Tracing policy movements: Methods for studying learning and policy circulation

Introducing learning as a research method

“Learning does not only happen when people are fretting and sweating about a challenge”, remarked one South African policy analyst when I asked him how he learns. “I see it as a more a diffuse process”. Such a declaration speaks to the broad aim of the paper to understand this momentous practical and conceptual issue of learning taking place amongst various local and global practitioners and policymakers. It is comfortably situated within the ballooning body of geographic literature committed to understanding policy models and the way in which they are assembled in one jurisdiction and arrive in other, often times, mutating through the mobility process (McCann and Ward, 2011; Peck and Theodore, 2015). The paper goes further however, connecting the messy and convoluted web of policy learning and adoption with our own journey through the academic netherworld by reflecting on learning both as a conceptual heuristic as well as a practical methodology. In so doing, it enables us to realize that the procedures for borrowing, interpreting and reworking policy knowledge are circular, uneven and at times, unpredictable.

The paper builds off my research of the circulation of bus rapid transit (BRT) across South African cities, which I argue is the product of extensive dialogue and exchange between South Africans and South Americans and amongst South African architects, urban designers, financial managers and transport engineers, politicians and other policy actors. Learning (sort of) goes back to July 2006 when international experts presented the attributes of BRT at a workshop at the Southern African Transport Conference (Wood, 2014b) and their subsequent visits to South African cities (Wood, 2014a); and in many ways, BRT adoption began when Johannesburg City Council became the first South African city to approve BRT in November 2006. South Africans proudly boast an association between their BRT systems and those in South America (for example, City of Johannesburg, 2011). They especially call upon the experiences of policymakers in Bogotá to defend their determination to improve the existing transport system in spite of initially staunch opposition from paratransit operators (Wood, 2015c). While it may appear a relatively straightforward process in which cities learned of and implemented BRT in a timely and efficient manner, policy circulation and adoption is gradual, repetitive and delayed (Wood, 2015b), and always subject to local political contestation (Wood, 2015a).

The case of BRT adoption reveals that policy circulation is not only the “purposeful, repetitive, programmable sequence of exchange and interaction between physically disjointed positions held by social actors in the economic, political, and symbolic structures of society” (Castells, 1996: 442) but a process of learning and exchange, which involves power and personalities (Healey, 1997; Healey and Upton, 2010; McFarlane, 2011a). These practices however have proven difficult to study since the exchanges rarely lead directly to uptake. Previous studies have traced the movement of knowledge through various “coordination tools” (McFarlane, 2011b: 364), which include consultancies, conferences, journals, multimedia, reports, study tours and workshops; and others have followed the transnational advocacy groups and learning forums that package, frame and legitimize global circulation (Theodore and Peck, 2011). Scholars also suggest that learning emerges through various voices, interests and expectations, translating and coordinating a multitude of information, including pre-existing knowledge, across asymmetrical power structures, and creating possibilities from the impossible (McFarlane, 2011a). This paper builds from these examinations by outlining a conceptual and practical framework for conducting policy mobilities research that attends to the plethora of ordinary practices – be it through engagements with fellow practitioners, with their toolbox of material solutions, or after a particular moment of discovery – that form the assemblages of learning. Such tactics appraise policy circulation through its retrospective trails, that is by looking at how a policy arrived once it is rooted somewhere, an approach that furthers Robinson’s (2013: 20) contention that “arriving at policies involves far more than assembling the discrete different entities, ideas, or objects which we can trace as they move from there to here” (see also, Robinson, 2011b). I propose a methodology that assimilates the researcher into the learning field by replicating the multilateral process of learning – relying on information exchanged through meetings with policy actors, hearing their anecdotes, reading their reports and attending their events – thereby exposing policy learning as more than just policymakers networking at a conference or a workshop promoting a particular international policy model but part of a lengthy and recurrent process of dialogue and exchange.

In addition to advocating for learning as a research method, this paper outlines a methodology for studying policy circulation that sets my own learning process alongside those of South African policy actors. My research, like their policy work, relies on the stories, conversations, experiences, opinions and decisions of those involved in metropolitan, regional and national government as well as non-governmental actors in consultancy, the academy and industry both inside and outside South Africa involved in policy mobilities and adoption. Just as I asked the various South African policymakers to divulge their processes of learning and circulation in the implementation of BRT in South African cities, they similarly acquired advice from their colleagues both inside and outside South Africa. I suggest that my own learning process is similar to that of my subjects in that we both learn mostly through our interactions with and by participating in networks alongside fellow actors, and we regularly validate this information through materials and at learning events. More than simply uncovering the movement of policy from place to place, actor to actor, such a methodology employs the intricate connections and dependencies between ideas, objects, people and places, as the unit of analysis.

