Civil Society as Networks of Issues and Associations: The Case of Food

Abstract

Scholars usually conceptualize civil society as both a discursive and an associational space. In the former, focus is on communicative practices; in the latter, attention shifts to the actors that cooperate or clash about the identification and production of collective goods. In this chapter, we sketch the contours of an approach to civil society that treats both dimensions in an integrated way. Looking at the role of food issues in urban settings as diverse as Cape Town, Bristol, and Glasgow, we borrow from social network analysis to explore first, how civic organizations combine an interest in food-related issues with attention to other themes, thus defining different, specific agendas; next, we ask if and how interest in food identifies specific clusters of cooperation within broader civil society networks.

Civil Society: Aggregative Versus Relational Perspectives

In this chapter, we provide an empirical illustration of how to apply a relational approach to the study of civil society. By this we mean an approach with which one can combine attention to the traits of civil society actors and to the relations between them, and use network data to explain and understand civil society, its collective processes, and its role in wider society. Such an integrated approach has not proved easy to develop, as we are striving to combine two fundamental dimensions of civil society, its communicative and organizational ones. In contrast, most analysts have either focused on the communicative/ideational elements (e.g., Alexander, 2006; Seligman, 1995) or the organizational/associational ones (e.g., Maloney & van Deth, 2008, 2010; for more discussions on defining civil society, see Calhoun, 2001; Edwards, 2004). Among the former approaches, analysts have portrayed civil society primarily as a discursive space, delineated by the communicative practices through which core societal values are defined, criteria of civility are established, collective goals are formulated, and collective identities are constructed (e.g., Alexander, 2006; Habermas, 1989). In the latter group, researchers have largely equated civil society to voluntary organizations’ contributions to the development of political capacity, the strengthening of social cohesion and the quality of democratic life, and the definition and production of collective goods (Anheier, 2004, 2007; Deakin, 2001; Maloney & van Deth, 2010).

These standard approaches share a further underlying problem for empirical research, namely, what we have called an “aggregative approach” to collective processes (Diani, 2015, Chap. 1). Researchers have often tended to treat civil society as a set of aggregated a priori properties or “traits,” frequently adhering to formal definitions of civility, by which different actors and organizations are said to be defined. Following this logic, they then describe the structure of civil society as the distribution of such actors’ traits, and gauge the strength of civil society in reference to the number of citizens who, for instance, value tolerance, rational debate, and the pursuit of the common good over private gain, those who express their trust in institutions (e.g., Norris & Inglehart, 2002; Putnam, 2000) or promote collective action on public issues (Maloney & Roßteutscher, 2006; Maloney & van Deth, 2010). Unquestionably, aggregative approaches have generated important insights. Yet, this has often been detrimental to the analysis of how the same actors relate to each other in complex patterns (for exceptions, see: Anheier & Themudo, 2002; Knoke & Wood, 1981; Laumann & Pappi, 1976).

In this chapter, we sketch the contours of a relational approach to civil society with which we attempt to address both difficulties: how to better integrate ideational and associational dimensions of civil society, and how to focus on the relational structures between civil society actors, rather than simply on their traits. Building on our previous work (Diani, 2015; Diani, Ernstson, & Jasny, 2018), we rely on socialnetwork analysis (henceforth, SNA) to explore civil society as a set of (a) multiple networks connecting a multiplicity of collective agents and (b) ideational elements that assign a specific meaning to collective action.¹

SNA provides a number of methodological tools to explore the connections between different elements of a population (Borgatti, Everett, & Johnson, 2013; Kadushin, 2012; Knoke & Yang, 2008). In contrast to standard statistical techniques, it does not require independence of cases; to the contrary, its utilizers focus on such cases’ interdependencies. Accordingly, civil-society analysts are able to go beyond the properties of civic agents to instead focus on the relations between them. Not only that: researchers may apply SNA to map the connections between non-agentic ideational elements, including symbols, words, concepts, or other cultural products (e.g., Carley, 1994; Diesner & Carley, 2011). Even network analysts have not commonly linked agents and ideas; most have focused either on networks of exchanges between specific organizations, or on networks of ideational elements (Ferguson, Groenewegen, Moser, Borgatti, & Mohr, 2017), with less genuine integration of these two levels (for exceptions: Basov & Brennecke, 2017; Oberg, Korff & Powell, 2017; von Atteveldt, Moser, & Welbers, 2017).

