPA, as was successfully conducted in a number of studies on how practices and regimes intersect. This entails an equal footing between levels of the MLP and practice elements and consists of finding the intersections between practices, niches and regimes. Seyfang and Gilbert-Squires (2019) demonstrate and visualize critical points of intersection between SPA and MLP, whereby: “(a) transitions in regimes (through niche development) are obstructed by embedded practices and (b) transitions in practices are obstructed by incumbent regimes. Our goal is to understand how to transform these points of constraint into points of opportunity, thereby informing policymakers and practitioners aiming to support sustainability transitions” (p. 231). Thus, the intersection-centred approach can aim for better comprehension of factors that lock in unsustainable practices and regimes. For this purpose, it focuses on how single practices and their constitutive elements cut across niches and regimes and how the difficulties in niche diffusion and mainstreaming result from entrenched practices resisting change. The horizontal analysis of practices enables researchers to address them as pervading multiple regimes, drawing attention to the necessity to avoid focusing exclusively on a single socio-technical system. This is seen as a way of further developing the MLP.

This type of MLP and SPA co-articulation quite easily enables a policy-relevant ambition: the identification of points of intersection with critical lock-in aspects, also termed “friction points” (Gismondi et al., 2015; Hoffman and Loeber, 2016) or “sticking points” (Watson, 2013). This allows researchers to reframe them as critical intervention sites to catalyse systemic transition (see also Morrissey et al., 2014). For example, in Gazulli et al. (2019), the MLP and SPA are combined to investigate how wood fuel systems failed to become part of an energy transition in Bamako, Mali. The MLP is used as a vertical approach to analyse interventions in the regime by setting new rules, and SPA is employed horizontally for considering the role of everyday practices embedded in local life, such as cooking and transportation. The study demonstrates how transition cannot be seen only as depending on technological innovation. In another example, a market-based, competition-dominated regime may hinder collaboration and networking practices between competitors in the private sector, thus creating a negative intersection and blockage (Paschen et al., 2017). However, various niche

innovations can lead to fractures in regime-level practices, which may subsequently either conform to the mainstream or result in a major transformation if the fracture deepens (O'Neill et al., 2019).

As Gazull et al. (2019) and Kokko and Fischer (2021) vividly demonstrate, regimes can be both formal and informal. Their studies focus on the Global South, where “splintered” (van Welie et al., 2018) or informal regimes are especially prominent. Such regimes differ from standardized and predictable Western systems for ensuring energy, food or mobility (the most commonly analysed systems in our sample, see above) in their extreme flexibility and adaptability to change, lack of formal rules and standards and fragmentary nature in terms of actors, service sustainability and predictability. They may not consist of neatly organized socio-technical systems that can be broken down into the “classical” system components of end users, markets, science and technology, and cultural factors (van Welie et al., 2018). Also, it may be that some regimes have both formal and informal dimensions, as are amply shown in studies on the Soviet economy, where under-the-table commerce and the black market constituted important parts of, for example, food or retail trade regimes (Keller, 2005). As existing studies show, SPA-inspired analysis emphasises the importance of local contexts while illuminating the nuances of such regimes.

Any policy or intervention attempt at sustainability transition should consider the degree of (in)formality of regimes, and where a regime is difficult to apprehend analytically, SPA can help to uncover the different practices that constitute it and how these practices interrelate.

SPA-inspired examination provides conceptual grounding as well as tools for analysing interactions between multiple practices, such as food provisioning and riding a bike, but also those that take place between different regimes, such as the food and mobility sectors. As noted in the transition literature, studying multi-regime interaction is an emerging field that is urgently needed because transformative challenges for developing sustainability extend beyond single socio-technical systems (Rosenbloom, 2020). Examining multiple-regime interactions is a fruitful research avenue, but they have been understudied thus far (Kanger et al., 2021; Geels, 2011, 2018a; Lazarevic and Valve, 2020).

Relationships between systems can be complex, and therefore, it can be challenging to delimit clear types of either competition or symbiosis. It can also be difficult to determine whether the interactions occur between several niches or between niches and regimes, particularly when everyday social practices are taken as the analytic point of departure. With regard to policy interventions, these ‘meeting points’ for different systems (e.g., the use of electric vehicles for electricity and transport systems) present challenges (e.g., to adopt new practices) both for users and other system actors. A practice-based analytic lens can prove beneficial to facilitate this kind of examination.