Bearing in mind these debates on assemblage, territoriality/relationality and actor-network theory, the next section theorizes scholarly studies of the practices for following mobility. Next, the paper considers the various processes through which policy actors in cities learn combining the literature on learning with research on mobile policymakers. These arguments unveil a set of procedures for unravelling the assemblage, first by “following the people” and their understandings of the mobile policy; second, by “following the materials” to experiment with a Latourian approach to materiality; and finally, by “following the meetings”, that is, the conferences, workshops, book launches, seminars and associated learning events mingle. In concluding this discussion on learning as a research method, the paper advocates that such procedures allow for personal reflection to further unfurl the assemblages that carve and sustain the pathways for the movement of knowledge and experiences.

Methods for following mobile policies

The physical, social and theoretical movement of ideas, objects, people and places can be difficult to study because they are constantly in motion, positioning and repositioning, erratically and sometimes, irrationally. It is not so much that policies-in-motion demand mobile methods but rather we need to develop a sensibility that caters to the ephemeral, ethereal and experiential aspects of this movement. Scholars have studied circulations by tracing policy knowledge and models (Freeman, 2012; Peck and Theodore, 2010, 2012), policy actors (Healey and Upton, 2010; Larner and Laurie, 2010; McCann, 2011; Roy, 2012), learning organizations (Saunier, 2001; Theodore and Peck, 2011) and the sites of learning (Clarke, 2010; McCann and Ward, 2012; Saunier, 2002), under the auspices that our research methods must be simultaneously “on the move” (Cresswell, 2006), evident through physical travel between research subjects and sites, and “moored” (Hannam et al., 2006) in the offices of policy actors. In such a mobile world, the challenge remains for research to be attentive to both the peripatetic nature of policy models and their affiliated policy actors as well as to the unpredictable character of espousal and emulation. The literature explored henceforth reveals that existing methods of tracing policy movement insufficiently attend to the unpredictable nature of mobilities and moorings; rather, to address the transitory nature of learning, we should follow the chains, paths, threads and intersections.

Within the recent mobilities turn (Cresswell, 2006; Sheller and Urry, 2006), two approaches pioneered by Marcus (1995) have emerged focusing on the moving process itself – “follow the thing” and “follow the people”. The former method has been used in a number of food studies to understand the relationship between the commodity produced, its consumers and everything in between: Freidberg (2001) follows the movement of the green bean along its commodity chains from the site of production in Africa to its European imports, and Cook and Harrison (2007) trace the relationship from one London family consuming hot pepper sauce to a group of farmers in rural Jamaica growing the requisite hot peppers; a similar multi-site ethnography was conducted by the Matsutake Worlds Research Group, a collection of cultural anthropologists conducting research on the movement of Matsutake mushrooms to show the global connection between human and non-human actants (Choy et al., 2009). Goldman (2005) takes up the charge of the latter following the inner workings of the World Bank from its headquarters in Washington, DC to its various project sites around the globe. He shadows the World Bank as a single actor involved in environmental, governance and social ventures to show its role in increasing global inequality. Using this methodology slightly differently, Roy (2010a) studies the poor and the powerless through the agents who manage poverty and their associated circuits of profit and investment. All of these studies recognize the way in which that particular assemblage of people, politics, produce and production animates the “thing” and/or the “people” and its movement. The concept of assemblage is particularly useful here in drawing out the relations between multiple actors and actants, simultaneously present and proximate, across socio-political and temporal divisions enabling a more thorough review of both the tangible and intangible aspects of learning and policy mobilities (Farias, 2010; Jacobs, 2012).