In our social network study, agents corresponded to large sets of voluntary organizations in three different cities, which all mobilized around variable combinations of service delivery and political advocacy. We conceived the ideational elements as issue priorities identified by those same agents. Admittedly, “issue priorities” might be regarded as a poor, partial proxy for cultural and ideational elements. However, claims about the issues that organizations regard as of primary relevance for them should not be dismissed as the mere identification of specific problems or the target of fleeting initiatives. Rather, surveying how each actor ranks issues’ relevance provides a core indicator of how organizations position themselves in the context of larger collective action fields—in other words, of how they distribute and prioritize scarce internal resources, or represent their activities to members, potential members, and the larger public. Furthermore, we approach issue priorities in relational terms, unpacking how issues never have single, uncontroversial meanings. To the contrary, and following a now consolidated tradition in the study of culture (DiMaggio, 1987; Mohr, 1998; Mohr & Duquenne, 1997), issues may be subject to different interpretations by different agents, depending on the symbolic context in which they are embedded, just like other cultural elements such as attitudes or beliefs. For example, concerns about the issue of urban pollution may take a very different meaning if actors connect it to global environmental problems than to upper-middle class concerns about status and urban lifestyle. By sampling a wider set of actors, researchers can unpack such contrasting interpretations through a relational network approach. Indeed, a strength with our approach is that we can identify how a sub-set of issues are central to a sub-set of civil society agents, with which we can in turn explain how issues are interconnected by those mobilizing on them, how single issues are woven into broader agendas, and how cultural frames for collective action (Snow, 2004; Snow, Rochford, Worden, & Benford, 1986) emerge out of the interaction between actors and issues.

Although civic organizations may be regarded as connectors between issues, issues may likewise be regarded as facilitators or obstacles to the activation of links between organizations. Accordingly, one might be led to believe that issues mapped on to organizing patterns, with organizations sharing some issues being automatically connected in distinctive, dense clusters of relations. However, reality has proven more complicated, with the authors of one study on civic-organization networks in British cities suggesting that although the presence or absence of the traditional traits of protest organizations did characterize specific network positions in some cities, this what not the case in others (Diani, 2015). Thus, whether identification with a set issues (or lack of it) creates boundaries that facilitate or discourage organizational alliances (Tilly, 2005) becomes a matter for empirical investigation. With our relational approach, we respond to this task in how we analyze issue priorities and network patterns. Rather than distinct and neat clusters of collaborative organizations that all share the same issues, what we can uncover is a more nuanced story of how civil society organizations create deeper-lying “modes of coordination” (Diani, 2015) that cut across the boundaries of specific groups or associations and develop complex cooperative networks on themes of common interest.

Our case study in this chapter is food, as it provides a good starting point to explore how civic organizations combine an interest in relatively specific issues with attention to other themes. We then explore if and how an interest in food defines specific clusters of cooperation within broader civil society networks (Levkoe, 2014; Levkoe & Wakefield, 2014; Luxton & Sbicca, 2021; Sumner & Wever, 2015). In doing so, we draw upon data from urban settings as diverse as Cape Town in South Africa and Bristol and Glasgow in the UK. This enables us to conduct a rare comparative analysis of organizational networks,² in contexts that differ substantially in terms of urban inequality as well as in the salience of major political cleavages.

Exploring Civil Society in British and South African Cities

Our evidence comes from two projects, “Networks of Civic Organizations in Britain,” conducted in Bristol and Glasgow between 2001 and 2003,³ and “Socioecological Movements and Transformative Collective Action in Urban Ecosystems,” conducted in Cape Town between 2012 and 2014.⁴ Admittedly, these data are far from recent. As we will point out in the discussion of our findings, the specific local agendas may well have changed substantially since data were collected. Still, we do not regard this as a problem, because this article is not an account of contemporary urban dynamics and should not be taken as such. Rather, we are illustrating an approach—and a method—to integrate the cultural and organizational dimensions of civil society. In such a context, data at different points in time can be useful, if they help analysts to explore network mechanisms in polities that differ on theoretically relevant grounds. In our case, this means looking at networks in settings with different levels of democratic consolidation and cleavage salience. In this regard, an exploration of South African and British cities may be treated as a most dissimilar design comparison (Dogan & Pélassy, 1984), despite their sharing a relatively similar institutional system by virtue of South Africa’s colonial past as part of the Commonwealth. On a smaller scale, significant differences in opportunities for civic activism may also be found between British cities, although they may appear quite homogeneous by comparison to urban areas elsewhere in the world (Diani 2015, pp. 46–47, pp. 194–198).

In relation to our particular focus, three dimensions are worth pointing out. First, although inequality has risen consistently across the globe in the last few decades, the depth of social divisions and their embeddedness in race and class have historical and deeper patterns in South Africa (Maharaj, 2020; Seekings, 2000). The country often scores among the most unequal countries in the world according to the Gini index, which has continued to deteriorate under the African National Congress (ANC) government, especially since its neo-liberal policy turn in the late 1990s (Ballard, Habib, Valodia, & Zuern, 2006b, pp. 13–14), and further under Zuma’s presidency 2009–2018.⁵ As for UK cities, despite a common shift from an industrial to a service economy, inequality and deprivation were still more pronounced in Glasgow than in Bristol at the time of the study (Diani, 2015, p. 30).