Most social practices tend to span multiple regimes that their “elements” draw upon (e.g., cooking is embedded in energy, agro-food, water/sanitation, and mobility systems). Practices can also criss-cross and permeate into different socio-technical systems (e.g., practices of legislation follow the same pattern within systems of energy, mobility, food, etc.). Thus, practices are, by definition, not confined to one regime; following their pattern and single elements (meanings or materials, for example) can reveal how different regimes interact and impede or foster change. A combined MLP and SPA analysis enables us to see how multiple regime interactions can obstruct the development of a proto-practice, as shown in the case of sustainable banking and retail systems “running into each other” examined by Seyfang and Gilbert-Squires (2019).

In addition, not only a single practice and its ‘travelling’ elements, but also larger interlocking practice configurations can shed light onto how different socio-technical regimes are joined together. Therefore, from the social practice standpoint, we could think of mundane, everyday, dispersed alignments as well as points of collision between systems that either facilitate or block the emergence of new practices. Here again, the concepts of stickiness and obduracy explain the ability of social practices to interlink into tight bundles with other practices across regimes or within a single system (Laakso, 2017; Laakso et al., 2021; Svennevik et al., 2020). Interlocked practices reveal multi-regime relations and open up avenues for practical coordination (breaking and creating links between regimes) in intervention design.

As noted above, an analysis that uses both the MLP and SPA needs to carefully delimit its terrain to retain empirical feasibility as well as practical applicability in intervention design. Practices intermesh in almost endless ways. Facing this complexity by recognizing that everything is connected to everything is not helpful. One way forward may be to mark types of practices that can aid in setting boundaries and selecting research foci. Most of the research inspired by SPA in relation to consumption and sustainability focuses on everyday practices that tend to be relegated to the realm of the domestic, as several authors note (Birchmeier, 2012; Edomah et al., 2017). Studies within the MLP tradition, on the other hand, tend to concentrate on institutional forces, governance procedures and technological transformations without fully accounting for the domains of consumption and everyday life. As Watson states, earlier attempts to articulate SPA with MLP have not significantly challenged a “corraling of the value of practice theory to be limited to understanding end user practices” (Watson, 2017, p. 350). If we are to take Schatzki’s notion of “plenum of practices” to heart, then policy, regulation, industry and other sectoral domains of socio-technical systems all entail specific social practices (Schatzki, 2002). The same has recently been emphasized by the transition researchers’ agenda: “There is a need for broader frameworks that bridge

production and consumption at system, technology and product levels" (Köhler et al., 2019, p. 32). Even though some studies have emerged in recent years that look at different professional practices, institutional entrepreneurship and changing business organisations (e.g., Verkade and Höfken 2019, Weissenfeld and Hauerwaas 2018), more conceptual and empirical work is required.

Related to this is the concern raised by Welch and Yates (2018) about the lack of theorisation and empirical analysis of collective action and collective practices that mobilize transition. The authors aptly remark that sustainability transitions research has "quarantined" the contributions of practice theory to lifestyles, consumption and domestic life when there is no reason to do so, as the literature on organisations and practices amply shows. Indeed, in organisation studies, a separate research stream of "strategy-as-practice" (focusing on organisational management) exists, yet this has not been in fruitful dialogue with practice-based consumption studies or MLP-based transition studies in any meaningful fashion (see Nicolini 2013). Everyday practices are mainly constrained by habit, lifestyle, socio-cultural conventions and the economic and material conditions of households. Yet professional practices are also, to a large extent, configured by standards, documented regulations (in addition to conventions, tacit rules, etc.), and both written and unwritten codes of conduct. Co-engaging the MLP and SPA allows researchers to identify the most relevant domestic, professional and institutional practices for a given social change issue and examine the interesting junctions between such practices.