Amidst this wave of mobile literature, one notable special issue edited by Allen Cochrane and Kevin Ward confronts the methodological challenges of researching geographies of policy mobilities (2012). Cochrane and Ward suggest at the onset that thinking through the movement of best practice is especially challenging because movement is often indirect and nonlinear, political and unrepeatable, and consequently difficult to trace. This compilation of research thus leads from diverse presumptions, objects of observation and analytical rigor and promises not to propose an academic template from which all like-minded research can emerge, but taken together it opens up a conversation on how best to confront this methodological challenge beyond the customary semi-structured interviews. Freeman (2012) for instance begins under the auspices that policy only exists as a mobile product as seen through the relationship between the meeting and the documents that emerge. He reasons that what matters for our understandings of mobile policies are those moments of friction in the translation from meeting to document and between documents. Roy (2012) picks up on this notion of complete-mobility focusing on what she calls, an “ethnography of circulations” of “middling technocrats” (Larner and Laurie, 2010), as the study of the practices of various actors and agencies and their involvement in mobilizing policy. Finally, McCann and Ward (2012) offer yet another starting point – the city – and a focus on the politics of place which enable the articulation of mobile knowledge. This paper amalgamates all of these ideas picking up where their research tails off, by taking the ordinary practices of policymakers and considering how they learn, and employing our reflections on learning within this mobile world as integral to research of policy mobilities.

Within this compilation, Peck and Theodore (2012) recommend a slightly alternative methodological starting point, what they dub, the “distended case study”, in which the researcher travels alongside the policy concentrating on the power and politics of policy translation. They, like Marcus (1995), suggest two analogous approaches – “follow the policy” (Peck and Theodore, 2010) and “follow the project” (Peck and Theodore, 2012), which combine multi-sited ethnography (Marcus, 1995) with the extended case method approach (Burawoy, 1998, 2009). This technique was first utilized by Mosse (2004) in the ethnography of aid policy and has been useful because “it is not always possible to ‘be here’, when in the study of global policy networks there is a constant imperative to also ‘be’ somewhere else” (Peck and Theodore, 2012: 25). Peck and Theodore go on to suggest that researching mobile policies need not always be a multi-sited venture but it will often necessitate “methodological travel, along the paths carved by the policies themselves” (2012: 24). This sensitivity towards mobility resonates with my findings which entail incorporating the site of origination with the site of adoption and the various stops along the way, all of which influence the uptake of a particular policy approach. It also calls for thoughtfulness for the retrospective character of mobility, questioning the extent to which transnational actors can be studied, especially if their involvement was transitory, and they have since moved on to other enterprises.

Thus, like Robinson (2011b, 2013), I root this strategy for research within the adopting locality and from there deliberate on the practices through which a policy from elsewhere was introduced, shaped and localized by various policy actors, their interactions with global advocates and inter-referencing across space and time (see also, Jacobs, 2012). Rather than observing or shadowing people’s movements, as Marcus suggests though in his understanding of “following the people”, this more grounded approach asks participants to interpret and reflect on their own decisions and learning processes. And, because a thorough investigation of policy mobilities does not entail pinpointing the origin of a particular policy or its movement across jurisdictional blockages, this methodology considers the connections between people and objects that interpret policy flows upon arrival. This spotlight on the adopting locality recalls Latour’s account of the Curies’ polonium as part of “a continuous chain of people using, testing and believing in polonium” as “necessary to maintain its existence”, concludes Latour, “but they are not same people” (1987: 138). He makes similar claims about the diesel engine explaining, the story “may be analyzed either by looking at the changing shape of the engine – tied to different people – or by looking at the changing type of people – linked to the engine. It is the same story viewed either from the standpoint of the enrolled people of Part A or from enrolling things of Part B” (1987: 138). Like the diesel engine, the story of learning and policy mobilities can be told either by examining “the people who are convinced, or by looking at the new associations made to convince them” (1987: 138). This would fulfil Robinson’s (2011a: 15) suggestion that “the connections themselves might well form the focus of comparison”.

All these cases of following suggest a sense of completed transfer in which the learning has concluded and the adoption accomplished. Temporality of course is a theme across the literature (Peck and Theodore, 2015; Wood, 2015b) but here it is important to consider how to track the interactions between actors and institutions across space and time progressively from the inception of the idea to the implementation of the policy model, or as I did, retroactively from the adoption process back to the initial learning. A retrospective trail allows for reflection by policy actors. Perhaps, in the course of the interview, a policy actor may decide to change the trajectory of future emulation, or he may appreciate the influence of distantiated sites thereby enticing future learning. Such ruminations further demonstrate the multifaceted and constantly mutating nature of policy circulation. The arguments set forth build from McCann and Ward’s (2012) call for researchers to move with the actors to understand how people, policy and place are made mobile. Rather than simply a physical movement between sites, this methodological approach moves temporally between the adoption of BRT and its initial arrival. We must follow the temporal trail to the beginning of the learning process to fully comprehend the process of policy circulation. In this spirit, the next section looks at the way in which policy actors learn and then utilizes their thinking to reflect on my own methods of learning.

