Introduction

After over two decades of research, the field of social entrepreneurship (SE) has evolved into an important research topic for researchers and has, as a result, gained attention in the mainstream literature (Bacq & Lumpkin, 2020; Cavalcanti, 2021; Dwivedi & Weerawardena, 2018; Gupta et al., 2020; McNally et al., 2020; Macke et al., 2018; Sengupta et al., 2018). This attention can be demonstrated by the number of researchers engaging in SE research from a variety of disciplines such as entrepreneurship, psychology, sociology, and economics (Hota et al., 2020; Saebi et al., 2019). Such growth has highlighted the need to advance the literature on learning in SE, in particular comparative studies involving, for example, different pedagogical tools and geographical contexts (McDougle et al., 2017; Otaka, 2017).

SE is a collaborative process between actors with the objective of applying business principles to solving social issues (Montgomery et al., 2012). In terms of SE business models, SE can be defined as non-profit organizations that solve social issues through new methods and innovation (Widjojo & Gunawan, 2020). To solve these problems and achieve social change, learning is a key component for SE. In a systematic literature review conducted by Macke et al., (2018), it was found that learning was a key driver of SE research, especially because learning is a phenomenon that increases the search for solutions to social problems.

While there is much research on learning in typical entrepreneurship settings (Arantes et al., 2018; Funken et al., 2020; Karataş-Özkan, 2011; Politis, 2005; Rae, 2017; Wang & Chugh, 2014), the literature on learning in SE is scarce. Little is known about the learning process in SE (Scheiber, 2016; Smith & Smith, 2017), which makes learning in SE a pre-paradigmatic and developing field. To date, most of the studies on learning in SE have taken a psychological approach, focusing on the individual level of analysis (Baden & Parkes, 2013; Faminow et al., 2009; McDougle et al., 2017; Scheiber, 2016). This misses the collective aspect of SE, despite the initial studies on collective or social dimensions such as collaborative actions, social interaction, and social relationships embedded in SE (Dacin et al., 2010; Montgomery et al., 2012) and learning (Wenger, 1998). Collective learning is key for successful SE activities (van der Horst, 2008). We maintain that learning in SE from the psychological perspective alone does not consider the collective and contextual aspects of organizations that are vital components for learning as the interpretive paradigm does (Tandon, 2014), which are critical to understanding learning in SE.

In this paper, context refers to two things, namely the geographical location and the field of work of an organization (Chandra & Kerlin, 2021; De Bruin & Lewis, 2015). The notion of learning in this study is based on communities of practice theory, in which learning refers to the social participation process and knowing where individuals engage in communities and are active participants by mutual engagement and shared repertoire (Lave & Wenger, 1991; Wenger, 1998), through which people learn.

Learning is a contextual phenomenon (Antonello & Godoy, 2011; Smith & Smith, 2017; Wenger, 1998). In fact, in terms of learning, context has been considered crucial in how SE is theorized (Chandra & Kerlin, 2021). In this sense, learning in SE has differences depending on the context in which it is investigated. Although some qualitative studies have investigated this topic in distinct contexts, they do not take into consideration diverse contexts, such as different countries. For example, these studies usually focus only on a single country or cultural context and similar fields of work (Howorth et al., 2012; Munoz et al., 2015; van der Horst, 2008). To move the study of context forward in the SE literature, it is necessary to, at a minimum, investigate it in different national contexts (Shaw & de Bruin, 2013).

Therefore, the primary research question of this study is: What are the learning practices in SE organizations in two different American countries? To answer the question, we conducted a qualitative study in two SE-related non-profit organizations in different countries using interviews. Because SE is highly influenced by contextual settings, some authors suggest that scholars should consider different contextual settings to conduct empirical research (e.g., De Bruin & Lewis, 2015; Gupta et al., 2020). As such, we chose two different geographical locations to conduct our research: Canada and Brazil. As we describe below, the choice of these two countries was driven by an examination of their key similarities and differences in SE practice. Thus, the aim of this study is to analyze SE learning practices in non-profit organizations in both Canada and Brazil.

Our research advances the literature on learning in SE in two ways. First, we investigate SE in two different national contexts, adding nuance to the contextual consideration of learning in SE environments. Second, we identify what the learning practices in SE are and what influences them in these contexts.

Following this introduction, this paper provides a synthesis of the literature exploring SE and learning in SE. We then explain and justify the research method employed in this study. Next, we present our findings and a discussion of our results in theoretical terms. Finally, we offer suggestions for future research.