Second, the salience of main political cleavages differed as well: in South Africa, proximity to or distance from the dominant ANC party and its partners in the so-called tripartite alliance (Sanco, the federation of anti-apartheid civic organizations, and Cosatu, the unions’ confederation) shaped the city’s alliance patterns (Diani, Ernstson, & Jasny, 2018; this despite ANC having lost Cape Town to its rivals since 2006). The same applied in a significant way to organizations close to or distant from Glasgow’s Labour party (the study was conducted before the growth of the Scottish National Party), although this was not the case in Bristol, where lines of political identification were multiple and not as consolidated (Baldassarri & Diani, 2007, p. 752).

The third important element to consider was the variable weight of contentious repertoires of action in the three cities. The most significant differences lay between the UK and South Africa, with the latter displaying exceptionally high levels of radical contention (Ballard, Habib, & Valodia, 2006a; McFarlane & Silver, 2017; Mottiar & Bond, 2012). This may be at least partially due not only to the stronger salience of cleavages in that country, but also to ANC’s electoral domination, akin to a one-party rule, coupled with the legitimacy gained by ANC, given its role in anti-apartheid struggles. This has, especially in the early post-apartheid period from 1994 to around 2010, made it recognizably harder to build wider multi-sectoral popular platforms outside the tripartite alliance to link localprotests into a broader national opposition of ANC’s neoliberal policies (Ballard, Habib, & Valodia, 2006a).⁶ Accordingly, the radicalization of repertoires of action may be related to a lingering lack of political opportunities for challenging groups, and their shortage of political resources, beyond the very local level (on radicalization, see e.g., Alimi, Demetriou, & Bosi, 2015). As for the UK, the continued perception of Glasgow’s civil society as more confrontational than Bristol depended to a large measure on the stronger ties between protest organizations in Scotland than in the South-West of England (Diani, 2012, 2015, Chap. 9).

The depth of social divisions, the salience of major cleavages, and the variation in action repertoires may all affect the structure of alliances within civil society (Diani, 1995, 2015). They may similarly influence the way in which issues are shaped and connected to each other. Food represents a notable case to illustrate these mechanisms. Like many large and multi-dimensional issues, it has been associated to quite different agendas and represents a focus of mobilization for highly diverse actors and coalitions. It may be part of classic environmental agendas, linking actions on the environmental impact of food production to the protection of the natural environment mainly based in a (new) middle class and moderate perspectives; but it may also be strongly connected to approaches focusing on inequality and social deprivation within affluent societies, possibly from a “right-to-the-city” perspective, as well as to global justice proponents arguing for a radical change in the relationship between food corporations and the small producers in the global south. Interest in food may drive attempts to transform the behavior of individual consumers and consolidate new markets, but it can also provide the basis for collective actions oriented to the practice of alternative lifestyles, as exemplified by environmental groups supporting food-growing allotment gardens in seeking a greener, more self-sustaining urban lifestyle. Attention to food as a public issue may be found across major political cleavages, involving quite diverse actors, from left-wing radical groups—as was historically the case with Black Panthers in US cities, who developed the Free Breakfast for Children programs as a mode of fighting structural injustice—all the way to liberal, conservative, and religious groups supporting charity “food banks” (Barthel, Parker, & Ernstson, 2013; Battersby & Haysom, 2019; Cherry, 2006; Forno, 2019; Halkier, 2019; Herring, 2014; Jallinoja, Vinnari, & Niva, 2019).

We collected our data among organizations that combined in a variable measure interest in social and ecological issues. In the UK, they focused on three main types of issues: environment, social exclusion and inequality, and minorities and migrants. Given the impossibility of mapping the whole of civil society, those issues were chosen because (a) they provided a sufficiently broad illustration of core urban problems and (b) they were distinct enough to be the target of specific campaigns or even nimby activism, yet could also serve to link into broader, more encompassing agendas. Apart from the major organizations operating on a city-wide scale, groups included in the study came from relatively deprived areas of the two cities: the Southside in Glasgow, characterized by a massive historical presence of working class, including neighborhoods such as Govan, Govanhill, Gorbals, and Pollokshields; and the neighborhoods of Easton, Knowles, Withywood, and Hartcliffe in Bristol, featuring a strong presence of ethnic minorities. Altogether, 124 organizations in Glasgow and 134 in Bristol were included in the study (Diani, 2015, Chap. 2).