Assemblages of learning

This section will incorporate the aforementioned conceptual discussion on mobile methodology with a broader research strategy for thinking about how policy actors learn new concepts and best practices. Much has been written about the knowledge and competitiveness of learning cities (Campbell, 2009) as well as the individuals learning within them (Rose, 2005; Sabatier, 1993). Healey (1997; Healey and Upton, 2010) introduces a mode for thinking about these processes that focuses on the planning elites involved in shaping the various urban processes and the way in which they mobilize ideas across borders. Healey’s 2010 edited volume focuses on the travelling experts moving ideas and technologies around the world from a global (Roy, 2010b) and historical (Ward, 2010) perspective taking account of the wider challenges of situating urban knowledge. This is one conceptual framework for understanding how policy actors learn, which remains a rather fringe focus still under-theorized by urban scholars. The city is a machine for learning, McFarlane (2011a, 2011b) asserts to describe a way of making sense of the city, and he goes on to unveil three processes through which learning is produced: translation, coordination and dwelling. This proposed research strategy combines the complex criss-crossing of Healey’s elites with the everydayness of McFarlane’s learning to interrogate how mobile knowledge is grounded and applied in alternative contexts.

The research strategy is designed in accordance with my overall thinking that policy actors learn of innovation in much the same way as any other agents – be it through engagements with fellow practitioners, through their toolbox of materials solutions, through a particular moment, or building upon existing knowledge and expertise. This appreciation for the agentful forces of individuals and their connections as well as their urban materialities and ephemeral engagements allow for a more thorough understanding of methods for studying policy movement across space and time as well as a conceptual understanding of the learning process as an assemblage driven by human dialogue, interactions and exchanges (i.e. face-to-face engagements), fuelled through the dissemination of various policy packages (i.e. policy documents, journal articles, documentaries and through the Internet) and steered by particular moments of discovery (i.e., conferences, ride-alongs). Just as the narrative of urban success form a peripatetic package of facts and figures, ideas and interpretations, and encounters and experiences able and ready to circulate, we too develop assemblages of learning in our own discovery process. These assemblages are then described, evaluated and abstracted in research journals, conferences and other academic engagements overlapping, intertwining and otherwise furthering our research journey. This more reflective approach to learning is a critical contribution to the literature on policy mobilities.

Follow the people

One device for unravelling these overlapping and intertwined processes is to survey the actors in the adopting locality and their learning process. This practice focuses on face-to-face engagements, typically taking the form of interviews, a device especially useful for probing beneath the socio-political exterior of the decision-making process. While interviews are often criticized for being somewhat staged and scripted, especially when they involve educated and articulate elites (Peck and Theodore, 2012), interviews are a relational process that exposes not just the achievements of the adopting locality but also the experimentation and failure associated with policy circulation. They are also the best means for understanding the connections and disconnections between actors fuelling the adoption process.

In South Africa, I set off as any policy actor embarking on a research inquiry, by depending on our previous personal and professional relationships to gain entry into meetings with key policymakers. In this case, I had previously worked with my first two interviewees at the South African Cities Network and I also contacted a professor at the university who I knew from previous engagements. Drawing on these professional experiences with policymakers connects this approach with the scholarship on researching elites (Cochrane, 1998; McDowell, 1998; Ward, 1999) especially those reflecting on insider/outsider status and positionality (Herod, 1999; Mullings, 1999; Sabot, 1999) proving that we are not detached and unsympathetic to our subjects but rather embedded within these wider economic, social and political networks. Though during fieldwork I did not occupy any formal position, I was not a neutral researcher in this study but a formerly active policy actor entering into circulation processes with local actors and recognized as part of their wider networks of learning. This strategy empowered me with “insider” access to important policy actors and an “outsider” outlook to critically appraise their learning. Our previous participation in our research sites enables us to “reconstruct a landscape in the eyes of its occupants” and to imagine the experiences of policy actors (Samuels, 1981: 129). This positionality provides the backbone for this research strategy endowing the researcher with greater empathy and understanding without limiting their ability to be critical. It also speaks to the wider argument of this paper that we as researchers often inhabit overlapping spaces.