Literature review and framing

Social entrepreneurship

SE is defined in several different ways in the extant literature. It has been identified as a new type of entrepreneurship (e.g., Apostolakis, 2011; Tracey & Phillips, 2007), a construct that itself has several different scholarly definitions. For example, some scholars define entrepreneurship narrowly as the launching of a firm (e.g., Fayolle & Gailly, 2008), while others use a broader lens by defining it as the ability to recognize and exploit opportunities in the market for financial gain in the business context (e.g., Shane, 2003) and beyond (Shane & Venkataraman, 2000). SE definitions share some similarities to those of entrepreneurship. For instance, SE and general entrepreneurship draw on the same skills and tools for spotting market opportunities and achieving financial viability. However, many scholars differentiate SE from entrepreneurship in terms of its typology (e.g., Apostolakis, 2011), which sees SE as focusing on solving important social problems by developing viable solutions for wide-scale change (Light, 2006; Martin & Osberg, 2007).

Some scholars suggest that SE differs from entrepreneurship primarily in that the former focuses on social change using non-profit approaches, whereas the latter has primarily for-profit motives (e.g., Martin & Osberg, 2007). In contrast, Dees (2001) highlights the adoption of a mission to simultaneously create and sustain private and social value as a key SE requirement as compared to entrepreneurship. This approach considers the traditional distinction between profit-driven and non-profit activities as irrelevant and instead emphasizes a hybrid model of for-profit and non-profit motivations.

In addition to the potentially large differences in profit motives, research suggests that there are differences between social entrepreneurs and entrepreneurs. For example, Dees (2001) suggests that social entrepreneurs act as agents of change and transformation through their economic activities with social impact. A social entrepreneur, then, can be considered a special type of entrepreneur who is distinguished by an appeal to goodwill, a socially or community driven mission (versus, for example, a purely profit-driven mission), and the intention of social value creation (Dees, 2001). According to these research findings, the main difference between an entrepreneur and a social entrepreneur lies in the social entrepreneur’s ethical and motivational commitment to societal benefits.

Taking the above definitions into account, while acknowledging there is no firm consensus in the literature about its formal definition (Saebi et al., 2019), in this paper we define SE as the type of entrepreneurship that has the primary objective of addressing important social problems by employing sustainable and innovative approaches (which may have for-profit and non-profit elements) that are beneficial to society. Through the process of consistently identifying and exploiting opportunities in challenging contexts, social entrepreneurs develop and deploy creative and innovative solutions to solve these problems.

We believe the above definition incorporates the notion that SE is a construct (Dwivedi & Weerawardena, 2018) that incorporates several heterogeneous phenomena (Saebi et al., 2019). These phenomena include community engagement (Dwivedi & Weerawardena, 2018), social change agents, institutional entrepreneurs, social ventures, social enterprise, social innovation (Mair, 2010), government sectors, non-profit and for-profit organizations (Austin et al., 2006), and hybrid organizations (Doherty et al., 2014).

To advance on the entrepreneurship construct, it is essential to identify all the various contexts in which SE can be constituted and practiced (Cappelli & Sherer, 1991; Johns, 2006). Context has been considered vital in how SE is explained (Chandra & Kerlin, 2021) and influences not only SE activities but also the SE at various levels of analysis (De Bruin & Lewis, 2015). As such, Welter (2011) suggests that attention should be given to the simultaneous multiplicity of contexts to understand how entrepreneurship happens. Furthermore, as no two contexts and no two enterprises are the same, one important way to consider contexts would be to research SE activities based on the country of origin to highlight characteristics coming from social and cultural variations (Gupta et al., 2020).

Learning in social entrepreneurship

Studies on entrepreneurial learning have been conducted since the 2000s and have contributed to the advancement of the literature on the topic (Arbaugh et al., 2021; Cope, 2003; Minniti & Bygrave, 2001; Oliveira & Cassandre, 2023; Pittaway & Cope, 2007; Politis, 2005; Rae, 2000; Wang & Chugh, 2014). However, the total number of studies on learning in SE is relatively small, despite the increasing research focusing on SE (Bacq & Lumpkin, 2020; Gupta et al., 2020; Saebi et al., 2019). The primary goal of this paper is to address this research gap by focusing on learning in SE.