The Cape Town study was part of a larger research program on urban ecology and urban political ecology (Ernstson, 2011, 2013; Ernstson & Sörlin 2019; Lawhon, Ernstson, & Silver, 2014), in which researchers conceived the urban environment broadly to include both ecological as well as social issues. The organizations surveyed ranged from classic environmental groups working on conservation issues to action committees addressing fundamental environmental justice themes such as access to energy, sanitation, and health, as well as the quality of food and housing. Barring the limited attention to migrants’ rights, the themes addressed by civic organizations in Cape Town are otherwise largely comparable to those addressed by UK civic organizations. Complete data were collected for 129 organizations in Cape Town, once again located in areas with a very diverse socio-economic status, from affluent Constantia to environmentally and socially deprived areas of Cape Flats (Diani, Ernstson, & Jasny, 2018).

In all three cities, respondents were asked to identify their priorities out of a long list of issues (about 50 issues in the UK, 30 in Cape Town), among which those linked to food attracted considerable attention (Table 8.1). In both Bristol and Glasgow, about one organization out of five expressed an interest in generic food issues, and a similar share combined this with a more specific attention to genetically modified food (henceforth, GM). In Cape Town, the wording of the issues was different, and so was the distribution of responses: 30% of organizations claimed an interest in “urban farming and food security,” but only three in GM food. Coupled with the fact that two of those groups also claimed an interest in the former, this low figure resulted in the Cape Town analysis only differentiating between organizations interested or uninterested in food issues, without further qualification.

Table 8.1 Interest in food-related issues

How did interest in food relate to other issues? As a preliminary step, we submitted the list of items in the three cities to a standard data reduction technique, principal component analysis (henceforth, PCA). In Bristol and Glasgow, we thus identified five underlying dimensions, which we labeled social exclusion, environment, minority citizenship, global justice, and housing (Diani, 2015, pp. 41–42; see also Table 8.11 in the appendix for details); in Cape Town, we identified four dimensions, grouping together issues linked to global environmental justice, urban conservation, right to the city, and issues aligned along the alternative between urbansustainability (i.e., a managerial approach to urban issues) and social rights (Diani, Ernstson, & Jasny, 2018; see also Table 8.12 in the appendix). We then proceeded to build networks of issues. We assumed a connection between two issues if a relatively high number of organizations indicated both among their priorities. We calculated the link’s strength through the Jaccard coefficient (Borgatti & Halgin, 2011, p. 421), which takes into account the fact that the strength of the connection between pairs of issues, sharing the same number of organizations interested in them, may actually be quite diverse, depending on the ties such issues may have to other issues. For example, the fact that ten organizations claimed an interest in both GM food and, say, animal rights would imply quite a different tie strength between the two if there were no organizations claiming an interest in GM food and urban pollution or environmental protection (or any other issue) than if many such existed. More specifically, we concentrated on the strongest ties, defined as those at least one standard deviation above the average strength of ties in any specific network. In the figures that follow, color and shape of nodes will correspond to the different sets of issues, identified through principal component analysis. This will offer us a preliminary way to start unpacking the two fundamental dimensions of civil society, its communicative and organizational dimensions, which we can now operationalize through comparing two criteria of group issues together in our data: correlation in the case of PCA, and co-occurrences in the issue networks.

The Structure of Issue Networks: Insights into the Discursive Space Produced by Civil Society

Starting with the Bristol case, the graph of the strongest connections between issues (Fig. 8.1) well matches the sets identified by PCA (Table 8.11 in the appendix): a set of environmental issues, which includes food along with nature conservation, pollution, forestry, energy, and transport among others, can be seen on the bottom right of the graph (white squares), whereas minority and migrants’ issues are on the left side (green triangles) and social deprivation (black triangles) and housing issues (white circles) are mostly at the network center. The only set of themes lacking a clear network position were those broadly associated with global justice (black circles); although correlated, they did not display consistently strong ties to each other. Interestingly, organizational representatives perceived GM food issues as connected to global justice themes (probably a reflection of their role in conflicts between strong corporations and producers in the global South) rather than to environmental ones, as was the case with generic food themes. However, even that connection does not appear to be particularly robust, as in terms of strong ties GM food was rather distinctive. This does not imply its marginality in local civil society, as interest in the theme was quite significant (see Table 8.1). Rather, it illustrates the difficulty to locate it within a specific discourse or a specific agenda.

Fig. 8.1

Issue network in Bristol (cut off point 0.41, one s.d. above mean). Green triangles: inequality and deprivation issues; black triangles: minorities and migrants’ issues; black circles: global justice issues; white squares: environmental issues; white circles: housing issues. Source: Design by authors

The position of generic food issues was quite different, as we show in Fig. 8.2, reporting their ego-network. On the one hand, Bristolian organizations perceived food as strongly connected to a broad range of environmental issues; on the other, they also assigned it some significant connections to social issues, such as access to higher education and housing, or globalization and migration dynamics. Apart from being embedded in distinctive sets of issues, food issues were also highly central in the whole issue network. In Table 8.2, we report two standard centrality measures, degree and betweenness.⁷ Out of 49 issues, food ranked eighth in terms of degree, with thirteen other issues being strongly connected to it (about one standard deviation above the average of eight); it was even more central in terms of betweenness, part of a very small set of issues with particularly high scores on that particular measure. As we also show in Fig. 8.1, food was often in an intermediate position on the paths connecting other issues in the network. We take this as a signal of how food could play a central role in constructing wider frames for collective action.