For this study of BRT adoption, the list of interviewees included those from each city’s transport officers, municipal politicians and engineers in the implementation agency, as well as those enthusiastic actors who pushed it through nearly a decade before it materialized in cities and those who conversely spoke against it. I targeted policy actors, not policy elites per se, as my subjects because I was eager to trace the learning through the perspective of those directly engaged with the application of mobile ideas. Cochrane (1998) reasons that focusing on elite actors rather than “middling technocrats” (Larner and Laurie, 2010) enables us to more readily see the enactment of power relations (see also, McDowell, 1998; Parry, 1998; Woods, 1998). While it is important to meet with the leading policy officials, as well as to hear dissident views, it is most vital to interview those who moved, shaped and adopted the particular mobile policy being reviewed. In many cases, those actors who eagerly embraced a somewhat unfamiliar notion of rapid buses and busways have since moved on to other ventures – e.g. from city official to private portfolio manager and from taxi operator to CEO of the bus operating company. One must be cautious though against merely “studying up”, by which Nader (1972) refers to studying the elites with power or “studying down”, by asking the powerless. Instead, I advocate for a process of “studying through” (Shore and Wright, 1997), which also enabled me to consider the way in which policy actors’ engagement with learning changed over time. “Following the people” is thus both multi-directional and multi-temporal penetrating beneath the official explanation for learning and instead unfolding the political framing and inter-referencing that furthers the adoption process.

Throughout the interview process, the researcher must remain alert to the various power hierarchies rooted in gender, class, race, nationality, politics, history and other socio-cultural characteristics including speech, dress and mannerisms that may shape the interview process as well as future interpretations of the data. Perhaps as a female researcher, I am less threatening and more approachable, and for some, my affiliation with a prestigious research university may have augmented my legitimacy. However, it would be impossible for me to determine the precise influence of these physical characteristics (McDowell, 1992; Schoenberger, 1992). Some scholars suggest that the illusion of power is reinforced by both the researcher and the respondent who create a particular version of themselves in interviews (Cochrane, 1998; McDowell, 1992). To minimize these imbalances, Cochrane (1998) recommends, “an active process of self-reflection”.

Respect for this more reflective approach was frequently raised by my interview respondents when they asked me, why I am doing this project?; why am I focusing on South Africa?; and sometimes, people turned my question around on me and asked how do I learn? In one of my interviews with another learning specialist, he rationalized that people learn best by doing ordinary things, which he enlightened, “is exactly what you are doing right now”. “When you ask me, how I learn, I want to turn that question around to you and ask you, how do you learn? How will you learn about BRT? Who will you speak to?”, he continued. “You will probably sit with them over a cup of coffee or a mango juice, just as we are now, and ask questions. You might ride the BRT. Or, better still, watch people ride. I don’t think you and I learn differently”. In “following the people”, I understood my role as an eager learner, who like my respondents was gathering empirical data and developing a more holistic understanding of my social and spatial environment. I was also cautious against becoming what Dunleavy (1980) disdains as the “inside dopester”, in which our research subjects and their knowledge and experiences become important by virtue of our researching them and in turn, access to these elites increases our legitimacy as researchers (see also, Cochrane and Ward, 2012). That is to avoid what, Cochrane (1998) explains as a process in which “we want to sell ‘our’ elites to everybody else, both because we have grown to know and understand them, and because getting others to recognize their important legitimates and confirms our own”. Thus as we become embroiled with our subjects, there is a risk of inflating their importance along with that of our own findings. The next section picks up here considering the conceptual and methodological role of materials in verifying interview transcripts and reducing agent-inflation, as well as understanding the assemblages of learning that further peripatetic planning.