The current literature on learning in SE focuses on such constructs as experiential learning (Baden & Parkes, 2013; Holtham & Rich, 2012), exploration and exploitation (Liu & Ko, 2012), single and double-loop learning (Mano, 2010), learning orientation (Baba, 2015; Choi, 2014; Mahmoud & Yusif, 2012), knowledge accumulation (Liu & Ko, 2012), knowledge management (Granados et al., 2017; Hume & Hume, 2016), knowledge spillover (Ko & Liu, 2015), socialization, externalization, combination, and internalization (SECI) framework (Granados et al., 2017), and service learning (Litzky et al., 2010; McDougle et al., 2017). These studies are typically rooted at the individual level to determine the psychological constructs associated with SE learning (Granados et al., 2017; McDougle et al., 2017; Scheiber, 2016). Though the psychological approaches are useful in identifying traits and learning orientations of SE practitioners, they do not provide a complete understanding of learning in SE because they miss important collective aspects that constitute learning (Wenger, 1998) and SE (Dacin et al., 2010; Montgomery et al., 2012). For example, collective learning is central to the successful fulfillment of SE activities (van der Horst, 2008), but is rarely studied as a context of learning in SE research.

Therefore, in this study, we define learning as a social (Wenger, 1998) and a collective (Brown & Duguid, 2001) phenomenon that is situated in social participation (Lave & Wenger, 1991). In this perspective, the focus of learning moves from psychological perspectives to the processes of participation and interaction in which the learning is situated (Gherardi et al., 1998), including communities (Howorth et al., 2012), cultural contexts, and operational contexts. Our learning choice provides an understanding of the context of organization and the collective dimension as key aspects of learning (Tandon, 2014) and SE.

According to Tandon (2014), from the communities of practice perspective, learning takes place in social structures that include formal and informal social configurations (i.e., communities and groups). Communities of practice can be defined as a “group of people who share a concern or a passion for something they do and learn” (Wenger, 2011, p. 1). These people are engaged in a system of collective learning in a shared domain of interest (Wenger, 2011). Communities of practice are the primary building blocks of the social learning system, and organizations are constituted in these social learning systems (Wenger, 2000). Therefore, a social definition of learning in this study implies mutual engagement and shared repertoire in learning systems.

Learning across these boundaries refers to the construction of a common understanding of shared knowledge sets (Tandon, 2014). As such, learning is a relational process located in social contexts (Smith & Smith, 2017). This means that learning practices have differences depending on the contexts in which they are operating and situated.

Considering that learning in SE is highly contextual, we chose to pursue a cross-cultural approach to our data collection. We believe that in so doing, we can highlight and expand upon key insights into learning in the current SE literature. Most literature on learning in SE focuses on a single country, missing the differences that might emerge from empirical studies in different national contexts.

Method

Research design

To analyze the SE learning practices, we adopted a qualitative approach to identify learning processes (Hennink et al., 2020) and to interpret how human beings construct meaning via their learning experiences (Patton, 2014). This approach helps us interpret the complex concepts of learning and knowledge in SE and gives us a deeper understanding of its aspects (Granados et al., 2017).

Context and participants

As discussed earlier, learning is contextual (Antonello & Godoy, 2011; Wenger, 1998), and SE is seen as a process resulting from the context (Mair & Marti, 2006), which is enriched by yet other contexts (De Bruin & Lewis, 2015). Thus, it is significant to study SE close to where things happen (Steyaert & Landström, 2011), such as learning practices. To do this, we conducted empirical research in two different contexts, Canada and Brazil, to highlight aspects of cultural and social differences (Gupta et al., 2020).

Context in our research is understood as (1) a geographical setting, such as a country in the spatial dimension, and (2) a field of work of the organization (Chandra & Kerlin, 2021; De Bruin & Lewis, 2015), such as technology and socio-environmental education, which are the focus of the non-profits selected in this study. In the communities of practice framework, we followed the suggestions of Nelson and Gopalan (2003), who studied different countries and different organizations with very distinct fields of work.

In terms of geography, the reasons why we chose the countries represented in this study rely on the following aspects: Canada and Brazil are each experiencing a significant increase in the number of social enterprises (Bosma et al., 2016; Scheiber, 2016), and in terms of institutional logic, both primarily utilize non-profit organizations as the main model of operation (Lopez, 2018; McMurtry & Brouard, 2015). The qualification of non-profits as SE is based on the Defourny et al.’ (2021) characterization of non-profits as one of the main social enterprise models. Thus, the volunteer participants in the two non-profit organizations in our study are regarded as social entrepreneurs. In 2018, over 24 million Canadians engaged in non-profits, and 79% of Canadians volunteered formally and informally to work for them (Hahmann, 2021). The number of non-profits in Brazil has grown substantially in the last decade, with numbers of up to 820,000 organizations (Lopez, 2018).