Fig. 8.2

Ego-network of food issues in Bristol (cut off point 0.41, one s.d. above mean). Green triangles: inequality and deprivation issues; black triangles: minorities and migrants’ issues; black circles: global justice issues; white squares: environmental issues; white circles: housing issues. Source: Design by authors

Table 8.2 Centrality of issues in Bristol (network dichotomized at 0.41 cutoff; issues marked by an asterisk are central in both UK cities)

Based on the structure of the issue network in Glasgow, we believe that foods occupies a more distinct and far less central position here than in Bristol, despite its overall popularity among civic organizations being very similar (Fig. 8.3). In Glasgow, food was part of a distinct component of the network, detached from the main one, and only consisting of three other heavily correlated issues: animal welfare, hunting, and science and technology. As for GM food, it was as isolated as in Bristol, at least in terms of strong connections, yet in a context in which most themes correlated with ideas of “global justice”—and indeed of environmentalism as well—were peripheral to the network. The other two components, distinct from the main one, again consisted of environmental themes (transport and energy) and those linked to global justice (globalization, Third World poverty, and asylum seekers). Social issues related to deprivation and community development, housing, and minority rights heavily dominated the main component. Similar considerations apply to the analysis of issue centrality. In both cities, most central issues referred to inequality and social exclusion. However, many central issues in Bristol also referred to environmental and global justice problems. This did not happen in Glasgow, where food was not among the most central issues (Table 8.3). In a nutshell, although similarly relevant in terms of appeal, food issues occupied very different positions in the two cities. In Bristol, they were at the intersection of several different agendas, combining different sets of issues; in Glasgow, they were strongly related to a small, distinctive set of themes, which combined in a very specific agenda.

Fig. 8.3

Issue network in Glasgow (cut off point 0.451, one s.d. above mean). Green triangles: inequality and deprivation issues; black triangles: minorities and migrants’ issues; black circles: global justice issues; white squares: environmental issues; white circles: housing issues. Source: Design by authors

Table 8.3 Centrality of issues in Glasgow (network dichotomized at 0.51 cutoff; issues marked by an asterisk are central in both UK cities)

Moving to the issue network in Cape Town provides still a different account of the position of food-related issues. It is not, it has to be said, a fully comparable account, as the list of issues submitted to organization representatives was different, and reflective of the project’s focus. In particular, whereas in Britain the reference was to generic “food issues,” in Cape Town it was more specific, to “urban farming and food security” (GM food was also represented, as in Britain). As a consequence, the search for macro-issues generates partially different factors from the ones identified in Britain (see Fig. 8.4): Whereas “urban conservation” and “global environmental justice” broadly correspond to the “environmental” and “global justice” factors in the UK, in Cape Townsocial inequality and community development issues combined under two different headings: one labeled “social rights,” addressing labor, gender, and youth conditions as well as health, and another labeled “right to the city,” combining community development issues with claims for the strengthening of urban democracy (Diani, Ernstson, & Jasny, 2018). In Cape Town, the issue network was split into two different components, one consisting of “global justice” issues, which included GM food (similarly to the UK), and another in which “right to the city” seemed to provide a bridge between “urban environmental conservation” and “social rights” themes.

Fig. 8.4

Issue network in Cape Town (cut off point 0.21, one s.d. above mean). Red circles: global environmental justice issues; green squares: urban conservation issues; black circles: social rights issues; white triangles: right to the city issues. Source: Design by authors

The centrality of the “urban farming and food security” issue closely matched that of generic food issues in Bristol in terms of the overall volume of connections, as the normalized degree scores were very similar (0.27 in Bristol vs. 0.23 in Cape Town: Table 8.4). However, the issue’s capacity to connect other types of themes and discourses seemed very limited, as its betweenness score was extremely low. It was, in fact, even lower than that of GM food, despite the latter’s peripheral position within a component limited to global justice issues (Fig. 8.4). The explanation for this apparent paradox lies in the structure of the ego-network of food security issues (Fig. 8.5): Although it was of a comparable size to the Bristolian one, it was more homogenous, consisting almost exclusively of other environmental issues, and most importantly, highly dense. This substantially reduced betweenness scores, suggesting that food issues be primarily contained within a fairly specific environmental discourse.