Follow the materials

A defining feature of this approach concentrates on the role of urban materialities – advertising notices, architectural models, blueprints, brochures, films, policy documents and websites – in redistributing agency between human and material actants in the dissemination of learning and best practice knowledge. The urban successes “come alive” in clever policy documents and stunning showrooms (Pow, 2014: 296) as seemingly mundane artifacts become “diverse technologies of seduction” that render policy models palatable for global consumption (Bunnell and Das, 2010). Harris (2013) notes how guidebooks and glossy magazines can be marshalled to promote world-class imaginaries of cities through transport infrastructure in cities like Mumbai; while, Freeman directs us towards the production of these documents, and in particular, the translation from meeting to document. He makes an astute connection between “where policy conceptions are built, tested, and rebuilt in communicative interaction” and the meeting in which the policy knowledge is pledged to paper (2012: 14). These communicative interactions, he reasons, are embodied through the meeting minutes which can be mobilized and dispatched to anybody (Freeman, 2008). This calls for the application of actor-network theory which provides “a certain sensibility towards the active role of non-human actors in the assemblage of the world, towards the relational constitution of objects” (Farias, 2010: 3; see also, Latour, 2005). In adhering to the recent mobilities turn, which posits the importance of human agency alongside the power of buildings, machines, objects, and so on, this methodology looks at the materials of circulation, however always remaining focused on the human decisions inspired by (and perhaps limited by) the materiality.

Practically speaking, we begin by gathering any and all materials used for the generation and translation of knowledge. There are dozens of technical guides, promotional pamphlets, academic assessments and newspaper exposés written by local city officials and various bus enthusiasts. It comes as no surprise then that Johannesburg publicly promotes their learning from Bogotá in their business plan for Rea Vaya on the Internet and there were plenty of newspaper articles to provide initial understandings of BRT implementation. In interviews, it was common for policy actors to show their personal learning materials – newspaper clippings from their personal archives, presentations from international learning events, reports they drafted many years earlier, vetoed transport plans and presentations by BRT advocates – which advanced policy adoption process. Thus the task of tracing policy movement through the materials involves significant sifting.

This dependency on materials of translation recalls McFarlane’s (2011a) comparison between the assorted composition of a house in a favela in Sao Paulo and the builder/occupier. McFarlane uses this heuristic to represent the process of learning as incremental. Both the house and its occupier are agglomerations of local circumstances and yet when connected, create something greater than each individually. In this methodology, the materiality is an expression of the assemblage of learning: whereas the materials of knowledge may be tactile and generally rooted in space and time, learning emerges through human practice and connections and is thus only temporarily affixed to the actor and the place. One BRT manager in Cape Town used the metaphor of a “toolbox” to explain these assemblages of learning. “You learn by increasing your toolbox of solutions”, he enlightened, and like many of his colleagues in South Africa, he learned by “finding out who has the best tools and then learning from them how to use them and why those tools work” (Personal Interview, April 2012).

For many, the toolbox contained Lloyd Wright’s BRT Planning Guide (2007), a technical manual for system construction and operations with several hundred pages detailing the engineering specifications supported by illustrations of operational systems. Almost all planners and politicians had a copy of the Guide promiscuously placed on their desks – like the Gideons Bible in a motel room nightstand – and while a great many South Africans praised the Guide as key to their learning, a surprisingly small number of policymakers admitted to reading it. Its importance seemed to supersede the words themselves and the materiality of the document became as much a justification for introducing BRT as an instructive technical manual. Such materialities are common in the learning process with learners finding solace in the tangibility of an otherwise impalpable process.

The Institute for Transportation and Development Policy’s film “Making Things Happen with BRT” (2004), another often cited material, is a dogmatic presentation of the political benefits of BRT in Curitiba, Bogota and Brisbane. For the leading transport politician in Johannesburg, the 15-minute film was fundamental in grabbing her attention. When I interviewed the politician six years after she initially learned about BRT, she still framed her intellectual odyssey through the powerful images depicted in the film. “What the video makes abundantly clear is that those cities who have invested in BRT systems has not been only about transport” she divulged, “it is really about quality of life interventions, economy, and environment” (Personal Interview, February 2012). This virtual engagement allows the viewer to contemplate the political aspects of BRT, and compels politicians to act through carefully crafted imagery of an efficient urban public transport system. For those too busy to read detailed technical documents like the BRT Planning Guide, virtual materialities facilitate a rapid reproduction of knowledge (see also, Rapoport, 2015).