There are also contrasts between Canada and Brazil. According to the Thomson Reuters Foundation (2019), it is relatively easy in Canada for social enterprises to access investment opportunities, while in Brazil SE, organizations find it difficult to access funding. SE organizations in Canada tend to enjoy a good deal of government support, whereas in Brazil, government support is lacking by comparison. Brazil has less favorable conditions to start and grow non-profit organizations than Canada (Thomson Reuters Foundation, 2019). Moreover, Canada is considered the best context for SE, while Brazil ranks much below, according to research (Thomson Reuters Foundation, 2019). Therefore, given the fact that the two American countries experience a significant increase in the number of social enterprises and, at the same time, they rank in extreme positions in these rankings, it seems interesting to conduct empirical research.

To select the non-profit organizations to compare, we needed a comprehensive set of SEs in both Canada and Brazil. To that end, we created a set of systematic criteria based upon the SE literature. We started seeking potential non-profits in databases in Canada and Brazil. We contacted SE organization founders by email and text message to recruit their participation. This resulted in the selection of two organizations, one in Canada and one in Brazil. In Canada, Community Tech (CT) was chosen. In Brazil, EcoHortas (EH) was selected. The primary field of work at CT is technology, and at EH, the focus is on socio-environmental education. Both non-profits were selected because of (1) the great relevance of the social impact they make in their communities, and (2) the heterogeneity of the field of work of the organization. We asked the founders of the above organizations to suggest names of potential social entrepreneurs to participate in our study. This resulted in the selection and interview of 13 participants, six at EH and seven at CT (see Table 1).

Table 1 Interviewees information
Full size table

The SE organizations represented by the participants in this study were all non-profits with a focus on solving various social problems. This is consistent with our definition of SE, which positions SE as having the primary objective of addressing important social problems by employing sustainable and innovative approaches (which may have for-profit and non-profit elements) that are beneficial to society. Our definition of SE is conceptualized as a type of entrepreneurship that has the primary objective of addressing important social problems by employing sustainable and innovative approaches that are beneficial to society. The two non-profits selected matched our understanding of SE since these two organizations have in their objectives and mission a social purpose to solve social issues through innovation.

Canadian context: Community tech

CT is a non-profit organization designed to help unite policymakers, technologists, and residents who are engaged in learning with each other and improving their communities. There are several CTs in various cities across Canada. For example, CT Fredericton is a non-profit organization in which its tech community, citizens, policymakers, and other non-profits make better places to live. Its slogan is “Technical solutions to social issues”. It facilitates knowledge sharing and collective work to create digital tools to solve social issues. All technical solutions built by the social entrepreneurs at CT are free for all members of the community to access.

CT Fredericton was founded in 2017 in the city of Fredericton, New Brunswick, a small province on the east coast of Canada. It is constituted by social entrepreneurs with different skills and volunteers. The organization is also strongly multicultural, having people from several different parts of the globe working in one place.

Different solutions get built at the same time at CT. Social entrepreneurs organize themselves into groups for each project, and inside these projects they divide work tasks. For example, there are generally five people working on a project at any given time, and each person works on a part of it. Team members take turns promoting and teaching courses to improve team members’ skills in a variety of ways. These aspects of sharing knowledge, different skills, and multiculturalism make CT a relevant place for the study of learning.

Brazilian context: EcoHortas

EH is a non-profit organization with a mission to connect people to promote sustainable cities through socio-environmental education, food, and agroecology in the Brazilian state of Goiás. The organization is located in the city of Goiânia, Brazil. It was founded in 2016 out of the desire to bring people together and act collaboratively toward social goals. In its first years, EH began producing urban gardens in public schools in Goiânia and cities around it.

EH has four pillars of action: education, food, environment, and community. Moreover, EH’s work is aligned with some of the sustainable development goals from the United Nations, such as zero hunger, good health and well-being, quality education, sustainable cities and communities, and partnership for the goals. There are nine working groups at EH, and each one has a coordinator and many volunteers. Each working group develops a part of a project. To become a volunteer, it is necessary to apply and participate in a selection process with résumés and interviews. To receive EH benefits, schools and companies must apply and go through a selection process.

As a non-profit, EH operates based upon donations from citizens, companies, and organizations in the public sector. It has also developed relevant partnerships, such as with the U.S. Embassy and Consulates, and it has won various awards. The work of the social entrepreneurs at EH has gained attention from the media, and they have appeared in important TV programs, magazines, and international institutions. EH has implemented 130 activities and estimates they have impacted the lives of more than 20,000 people.

Therefore, the characteristics discussed in the sections about the two non-profits make them significant contexts for empirical research in learning in SE.