Table 8.4 Centrality of issues in Cape Town (network dichotomized at 0.21 cutoff)

Fig. 8.5

Ego-network of Food security issues in Cape Town (cut off point 0.21, one s.d. above mean) Green squares: urban conservation issues; black circles: socialrights issues; white triangles: Right to the city issues. Source: Design by authors

A few comparative comments may be in order before shifting the focus to the relationship between food issues and alliance structures. First, the popularity of GM food issues differed substantially, being high in the two British cities and very low in Cape Town (although the size of that difference may have been partially due to differences in research design). When it comes to embeddedness in broader agendas, however, the only meaningful cluster of issues comprising GM food was actually found in Cape Town, in the context of global environmental justice initiatives. In Bristol and Glasgow, GM food seemed to stand out as an issue with a peculiar profile, that was difficult to connect systematically to one specific agenda. The three cities were more similar in the popularity of other food issues (generically defined in the UK study, linked to urban farming in the case of Cape Town), with 30% to 40% of organizations claiming an interest in them. However, the three cities differed substantially in the centrality of food issues in relation to broader agendas. In Glasgow, these were part of an isolated component. In Cape Town, they had high centrality, but this depended largely on their embeddedness in environmental agendas, and was not matched by strong links to other issues. Only in Bristol did they seem to play a central role in establishing connections between different agendas within civil society. The finding about Cape Town is particularly intriguing: given the city’s high levels of deprivation, one might have expected a stronger connection between food and social inequality issues. This would also be consistent with the very high number of groups and organizations that were documented to be active on food issues just after our fieldwork (Battersby et al., 2014), a paradox we will return to in the conclusions.

The Structure of Alliance Networks: Insights into the Associational Space Produced by Civil Society

In the previous section we explored the connections that organizations create between different issues by including them among their priorities. Here we reverse the perspective and ask if and to what extent organizations interested in food-related issues occupied distinctive positions within civil society networks. In the case of the UK, one can differentiate between organizations that did not identify food as a priority, those who were interested in generic food issues, and those who combined such interest with a more specific attention to GM food. Studying the graph showing inter-organizational collaborations in Bristol, of any intensity, one can see that groups with an interest in GM food (triangle-shaped nodes in Figs. 8.6 and 8.8) were more densely interconnected. Groups only interested in generic food issues also displayed some level of connectedness, but with a higher proportion of unlinked organizations and rather engaged in alliances with groups focused on other issues (Figs. 8.7 and 8.8).

Fig. 8.6

Inter-organizational cooperations in Bristol (white circles: not interested in food; green squares: interested in food issues, but not in GMO; black triangles: interested in both types of issues). Source: Design by authors

Fig. 8.7

Cooperations in Bristol between organizations interested in food issues but not in GMO. Source: Design by authors

Fig. 8.8

Cooperations in Bristol between organizations interested in both generic food issues and GMO. Source: Design by authors

A more formal test of the distribution of ties between three types of organizations confirms the visual impression (Table 8.5). Ties between organizations interested in food but not in GMOs were slightly denser than a random distribution would suggest (the ratio between observed and expected ties under conditions of independence was about 1.5), whereas ties between organizations also focused on GM food were more than three times above what one should expect if the issue had no effect whatsoever on alliance patterns. These differences are highly significant, suggesting that attention to food issues actually characterized the local civic network in some meaningful ways. This was not the case, however, if we concentrated on the strongest ties, in other words, those that combined exchanges of resources with the deeper connections created by shared core members or strong interpersonal ties (“social bonds”: Baldassarri & Diani, 2007). Here, no significant discernible pattern emerged. If anything, organizations sharing a similar position on food issues seemed less, rather than more, likely to be connected by strong ties (Table 8.6).

Table 8.5 The salience of food issues in the Bristol civic network

Table 8.6 The salience of food issues in the Bristol civic network (strong ties only)

A broadly similar pattern can be detected in Glasgow, if slightly less pronounced, and with a more similar structure of ties among the two sets of organizations with interests in food (Figs. 8.9, 8.10, and 8.11). As in Bristol, inter-organizational connections were most likely among groups with an interest in GM food (Table 8.7); in contrast to Bristol, the ratio between observed and expected ties remained higher for groups mobilizing on GM food also in the case of the strongest “social bonds.” However, dense connections also linked these organizations to groups with no interest in food whatsoever, which made it difficult to identify a salient role for food issues in the strong ties network (Table 8.8).