These materials as well as many other manuals, reports and presentations were instrumental in exposing practitioners to BRT and were similarly instrumental in shaping my own learning and research. These materials provided me with the official accounts of such varied experiences as study visits and private meetings. I used Freeman’s (2012) approach to find the intersection between these materials and my interview transcripts, focusing on the friction, contestation and resistance as much as the smoothness of policy adoption. This practice of following the materials draws upon ideas of actor-network theory to interpret the extent to which policy actors utilize these materials as tools of the learning process. In moving beyond the conceptualization that all actors in the network are equal and apolitical, these assemblages of learning also attend to the capacity of human subjectivity to interpret the non-human materials. This methodology will now consider the role of conferences and meetings as moments where policy actors and materials intersect.

Follow the meetings

A third useful tactic for thinking through mobile learning practices is following the frequent conferences, workshops, book launches, seminars and associated learning events, which have become integral to our social life. These gatherings animate the various guides and reports, endowing certain models with the energy to circulate. Learning often happens in these “globalizing microspaces” (Larner and Le Heron, 2002: 765), which McCann (2011: 118–119) defined as the meeting rooms, hallways, cafes, bars and restaurants of conferences where best practice is discussed and debated (at times informally) by both policy gurus and novices; through these conversations, the standing of cities are made and unmade.

Conferences are one critical moment in which people, materials and learning intersect, albeit briefly. These microspaces have been studied as part of career development in which the conference-as-a-thing is seen as a means of controlling academic, managerial and professional employees (Bell and King, 2010; Ford and Harding, 2008). Previous studies rely on ethnographic observations, field notes, interviews, conference brochures, exhibition samples and press releases, and follow a Lefebvre (1991) approach in which space is produced and reproduced through the social interactions within the physical locale. They provide useful accounts of the way in which conferences are both a practical opportunity for career development as well as social spaces in which the culture of learning is reproduced. Cook and Ward (2012) take us one step further, seeing conferences as “trans-urban policy pipelines” in which face-to-face interactions enable the exchange of verbal, visual and symbolic information that later facilitate the movement and attachment of certain policy models. These conferences together with the presentations and other exhibition materials, promotional websites, proceedings and reports, and institutions affiliated with the event, become assemblages of knowledge circulation, holding together perhaps only temporarily, perhaps for far longer, but always subject to the fluidity of learning. “This infrastructure is both the cause and effect of wider transformative processes”, reason Cook and Ward (2012), suggesting that it is the cause in that it provides opportunity for the formation of policy models as well as their distribution, and the effect because it has led to the establishment of various new agencies and associations that plan such events.

Oftentimes, following the meetings requires the researcher to be both mobile and moored, demanding considerable travel time as well as long stays and embeddedness elsewhere. There are many noteworthy nods to the importance of individual mobility in learning – conference attendance (Cook and Ward, 2012; Ford and Harding, 2008), study visits (Cook et al., 2014; Cook and Ward, 2011; González, 2011; Wood, 2014a) and work-related journeys (Larner and Laurie, 2010; McCann, 2011), and many of these studies have garnered astute reflections on the challenges and opportunities for travel and ethnography. Roy (2012) for instance ruminates on her own status as an “interloper in the crowd of poverty experts”; Shurmer-Smith (1998) recalls how she crossed over from academic researcher of, to professional involvement with, the Indian Administrative Service; and Parry (1998) reflects on her experiences become so incorporated into her research network that her own name began appearing on the circulating list of members. Rather than seeing these overlaps as restricting the research, I argue, they enhance it by providing an opportunity for the researcher to bring the physical, mental and social sites of learning together thereby reinforcing her own learning and reflection.

All these methods require a willingness to move, but none are as mobile as the “go-along”, which tends be either a walking interview (Carpiano, 2009; Evans and Jones, 2011) or a driving interview (Laurier, 2004, 2008). In South Africa, they ranged from wandering around the bus depot to organized driving tours across the city. In the literature, the go-along correlates what people say with where they say it and many researchers indicate that this process promises sounder results than sedentary interviews by prompting interviewees to connect with the surrounding landscape rather than the interviewer. Evans and Jones (2011) make an important distinction between mobile methods and “sedentary methods in motion”, suggesting an incongruity between interviewing participants about the corporeal experiences of mobility while moving in a car (Laurier, 2004) or on a train (Bissell, 2010), which could be dubbed, the “ride-along”, and walking with research participants as they show you around their neighbourhood (Carpiano, 2009), also known as “walking-along”. Such a methodology reflects Simmel’s (1950) perspective on social space, which he sees as the context for the creation of particular personalities and interactions (e.g. the stranger), and provides great insight into the learning process by allowing the researcher to examine participant’s knowledge and experiences as well as the way in which they engage with their social and spatial surroundings. While the official purpose of these extraordinary experiences was to see first-hand the technical features of the BRT, like study tours in general, the go-along helped build relationships between the researcher and the subject, connections that enabled me to attain a deeper understanding of learning in South Africa.