Data collection

To understand the representations of the participants regarding the phenomena inherent in social participation, which culminate in learning, some strategies were applied in the data collection: a combination of in-depth interviews (Yeo et al., 2013) and observation (Musante & DeWalt, 2010). The multiple sources of evidence (Yin, 1994) or triangulation (Hartley, 2004) of these two instruments brought more robustness and validity to our study (Simón-Moya & Rodríguez-García, 2021). A summary of the data collection associated with this study can be found in Table 2.

Table 2 Data collection
Full size table

In-depth interviews make it possible to investigate issues in detail and determine where the phenomenon is located (Lewis & Nicholls, 2013). We developed a semi-structured interview guide containing approximately 15 questions based on the communities of practice theory (Wenger, 1998). Two semi-structured guides were created, one in English for the participants in Canada and the other in Portuguese for the Brazilian participants. At CT, all interviews were conducted in English except for one, which was conducted in Portuguese. At EH, all interviews were conducted in Portuguese.

Before collecting, we conducted a pilot test (Yin, 1994) with the two non-profits to adjust and improve the guide questions as needed. The two interviews were not included in the final analysis of the results in this paper. When the guide was ready, a two-stage interview process (Scheiber, 2016) was carried out with the social entrepreneurs. All interviews were recorded using a professional voice recorder.

The interviews with the participants from CT were conducted in CT meeting rooms, coffee shops, and other meeting rooms. In EH, the interviews took place in coffee shops, restaurants, and participants’ houses. Observations of collaborative behavior provided our secondary data. One of the authors of this paper made the observations and wrote the associated field notes. The observations in the Canadian context took place mostly in the CT meetings. At EH, the observations were made in meetings and in the activities performed by the social entrepreneurs.

Data analysis

We adopted thematic analysis based on Braun and Clarke (2006) since it is a useful and flexible technique that helped us identify and analyze the learning practice themes within the data. Initially, we used Logus Academy software to assist in transcribing the interviews. We then analyzed the full interview transcripts and all relevant field notes using Nvivo version 11, a software program that enables systematic data coding, information arrangement, and the extraction of patterns (Richards, 1999; Welsh, 2002). We combined all the data (Simón-Moya & Rodríguez-García, 2021), transcripts, and field notes. Next, we checked the transcripts three times, listening carefully to the recorded narratives to ensure high fidelity between the recordings and transcripts associated with this research and to familiarize ourselves with our data (Braun & Clarke, 2006).

The next step was coding to investigate the learning practices. This coding process had three stages. First, we had two categories derived from communities of practice theory: mutual engagement and shared repertoire (Wenger, 1998), which we used to fit within the data by analyzing the interview data and field notes. These are our a priori deductive categories. The second coding stage took place by analyzing the two a priori categories in the context of SE. We analyzed the codes of the two categories, and we identified the emerging category: social interactions. The third stage was similar to the second; however, here we analyzed all codes within each category and then between the categories to check what the data would reveal. By analyzing the three categories together, an axial category emerged: power relations. After we identified power relations, we also reviewed the entire data set, collating data that might inform key insights (Howorth et al., 2012) as well as divergences on social interaction and power relations.

The new categories, social interactions, and power relations are inductive because they come from data. Therefore, this study has a deductive-inductive approach. The communities of practice theory played a role in this study by providing us with two a priori categories to discover the emerging category and the axial category and to build our analytical model. The three categories, also called themes in this study, are the learning practices in the context of SE, and the axial is what is within these practices. Figure 1 shows the three learning practices we identified, along with the axial category.

Fig. 1
figure 1

Learning practices in SE. Microsoft word was used to create the Fig. 1

Full size image

Findings

Community tech

A priori category 1: mutual engagement

In this first a priori category, based on Wenger (1998), our data reveals that the basis of mutual engagement at CT is related to its work organization. First, there are general meetings with all social entrepreneurs to discuss ideas. Second, they organize themselves into small groups to work on projects. Mutual engagement at CT happens in the various activities that the social entrepreneurs participate in. For example, it occurs when individuals go to the CT meetings and interact by joining a workgroup, when they vote to select ideas, or when they are engaged in a project together solving a technical issue of software (see Quote 1 in Table 3, for example). It also happens when they attempt to define themselves as social entrepreneurs and how to work collectively (see Quotes 2 and 3).