Fig. 8.9

Inter-organizational cooperations in Glasgow (white circles: not interested in food; green squares: interested in food issues but not in GMO; black triangles: interested in both types of issues). Source: Design by authors

Fig. 8.10

Cooperations in Glasgow between organizations interested in food issues but not in GMO. Source: Design by authors

Fig. 8.11

Cooperations in Glasgow between organizations interested in both food issues and in GMO. Source: Design by authors

Table 8.7 The salience of food issues in the Glasgow civic network

Table 8.8 The salience of food issues in the Glasgow civic network (strong ties only)

It is worth noting that food issues seemed to have the same salience in the two cities despite being located in so different positions within the issue network. As we described, food was fairly central in Bristol, but quite peripheral in Glasgow. Yet, this did not result in differences in the issue’s salience in the inter-organizational network. In both cities, this was significant in terms of generic resource exchanges, but not in terms of the strongest “social bonds.” As we have shown in our previous work, strongest ties are most likely to connect organizations involved in social movement dynamics, or social movement “modes of coordination” (Diani, 2015). In this case, however, British cities displayed network patterns that suggested food was primarily the object of initiatives taking a coalitional form: in other words, relatively dense exchanges of resources, but much lower levels of solidarity and shared identity between organizations interested in the issue.⁸ The low ratio between observed and expected ties among organizations that were uninterested in food also points to the fact that food issues had a modest capacity to stir emotions and to generate strong oppositions. This finding should not be overemphasized: after all, it is fairly normal that ties be denser among actors with a specific interest than among those who only share their disinterest in that particular issue. However, empirical exploration of civicnetworks suggests that some issues may be more polarizing than others. For example, in our work on Cape Town civic networks (Diani, Ernstson, & Jasny, 2018) we found “urban conservation” and “global environmental justice” issues to be more polarizing than “right-to-the-city” issues.

Civic networks in Cape Town present a different profile on several grounds. The network consisting of all resource exchanges (Fig. 8.12) suggests quite frequent ties between organizations that differed in their attention to food. Although the majority of organizations interested in food were connected, many others were disconnected in that particular sub-network (Fig. 8.13). As it happens, only four organizations interested in food (indicated by black triangles) were isolated in the full civic network (left-hand side of Fig. 8.12), but 15 were isolated in the network only consisting of actors mobilizing on food (Fig. 8.13). In Table 8.9, we have confirmed the impression generated by the visual inspection of the graphs: The propensity of groups interested in food to exchange resources or collaborate with each other was only marginally above a random distribution, certainly much lower than in the two British cities (Tables 8.5 and 8.7).

Fig. 8.12

Inter-organizational cooperations in Cape Town (white circles: not interested in food; black triangles: interested in food). Source: Design by authors

Fig. 8.13

Cooperations in Cape Town between organizations interested in food. Source: Design by authors

Table 8.9 The salience of food issues in the Cape Town civic network (any tie)

The effect of interest in food on inter-organizational exchanges in Cape Town seems to follow an opposite pattern to what we found in the UK: If we focus only on strong ties, we find significant effects of interest in food on the structure of the network (Table 8.10). The probability of a strong tie between organizations interested in “urban farming and food security” was more than twice what one should expect in case of a random distribution. Again, it was not a polarizing issue (disinterested groups were not strongly connected to each other), yet it seemed to have the capacity of shaping the stronger ties, those that imply some higher level of mutual solidarity, rather than ties that were less demanding and less symbolically charged.

Table 8.10 The salience of food issues in the Cape Town civic network (strong ties only)

Conclusions

In this chapter, we have proposed an approach to better integrate the ideational and associational dimensions of civil society, which researchers usually treat disjointedly (Edwards, 2004). In doing so, we have attempted to move from an aggregative to a relational view of civil society (Diani, 2015), examining the interactions between its different components rather than focusing exclusively on their traits or properties. More specifically, taking interest in food as our case study, we have explored the relation between ideational elements and associations from two complementary perspectives. On the one hand, we have suggested that by making claims about their priorities and combining them, members of civic organizations define specific systems of meanings, and shape civil society agendas. We have shown that issues do not have an objective meaning, but take different meanings depending on the other themes to which they may be related (DiMaggio, 1987; Mohr, 1998; Mohr & Duquenne, 1997; Pachucki & Breiger, 2010). On the other hand, we have explored the extent to which interest in food issues characterized specific structural positions within civic organizations’ networks in different local settings.

It is important to be clear about the limits of the exercise. Admittedly, our treatment of food issues was quite superficial. With the exception of GM food, both studies relied on generic definitions of the issue. In the UK, respondents were only asked about their interest in “food,” without further qualification. In Cape Town, the question was more specific, referring to “urban farming and food security,” but still broad. A more detailed treatment of the multiple aspects of food (e.g., as part of alternative lifestyles, as dimension of domestic inequality, or as a global issue linked to multinational corporate capitalism) would have certainly sharpened our analysis. Even so, our findings still highlight some of the main differences in the insights that an aggregative and a relational approach to civil society may generate.