Another avenue for classifying social practices would be to analytically delimit those that are emergent and vulnerable, i.e., "proto-practices" (Seeyang and Gilbert-Squires, 2019), and those that are most entrenched, obdurate or sticky. Whether a novel emerging social practice (e.g., Nordic walking, Shove and Pantzar 2005) constitutes a niche on its own or whether we can also observe proto-practices at the regime level (e.g., solar PV among cooperatives, public authorities, and energy utilities, Hirt et al., 2021) are additional empirical questions. Also, not all emergent practices may be favourable for sustainable transitions, as is the case of certain emergent "green economy" practices that may have significant social and environmental impacts over the long term. For example, Lane et al. (2018) argue that the rapid dissemination and uptake of household IT is fundamentally unsustainable in its current form, and they examine some of the issues of current research focusing primarily on green technology development. They call for more attention to be directed towards these unsustainable practices and envisage development of new niche-level material- and energy-efficient practices.

As transition entails the emergence of new ways of doing, be it eating or banking, a zoom in to practices whose elements have not yet fully coagulated into an integrated whole can reveal important points where intervention is needed to nurture and support such practices. However, as cases of transition failure or limited success convincingly demonstrate, entrenched and sticky everyday practices that resist change can be valuable in explaining why a change effort stalls or derails, regardless of whether these practices occur in domestic or professional settings and are carried out by households or institutional actors. For example, incumbent firms with considerable power can engage in activities of retaliation and resistance against new practices associated with niches, as in the case of solar PV uptake in New Zealand. Also, sedimented domestic conventions and routines that underlie the culture of a system may prove to be resistant and capable of "bouncing back" (Ford et al., 2017). Thus, we may say that when practices become sticky bundled together or change-resistant in their element alignments, this does not happen in isolation. Instead, the particular ways that practices are embedded in systems help to explain the sources of obduracy.

Kokko and Fischer (2021) explore the case of a "radical innovation" provided by Peepoole, a Swedish social entrepreneurial venture, which launched a new service for a sanitation regime. The study concerns the use of the Peepoobag – a biodegradable and self-sanitising bag for urination and defecation – as a 'niche' service that was introduced into an informal settlement in Nairobi. The authors use its apparent failure to bring about a lasting change in sanitary practices to uncover the factors that support or hinder changes through innovation. They identify funding as the main issue, as well as the failure of a circular business model to materialise that would create value from human waste. Also, the innovation was not aligned with policy measures in place at the regime level. It seems to be a fair assumption that most practices have a degree of obduracy (Stanković et al., 2020), and a nuanced analysis of practice elements is in order to shed light onto the origins of this "stickiness" and the potentialities for unlocking opportunities for change.

Within the MLP, landscape is less theorized compared to niches and regimes. It tends to account for exogenous factors, natural phenomena, big events like wars, and international political developments, among others. Also, discourses, narratives and large-scale and long-term changes in public sentiment can be categorized within landscape. Landscape issues are understudied, especially when it comes to sustainability policy and interventions. A recent literature review (Kanger et al., 2020) on possible intervention points within niches, regimes and landscapes demonstrates how in policy-relevant literature using the MLP (using a sample of 55 articles), studies on "tilting the landscape" are virtually absent (appearing only 3 times). Two insights from practice theory can be helpful in this context.

"Teleoaffective formations", a concept offered by Welch (2020) and originally inspired by T. Schatzki, embrace several social practices. These phenomena are defined as "configurations across multiple practices that enjoin those practices to common ends, ordering their affective engagements and offering general understandings through which participants make sense of the projects they pursue" (Welch, 2020, p. 61). Thus, teleoaffective formations are large-scale nexuses of discourse and practice that from the socio-technical systems' point of view, can only be understood on the landscape level as exogenous "macro-level" forces that affect the meanings of individual practices.

Also, as several studies point out, teleoaffectivities (goals and affective states) can be analysed as occurring on the regime level. More abstract landscape-level categories contribute to and impact reconfiguration in practices, as they are practically enacted

embodied and discursive realities, not just “large narratives” in the abstract (Laakso et al., 2021). For example, in the Kokko and Fischer (2021) study, one conclusion is that interventions must target meaning-making at the level of regimes through the identification of “transition intermediaries” who can contribute to public communication with various audiences.