Table 3 Representative quotations of SE learning
Full size table

Through the process of solving social issues, collaboration takes place from the beginning of each project to motivate social entrepreneurs to develop new solutions or to determine a solution that exists elsewhere (for example, see Quote 4). This engagement in searching for solutions leads to collective learning. As our data show, mutual engagement at CT starts with participation in the meetings, and it becomes a source for creating shared knowledge and discourses, the shared repertoire.

A priori category 2: shared repertoire

Mutual engagement is the base for the development of a shared repertoire and includes elements (Wenger, 1998) such as shared knowledge and discourses. The shared repertoire starts gathering the social entrepreneurs. In this process of gathering people, workers build and share knowledge and learn from each other (see Quote 1). Thus, collaboration is critical to the development of a shared repertoire.

Another way that they learn is through shared discourse. For example, when one member of a learning group asks for a definition of a shared work-related concept and all members reply with a similar, correct answer, the shared discourse has occurred (see Quote 2). This unity allows people to engage mutually in solving social issues. They also need to have the same understanding of CT`s aims, projects, and solutions to solve social issues (see Quote 3).

Emerging category: social interaction

Analyzing the two a priori categories, mutual engagement and shared repertoire, a new category emerged: social interaction. This suggested that mutual engagement and shared repertoire occur due to social interactions in the SE environment. At CT, social interactions are common and occur regularly in meetings, observations, and many other elements of the work process. For example, for some people, going to CT is seen as an opportunity to learn new things (see Quotes 1 and 2) and develop new skills (see Quote 3). Interaction is also encouraged virtually at CT (see Quote 4) and with outsiders (see Quote 5).

Axial category: power relations

From the analysis of the a priori categories and the emergent category, we found evidence of another emergent theme: power relations. Power relations, the axial, is what intermediates the learning process through the practices of shared repertoire, mutual engagement, and social interactions. Learning at CT implies distinguishing who is speaking because it involves power to speak from the experiences and from the knowledge that the person has (Quote 1). At CT, there is recognition of who knows more and who is an expert in a certain subject or domain. Social entrepreneurs recognize and follow those who have more knowledge on a subject. They listen to each other's opinions, reflect on what path should follow, and take decisions together.

Our interviews and observations at CT revealed a strongly democratic process in all its activities and decisions. For example, to select the next social problem to solve, all social entrepreneurs submit votes, with the topic with the most votes getting pursued. Everyone is in the same position at CT. Despite the fact that Gisele is the founder and organizer, she does not make the decisions by herself (see Quote 2). Our observations confirmed that many decisions were taken by this discussion/voting procedure.

Power relations at CT emerge not from positions, but from knowledge and experience, and most significant, from democracy (see Quote 3). Knowledge and experience are sometimes necessary to make decisions about specific aspects of the chosen solutions (e.g., a digital tool). Thus, power emerges from democracy at CT since all social entrepreneurs are involved in the decision-making process.

EcoHortas

A priori category 1: mutual engagement

At EH, data show mutual engagement (Wenger, 1998) is driven more by a membership model than the largely democratic process observed at CT. Although all EH social entrepreneurs are volunteers, they must become members of the organization via an established process before having a say in the decision-making processes. This process is comprised of a formal application consisting of résumé analysis and interviews. As such, only candidates who are selected into EH can engage in mutual engagement, as all other candidates are weeded out before having the chance to do so.

The social entrepreneurs at EH perform various activities together, such as planning events, growing gardens, conducting research (see Table 3, Quote 1), and creating work methodologies (see Quote 2). They see EH as a mutual engagement in terms of making a social impact that brings learning opportunities along with it (see Quote 3). Several social entrepreneurs reported joining EH as a way to learn things related to agroecology, sustainability, gardens, and agronomy. Others joined because they wanted to gain volunteer experience to help them apply for jobs and universities abroad.

A priori category 2: shared repertoire

According to communities of practice theory, mutual engagement is the source of the development of the shared repertoire (Wenger, 1998). The shared repertoire at EH includes ways of doing things, routines, and tools such as books and materials about how to make and care for vegetable gardens. Social entrepreneurs share these tools with schools, society, and each other. For example, the EH website and blog provide information about the activities and events that EH performs (see Quote 1). Ancestral knowledge and awareness that they are a community are also shared among the social entrepreneurs (see Quotes 2 and 3).

Emerging category: social interaction

From the analysis of mutual engagement and shared repertoire at EH, social interaction is interpreted here as an emerging category of EH. Again, mutual engagement and shared repertoire happen through social interactions. Social interaction at EH happens among the social entrepreneurs both inside and outside of the organization. Inside, team meetings at EH do not happen every week with all social entrepreneurs together, as was observed at CT. They meet according to their working groups, with each group deciding upon its own priorities. Social interaction with all members happens when there are events (see Quotes 1 and 2).