While the abovementioned methodology presents a well-defined process for thinking about data collection and analysis, learning is the product of constant and continuous interactions and exchanges between people, place and products. This analysis of the various learning practices offers an important opportunity to address the usefulness of mobility within debates regarding the role of policy experts and the power structures employed to adopt policies from elsewhere. I call on researchers to look at both the social and material assemblages of mobilities, always considering their own actions, decisions, and motivations as well as those of the actors directing and furthering both the policy circulation and policy adoptions processes. Such an approach allows for a retrospective review that considers the implication of the movement on the adopting localities.

Conclusion

In concluding, let us return to the remarks by the South African policy analyst that opened the paper: he concluded that there is an “ether” in learning; that one rarely learns consciously and systematically but rather that the most astute and localizable concepts diffuse through less measurable means. “Imagine, two policy actors are talking about transit-oriented development over refreshments after a seminar in an unfamiliar city”, he goes on. “Neither actor uses that precise wording but both seem to implicitly share a common understanding of good urban form”. This, he enlightened, is how people learn. But, how do we study such ethereal processes? How do we analyse the international learning expeditions alongside the ordinary moments of dialogue, and how do we evaluate which lead to urban transformation and which appear as miscarried mobilities?

This paper develops a practical research methodology to advance the wider conversation on how and why policy actors assemble, circulate and adopt global best practices and universal solutions. The arguments draw on the relational/territorial debate understanding policy as fundamentally territorial, in that it is bound up in a whole set of locally dependent interests, and similarly relational, as it moves through relations of actors and is subject to interpretation upon arrival. I also discuss the literature on learning and specifically consider the implications of tracing learning through the people, product and pathways in which policy flows to understand how it is territorialized elsewhere. These methods are grounded by a sensitivity for the way in which practices of citation and comparison are socially constructed processes embedded in global-local power relations and animated by both supply-side innovativeness and demand-side neediness. In thinking through the way in which these various connections and disconnections form pathways for policy movement, the proposed methodology is attuned to the “immaterial infrastructures” and the way in which these exchanges materialize through an assemblage of policy actants and artifacts that render mobile policies detectable and debatable. These views constitute an attempt to unravel the various practices of learning and in so doing, propose one possible strategy for studying policy mobilities and circulation.

Our expedition through policy movements goes further however, challenging researchers to reflect on their own processes of borrowing, interpreting and localizing knowledge. The paper illuminates on broader conceptual conversations by suggesting that our own learning process is similar to that of our subjects in that we both unfurl the pathways and processes of erudition alongside fellow policy actors through trusted engagements with their proven policy products. A more reflective approach to learning, one that orients our own learning as researchers in reference to that of our subjects, their networks, policy documents and conferences, thickens our scholarly interpretations of the multilateral flows of information. Rather than fashioning an agenda in which all learning takes place through particular practices, this paper develops one set of tactics for understanding the complexity of policy mobilities as an element of contemporary policymaking. Such an intellectual endeavour expands our understanding of how cities innovate and sharpens theorizations of the actors mobilizing material and nonmaterial forms of knowledge.

The landscape of learning invites us to reflect on the methodological implications of conducting this type of research. For methodological-practice, the paper advocates for the utilization of a variety of methods to understand how policy flows through a multitude of mobile and complex interactions between people, materials and meetings and in so doing furthers the discussions on learning, mobilities and particularly policy mobilities. In terms of methodological-theory, such tactics allow research to deliberate on practice as just that – an incident of learning exchange, annual attendance at a conference, a subscription to an academic journal – perpetual and perennial habits and systems alongside ephemeral and interim discussions that repeatedly conclude without noticeable results but are no less indispensable in the learning process. We tend to assess the efficacy of a learning experience according to its outcomes and for the most part, this is because it has been difficult to locate these imperceptible policy movements. A methodology that not only locates these elusive moments but also interrogates their circular and circuitous, prolonged and protracted influence on policymaking makes an important contribution to the wider discussion on policy mobilities.