In particular, we have found that an issue’s popularity (measured by the number of organizations that regard it as a priority) does not necessarily correlate with its location in broader agendas: highly popular issues are not necessarily central to the formation of wider comprehensive frames for collective action. This is an important finding because proponents of an aggregative approach would necessarily take a popular issue to be central for wider collective action. Indeed, utilizing an aggregative approach to the interest in GM food shows that its popularity was much higher (at least at the time of the surveys) in British cities than in Cape Town. Although this is an interesting finding in its own right, suggesting that the topic was more easily addressed by organizations operating in more affluent settings, our relational approach paints a richer picture. Despite its significant appeal to civic organizations, GM food’s structural position in the larger issue network was one of isolation in both British cities. Ironically, it was in Cape Town, where its weight was more limited, that GM food was linked into a distinctive cluster of issues. However, that was a set of global themes, isolated from the rest of the issue network—unable, in other words, to connect in a significant way to agendas more closely addressing local issues, whether from a social or an environmental perspective.

A relational approach is similarly rich in insight if one examines generic food issues. Using an aggregative approach, one would conclude that their popularity was pretty constant across the three cities. If, however, one looks at the patterns of relations between issues—or, in other words, at the structures of civic agendas—in different cities, a finer-grained picture emerges. In all three cities, generic food issues were primarily connected to broader environmental concerns at the time of the surveys, rather than to socialinequality and welfare agendas. But the extent of such connection differed: in Bristol, namely, in the city closest to a post-industrial economy based on high-tech research and an advanced service sector, food issues were at the intersection of environmental and broader social agendas; in Glasgow, they were far more peripheral, as a focus on deprivation still seemed to influence local public discourse in a significant way; ⁹ in Cape Town, they were firmly located within an environmental conservation agenda.

If utilizing a relational approach to issues results in a different story from the one that utilizing an aggregative perspective would, the same applies to alliance patterns between organizations interested in food. We had a similar number of such organizations in the three cities, yet the probability that they worked together was not the same but varied depending on the type of relation. In Bristol and Glasgow, the probability of a collaboration was significantly higher when “collaboration” meant the exchange of resources between two organizations, regardless of the depth of such connection. However, we found no significant difference when we focused on the network consisting of the strongest links, those that also implied sharing core activists and/or strong personal ties between core members of two organizations. In Cape Town, in contrast, organizations interested in food issues represented a distinct cluster within the larger network only when we took the strongest ties into account. Cape Town was the only city in which interest in food issues seemed to characterize clusters of organizations linked by the strong, multiplex ties that are closest to a “social movement mode of coordination” (Diani, 2015).

To further illustrate the power of a network analytic approach, we will close by returning to the paradox of food issues in Cape Town. With our combined findings about the issue and organizational networks, we have generated a profile of the situation in Cape Town that people familiar with the area might find puzzling. Given the amount of deprivation in some communities within the city, the persistent segregation across race and class lines, and the role that urban farming might play in addressing at least partially the link between inequality and poor diet, one might have plausibly expected “urban farming and food security” to be more strongly linked to “social rights” or “right-to-the-city” type of issues, and thus for food to be part of wider-spanning agendas for collective action. As a matter of fact, there is ample evidence that activism on food issues has intensified over the last few years, with major civil society, policy, and scholarly activities around food security being promoted in the city, including in areas that we already widely covered in our study, like the Philippi Horticulture Area (Battersby et al., 2014; Kanosvamhira, 2019). This means that our network survey, if carried out today, might show that food has become a more integrative issue. However, since our goal was to illustrate a logic of analysis, not to provide an up-to-date account of urban politics in specific settings, we can make a more useful final point on how one could expand the relational analysis further.

The fact that food in Cape Town was most strongly linked to conservation issues does not mean that it had no connection to other issues. It simply means that more organizations stressed a link between food and environmental conservation than between food and social inequality issues. This may be due to deep differences in resources within civil society; organizations with a (new) middle-class membership, richer in resources, may find it easier to engage on multiple issues, combining attention to urban farming with conservation issues, whereas grassroots groups representing the most dispossessed communities may be forced to concentrate on specific issues because of their limited resources. To explore this hypothesis within a narrow aggregative approach would be difficult, if not impossible. A relational approach would simply looking separately at the issue networks created by organizations operating in affluent or deprived environments. If many affluent (mostly white) groups indicated food as a priority alongside conservation issues, whereas deprived (black/colored) groups focused mostly on single-issues, that might account for food’s more solid link to conservation than to “right to the city” issues. Of course, this is just a working hypothesis. What we have illustrated in this piece, nonetheless, is the power of a network analytic approach to public issues when it comes to identifying non-obvious patterns and new, challenging research questions.

Appendix

Table 8.11 Issues addressed by civic organizations in Bristol and Glasgow, and their popularity

Table 8.12 Issues addressed by civic organizations in Cape Town, and their popularity