Axial category: power relations

Power relations emerged as an axial category of the three learning practices at EH: mutual engagement, shared repertoire, and social interaction. Data showed strong evidence of power relations at EH due to its organizational configuration. Each working group at EH is led by an assigned social entrepreneur coordinator, who, in turn, reports to EH’s executive directors. The social entrepreneurs in the lowest positions are assigned activities by the coordinators (see Quote 1). All final project decisions are determined by the executive directors at EH (see Quote 2).

Synthesis of contextual learning practices

In this section, we compare the learning practices that we identified in the two non-profits in Canada and Brazil (see Table 4).

Table 4 Contextual singularities
Full size table

Mutual engagement at CT happens in informal and organic ways. New members are recruited by existing members to join the organization, and all social entrepreneurs are welcome. The process of mutual engagement at EH, in contrast, is based on a formal process. Newcomers must first be chosen to join the organization through a selection process. Social entrepreneurs from the highest positions at EH then select the new social entrepreneurs. This represents a trait of some Brazilian organizations. For example, formalism and rules are characteristics of Brazilian culture (Santos, 1996; Silveira & Crubelatte, 2007). Also, social entrepreneurs at EH are engaged to find work methodologies for themselves and to replicate this model in other non-profits, while at CT, social entrepreneurs are trying to build their identity as social innovators before themselves, society, other non-profits, and governors.

The shared repertoire at CT is mainly related to a unified understanding built by learning, resulting in a shared learning framework. In contrast, shared repertoire at EH refers to ancestral knowledge as a means of learning that social entrepreneurs can access through social interaction. Social interaction happens more in CT than in EH because all social entrepreneurs meet weekly, in person, to discuss solutions to social problems. Our data revealed that in the context of CT, social interaction takes place with the perceived freedom and collaboration of social entrepreneurs. The social entrepreneurs are free to share their ideas and to help each other. At EH, social interaction is a complex process of collaboration and tension. Social entrepreneurs help each other, but there are tensions between people in high and low positions.

The three learning practices of CT and EH are constituted by the dimension of power, since, in a community of practice, learning always implies power relations (Farnsworth et al., 2016). Our findings do not show significant signs of levels of power or even many conflicts at CT. Studies on power in Canadian organizations show low levels of power distance between employees (Aycan et al., 1999; Hofstede et al., 2010). The democratic dimension identified has nothing to do with just a unique aspect of CT, such as the way social entrepreneurs organize themselves. However, it is also a piece of what happens in the Canadian context. Canada is considered one of the most democratic countries in the world (Unit, 2021).

In contrast, our data demonstrate high levels of power concentrated in EH. It is first due to the collective configuration in EH, which has authority and few democratic practices. These results are also consistent with national studies that indicate the Brazilian style of management has traits such as concentration of power (Prates & Barros, 1997). The high level of power implies a low climate of openness (Dastmalchian et al., 2000), and control constrains learning as social entrepreneurs are unable to take decisions by themselves to learn by doing (Cope & Watts, 2000). Therefore, the singularities of the learning practices in this study primarily arise from two factors: (1) the collective configuration of each non-profit, and (2) the national contexts. The national context or culture is a crucial factor that differentiates one organizational context from another (Motta & Caldas, 1997).

Discussion

According to Wenger (1998), mutual engagement is the source for the development of shared repertoire, and this later reflects the history of mutual engagement. However, our findings show that not only mutual engagement led to shared repertoire, but shared repertoire also led to mutual engagement. In other words, social entrepreneurs together create a shared repertoire from mutual engagement, and from the shared repertoire, they engage mutually to solve social issues.

Data revealed that mutual engagement and shared repertoire are governed by a process of social interaction between the social entrepreneurs of each non-profit and outsiders. Thus, members work together and exchange information (Wenger, 1998) while trying to find solutions to make a social impact in their communities, and they learn as a result. Interactions such as these are associated with innovation and diverse learning (Amin & Roberts, 2008).

Our analyses of interview data indicated that the three learning practices operate on a similar logic, with the axial category of power relations emerging from all three. Power is related to forms of discipline maintained by discourses that establish knowledge and truth (Foucault, 1980), which have symbolic aspects such as hegemony (Gramsci, 1957) and democracy. Power is inherent in learning (Farnsworth et al., 2016). According to Gherardi et al. (1998), knowledge involves power relations in a social context, and social relations at work do not only involve support and friendship but also conflicts and rivalries between people, as we found in our study.

At CT, the level of concentrated power is low and appears to emerge from democratic processes, even at various levels of expertise. In this sense, the three learning practices at CT are permeated by collegiality, expertise, pleasure, and trust, which is consistent with communities of practice (Wenger, 1998). These collective characteristics also reflect what happens in the national context. In studies involving power, the Canadian context is usually characterized by low levels of power distance (Aycan et al., 1999; Hofstede et al., 2010), paternalism (Aycan et al., 1999), rigidity, and control, and organizations report high scores on climates of openness (Dastmalchian et al., 2000). In this context, power relations promote learning through the three learning practices.

At EH, we found that power tended to be concentrated at the top executive levels. Questions from people at the lowest positions in some non-profit organizations seem to be a simple way of obtaining information to do things right. However, our data suggest that asking is not only a way to obtain information but also a type of permission and liberty to act. The fear of making mistakes is not just because the volunteers will do something wrong but also because of the power of high positions. The learning practices at EH happen in an environment with contradictions such as collegiality, ease, pleasure and deprivation, authority, tensions, and conflicts (Wenger, 1998). At EH, power relations seem to act as inhibitors of freedom and autonomy, which might inhibit learning opportunities since high levels of autonomy improve learning opportunities (van Ruysseveldt & van Dijke, 2011).

The high levels of power at the top levels are characteristic of the collective dimension of EH, but it is also the broader Brazilian context. Research on managerial style and power relations in Brazil shows a high concentration of power (Hofstede et al., 2010; Silveira & Crubelatte, 2007), hierarchy (Freitas, 1997), authority (Carbone, 2000), and paternalism (Carbone, 2000; Silveira & Crubelatte, 2007) in the organizations. Canada is considered one of the most democratic countries in the world, while Brazil’s democracy is often viewed as flawed (Unit, 2021).

The singularities identified in the categories of the two non-profits studied here show characteristics of the collective configuration of each non-profit. However, results evidence that singularities also come from the broad context of the two countries. Although national context and culture tend to have strong influences on organizational aspects, these organizational contexts are shaped by the cultural preferences of the members (Dastmalchian et al., 2000). As such, social entrepreneurs decide what national traits they will reproduce inside their non-profits.

Theoretical & practical implications

This study contributed to SE and learning in SE literature in three main ways. First, it empirically investigated SE in different geographic contexts and distinct organizational contexts of fields. Data suggested that the two non-profits in our study reproduce their national context traits in their management styles. The Canadian non-profit showed low levels of power and high democratic processes. In contrast, in the Brazilian context, the non-profit demonstrated high levels of power and fewer democratic practices as compared to the Canadian context.

In terms of the literature on learning in SE, our study contributes by showing the emergence of learning practices in SE and how they operate in different cultural contexts. Our results show that the learning practices in SE are mutual engagement, shared repertoire, and social interactions, which are each configured by power relations. We also show that learning methods in SE can be promoted or repressed depending on the level of power in non-profits. If a non-profit or a group has low levels of power, learning tends to be more likely to occur, rather than in the context of high levels of power.

As we showed, the findings of this study suggest that power can encourage or inhibit learning practices in SE organizations. Power encourages learning when social entrepreneurs can speak freely, participate, and take decisions together, especially in a democratic environment. Power can inhibit learning in contexts when social entrepreneurs feel they cannot openly participate in sharing ideas and solutions and taking decisions to solve social issues. This is consistent with communities of practice theory. For example, Lave and Wenger (1991) argue that it is difficult to learn when power relations are constituted in an impeditive way. In this sense, from a practical perspective, to promote learning practices in SE, we suggest that social entrepreneurial leaders create an open climate with possibilities to give solutions, have autonomy, and take decisions collaboratively.

Limitations & future research

Despite its promising findings, as with all research, our study has limitations. For example, although we found that SE non-profits tend to reproduce their national political contexts in their operations, we did not explore why this occurred. Future studies might analyze why and how this cultural reproduction occurs. For example, a quantitative study can measure and test the level of influence of national context traits on non-profits.

Our results indicated that the three learning practices we identified are configured by power. However, we did not focus on specific SE learning methods, and we did not deepen the discussion of how power promotes or impedes learning beyond identifying that it does. While these questions were beyond the scope of this study, we suggest that future studies analyze how power relations influence SE learning to promote or impede specific SE learning methods. Quantitative studies can also test aspects of power and authoritarianism and how they interact with learning possibilities in different countries. Finally, future researchers should have a larger number of non-profit organizations to examine learning practices more closely.